Lecture 1: History of Aerial Photography

Prof. John E. Estes Remote Sensing Research Unit Department of Geography University of California Santa Barbara, California 93106-4060 Telephone: (805) 893-3845 Fax: (805) 893-3703

Introduction
There are many interesting events in the history of aerial photographic interpretation/remote sensing. The First Edition of the Manual of Remote Sensing of the American Society of Photogrammetry and Remote Sensing has a good chapter (chapter 2); while the Second Edition of the "Manual" has only one part of Chapter One covering history. Some of the material that follows was taken from a wide variety of sources including the Third Edition of the Manual of Photogrammetry; books such as Deep Black by William Burrows; and Air Spy, by Constance Babington Smith; and many Technical Reports and some newspaper articles. The chronology shows that this technology has: matured relatively recently; been built upon the inputs of a wide variety of individuals, some of whom they have heard of before; been driven by both the military and the commercial marketplace; and is continuing a rapid technological advance on a global scale (e.g. SPOT, France; Radar-Sat, Canada; JERS-1, Japan; IRS, India; and all the U.S. commercial satellites).

Chronological History of Aerial Photography and Remote Sensing

Important dates in the chronological history of photography, aerial photographic interpretation, and remote sensing: Circa 300 BCE - Greece, Aristotle philosophizing at some length about the nature of light, envisions light as a quality and not as an actual substance; as it was thought of by many at the time. He observed that some objects have the potential for transpa rency but this state is only rendered actual by the presence of light. He then defined light as the act of, or energy of, a transparent body as such. 10th Century - Al Hazan of Basra credited with the explanation of the principle of the camera obscura. 1666 - Sir Isaac Newton, while experimenting with a prism, found that he could disperse light into a spectrum of red, orange, yellow, green, blue, indigo, and violet. Utilizing a second prism, he found that he could recombine the colors into white light.

1802 - Thomas Young puts forth the basic concepts of the Young-Von Helmholtz theory of color vision: Three separate sets of cones in the retina of the eye, one attuned to red, one to blue, and one to green. 1827 - Joseph Nicephore Niepce takes the first picture of nature. (Exposure time was 8 hours, emulsion was bitumen of Jedea.) 1829 - Joseph Nicephore Niepce and Louis M. Daguere signed their partnership agreement (Nicephore Niepce had been working on Heliography, or sun drawing; Daguerre on dioramas, which he constructed with the aid of a camera obscura.) 1839 - Daguerre announces the invention of Daguerrotype (Niepce had died). Daguerre had discovered that mercury vapors could bring out an image on a silver plate and that sodicum thiosulfate ("hypo") could fix the image and make it permanent. 1939 - William Henry Fox Talbot describes a system of imaging on silver chloride paper using a fixative solution of sodium chloride. Talbot later found that the latent image could be developed in a solution of gallic acid, and he was teh first person to employ a negative/positive process "Calotype" laying the groundwork for modern photography. 1830s - Invention of the stereoscope by the Germans. The device was used during the Victorian era for amusement. 1855 - Scottish physicist James Clark Maxwell, postulates the color additive theory for the production of color photographs. 1858 - First known aerial photograph is taken from a captive balloon from an altitude of 1,200 feet over Paris by Gaspar Felix Tournachon Nadar. 1861 - With the help of photographer Thomas Sutton, Maxwell demonstrates his techniques using a bow of multicolored ribbon. (Red filter - sulfo-cyanice of iron, blue filter - ammoniacal sulfate of copper, green filter - copper chloride, a fourth filter of lemon-colored glass was also used.) 1860s - Use of aerial observations from captive balloons in American War. Balloons used to map forest in 1862 not aerial photo though. 1870s - Pictures taken from greater heights, 33,000-34,000 feet, from free balloons. 1873 - Herman Vogel found that by soaking silver halide emulsions (which are naturally sensitive to only blue light) in various dyes, he could extend their sensitivity to longer and longer wavelengths, paving the way for photography in the near infrared. 1879 - S.P. Langley begins work to find a superior radiation detector. 1887 - Germans began experiments with photography for forestry. 1899 - George Eastman produced a nitrocellulose-based film which retained the clarity of the glass plates which had been used to that time. 1903 - Julius Neubronne patents breast mounted camera for pigeons. 1906 - Albert Maul takes first aerial photograph using a rocket propelled by compressed air which rose to a height of 2,600 feet and took pictures and then parachuted the camera back to earth. 1906 - G.R. Lawrence who had been experimenting with cameras for some time (some of which weighed more than 1,000 lbs.) which were hoisted into the air with the aid of balloon-kites and associated controls, takes pictures of San Francisco earthquake and f ire

damage from an altitude of some 600 meters. Many people have thought that these photos were taken from airplanes. Lawrence's camera alone weighed more than the Wright Brothers plane and its pilot combined. 1909 - Wilbur Wright takes first aerial photograph from an airplane of Centrocelli, Italy. WWI produced a boost in the use of aerial photography, but after the war, enthusiasm wanted. 1914 - Lt. Lawes, British Flying Service, first takes airplane over enemy territory. 1915 - Cameras especially built for aerial use are being produced. Lt. Col. J.T.C. More Brabazon designed and produced the first practical aerial camera in collaboration with Thornton Pickard Ltd. 1918 - By this time in WWI, French aerial units were developing and printing as many as 10,000 photographs each night, during periods of intense activity. During the MeuseArgonne offensive, 56,000 prints of aerial photograpy were made and delivered to A merican Expeditionary Forces in four days. 1914-1919 - WWI produces boost in the use of aerial photography, but after war interest wanes. 1919 - Canadian Forestry Mapping Program begins. 1919 - Hoffman first to sense from an aircraft in thermal IR. First books: Lee 1922; Joerg 1923 (urban); Platt & Johnson 1927 (archaeology). 1924 - Mannes and Godousky patent the first of their work on multi-layer film which led to the marketing of Kodachrome in 1935. 1931 - Stevens development of an IR sensitive film (B&W). 1934 - American Society of Photogrammetry founded. Photogrammetric Engineering is first published. This journal of the American Society of Photogrammetry was later renamed Photogrammetric Engineering and Remote Sensing. The Society is now named the Ame rican Society of Photogrammetry and Remote Sensing. 1936 - Captain Albert W. Stevens takes the first photograph of the actual curvature of the earth - taken from a free balloon at an altitude of 72,000 feet. 1920s-1930s - Interest in the peaceful uses of aerial photography increases (ISDA, USAF, TVA). WWII brought about more sophisticated techniques in API. 1941-1945 - WWII brings about the development of more sophisticated techniques in aerial photographic interpretation (API). American, British and Germans all produce promising TIR devices. 1942 - Kodak patents first false color IR sensitive film. 1946 - First space photographs from V-2 rockets. 1950s - Advances in sensor technology move into multi-spectral range. 1954 - Westinghouse develops first side-looking airborne radar system. 1954 - U-2 takes first flight. 1956 - Lu Meuser makes first TIR motion picture employing an AN/AAS-4, a devise for air to ground strip mapping ("...features and vehicles move like an old keystone cops movie.")

1960 - U-2 is "shot down" over Sverdlovsk, USSR. 1960 - TIROS 1 launched as first meteorological satellite. 1960s - U.S. begins collection of intelligence photography from Earth orbiting satellites, CORONA and KH programs. 1962 - Zaitor and Tsuprun construct prototype nine lens multispectral camera permitting nine different film-filter combinations. ITEK employs camera to explore the potential value of multispectral photography. 1964 - SR-71 shown to the press in the Presidential campaign between Goldwater and LBJ. Late 1960s - Gemini and Apollo Space photography. 1968 - Hemphill describes first use of laser for airborne sensing. 1972 - Launch of the first Earth Resources Technology Satellite (ERTS-1). This system is later renamed Landsat-1. ERTS carries a return beam vidicon (RBV) and a multispectral scanner (MSS). 1972 - Photography from Sky Lab precursor of manned space station whos first element launch is currently scheduled for 1998. 1975 - Launch of Landsat 2. 1978 - Launch of Landsat 3 (March 5). 1978 - Launch and failure of Seasat. First civil SAR satellite. 1978 - Launch of Nimbus 7 (Coastal Zone Color Scanner). 1978 - Launch of NOAA 6 (aka TIROS-N), first satellite to carry the advanced very high resolution radiometer (AVHRR) on board. 1981 - Launch of SIR-A (Space Imaging Radar - A). 1982 - Launch of Landsat 4 (Thematic Mapper and MSS). 1984 - Launch of SIR-B. 1984 - Launch of Landsat 5. 1985 - Landsat Commercial contract awarded to EOSAT. Vendor takes over operation of the satellites and rights to Landsat data. 1986 - Launch of SPOT-1, French Earth Resources Satellite (Systeme Probatoire de la Observation de la Terre. 1988 - Indian Remote Sensing Satellite (IRS) launched. 1990 - Launch of SPOT-2. 1991 - Launch of ERS-1, European Radar Satellite, primarily designed for oceanographic applications. 1991 - Second Indian Remote Sensing Satellite launched. 1992 - JERS, Japanese Earth Resources Satellite launched with L-band radar and visible and infrared radiance/reflectance recording devices on-board.

1992 - Land Remote Sensing Act of 1992 brings Landsat back under U.S. Government control. EOSAT retains data rights to some Landsat data for up to ten years from acquisition. 1993 - Launch of SIR-C. 1993 - Launch of SPOT-3. 1994 - Landsat 6 fails to achieve orbit. 1995 - Third Indian Remote Sensing Satellite launched. 1995 - Canada launches RADARSAT. 1995 - Early CORONA and KH satellite data are declassified by an Executive Order signed by President Clinton on 23 February. This order authorizes the declassification of intelligence satellite photography acquired in the 1960s. 1995 - Launch of ERS-2. 1995 - First indication that a new class of intelligence satellite is being developed appears in the press. The new satellite code name 8x is said to be a major upgrade of the KH-12 spy satellite. The satellite which may weigh as much as twenty tons wil l be able to acquire "intricately detailed images of an area as large as 1,000 square miles of the Earth's surface...with roughly the same precision as existing satellites," according to an article in the September 28 Los Angeles Times. The Time article goes on to say that the current generation of photographic satellites photograph areas about 10 miles by 10 miles (100 square miles) typically showing details as small as six inches. 1997 - Proposed launch date of SeaWiFs, replacement for the coastal zone color scanner. 1997 - Proposed launch of SPOT-4. 1998 - Proposed first launch of the Earth Observing System's (EOS) AM-1 series on a Polar Orbiting Platform (POP). 1998 - Proposed launch date for Landsat-7. 2000 - Proposed launch date of EOS PM-1 series on POP.

References
Babington-Smith, Constance. Air Spy: The Story of Photo Intelligence in World War II. New York: Harper, 1957. Babington-Smith, Constance. Air Spy: The Story of Photo Intelligence in World War II. Falls Church, Virginia: American Society for Photogrammetry and Remote Sensing, 1985. (reprint) Burrows, William E. Deep Black: Space Espionage and National Security. New York: Random House, 1986. Digital Photogrammetry: an addendum to the Manual of Photogrammetry. Cliff Greve, editor. Bethesda, Maryland: American Society for Photogrammetry and Remote Sensing, 1996. Manual of Photogrammetry. Preliminary edition. P.G. McCurdy, et al., editors. New York: Pitman Publishing Corp., 1944. Manual of Photogrammetry. Second edition. Washington, DC: American Society for Photogrammetry and Remote Sensing, 1952. Manual of Photogrammetry. Third edition. Morris M. Thompson, Robert C. Eller, William A. Radlinski, and Julius L. Speert, editors. Falls Church, Virginia: American Society for Photogrammetry and Remote Sensing, 1966 Manual of Photogrammetry. Fourth edition. Chester C. Slama, Charles Theurer, and Soren W. Henriksen, editors. Falls Church, Virginia: American Society for Photogrammetry and Remote Sensing, 1980. Manual of Remote Sensing. First edition. Robert G. Reeves, Abraham Anson, and David Landen, editors. Falls Church, Virginia: American Society for Photogrammetry and Remote Sensing, 1975 Manual of Remote Sensing. Second edition. Robert N. Colwell, editor. Falls Church, Virginia: American Society for Photogrammetry and Remote Sensing, 1983.

Lecture 2: Elements, Aids, Techniques and Methods of Photographic/Image Interpretation

Prof. John E. Estes Remote Sensing Research Unit Department of Geography University of California Santa Barbara, California 93106-4060 Telephone: (805) 893-3845 Fax: (805) 893-3703 2.1 Review - Definitions
Photo Interpretation - The act of examining aerial photographs/images for the purpose of identifying objects and judging their significance. Photography - The art or process of producing images on a sensitized surface by the action of light or other radiant energy. Image - A reproduction or imitation of the form of a person or thing. The optical counterpart of an object produced by a lens or mirror or other optical system. Photogrammetry - The science or art of obtaining reliable measurements by means of photography.

2.2 Activities of Airphoto/Image Interpreters

2.2.1 Detection/Identification - The primary task of the interpreter is the detection and identification of objects, features, phenomena and processes. This is primarily a stimulus and response activity. The interpreter conveys his or her response by labeling. These labels are often expressed in qualitative terms, e.g. likely, possible, probable, or certain. 2.2.2 Measurement - As opposed to detection and identification, the making of measurements is primarily quantitative. Techniques used by interpreters typically are not as precise as those employed by photogrammetrists who use sophisticated instruments in making their measurements. 2.2.3 Problem Solving - Interpreters are often required to identify objects from a study of associated objects that they can identify; or to identify object complexes from an analysis of their component objects. Analysts may also be asked to examine an image which depicts the effect of some process and suggest a possible or probable cause. As we will see, a solution may not always consist of a positive identification. The answer may be expressed as a number of likely scenarios with statements of probability of correctness attached by the interpreter.

2.3 Elements of Image Interpretation

2.3.1 Basic, first order 2.3.1.1 Tone/Color - Tone can be defined as each distinguishable variation from white to black. Color may be defined as each distinguishable variation on an image produced by a multitude of combinations of hue, value and chroma. Many factors influence the tone or color of objects or features recorded on photographic emulsions. But, if there is not sufficient contrast between an object and its background to permit at least detection, there can be no identification. While a human interpreter may only be able to distinguish between ten and twenty shades of grey; interpreters can distinguish many more colors. Some authors state that interpreters can distinguish at least 100 times more variations of color on color photography than shades of gray on black and white photography. Examples of tone: Examples of color:

2.3.1.2 Resolution - Resolution is defined as the ability of the entire photographic system, including lens, exposure, processing, and other factors, to render a sharply defined image. An object or feature must be resolved to be detected and/or identified. Resolution is one of the most difficult concepts to address in image analysis. Resolution can be described for systems in terms of modulation transfer (or point spread) functions; or it can be discussed for camera lenses in terms of being able to resolve so many line pairs per millimeter. There are resolution targets that help to determine this when testing camera lenses for metric quality. Photo interpreters often talk about resolution in terms of ground resolved distance, the smallest normal contrast object that can be detected and identified on a photo. 2.3.2 Second order -- Geometric Arrangements of Objects 2.3.2.1 Size - The size of objects can be important in discrimination of objects and features (cars vs. trucks or buses; single family vs. multi-family residences, brush vs. trees, etc.). In the use of size as a diagnostic characteristic both the relative and absolute sizes of objects can be important. Size can also be used in judging the significance of objects and features (size of trees related to board feet which may be cut; size of agricultural fields related to water use in arid areas, or amount of fertilizers used; size of runways gives an indication of the types of aircraft that can be accomodated). 2.3.2.2 Shape - The shape of objects/features can provide diagnostic clues that aid identification. The Pentagon building in Washington is a diagnostic shape. Man-made features have straight edges that natural features tend not to. Roads can have right angle turns, railroads do not. Examples of shape:

2.3.3 Second order -- Spatial Arrangement of Tone/Color: It must be noted here that pattern is scale dependent. An orchard on a low altitude aerial photo may be an area of rough texture on a high altitude air photo or an area of medium or rough texture on a satellite image. This pattern may be thought of as being scale dependent. Examples of pattern and texture being scale dependent:

2.3.3.1 Texture - The frequency of change and arrangement of tones. This is a micro image characteristic. The visual impression of smoothness or roughness of an area can often be a valuable clue in image interpretation. Water bodies are typically fine textured, while grass is medium, and brush is rough, although there are always exceptions.

Examples of texture:

2.3.3.2 Pattern - Pattern is the spatial arrangement of objects. Pattern can be either manmade or natural. Pattern is a macro image characteristic. It is the regular arrangement of objects that can be diagnostic of features on the landscape. An orchard has a particular pattern. Likewise, the network or grid of streets in a subdivision or urban area can aid identification and aid in problem solving such as the growth patterns of a city. Pattern can also be very important in geologic or geomorphologic analysis. Drainage pattern can tell the trained observer a great deal about the lithology and structural patterns in an area. Dendritic drainage patterns develop on flat bedded sediments; radial on/over domes; linear or trellis in areas with faults or other structural controls. Examples of pattern:

2.3.4 Third order -- Locational or Positional Elements 2.3.4.1 Site - How objects are arranged with respect to one another; or with respect to various terrain features, can be an aid in interpretation. Aspect, topography, geology, soil, vegetation and cultural features on the landscape are distinctive factors that the interpreter should use when examining a site. The relative importance of each of these factors will vary with local conditions, but all are important. Just as some vegetation grows in swamps others grow on sandy ridges. Agricultural crops may like certain conditions. Man made features may also be found on rivers (e.g. power plant) or on a hill top (observatory or radar facility). Examples of site:

2.3.4.2 Association - Some objects are so commonly associated with one another that identification of one tends to indicate or confirm the existence of another. Smoke stacks, step buildings, cooling ponds, transformer yards, coal piles, railroad tracks = coal fired power plant. Arid terrain, basin bottom location, highly reflective surface, sparse vegetation = playa. Association is one of the most helpful clues in identifying man made installations. Aluminum manufacture requires large amounts of electrical energy. Absence of a power supply may rule out this industry. Cement plants have rotary kilns. Schools at different levels typically have characteristic playing fields, parking lots, and clusters of buildings in urban areas. Large farm silos typically indicate the presence of livestock. 2.3.5 Third order -- Interpreted from lower order elements Height - Height can add significant information in many types of interpretation tasks; particularly those that deal with the analysis of man-made features. How tall a tree is can tell something about board feet. How deep an excavation is can tell something about the amount of material that was removed (in some mining operations excavators are paid on the basis of material removed as determined by photogrammetric analysis). Shadow - Geologists like low sun angle photography because shadow patterns can help identify objects. Steeples and smoke stacks can cast shadows that can facilitate interpretations. Tree identification can be aided by an examination of the shadows thrown. Shadows can also inhibit interpretation.

Examples of shadow:

2.4 Aids to and Techniques of Image Interpretation

2.4.1 Collateral Material A review of existing source material concerning a given area, process; type of facility, or object, can aid in the interpretation of remotely sensed data. The use of collateral material may also result in a better definition of the scope, objectives, and problems associated with a given project. Collateral material can be data of many types and can exist in either analog or digital form. Collateral material has also been called in the literature ancillary data, and site scientific literature. Collateral material may come in the form of text, tables, maps, graphs, or image data/information (metadata). Census data, a map of flora of a given area, a land use map of an area, meteorological statistics, or agricultural crop reports can all be used in support of a given interpretation. Basically, collateral material represents data/information that an intepreter may use to aid in his/her accomplishment of a given analysis task. Material contained within a Geographic Information System (GIS) that is used to assist an interpreter in an analysis can be considered collateral data. Two classes of collateral data deserve special mention here. These are photo/image interpretation keys and field verification. 2.4.1.1 Image Analysis Keys - A photo/image interpretation key is a set of guidelines used to assist interpreters in rrapidly identifying photo/image features. Determination of the type of key and the method of presentation to be employed will depend upon: 1.) The number of objects or conditions to be identified; and 2.) the variability typically encountered within each class of features or objects within the key. Some authors say that as a general rule, keys are more easily constructed and used for the identification of manmade objects and features than for natural vegetation and landforms. For analysis of natural features, training and field experience are often essential to achieve consistent results. Basically, an interpretation key helps the interpreter organize the information present in image form and guides him/her to the correct identification of unknown objects. Keys can be used in conjunction with any type of remotely sensed data. That is, a key can be developed to aid the interpreter as he/she analyzes any type of imagery. Such keys can differ from those employed in other disciplines in that they can consist largely of illustrations, e.g. landforms, industrial facilities, military installations. Many types of keys are already available, if you can find or get your hands on them (this can be very difficult and a reason why people develop their own keys). All keys are, however, based upon the diagnostic characteristics of objects or conditions to be identified. Depending upon the manner in which the diagnostic features are organized, two types of keys are

generally recognized. Selective keys (Avery, 3rd ed., pg. 335, Industry, vegetation; 236, 237) are arranged in such a way that an interpreter simply selects that example that most closely corresponds to the object he/she is trying to identify, e.g. industries, landforms, etc. Elimination keys (Paine 387) are arranged so that the interpreter follows a precise step-wise process that leads to the elimination of all items except the one (ones) that he/she is trying to identify. Dichotomous keys are essentially a class of elimination key. Most interpreters prefer to use elimination keys in their analyses, although studies have revealed no significant difference between the results achieved from the use of the two types of keys as long as the material within each key is well organized. 2.4.1.2 Field Verification - Field verification can be considered a form of collateral material. Field verification is typically conducted to assist in the analysis of the data to be analyzed. Essentially, this is familiarizing the interpreter with the area or type of feature or object to be interpreted. This type of verification is done prior to the interpretation. After an interpretation, field verification can be accomplished to verify the accuracy of the interpretation conducted. Field work can be very expensive, thus this type of activity must be carefully planned. Field work is sometimes calculated as being three times as expensive as lab analysis. (This is why good interpreters can be so valuable). The nature, amount, timing, method of acquisition, and data integration procedures should be carefully thought out. Will you use windshield surveys, point or transect sampling? Will the sampling be random or systematic? Will demographic information be collected and will a human subjects release form be necessary? Will photos from light aircraft or higher resolution air or satellite photos be used to verify analyses (e.g. Thematic mapper to verify AVHRR or Airphotos to verify TM). The amount and type of field work required for a given project may vary greatly and is generally dependent upon: The type of analysis involved; Image quality, including scale resolution and information to be interpreted; The accuracy requirements for both classification, and boundary delineation; The experience of the interpreter and his/her knowledge of the sensor, area, and subject to be interpreted; and, The terrain conditions and the accessibility of the study area. For various reasons, an area may be inaccessible and the existence and/or availability of other source material limited. 2.4.2 Handling of Imagery Although a good deal of photo interpretation is still accomplished using paper prints, the use of transparencies is increasing. Transparencies can be used either as single frames or as a roll. Care should be taken when handling transparencies so that they are not marred. An orderly procedure for the handling of either prints or transparencies should be developed and adhered to in any interpretation project. Basically prints typically are

numbered and should be kept in order in so far as practical. Flight lines might be kept separate. Different dates should be kept separate, etc. When transparencies are on rolls, they are easy to keep in order. If individual frames are cut from the roll, the task of keeping things in order becomes more difficult. Try to keep track of where frames come from and who has them. Any time transparencies are used, surfaces should be as clean as possible. The interpreter should either wear cotton gloves or be sure not to touch the emulsion surface as skin oils can cause image deterioration. 2.4.3 Stereoscopic Viewing Binocular vision is an important part of most of our daily lives. It needs to be fully understood and consciously exploited by the image interpreter. Although many remote sensor systems can be employed to acquire stereo data; most stereoscopic viewing for interpretation purposes is done from vertical or near vertical aerial photography acquired by conventional aerial camera systems. When learning to interpret stereo data, the interpreter should: Make certain that the photos are properly aligned at all times, preferably with the shadows falling toward the viewer. Be careful to keep the eye base and the long axis of the stereoscope parallel to the flight line at all times. Maintain an even glare free illumination on the prints or transparencies being studied. Arrange for comfortable seating. Keep the lenses of the stereoscope clean, properly focused and separated to your interpupillary distance. The novice interpreter should not work with stereo more than thirty minutes out of any hour period. Interpreters who have difficulty with stereo should be aware of the following: A person's eyes may be of unequal strength. If a person normally wears glasses for reading or close-up work, one should also wear glasses when using the stereoscope. Poor illumination, misaligned prints or uncomfortable viewing positions may result in eye fatigue. Illness or severe emotional distress may create sensations of dizziness in one using a stereoscope. Reversal of prints may cause pseudo-stereo. A similar problem may occur if prints are aligned with the shadows falling away from rather than towards the interpreter. Objects that change positions between exposures cannot be viewed in stereo. In areas of steep topography, scale differences in adjacent photographs may make it difficult to obtain a three dimensional image.

 Dark shadows or clouds may prohibit stereo viewing of an area by obscuring an object on one photo. Individuals who have continued difficulties using stereoscopes should not attempt to master the art of stereoscopic vision with the unaided eye. 2.4.4 Use of Multiple Images Multi-Station - Not to be confused with multi-stage. This is the successive overlapping of photographs taken along a given flight line as being flown by an aircraft or a satellite. Multi-Band - Often seen to have an overlapping meaning with the next term, multispectral. As used by here, multiband indicates individual spectral bands within a given region of the EM spectrum (e.g. the red, green, and blue bands of the visible portion of the EM spectrum). Multi-Spectral - The use of images from various regions of the EM spectrum (e.g. visible, infrared, and microwave). Multi-Date - The use of images taken over time of a given area. Multi-Stage - This typically means the acquisition of images from platforms flying at different altitudes (e.g. the use of data flown by low flying aircraft, high flying aircraft, and satellite data in a given study). It has also been applied to sampling strategies. A multi-stage sampling scheme as used in statistics is one where progressively more information is obtained for progressively smaller subsamples of the area being studied. Multi-Polarization - Objects in the environment exhibit different qualities with respect to the ability to rotate a signal returned to a sensor. Multi-Direction - Most sensors employed in remote sensing are aimed vertically (SAR being the major exception). There are times when more information can be obtained using viewing angles other than vertical. Multi-Enhancement - There are many types of enhancement available to the image analyst today (e.g. optical, electronic, computer assisted, multi-date, multi-band, multispectral). Multi-Disciplinary - Basically, no one interpreter can know everything about the Earth system. By using teams of interpreters with expertise in different disciplinary areas, more information may often be gained for a given application. Multi-Thematic - Remote sensing images are a one time write, many time read. Many different themes (e.g. hydrology, vegetation, transportation, urban areas, etc.) can be extracted from a single set of images. Multi-Use - Many types of individuals from environmental planners and resource managers to public policy decision makers can use the output of an image analysis task. 2.4.5 Methods of Search The interpreter should always keep in mind the basic qualities of the imagery he/she is dealing with, e.g. film filter combinations; season of acquisition; and time of day of acquisition; scale, etc. In addition, the interpreter should always remember to examine all the title information (peripheral information) on an image. Interpretation should be conducted logically one step at a time. Interpretation should begin with the general and proceed to the specific.

 Interpretation should proceed from the known to the unknown. 2.4.6 Convergence of Evidence Image interpretation is basically a deductive process. Features that can be detected and identified lead the interpreter to the location and identification of other features. Deductive interpretation requires either the conscious or unconscious consideration of all of the elements of image interpretation that we have been discussing. The completeness and accuracy of an interpretation is to some measure, proportional to the interpreter's understanding of the how and the why of the elements, techniques, and methods of interpretation that we have been discussing.

References
Avery, T.E., and G.L. Berlin, 1992. Fundamentals of Remote Sensing and Airphoto Interpretation. Fifth edition, New York, Macmillan Publishing Company, 472 p. Estes, J.E., E.J. Hajic, and L.R. Tinney (Author-Editors), "Fundamentals of Image Analysis: Analysis of Visible and Thermal Infrared Data," Chapter 24, in Manual of Remote Sensing, Second edition, Falls Church, Virginia, American Society of Photogrammetry, pp. 987-1124. Paine, D.E., 1981. Aerial Photography and Image Interpretation for Resource Management, New York, John Wiley and Sons, 571 p.

Lecture 6: Geometry of Aerial Photography

Ahmed Fahsi, Ph.D. Alabama Agricultural and Mechanical University Department of Plant and Soil Science Center for Hydrology, Soil Climatology and Remote Sensing Normal, Alabama 35762
Objectives: 1. Differentiate among vertical, high oblique, low oblique, and horizontal photographs and state the advantages of one over the other. 2. Determine the relationship between the focal length, the angle of coverage and the ground area photographed and understand the use of a specific focal length under specific circumstances. 3. Be able to identify the fiducial marks on an aerial photograph, draw the coordinate axes and locate the three photo centers on a tilted vertical aerial photograph in areas where tall objects are present and clearly visible. 4. List the different types of photo distortion or displacement and define the difference between these two effects, state their causes, and provide solutions for their corrections. 5. List the types of distortion or displacement that radiates from the three photo centers and know how to correct for or avoid them. 6. Describe the relief, tilt, and their combined effects on the object image displacement and state inferences that can be made from the relief displacement equation for topography and solve problems based on these inferences. 7. List the factors that influence relief displacement on a single aerial vertical photograph. 8. Determine object image displacement and object heights due to relief based on the relief displacement equations. 9. Understand the geometry of the different types of aerial photographs and be able to determine ground distances, orientations, coordinates, and other related parameters from single aerial photographs (untilted, tilted, or oblique). 6.1. Classification of photographs A number of systems have been used to classify photographs. The most common system is the one that separates photographs into terrestrial and aerial (Figure 6-1).

Figure 6-1. Classification of Photographs (from Paine, 1981) Terrestrial photos are taken on the ground, while aerial photos are taken from a platform in the air. Aerial photographs may be either vertical or oblique. Vertical photographs can be truly vertical, or slightly tilted (less than 3o from the vertical). Most aerial photos are tilted to some degree. Therefore, the use of the term vertical photographs in this chapter assumes truly vertical photographs, while in reality they might be tilted up to 3o. Oblique aerial photos are photographs purposely taken with an angle between 3 and 90o from the vertical. They can be low oblique (if the horizon is not visible) or high oblique (if the horizon is visible) (Figure 6.2, Low oblique versus high oblique aerial photographs). Terrestrial photos are usually oblique or horizontal, where the axis of the camera is tilted about 90o form the vertical.

Figure 6-2. Low oblique versus high oblique aerial photographs Whether they are vertical or oblique, aerial photographs may be obtained in various formats and sizes depending on the purpose and the type of application they are used for. The format of most aerial photographs is square and sometimes rectangular depending on

the camera but the most common format is 23 by 23 centimeters (9 by 9 inches). Vertical and oblique aerial photographs also have advantages over one another in respect to their usage.

6.1.1. Advantages of vertical over oblique aerial photographs

1. Vertical photographs present approximately uniform scale throughout the photo but not oblique photos. It follows that making measurements (e.g., distances and directions) on vertical photographs is easier and more accurate. 2. Because of a constant scale throughout a vertical photograph, the determination of directions (i.e., bearing or azimuth) can be performed in the same manner as a map. This is not true for an oblique photo because of the distortions. 3. Because of a constant scale, vertical photographs are easier to interpret than oblique photographs. Furthermore, tall objects (e.g., buildings, trees, hills, etc.) will not mask other objects as much as they would on oblique photos. 4. Vertical photographs are simple to use photogrammetrically as a minimum of mathematical correction is required. 5. To some extent and under certain conditions (e.g., flat terrain), a vertical aerial photograph may be used as a map if a coordinate grid system and legend information are added. 6. Stereoscopic study is also more effective on vertical than on oblique photographs.

6.1.2. Advantages of oblique over vertical aerial photographs

1. An oblique photograph covers much more ground area than a vertical photo taken from the same altitude and with the same focal length. 2. If an area is frequently covered by cloud layer, it may be too low and/or impossible to take vertical photographs, but there may be enough clearance for oblique coverage. 3. Oblique photos have a more natural view because we are accustomed to seeing the ground features obliquely. For example, tall objects such as bridges, buildings, towers, trees, etc. will be more recognizable because the silhouettes of these objects are visible. 4. Objects that are under trees or under other tall objects may not be visible on vertical photos if they are viewed from above. Also some objects, such as ridges, cliffs, caves, etc., may not show on a vertical photograph if they are directly beneath the camera. 5. Determination of feature elevations is more accurate using oblique photograph than vertical aerial photographs. 6. Because oblique aerial photos are not used for photogrammetric and precision purposes, they may use inexpensive cameras.

6.1.3. Advantages of aerial photography (vertical or oblique)

Aerial photographs have the advantage of providing us with synoptic views of large areas. This characteristic also allows us to examine and interpret objects simultaneously

on large areas and determine their spatial relationships, which is not possible from the ground. Aerial photographs are also cost effective in interpreting and managing natural resources. They have played a significant role in map making and data analysis. The basic geometry of a vertical aerial photograph is illustrated in figure 6.3. Incoming light rays from objects on the ground pass through the camera lens before they are imaged on the film in the focal plane. The distance between the lens and the focal plane is termed focallength. The x coordinate axis is arbitrarily assigned to the imaginary flight line direction on the photograph and the y-axis is assigned to a line that is perpendicular to the x-axis. These two axes usually correspond to the lines connecting the opposite fiducial marks recorded on each side of the print (i.e., positive image).

Figure 6.3. Basic geometry of a vertical aerial photograph. The photocoordinate origin, o, is considered to be the intersection of the two lines joining the fiducial marks and, therefore, can be assumed to coincide exactly with the physical center (i.e., the geometric center) of the photograph, called principal point. Points to the east side of y-axis have positive x coordinates and points to the west side have negative x coordinates. Similarly, points north of the x-axis have positive y coordinates and points south of the x-axis have negative y coordinates. 6.2. Focal Length and Angle of Coverage Beside the quality of the lens, the focal length is one of the most important components of an aerial camera. Focal length is defined as the distance from the focal plane to the center of the lens when focused at infinity (figure 6.4). The angle of coverage corresponds to the angle of the cone of light rays reflected from objects on the ground and passing through the lens of the camera before they expose the film in the focal plane. Figure 6.5 shows that the angle of coverage increases as the focal length decreases. It

follows that the narrower the angle (i.e., the longer the focal length), the smaller the area covered on the ground. The most common focal lengths used in aerial photography are the 152.4 mm (6-in.), the 210 mm (8-in.), and 305 mm (12-in.). The choice of the most appropriate focal length depends on the type of application, the area to be covered, the topography of the region, and the funds available. Because wide angle lenses (i.e., short focal length) exaggerate the displacement of tall objects, they are better suited for photographing flat than mountainous terrain. There is always a trade-off when choosing short or long focal length lenses. This will be discussed later in this chapter (see also chapter 8).

Figure 6.4. Focal length of a simple lens. (From Paine, 1981)

Figure 6.5. Relationship between focal length and ground coverage; a short focal length covers larger areas. 6.3. The Centers of an Aerial Photograph Considering that most aerial photographs are not perfectly vertical, there are three different photo centers: the principal point, the nadir, and the isocenter. Each one of these centers plays a specific role and is of great importance to the photogrammetrist because different types of distortion and displacement radiate from each of these points. Theoretically, if an aerial photograph is perfectly vertical, the three centers coincide at one point (i.e., the principal point), which is the geometric center of the photograph defined by the intersection of lines drawn between opposite fiducial marks (figure 6.6). The principal point is the optical or geometric center of the photograph. It is the image of the intersection between the projection of the optical axis (i.e., the perpendicular to the center of the lens) and the ground (figure 6.6). The principal point is assumed to coincide with the intersection of the x and y axes. We can locate the principal point (PP) on a single photo by the intersection of lines drawn between opposite side or corner fiducial marks. When marking up vertical aerial photographs, it is convenient to mark the PP by pricking through the surface of the photo using a pin, then draw a small circle (about 3-mm diameter) around the pin-point with a faint red ink pen. This PP is then transferred stereoscopically onto the adjacent (left and right) photographs of the same flight line by pricking through its transferred positions and marking them with a circle of identical diameter. These transferred points are called transferred principal points or conjugate principal points (CPP). The line segment joining the principal points and the conjugate principal points constitute the flight line of the aircraft, also called base line or air base.

Figure 6.6. Diagram of a tilted photograph illustrating the location of the principal point (PP), the nadir (n), the isocenter (i), the axis of tilt, and the direction of tilt (up and down sides). (From Ulliman, 1995). The air base is important as it is used for lining up a photograph to adjacent photographs on the same flight line to correctly see in stereoscopy, determine height and difference in elevation of objects, precisely perform photogrammetric measurements, and to prepare maps. Because of distortions and image displacement (discussed later in this chapter), the distance between the PP and the CPP of the adjacent photograph will often be different. The nadir, also called vertical point or plumb point, is the image of the intersection between the plumb line directly beneath the camera center at the time of exposure and the ground (figure 6.7). The nadir is important because relief displacement is radial from this point and is a function of the distance of the displaced image from it. Unlike the principal point, there are no marks on the photograph that permit to locate the nadir. Locating the nadir on a tilted aerial photograph usually requires sophisticated stereoscopic plotting techniques involving expensive instruments and ground control information. However, because relief displacement is radial from the nadir, in areas where tall and perfectly vertical objects (e.g., towers, smokestacks, electric poles, tall buildings, etc.) are clearly located on the photograph, the nadir point may be determined by projecting lines along the displaced edges of these buildings as shown in figure 6.8. Notice that because the photograph is not perfectly vertical, the nadir and the principal point have different positions. Notice also that the nadir is always on the down side of the tilted photograph. If we locate the nadir and the principal point and if we know the direction of the flight, then we can determine the direction of tilt of the aircraft (figures 6.6 and 6.8).

Figure 6.7. Locations of the principal point (PP), the nadir (n) and the isocenter (i) on a tilted vertical aerial photograph. (From Paine, 1981).

Figure 6.8. Determination of the nadir point from perfectly vertical tall objects on a tilted vertical aerial photograph. (From Ulliman, 1995). The isocenter is the point halfway between the principal point and the nadir and on the line segment joining these two points on the photograph (figure 6.6). It is a point intersected by the bisector of the angle between the plumb line and the optical axis (figure 6.7). The intersection of the plane of the tilted photograph and the plane the photograph would have occupied had it been free of tilt forms a line which lies at right angles to the line connecting the principal point (PP) and the nadir (n). The isocenter is at

the intersection of these two lines (figure 6.6). The isocenter is the focus of the tilt displacement for it is the point from which tilt displacement is radial. On a true vertical photograph, the principal point, the nadir, and the isocenter coincide at the geometric center (i.e., PP) of the photograph. In this case, there is no tilt displacement, only relief displacement and radial from the geometric center of the photograph. 6.4. Distortion and Displacement Because of the optical characteristics inherent in a vertical aerial photograph and the anomalies from the camera components, a vertical photograph is not a map. According to Paine (1981), distortion in aerial photography is defined as any shift in the position of an image on a photograph that alters the perspective characteristics of the image and displacement is any shift in the position of an image on a photograph that does not alter the perspective characteristics of the photograph. Displacement results mainly from the perspective viewing of the camera resulting in a perspective or central projection on the photograph. In contrast, a map is the product of an orthographic projection (figure 6.9). Furthermore, a photograph is subject to more distortions than a map because a map is produced on a stable base material. 6.5. Sources of Distortions and Displacement On an accurate planimetric map, all features on the ground surface are portrayed at their correct horizontal positions. The observer thus has a truly vertical view of every detail shown and the spatial relationship among features is accurate. Unfortunately, this standard cannot be met by aerial photographs because of various sources of distortion or image displacement. Images on vertical aerial photographs are generally displaced from their true plane position. The main sources of displacement and distortion are the optical or photographic deficiencies (film and paper shrinkage, lens aberrations, filter aberrations, failure of the film-flattening mechanism in the camera focal plane, shutter malfunction), image motion, atmospheric refraction of light rays, curvature of the earth, tilt, and topography or relief. Fortunately, modern camera systems are built with high precision optics and mechanism that, if operated by experienced flight crews, distortions related to optical or graphical deficiencies may be negligible.

Figure 6.9. Central projection (photo) versus orthogonal projection (map). Film and print shrinkage or expansion: the quality of the film and paper print is very important to the quality of data storage and accuracy. Dilatation or shrinkage of film and print under heat or cold may change the scale of the photographs and the actual position of the objects on the photographs. Fortunately, most photographic films used for photogrammetric purposes have excellent dimensional stability. Therefore, film dimension changes are relatively small, amounting to about 0.025 mm (less than 0.001 inch) in most cases, but photographic papers are far less stable. A stable film and paper would minimize these distortions. For this reason, diapositives printed on film base are preferred for precision mapping and glass plates are used rather than film whenever extreme accuracy is desired. To determine the shrinkage or expansion of a paper we compare the measured distances between opposite fiducial marks on a print with their corresponding values obtained during camera calibration. Then, a correction factor is computed by dividing the calibrated distances by the measured distances and the resulting factor value is applied to each x and y photocoordinates employed in any subsequent analysis. The computation equations are expressed as:

where: x is the corrected photocoordinate along the x-axis for a point a, y is the corrected photocoordinate along the y-axis for a point a, xc is the calibrated fiducial distance along x-axis,

yc is the calibrated fiducial distance along y-axis, xf is the measured fiducial distance along x-axis, yf is the measured fiducial distance along y-axis, xm is the measured photocoordinate for point a along the x-axis, ym is the measured photocoordinate for point a along the y-axis, and xc/xf and yc/yf are the correction factors along x-axis and y-axis respectively. Example 6.1 Suppose that the calibrated distances between the fiducial marks on the camera are 23.25 cm along x-axis and 23.30 cm along y-axis. The corresponding distances measured on a photographic print from the same camera are 23.33 cm and 23.36 cm. If the photocoordinates, x and y, of a point measured on the print are 8.15 cm and 11.04 cm, what are the corrected photocoordinates of the point. Solution:

Lens aberrations: lens distortion radiates from the principal point, which causes object displacement either toward (closer to) or away (farther) from the principal point (the optical or geometric center) of the photograph than it actually is. Because this distortion is radial from the principal point, objects near the edge of the photograph are more distorted. This anomaly can be corrected by calibrating the lens. One way of doing this is to fly a plane area at an accurate flying height and computing the exact scale between several known points located radially between the principal point and the edge of the photograph. Through this technique, we can obtain a distortion curve that shows us how the distortion varies with the radial distance from the principal point and allows us to make corrections accordingly. Correction for these systematic distortions is usually necessary for precise mapping. In the United States, the Bureau of Standards usually performs the calibration operation and certifies the lens. In most aerial camera lenses, distortions near the center of a vertical photograph are negligible; in many lenses it is less than 0.01 mm (0.0004 inch) halfway out to the edge; and in the new low-distortion lenses, it is less than 0.01 mm at the edge of the photograph. Other lens anomalies may include random distortions throughout the lens, or scratched lens. In this case, more advanced techniques may be used. Such techniques may include high density of points sampled on a known scale photograph (as described in the previous paragraph). Instrumental errors may also be introduced if the camera and lens are incorrectly assembled or if the film is not held absolutely flat at the moment of exposure. Filter aberration may present similar distortions and corrections for these anomalies may be done similarly.

Image motion: Image motion is the smearing or blurring of imagery on an aerial photograph due to the relative movement of the camera (i.e., aircraft) with respect to the ground during the exposure. It is influenced mainly by the film speed and the shutter speed. Because the airplane, and hence the camera, is in motion during the time the shutter is open, each discrete point in the photograph will be imaged as a line. The faster the aircraft, the longer will the line of the imaged point be and the blurrier the photograph. Similarly, the slower the shutter speed (i.e., longer exposure) the longer the line of the points imaged. Image motion (M) can be quantified using either of the following formulas:

Where: M is the image motion (movement) on the photograph (in millimeters in equation (6.3) and in inches in equation (6.4)), 0.2778 is a constant, with units: meter hours per kilometer second (in equation (6.3)) and 17.6 is a constant, with units: inch hours per mile second (in equation (6.4)), is the ground speed of the plane in kilometers per hour in equation (6.3) and in miles per hour in equation (6.4), t is the shutter speed in seconds, f is the focal length of the camera lens (in mm in equation (7.3) and in feet in equation (6.4)), HD is the flying height of the aircraft above the datum (in meters in equation (6.3) and in feet in equation (6.4)). In both equations the term f/HD corresponds to the photo scale. Therefore, the equations above may be rewritten as:

Where: PSR is the photo scale reciprocal (1/photo scale or 1/(f/HD)). Example 6.2 Suppose that an airplane was flying 3000 meters (9840 feet) above the ground at 500 kilometers (about 310 miles) per hour . Suppose also that the camera was taking photographs with a 305-mm (12-in.) focal length camera lens and a shutter speed of 1/40th of a second (0.025 s). What would be the image motion?

Solution:

For good interpretation of aerial photographs, the acceptable limit for image motion is about 0.05 mm (0.002 in) and for precise mapping, this limit is half as much. Therefore, image motion of 0.353 mm (0.014 in) would result in a definite blur on the photograph. By examining the image motion formula (equation 6.3 or 6.4), we may understand that there are many ways of reducing image movement or blur. One method would be to use a faster shutter speed (t), which would require either a large aperture or a faster film. Another solution would be to use a slower-flying aircraft (v). A third solution would be to fly at higher altitude (HD) above the ground. Another option would be to uses shorter focal length (f). Image-motion problem sometimes makes it difficult to obtain satisfactory large-scale photography, especially when using fast aircraft and older color films that have relatively slow film speeds. Example 6.3 Suppose this time that an aircraft was flying at a speed of 450 km (about 280 miles) per hour and taking photographs with a shutter speed of 1/125th of a second using a focal length of 152.4 mm (6-in.). What should be the flying altitude of the aircraft above the ground to assure the acceptable image motion of 0.05 mm (0.002 in) on the photographs? Solution: Using equation 7.3 (or 7.4) again and rearranging the terms, we obtain:

Notice that the two results are slightly different due simply to data conversion between the English and the metric systems. In fact, 450 km = 279.6768 miles (not 280) and 0.05 mm = 0.0019685 in (not 0.002 in). If we use these two values (279.6768 miles and 0.0019685 in instead of 180 miles and 0.02 in ), we will find the exact number (about 10,000 ft) as in the metric equation. This is another indication that your measurements

and your input data (aircraft speed, shutter speed, and focal length) must be as accurate as possible to obtain reliable and satisfactory photographs. 6.6 Displacement Due to Topography

Figure 6.15. Effect of camera focal length on relief displacement. (From Ulliman, 1995). Example 6.6 Two towers were identified on a perfectly vertical photograph taken from 2500 m above the datum. The distances from the base of the towers to the photo center are equal and are measured to be 8.35 cm. If the height of tower1 is 120 m and that of tower2 is 85 m above the datum, find the relief displacement of the summit of these towers on the photograph? Conclude. Solution: First of all, because the photograph is perfectly vertical, the photo center is the nadir point using equation 6.12, we obtain:

for tower 1:

for tower 2: Conclusion: Relief displacement varies directly as the height of the object. Because tower1 is higher than tower2, its image is displaced more.

Example 6.7 In example 6.6, suppose that tower1 and tower2 have the same height of 100 m above the datum but the distances from their summit to the center of the photograph were found to be 6.55 cm and 9.21 cm respectively for tower1 and 2. Find the relief displacement for both towers and conclude. Solution: Using equation 6.11, we obtain:

for tower1:

for tower2: Conclusion: Relief displacement varies directly as the distance of the object from the nadir Because tower2 is farther from the nadir than tower1, its image is displaced more. In short, we may conclude that the amount of relief displacement is directly correlated with the height or depth of the object and the distance of the object from the nadir. This displacement is inversely correlated with the flying altitude of the aircraft above the datum and the focal length used. Higher altitude photography will produce less relief displacement than lower altitude photography. In addition, relief displacement is radial and either away from (for objects extending above the datum) or toward (for objects extending below the datum) the nadir. Even though relief displacement constitutes a source of errors in measuring horizontal distances on vertical aerial photographs, it is not necessarily a nuisance; because of relief displacement, we can determine the height of objects (or difference in elevation between objects) and to see in three-dimension by viewing stereoscopic pairs of aerial vertical photographs. Relief displacement on aerial vertical photographs also allows us to make topographic maps.

6.6.3. Measuring height of displaced objects on a single vertical photograph

The displacement formula (equation 6.11 or 6.12) can also be used to determine the height of an object from a single vertical photograph if the amount of displacement (d) can be accurately measured on the photograph. Height determination is more accurate for tall features imaged near the edge of the photograph. This is because relief displacement at this location is more exaggerated and separation between the top and the base of features is clearly visible on the photograph, thus, the distance, d, between the top and the base is more precisely measured. If we know the flying altitude (HD) of the aircraft above the datum, if the bottom and the top of the object are clearly visible to precisely measure the distance d, then object heights (ho) may easily be determined from equation 6.11 (or

6.12) as follows:

from

or from Where: ho is the height of the object under consideration, d is the length of the displaced object measured on the photograph, r is the radial distance from the nadir to the top of the displaced object measured on the photograph, r' is the radial distance from the nadir to the base of the displaced object measured on the photograph, and HD is the flying altitude of the aircraft above the base of the object (i.e., the datum). Example 6.8 A tower was identified on a "perfectly" vertical photograph and the distance between its top and its bottom was measured to be 14.3 mm and that from the photo center to the top of the displaced tower was measured to be 85.6 mm. If the flying height of the aircraft is 1500 m above the mean sea level (MSL) and the base of the building is 400 m above MSL, how tall is the building? Solution: Substituting in equation (6.13), we obtain:

Example 6.9 In figure 6.12, assume that the relief displacement for the summit of the tower is 5.3 mm (measured from the bottom, b, to the summit, s, of the tower on the photograph) and the radial distance measured from the photo center (assuming a true vertical photograph) to the base (b) of the tower is 59 mm. If the scale of the photograph is 1:10,000, as printed on the photograph, and the focal length used to take this photograph is 152.4 mm, how tall is the tower? Solution: From equation 6.14, we notice that in order to solve for ho, we first need to determine the flying height (HD) of the aircraft above the datum. The formula for the photo scale is f/HD. Therefore, 1:10,000 = f/HD, thus: HD = 10,000 x f = 10,000 x 152.4 mm = 1524 m. Finally, substituting in equation 6.14, we obtain:

Notice that d, r, and r' are expressed in the same units (usually in mm, cm, or inches) and HD is expressed in the units desired for the height, ho, of the object (usually in meters or feet). The accuracy of object heights determined by this technique is highly dependent upon the verticality of the photographs, precise values for flying heights, and very careful measurement techniques of d and r (or r'). Therefore, if the measurement instruments and/or techniques used to determine d and r (or r')are not accurate or if the flying altitude of the aircraft above the datum (HD) is not precisely known, then resulting object heights (ho) must be regarded as approximations only. In example 6.9, for example, if the displacement of the tower were measured 5.5 mm instead of 5.3 mm (an error of 2 tenth of a mm), then the height of the tower would have been 130.36 m instead of 125.62 m (an error of about 5 m ). Hence, small measurement errors on the photograph may produce large errors in determining ground object heights. This situation is even more accentuated as the scale of the photograph becomes smaller or the flying altitude of the aircraft becomes higher.

6.6.4 Determining horizontal ground distances, directions, and angles from photocoordinates on an aerial photograph taken over varied terrain
When a photograph is taken vertically over flat terrain, it will accurately represent the geometry of the imaged terrain and the scale will be constant throughout the entire photograph. In this case, the photograph may be used as a map and distances measured on the photograph may be corrected to ground distances by simply multiplying the measured values by the photo scale reciprocal (i.e., 1 divided by the photo scale). Similarly, angles may be directly measured on the photograph as accurately as they would be on a map. However, over a terrain of varied relief, photo scale is not constant throughout the photograph and ground distances and directions are not determined as in they would be on flat terrain. Figure 6.16 is an illustration of a vertical photograph taken over varied terrain. In figure 6.16a, if all ground points were at the datum plane (i.e., the plane containing points A', C, B', and O), points A and B would be located at A' and B' and their images would be at points a' and b' on the photograph. In this case, there would be no relief displacement. However, due to the varied relief, the image of point A is shifted radially outward from the nadir (assumed to coincide with point o) to point a on the photograph because point A extends above the datum and the image of point B is shifted radially inward toward the nadir to point b because B extends bellow the datum. This displacement of images for points A and B is the relief displacement effect. If points A and B were located at points A' and B' (i.e., the same elevation as the datum), then, their images would have been at points a' and b' on the photograph and the segment a'b' would accurately represent the scaled horizontal distance and orientation of the ground line AB. Because A and B are at different elevations than the datum, their images at a and b on the photograph are displaced from their true positions and the resulting line ab has a considerable altered distance and orientation. Relief displacement does not only alter distances and orientations, it also distorts angles. In figure 6.16b, the accurate (true) representation of the horizontal ground angle ACB is the angle a'cb' on the photograph. Therefore, the imaged (observed) angle acb on the photograph is distorted due to relief displacements. Because relief displacement is radial

from the nadir (assumed to coincide with the photo center when tilt is less than 3o), angles that have their summit at the photo center (o) are not distorted. For example, angles aob, aoc, and boc on the photograph accurately represent their corresponding angles AOB, AOC, and BOC on the ground.

Figure 6.16. Relief displacement on an aerial vertical photograph taken over varied terrain. (a) Displacement of terrain points. (b) distortion of horizontal angles. (Adapted from Lillesand and Kiefer, 1987) To determine distances and angles of points that are at different elevations on the ground, we must first convert the photocoordinates of points on the photograph to a datum-level ground coordinate system. From similar triangles LOAA' and Loa' in figure 6.17, we obtain:

Figure 6.17. Determination of ground coordinates from measured photocoordinates. (From Lillesand and Kiefer, 1987) And from similar triangles LA'A and La'a, we obtain:

where: HD is the flying altitude of the aircraft above the datum, hA is the ground elevation of point A above the datum, xa is the photocoordinate of point A measured on the photograph along the x-axis, ya is the photocoordinate of point A measured on the photograph along the y-axis, XA is the ground coordinate of point A along X-axis, and YA is the ground coordinate of point A along Y-axis. Equations 6.15 and 6.16 suggest that ground coordinates of any point on the photograph are obtained by simply multiplying the photocoordinates of that point (measured on the photograph) by the photo scale reciprocal (i.e., (HD - hA)/f, because photo scale is f/(HD hA)). Therefore, ground coordinates for point B can be expressed as:

Similarly, the ground coordinates for any point on the photograph can be determined as:

Note from these equations that this procedure of obtaining ground coordinates from photocoordinates requires knowledge of the terrain elevation (h) of these points. Axes (X, Y) define an arbitrary ground coordinate system that is directly beneath the photographic x and y axes and at the same elevation as the datum plane. Once the ground coordinates of points A and B are determined, the horizontal distance between A and B may be mathematically computed using the Pythagorean theorem. This theorem stipulates that the hypotenuse of a right triangle is equal to the square root of the sum of the squares of the other two sides of the triangle. In figure 6.17, the hypotenuse is represented by the line AB and the other two sides are defined by (XA - XB) and (YA - YB) as shown in figure 6.18. It follows that:

Figure 6.18. Determination of distances and directions from ground coordinates. (From Lillesand and Kiefer, 1987). Example 6.10 In figure 6.17, assume that the aircraft was flying at an altitude of 1500 m above the datum taking vertical photographs with a 305 mm focal length camera. Two image points a and b (of A and B on the ground) were depicted on a photograph and their photocoordinates were measured to be xa = 65.3 mm, ya = 71.5 mm, xb = -29.5 mm, and yb = -52.3 mm. If point A is 163 m and point B is 198 m above the datum, what is the horizontal distance on the ground between A and B?

Solution: Using equations 6.15 through 6.21, we obtain:

Note that XB and YB, in this case, have negative values and, therefore, -XB and -YB are positive values. This is because, as mentioned earlier, points that are below the X-axis have negative Y coordinates and those above the X-axis have positive Y coordinates. Similarly, points that are to the left of Y-axis have negative X coordinates and those to the right have positive X coordinates. The ground coordinates can also be used to determine the direction of the ground line passing through points A and B. If is the angular direction of line AB measured clockwise from
the +Y- axis (figure 6.18), it can be computed trigonometrically from the ground coordinates as:

Example 6.11 Using the ground coordinates computed for A and B in Example 7.10, find the direction from B to A with respect to the +Y ground axis. Solution:

Direction from A to B is computed by simply subtracting from 180o.

6.7. Displacement due To Tilt Many basic photogrammetric procedures involve the use of vertical aerial photographs. An aerial photograph is considered truly vertical when the optical axis of the camera is in perfect alignment with the plumb line. Because of unavoidable angular tilts of the aircraft during the exposure, virtually all aerial photographs are tilted to some degrees for the perfect aerial camera stabilizer has yet to be developed. A photograph is considered tilted when the angle between the perpendicular projection through the center of the lens and the plumb line is greater than 3o (figure 6.19). Tilt is the effect of lateral tip of the wings of the aircraft or the dipping of the aircraft fore and aft. Some use these two effects separately as tip, in the first case, and tilt, in the second case, whilst others refer to both of them simply as tilt, combined tilt, or resultant tilt. The tilt of the axis of the camera at the moment of exposure causes marked displacement of images on the photograph.

Figure 6.19. diagram illustrating the angle of tilt as defined by the optical axis of a tilted camera and the plumb line (i.e., vertical line).

6.7.1. Effect of tilt

An aircraft or airborne not perfectly horizontal causes a rotation of the camera about the x-axis or about the y-axis. Rotation about the x-axis causes lateral tilt or y-tilt, which is due to the aircraft being wing-up-wing-down and displacing the nadir point along Y-axis. Rotation about y-axis causes longitudinal tilt or x-tilt or list, which is due to the nose of the aircraft being up or down causing the nadir point to be displaced along X-axis. The two types of tilt radiate from a point on the photograph referred to as the isocenter (defined earlier in this chapter) and cause the objects located on the up side of the tilted photograph to be displaced radially toward the isocenter and those on the down side to be displaced radially outward or away from the isocenter (figure 6.20). Along the axis of tilt

(see figures 6.6 through 6.8), there is no displacement relative to an equivalent untilled photograph as this is the line where a tilted photograph and an equivalent vertical photograph would match and intersect on another. Notice also that the nadir is always on the down side of the axis of tilt and opposite the principal point from the isocenter.

Figure 6.20. Image topographic displacement relative to the center of the photograph; images on the up side of the tilted photo are displaced toward the center of the photo and those on the down side are displaced away from it. (After Ulliman, 1995). The effect of tilt can be most simply demonstrated by assuming that the surface photographed is perfectly plane; in other words, there is no relief displacement. From figure 6.21, point A on the ground is imaged on the down side of the tilted photograph at point a instead of a' had the photograph been free of tilt. This figure also shows that point a is registered farther from the isocenter (and the center) of the photograph (p) than point a'. Similarly, ground point B is imaged at point b on the up side of the tilted photograph, which is closer to the isocenter (and the center) than where it would have registered (b') had the photograph been free of tilt.

Figure 6.21. Image topographic displacement on a single tilted photograph. Figure 6.22 is another illustration of tilt displacement. T1 and T2 are two identical trees of the same height, h, and at the same distance, D, from the nadir (N) on the ground. The displaced images of these two trees are represented by t1 and t2 on the tilted photograph. Because the image of T1 (t1 on the photograph) is on the down side of the tilt (on the nadir side), it is displaced radially and outward from the isocenter (i) on the photograph. However, the image of T2 (t2 on the photograph) is displaced radially inward toward the isocenter because it is on the up side of the tilted photograph. Notice also that the size of image t2 is more compressed and closer to the isocenter than t1. Because features on the up side of a tilted photograph are imaged closer to the photo center (i.e., isocenter) and those on the down side are imaged farther from the photo center, the average scale of the photograph is altered. This effect also causes the overlap between adjacent photos (i.e., endlap) and between adjacent lines (i.e., sidelap) to be decreased on the down side of the tilt and to be increased on the up side of the tilt. Furthermore, while displacements due to tilt are usually less in magnitude than those produced by relief, they are far more difficult to detect, to calculate, and to correct. This type of displacement may cause large errors in determining scale and distances on aerial photographs. For these reasons, the safest ways to ovoid large errors due to tilt in the calculation of nominal scale values is to use two known or measurable ground distances that are about the same elevation, equal distances from the photo center, and diametrically opposite from the center. In this way, errors due to tilt (which may be present but not apparent) tend to be somewhat compensating.

Figure 6.22. Diagram illustrating the effect of tilt on the radial image relief displacement.

6.7.2. Determining directional errors on tilted photographs

Oftentimes, the interpreter rarely knows the angle and direction of the tilt, which makes the process of depicting the precise location of the isocenter hard and tedious. Furthermore, the presence of slight amounts of tilt is often ignored or goes undetected. However, because, in most applications, only the central portions of photographs are used for interpretation, angular tilt less than 3o can usually be ignored without serious consequences. In such cases, photographs are considered truly vertical and the three centers (i.e., the principal point, the nadir, and the isocenter) are assumed to coincide at the principal point, which is easily located. In nearly vertical photograph, the extent of tilt displacement is not usually great. With a 209-mm (8-inch) focal length camera, a 2-degree tilt will produce a displacement of about 2.5 mm (about 0.1 inch) at the edge of the photograph and about 0.76 mm (0.03 inch) half way to the center. At the edge of a 23-cm by 23-cm (9-in by 9-in) photograph, a 209-mm (8-inch) focal length may introduce an error of 1% in scale for each 1o tilt and with a 152.4-mm (6-in) focal length, this error is increased by 25% for each 1o of tilt. Therefore, the determination of tilt in aerial photographs is very important for precision mapping from aerial photographs. The angle of tilt ( ) can be determined from the right triangle nLp in figure 6.21 as:

where is the angle of tilt in degrees, d is the distance between the nadir and the principal point, measured on the photograph in mm, and f is the focal length in mm (note that d and f must be in the same units -mm, cm, or inches). If the nadir can precisely be determined on a tilted photograph, it is possible to compute the angle of tilt using

equation 6.23. Similarly, if the angle of tilt (or its complement, 90o - , the depression angle) is known, it is possible to determine the distance, d, between the principal point and the nadir from equation 6.23 as: Example 6.12 On a tilted photograph, the nadir point was determined by intersecting lines passing through perfectly tall and clearly visible features on the photograph. The distance between the nadir and the principal point was measured to be 4.5 mm. What was the angle of tilt the camera at the time of exposure if a 209-mm focal length lens was used?

Solution: From equation 6.23: Example 6.13 Assume that the optical axis of a camera was 3o from the vertical when taking a photograph with a 152.4-mm focal length. What will the distance between the nadir and photo center be on the photograph. What is the corresponding distance on the ground if the aircraft were flying at 3500 m above the datum?

Solution: From equation 6.24:

From figure 6.21, it can be seen that: where LN is the flying altitude (HD) of the aircraft above the datum and ON is the ground distance (D) between the nadir and the principal point.

Therefore, correction for tilt effect can easily be done if it can be observed on the photograph. In figure 6.21, the distance between a and b is approximately equal to the distance between a' and b' when the tilt is small (less than 3o). Mathematically, the amount of tilt varies approximately as the square of the distance of the image from the isocenter (i) of the photograph. Correction for tilt can also be done by rectifying tilted negatives, if the amount of tilt is known, by recreating the relative tilt between the negative and the printing paper. This technique, called rectification, is used by photogrammetrists in map making and other processes requiring accurate outputs. This technique involves expensive vertical and horizontal ground-control measurements.

The scale of a tilted photograph is altered increasingly from the isocenter to the edge of the photograph. For example, in figure 6.22, if the image (t3) of a third tree, T3, identical to T1 and T2, is located at the edge of the photograph, its image displacement (t3) will not be the same as for T2; it will be greater because tilt effect is more exaggerated towards the edge of the photo. These changes in scale occur in a regular manner throughout the photograph, but there is no change along a line perpendicular to the direction of tilt, such as lines p'p" and i'i" in figure 6.23.

Figure 6.23. Geometry of a tilted photograph. (From Howard, 1970). When the tilt is slight in an aerial photograph, the scale (S) of any image on that photograph can be calculated as:

where: y is the distance of the image from the isocenter (or from the principal point when the tilt is less than 3o), H is the flying altitude of the aircraft above the mean sea level (MSL), h is the elevation above MSL of the point at which the scale is being calculated, f is the focal length of the camera used when taking the photo, is the angle of tilt (the angle between the optical axis and the plumb line). When the tilt is significant ( > 3o), equation 6.25 is not applicable; because the geometry would be different, equation 6.25 would be modified (see RSCC Volume 1, Lecture 7). Example 6.14 A 209-mm (8-in.) focal length camera was used on an aircraft flying at 8,000 m (26240 ft) above the mean sea level (MSL) to take a photograph of an object 1,500 m (4920 ft)

above MSL. The tilt on the photograph was found to be 3o. What is the photographic scale at: a) the isocenter (assumed to coincide with the principal point) of the photograph, b) 15 cm (5.9 in.) from the principal point on the upper side of tilt, and c) 15 cm (5.9 in.) from the principal point on the lower side of tilt. Solution: From equation 6.25, we obtain:

6.7.3. Determining ground coordinates from a tilted photograph photocoordinates

When the angle and the direction (swing) of tilt are known, the ground coordinates of a point on a tilted photograph may be computed. The swing angle (s) is the angle measured in the plane of the photograph from the positive photographic y-axis clockwise to the nadir point side of the principal line as shown in figure 7.24b. If we arbitrary select a coordinate system such that Y-coordinate axis lies in the principal plane and connecting the principal point, the nadir, and the isocenter on the ground and the X-coordinate axis lies perpendicularly to Y-axis and passing through the ground nadir point, as illustrated by figure 6.24, then, N is the ground coordinate origin for (X,Y) ground coordinate system.

Figure 6.24. Determination of ground coordinates from a tilted photograph photocoordinates. (Adapted from Moffitt and Mikhail, 1980). A point A at a ground elevation hA above the MSL will be imaged at point a on the tilted photograph and its photocoordinates x and y can be measured with respect to the coordinate system defined by the fiducial marks. For the (x', y') coordinate system to coincide with the (x, y) system on the photograph, it needs to be rotated and translated. The rotation is defined by the angle as shown in figure 6.24b and the translation is defined by the distance no, between the nadir and the principal point. By rotating the (x', y') system, the photocoordinates become: x' = x cos + y sin y' = -x sin + y cos (6.26)

At this point, y' coincides with y but x' does not coincide with x; axis nx' is parallel to axis ox but separated by a distance equal to no. For n to coincide with o, the (x', y') system needs to be translated by the distance no = f tan . After translation, equation 6.26 becomes: x' = x cos + y sin (6.27) y' = -x sin + y cos + f tan Since Ln is a vertical line, the perpendicular line mw is a horizontal line. And since aw is perpendicular to the principal line, nw (y'), it is also horizontal. Therefore, the plane containing a, m, and w is a horizontal plane. Thus, from similar triangles Lma and LNA we obtain:

Because Lm = Ln - nm = f sec - y' sin , then:

We know that the ground coordinates of any point on the photograph at an altitude h above the MSL is equal to the coordinates of that point on the photograph multiplied by the photo scale reciprocal. In figure 6.24a, photocoordinate along x-axis for point a is the photo distance wa = x' and the corresponding ground distance is WA = XA. Therefore:

Similarly,

Example 6.15 An aircraft was flying 3500 m above mean sea level with a swing angle of 250o. A 152.4mm focal length camera tilted at 2.9o from the vertical was used to take photographs. Two points a and b were identified on a photograph and their ground elevations were determined from a topographic map as hA = 100 m and hB = 300 m. From the photocoordinate system defined by the lines connecting opposite fiducial marks on the photograph, the photocoordinates of the two points were measured to be: xa = 7.15 cm, ya = 2.10 cm, xb = 6.53 cm, and yb = 3.94 cm. Find the coordinates of the corresponding points A and B on the ground and the length of line AB. Solution:

From equation 6.27, x'a = xa cos + ya sin y'a = -xa sin + ya cos + f tan and x'b = xb cos + yb sin y'b = -xb sin + yb cos + f tan From figure 6.24b, we can see that = 180o - s = -70o cos = 0.34202 sin = 0.05059 tan = 0.05066 sin = -0.93969 cos = 0.99872 sec = 1.00128 Then, x'a = (7.15)(0.34202) + (-2.10)(-0.93969) = 4.42 cm y'a = (-7.15)(-0.93969) + (-2.10)(0.34202) + (15.24)(0.05066) = 6.77 cm, and x'b = (-6.53)(0.34202) + (-3.94)(-0.93969) = 1.47 cm y'b = (-6.53)(-0.93969) + (-3.94)(0.34202) + (15.24)(0.05066) = 5.56 cm Using equations 7.28 nd 7.29, we obtain:

Finally,

6.8. Combined effects of Relief and Tilt At this point, we can clearly say that most aerial photographs are normally tilted and displaced radially both from the photograph nadir as a result of relief and from the isocenter as a result of tilt. Since these displacements are radial from two different points, their combined effect in a single photograph will result in lateral as well as radial displacement of images. In other words, a radial from the principal point through the image of an object on an aerial photograph will not pass through the true map position of that object but beside it. This lateral displacement may be significant when photographs of moderate tilt are taken over rough terrain or when photographs are taken over moderate relief but with large angle tilt. Figure 6.25 illustrate the combined effect of relief and tilt. Point P on the graph is the true image position of an object at a lower ground elevation than the nadir. If the photograph

were free of tilt, the object image would be displaced radially inward toward the nadir at point R because its elevation is lower than that of the nadir. However, when there is tilt displacement (which is usually the case), the object image will be displaced radially away from the isocenter because, in this case, it is located on the down side of the photo positive. The combined effects of relief and tilt will then place the object image at point (R+T). When the object is higher in elevation than the nadir, the geometry of the combined effects of relief and tilt different.

Figure 6.25. Illustration of lateral displacement caused by the combined effects of relief (R) and tilt (T). (From Paine, 1981). The amount of lateral displacement for various combinations of relief and tilt at a mapping scale of 1:15,840 is shown in figure 6.26. This graph shows that a 305-m (1000 ft) difference in elevation combined with 2 degrees of tilt will produce a lateral displacement of about 0.7 mm (more than 0.026 inch), which, at 1:15,840 scale, represents 11 m on the ground, an amount large enough to produce significant errors on a map. The mathematical aspects of the combined effects of relief and tilt in image displacement were elaborated by Wood (1950) and Faulds (1959). Because relief and tilt displacements radiate from the nadir and the isocenter respectively, the principal point has no geometric significance as a radial center. In fact, when an aerial photograph is affected by both relief and tilt displacements, there is no single point on the photograph that can correctly be used as the radial center. Therefore, in order to evaluate relief and tilt displacements, the relationships between directions from the three photo centers (principal point, nadir, and isocenter) and corresponding directions on the ground are elaborated under varying conditions.

Figure 6.26. Amount of lateral displacement for various combinations of relief and tilt at a mapping scale of 1:15,840. (from Spurr, 1960).

6.8.1. Evaluating relief and tilt distortions about the nadir point
Let us consider a tilted photograph taken over varied terrain (figure 6.27), on which the principal point (o) and the nadir (n) were located. If a is the image on the tilted photograph of a ground point A, the direction of the radial line na is defined by the angle anw (designated as ), measured clockwise in the plane of the photograph between the principal line (nw) and the radial line from n through a. Since line wa is perpendicular to the principal line, it is therefore horizontal. The projection of the nadir point (n) to the horizontal plane lies at point k and the triangle akw is therefore horizontal. The angle nwk is equal to the tilt angle ( ) since the triangle knw is a right triangle in the principal plane. As we have seen in section 6.6.4., irrespective of the ground altitude of the corner points, angles that have their summit at the nadir point are not distorted. therefore, the angle ANW is equal to the projected angle of these points on the horizontal plane. That is, the angle ANW = A'N'W' = in figure 6.27. And from right triangle akw,

In triangle wnk,

Figure 6.27. Relationship between directions radiating from the nadir of a tilted photograph and their corresponding directions on the ground. (adapted from Faulds, 1959). Likewise, in right triangle anw, in the plane of the tilted photograph,

Substituting for aw and nw (equations 6.30 and 6.31), we obtain:

6.8.2. Evaluating relief and tilt distortions about the isocenter

Figure 6.28 illustrates the same geometry as in figure 6.27, except that, in this case, the distorted angle, , is evaluated about the isocenter (i) instead of the nadir (n) and its corresponding angle, , on the ground has its summit at the projection of the isocenter on the ground (I'). It should be noted that, unlike the situation in section 6.8.1, the angle AIW, in the horizontal plane, and the angle A'I'W', in the ground plane, are not equal because their summits are not at the nadir but at the isocenter.

Figure 6.28. Relationship between directions radiating from the isocenter of a tilted photograph and their corresponding directions on the ground. (adapted from Faulds, 1959). Let us first consider the situation where points A, I, and W are in the horizontal plane. In this case, right triangles alw and AIW are similar. It follows that alw = AIW = and tan = aw/lw. From triangles alw and aiw we have

Because triangle ilw is isosceles, iw = lw and

Now, because of varied terrain, the projection of the isocenter is at a position I' (at an elevation of h below the ground point A) instead of that on a horizontal plane, I. Therefore, the true angle on the ground is A'I'W' (designated as ') and the direction of line ia measured clockwise from the principal line (iw) is a function of this ground angle, '. This direction is defined by the angle and it is equal to the horizontal plane direction as shown above. Therefore, the relationship between angle and angle

is the same as that between and '.

In right triangle alw:

and in right triangle AIW: where Y = h tan( /2) (6.34) From right triangle A'I'W', X = sin ' and Y = cos ' Therefore,

6.8.3. Evaluating relief and tilt distortions about the principal point
As in figures 6.27 and 6.28, figure 6.29 illustrates the same geometry but about the principal point. In this case, angle and its corresponding ground angle have their summits at the photographic center (o) and its corresponding point (O) on the ground, respectively. Like in the previous cases, let us first assume that the ground surface is flat. Under, this condition, the ground angle is defined between the principal line OW and the radial line through A, OA, and its image on the tilted photograph is the angle , defined between the principal line nw and the radial line through a, oa. The effect of tilt can be evaluated by determining the relationship between angle on the tilted photograph and its corresponding true angle, , on the ground. Because aw is perpendicular to the principal line ow and m is the horizontal projection of the principal point, o, along the optical axis to the horizontal plane containing line aw, the right triangle awm is horizontal and similar to triangle AWO on the ground. Therefore, AOW = amw = and

Also, from right triangle wom,

and from right triangle aow in the plane of the tilted photograph,

Figure 6.29. Relationship between directions radiating from the principal point on a tilted photograph and their corresponding directions on the ground. (adapted from Faulds, 1959). Substituting for aw and ow in equation (6.38), we obtain Equation 6.39 represents the relationship between the distorted angle on the tilted photograph and its corresponding angle, , on a flat ground surface. But, in most cases, ground surface is not flat and we need to establish the relationship between the distorted angle on the tilted photograph and its true corresponding angle, ', on a non-flat surface, represented by the true horizontal angle A'O'W' on the ground. On a flat ground surface,

and from right angle A'O'W' X = sin ' and Y = cos

Therefore, Substituting equation 6.37 in equation 6.36, we obtain

To be able to understand the combined effects of relief and tilt, we need to clearly understand the effect of each of these phenomena separately. In other words we need to understand that because of relief effect, objects that are higher in elevation than the nadir are displaced radially outward away from it and those lower than it are displaced inward toward it. Similarly, because of tilt effect, objects on the up side of the tilted photo are displaced radially toward the isocenter and those on the down side are displaced radially outward from the isocenter. Depending on the elevation and the position of the objects on the photograph and in respect to the tilt axis, we get different results. This suggests that the combined effects of relief and tilt are sometimes accumulative and sometimes compensating. In any situation, there is no tilt displacement along the axis of tilt, the only displacement is from topographic effect. Also, there is no lateral shift at right angles to the axis of tilt on a line through the three photo centers. In conclusion, displacements due to combined effects of relief and tilt are variable and complicated. 6.9. Geometry of Oblique photographs Technically, a tilted photograph is a low oblique photograph. In fact, an oblique photograph may be considered as a "vertical" photograph with deliberate tilt of known angle between 3o and 90o. It follows that the geometry of an oblique photograph is similar to that of a vertical photograph and the same principles apply. In an oblique photograph, however, because the tilt is substantial, the combined effects of relief and tilt are intensified and the relationship among the three photo centers (i.e., principal point, nadir, and isocenter) is more apparent.

6.9.1. Geometry of high oblique photographs

Figure 6.30 illustrates the principal plane section of a high oblique photograph and the resulting print. The geometry of a high oblique photograph requires the knowledge of three principal parameters: the focal length of the lens (f), the angle of tilt ( ) or its complement, the depression angle ( ), and the flying altitude of the aircraft above the datum (HD). The angle of tilt, , is the angle between the perpendicular projection through the center of the lens (LO) and the plumb line (LN) as shown in figure 6.30a. The complement angle (90o - ) is the depression angle ( ) of the photograph, which is the angle between the optical axis (LO) and the true horizon (Lz). The focal length, f, is the distance between the center of the lens (L), where the photograph was taken from, and the principal point (o) of the photograph. The flying height of the aircraft (HD) is the vertical altitude above the datum, which is equal to the distance LN in figure 6.30a.

Figure 6.30. Geometry of an aerial high oblique photograph; (a) the principal plane section and (b) the resulting image. Other pertinent parameters to the geometry of a high oblique photograph are the true horizon, the apparent horizon, the apparent depression angle, and the dip angle, and shown in figure 6.31. The true horizon line is the intersection of the horizonal plane containing the camera lens and the plane of the oblique photograph. It is an imaginary line and, therefore, does not appear on the photograph. The apparent horizon line is the actual line where the earth surface appears to meet with the sky. This line is not imaginary, it is visible on the high oblique photograph and due to the curvature of the earth, its image will be slightly curved on a small scale high oblique photograph.

Figure 6.31. Diagram illustrating the true horizon, the apparent horizon, and the dip angle. The dip angle, d, is the angle measured in the principal plane between the true and the apparent horizons. The dip angle is produced by the height of the aircraft (i.e., camera) above the ground and it increases proportionally as the height of the aircraft. From figure 6.31, if HD is the flying height above the datum, R is the radius of the earth, LZ is the true horizon, and LZ' is the apparent horizon, then, d can be determined as:

where LZ' is determined by the Pythagorean theorem as: Substituting for LZ',

Since HD is small compared to R, this equation reduces to:

Because atmospheric refraction reduces the value of d, this latter needs to be multiplied by the constant 0.9216 to account for this effect. Further, because d is always small, tand d in radians. Converting radians to seconds and substituting the mean earth radius

(20.9 x 106 ft) into the equation, equation 6.42 is reduced to: if HD is expressed in feet, or if HD is expressed in meters. As seen from equation 6.43, the dip angle, expressed in minutes, is approximately equal to the square root of the flying height above the ground in feet:

or, when HD is expressed in meters. Apparent depression angle is the angle, a, measured in the principal plane, between the optical axis of the camera and the apparent horizon (see figure 6.30a). Since the apparent horizon is visible, and usually traced, on a high oblique photograph, the distance oz' can be measured. Further, because oz' is perpendicular to Lo (figure 6.30), a can be calculated as:

Due to the curvature of the earth and the flying height of the aircraft above the ground, the apparent horizon falls below the true horizon. Therefore, the apparent depression angle should always be increased by the dip angle to obtain the true depression angle: Example 6.16 Assume a 152.4-mm focal length camera was used to take an oblique photograph at a flying height of 3000 m above datum. On the tilted photograph, the distance between the principal point and the apparent horizon was measured to be 9.5 cm. Find the true depression angle.

Solution: From equation 6.44,

and from equation 6.47, Finally, from equation 6.48, = a + d = 33.56 degrees, or 33o33'28"

When the true depression angle is determined, the correct position of the true horizon, in respect to the principal point, may then be calculated as follows: In figure 6.30a, where oz is the distance between the true horizon (z) and the principal point (o), f is the focal length, and is the depression angle of the photograph from the true horizon. At this junction, we should note that the oblique photograph considered in figure 6.30 was taken from an aircraft being nose-up-tail-down. In this case, only the longitudinal tilt (i.e., x-tilt) is assumed to be affecting the geometry of the oblique photograph. However, the lateral tilt, or y-tilt, (aircraft being wing-up-wing-down), as in the case of a vertical photograph, may also have significant effect on the geometry of an oblique photograph. When the y-tilt is significant, the geometry of an oblique photograph is different from that of figure 6.30. Since the y-tilt is usually very slight compared to the x-tilt, it is often insignificant, thus usually ignored. Because of large tilt angle on high oblique photographs, the nadir point (n) may lie outside the photograph limits at the intersection of the vertical (plumb line) from the camera lens (L) and the plane of the tilted photograph (i.e., the plane passing through o and i in figure 6.30a). Notice that the principal point (o), the nadir (n), and the isocenter (i) are very distinct. Had the photograph been free of tilt, the optical axis would have intersected the photograph plane at the isocenter (i), in which case, the three photo centers would coincide. Figure 6.30b portrays the diagrammatic representation of the photograph produced by the geometric settings in figure 6.30a. On this diagram, the principal meridian of the photograph transects the four centers of radial displacement: the nadir (n), the isocenter (i), the optical center or principal point (o), and the horizon point (z). When the oblique photograph is assumed free of y-tilt and either the tilt angle ( ) or the depression angle is known, the position of each of the four points (o, n, i, and z) on the principal meridian can be determined. Conversely, if either z, i, or n can be located with respect to o on the photograph, the tilt angle ( ) or it complement ( = 90o - ) may be computed readily from the same formulas . This is most often possible in the case of high oblique photographs where the position of the apparent horizon is pictured on the photograph. Once the position of the true horizon is located on the principal meridian (equation 6.49), the position of the other three points (o, n, and i) can readily be determined. Since the true horizon was determined in respect to the principal point, the same formula (equation 6.49) is used to locate the principal point in respect to the true horizon (z). Similarly, the positions of the nadir point (n) and the isocenter (i) from the principal point (o) may be computed as: where no is the distance between the nadir and the principal point, io is the distance between the isocenter and the principal point, and is the angle of tilt of the photograph. Example 6.17 A high oblique photograph was taken at a flying height of 3000 m above datum with a 152.4-mm focal length camera. The photo distance between the principal point and the

apparent horizon, measured along the principal line, is 7.50 cm. Find the true depression angle and the positions of the true horizon, the nadir, and the isocenter with respect to the principal point. Solution: The position of the true horizon with respect to the principal point is determined by the depression angle, which is the sum of the apparent depression angle and the dip angle. The apparent depression angle is obtained by equation 6.47: and the dip angle is obtained by equation 6.44: The true depression angle is then, 26o 12' 11" + 1o 37' 13" = 27o 49' 24" or 27.88 degrees Using equation 6.49, the position of the true horizon, with respect to the principal point is: oz = 152.4 mm tan 27.88 = 80.63 mm and the positions of the nadir and the isocenter are obtained from equations 6.50 and 6.51, respectively as: on = 152.4 mm tan (90o - 27.88) = 288.04 mm oi = 152.4 mm tan [(90o - 27.88)/2] = 91.78 mm Because of the exaggerated perspective distortions on an oblique photograph, figure 6.30b shows that equidistant lines drawn perpendicularly to the principal meridian appear to be imaged closer and closer as their distance from the nadir increases. Similarly, lines which are actually parallel to the principal meridian appear to converge at point z on the true horizon. Consequently, the size of equal squares of a grid constructed by these lines on the ground would appear smaller and smaller as they get farther from the nadir. The techniques of constructing, on the photograph, the equidistant lines and those converging at the true horizon point are described by Spurr (1960).

6.9.2. Geometry of low oblique photographs

As seen in the case of a high oblique photograph, the development of the geometry of an oblique photograph depends on the visibility of the apparent horizon on the photograph. The lack of a horizon on a low oblique photograph is the major difference between this type of photograph and a high oblique photograph. Because of the absence of a horizon on a low oblique photograph, the determination of the former becomes dependable on the type of features imaged on the photograph. Two possibilities may be used to locate the position of the horizon on a low oblique photograph. First method: If ground features of horizontal parallel lines (e.g., limits of parcels) are clearly visible on the photograph, such as in figure 6.32, the projection of these lines will intersect at their corresponding so-called vanishing points on the true horizon. From these vanishing points, the location of the true horizon point (z) can be constructed. Once the true horizon point is positioned, the same geometry developed for a high oblique photograph applies. The depression angle can be determined by measuring the distance, oz, between the principal point and the horizon point and applying equation 6.47.

Second method: This method depends on the presence of perfectly vertical tall objects (e.g., towers, trees, poles, buildings) on the oblique photograph (figure 6.33). If these objects are clearly visible, the vertical lines in perspective will converge at the nadir point. Once the nadir point is located, the distance, no, between the nadir and the principal point can be measured and the depression angle can be computed from equation 6.49 ( = 90o - ). 6.32. Determination of the true horizon from horizontal parallel lines.

6.33. Determination of nadir point by convergence of vertical lines. REFERENCES Abdel-Aziz, Y. I. 1975. Asymetrical Lens Distortion. Photogrammetric Engineering and Remote Sensing 41 (3):337-346. American Society of Photogrammetry. 1980. Manual of Photogrammetry, Fourth Edition, Falls Church, Virginia, Chapters 2 and 4. Avery, E. T. and L. G. Berlin. 1985. Interpretation of Aerial Photographs. Fourth Edition, Burgess Publishing Company, Minneapolis, Minnesota, Chapter 3. Bertram, S. 1966. Atmospheric Refraction. Photogrammetric Engineering 32 (1):76-84. Brock, G. C. 1952. Physical Aspects of Air Photography. Longmans, Green and Co. New York, N. Y., Chapter 8. Brown, D. C. 1966. Decentering Distortion of Lens. Photogrammetric Engineering 32 (3):444- 462.

Faulds, Arthur H. 1959. Some notes on the displacement of photographic images caused by tilt and relief. Photogrammetric Engineering 25 (1):110-115. Forrest, R. B. 1974. Refraction Compensation. Photogrammetric Engineering 40 (5):577582. Friedman, S. J. 1957. Distortion Tolerance Specification for Mapping-Camera Lenses. Photogrammetric Engineering 23 (1):39-48. Gay, Sylvester P., JR. 1957. Measurement of vertical heights from single oblique aerial photographs. Photogrammetric Engineering 23 (5):900-908. Gay, Sylvester P., JR. 1956. Nomographic solution to oblique photo mensuration. Photogrammetric Engineering 22:674-692. Hallert, B. 1963. The Method of Least Squares Applied to Multicollimator Camera Calibration. Photogrammetric Engineering 29:836-849. Hallert, B. 1956. Some Preliminary Results of the Determination of Radial Distortion in Aerial Pictures. Photogrammetric Engineering 22:169-173. Howard, J. A. 1970. Aerial photo-ecology. American Elsvier Publishing Company, Inc., New York. N. Y., Chapters 9 and 10. Johnson, Evert W. 1962. The Effect of Tilt on the Measurement of Spot-Heights Using Parallax Methods. Photogrammetric Engineering 28 (3):492-508. Johnson, Evert W. 1956. Photographic determination of pulpwood volume in Rickpiled storage yards. Alabama Polytechnic Institute, Bulletin 229, 51 pp. Katz, Amrom H. 1950. Contributions to the theory and mechanics of photo-interpretation from vertical and oblique photograph. Photogrammetric Engineering 16:339-386. Kawachi, D. A. 1966. Geometry of Vertical and Oblique Panoramic Photography. Photogrammetric Engineering 32 (2):298-306. Kawachi, D. A. 1965. Image Motion and its Compensation for the Oblique Frame Camera. Photogrammetric Engineering 31 (1):154-165. Kraus, K. 1972. Film Deformation Correction with Least Squares. Photogrammetric Engineering 38 (5):487-. Landis, G. H. and H. A. Meyer. 1954. The Accuracy of Scale Determination on Aerial Photographs. Journal of Forestry 52:863-. Lane, Benjamin B., JR. 1950. Scales of oblique photographs. Photogrammetric Engineering 16: 409-414. Legterink, G. H. 1972. Film-Glass Differences. Photogrammetric Engineering 38 (2):269-273. Lillesand, Thomas M. and R. W. Kiefer. 1987. Remote Sensing and Image Interpretation. Second Edition, John Wiley & Sons, New York, N.Y., Chapters 2 and 5. Moffitt, Francis H. and E. M. Mikhail. 1980 . Photogrammetry. Third Edition. Harper & Row, Publishers, New York, Chapters 4, 6, 7, and 10. Paine, D. P. 1981. Aerial Photography and Image Interpretation for Resource Management. John Wiley & Sons, New York, N.Y. Chapters 1, 2, 7, and 12.

Pryor, William T. 1959. Relationship of Topographic Relief, Flight Height, and Minimum and Maximum Overlap. Photogrammetric Engineering 25 (4):572-589. Raasveldt, Henri C. 1959. Determination of the Angle of Dip of Seemingly Vertical Strata on Vertical Aerial Photographs. Photogrammetric Engineering 25 (1):49-53. Scherz, J. P. 1974. Errors in Photogrammetry. Photogrammetric Engineering 40 (4):493500. Scholer, H. 1975. On Photogrammetric Distortion. Photogrammetric Engineering and Remote sensing 41(6):761-769. Spurr, S. H. 1960. Photogrammetry and Photo-interpretation. Second Edition. The Ronald Press Company, New York, Chapter 6. Tayman, W. P. 1974. Calibration of Lenses and Cameras at the U.S.G.S. Photogrammetric Engineering 40 (11):1331-1334. Trorey, Lyle G. 1950. Handbook of Aerial Mapping and Photogrammetry. Cambridge University Press, London. 178 pp. Ulliman, J. J. 1995. Aerial Photo Interpretation of Renewable Natural Resources: Course Notes and Workbook, 170 p. Department of Forest Resources, University of Idaho, Moscow, Idaho 83844. Underwood, P. H. 1947. The Determination of Tilt from Scale Check Lines. Photogrammetric Engineering 13 (1):143-. Van Roessel, J. 1970. Estimating Lens Distortion with Orthogonal Polynomials. Photogrammetric Engineering 36 (6):584-588. Wolf, Paul R. 1983. Elements of Photogrammetry with Air Photo Interpretation and Remote Sensing. Second Edition. McGraw-Hill Book Company. New York, N.Y., Chapters 5, 6, and 17. Wood, Edward S. JR. 1950. The uses of high altitude photography for mapping and reconnaissance: a discussion. Photogrammetric Engineering 16:613-618. Ziemann, H. 1971. Image Deformation and Methods for its Correction. Canadian Surveyor 25 (4):367-.

PRACTICE PROBLEMS
Part I: General
6.1. Define an aerial vertical photograph, an aerial tilted photograph, a low oblique aerial photograph, and a high oblique aerial photograph. 6.2 In question 6.1, state the advantages and disadvantages of each type of aerial photograph in respect to the others. 6.3. Define the following: Focal length Principal points Nadir point

Isocenter Flying height Datum Plumb line Optical axis Fiducial marks Ground elevation

6.4. Draw a diagram showing the geometric relationship among all the elements in question 6.3. 6.5. What are the major sources of distortions and displacements affecting aerial photographs? 6.6. Describe relief displacement and its effect on a vertical aerial photograph and state its benefits and drawbacks. 6.7. What is image motion and what causes it? 6.8. Suppose we measured the distances between the fiducial marks of a vertical photograph to be 23.37 cm along x-axis and 23.35 cm along y-axis. The calibrated measurement on the same camera used to take the photograph were found to be 23.15 cm and 23.17 cm, respectively along x-axis and y-axis. Find the correct values for the following measured photocoordinates of points a, b, c, and d: xa = -8.93 cm xb = -10.15 cm xc = 9.37 cm xd = 6.92 cm ya = -9.25 cm yb = 7.67 cm yc = 10.13 cm yd = -10.41 cm 6.9. Assume that we know the correct photocoordinates of the following points on a vertical photograph: xa = 83.74 cm xb = -76.35 cm xc = -24.83 cm xd = 50.52 cm ya = -90.92 cm yb = 86.80 cm yc = 5.09 cm yd = 54.27 cm The photocoordinates of these points were measured on the photograph as: xa = 84.08 cm xb = -76.67 cm xc = -24.96 cm xd = 50.74 cm ya = -91.47 cm yb = 87.33 cm yc = 5.19 cm yd = 54.62 cm If the distances between the fiducial marks of the photograph were measured to be 233.6 mm along x-axis and 233.4 mm along y-axis, what are the corresponding calibrated distances on the camera?

Part II: Relief and tilt displacement

6.10. Assume you want to take vertical aerial photographs over a mountainous terrain. Would you use a 152.4-mm or a 209-mm focal length camera? Why? 6.11. In question 7.8, which flying height would be more suitable to take better photographs, 2000 m or 3500 m above datum? Why? 6.12. Draw a diagram showing relief displacement of an electric pole extending above datum and a ditch extending below datum. On the photographic positive image, draw the displaced image of each feature and use an arrow to indicate the direction of displacement in respect to the photo center (assume a truly vertical photograph).

6.13. Select the correct word(s) in the brackets: a. Topographic displacement varies [directly, inversely] with the focal length of the camera, [directly, inversely] with the flying height of the aircraft above datum, [directly, inversely] with the height of the object, and [directly, inversely] with the distance of the object from the [principal point, nadir, isocenter]. b. Objects above datum are displaced radially [inward toward, outward away from] the [principal point, nadir, isocenter] and those below datum are displaced radially [inward toward, outward from] the [principal point, nadir, isocenter]. c. Tilt displacement is radial from [the principal point, nadir, isocenter]. 6.14. Describe tilt and its effect on vertical aerial photographs and state its causes. 6.15. Draw a 5 cm by 5 cm vertical photograph, place the four fiducial marks, and find the principal point of the photograph. Suppose an aircraft was flying in the East-West direction (East is to the right on your 5x5 cm photo) with the left wing being up and the right wing being down. Sketch the position of the nadir point and the isocenter with respect to the principal point, then draw the tilt axis and indicate the up and down sides of the tilt. 6.16. From problem 6.13, assume a tall tower was located on the up side of the tilted photograph. Sketch the actual position of the tower on the photograph as affected by the combined effect of relief and tilt and sketch the true position of this tower had the photograph been free of tilt and relief displacement (see figure 6.25 for help). 6.17. Assume you have located the principal point on an aerial vertical photograph. Assume also that the photograph was taken from an aircraft 3000 m above datum and using a 152.4-mm focal length camera tilted 2o from the vertical. If a 100-m tower is located 9.8 cm from the principal point, on the down side of the tilt, and on the line passing through the principal point, what is the image relief displacement on the photograph?

Solution:

Because relief displacement is radial from the nadir and because all measurements are from the nadir point, we first need to find the position of the nadir. We know that the nadir is always on the down side of tilt from the principal point as illustrated by the figure above. Since the tower is on the same line connecting the p and the nadir (n), the position of the nadir can be determined from the right triangle onp as: np = op.tan2o, where op is the flying height above the datum (3000 m) np = 3000 m (0.03492) = 104.76 m, which is the distance between the principal point and the nadir. From similar triangles on'p' and onp, we can write:

therefore, the distance n'p' on the photograph may be calculated as: n'p' = (f . np)/HD = (0.1524 x 104.76)/3000 = 5.32 mm It follows that the distance, r', from the nadir to the tower is then: r' = tp - np = 98 mm - 5.32 mm = 92.68 mm

and the displacement, d, of the tower can be computed using equation 7.12 as:

6.18. An aircraft was flying at 5000 m above datum taking aerial vertical photograph. If the optical axis of the lens was tilted 4o from the vertical, how far would the nadir be from the principal point on the ground and on the tilted photograph if the focal length of 152.4 mm was used. How far would the isocenter be from the principal point and the nadir? 6.19. A truly vertical photograph was taken at a flying height of 6000 m above datum. A 110-m- tall building was located on the photograph and the distance between its base and the principal point was measured to be 8 cm. Calculate the displacement of the building on the photograph and state the direction of this displacement in respect to the photo center. 6.20. In the previous problem, another tower was identified on the photograph and the distances between the photo center and the base and the top of the tower were measured to be 7.53 cm and 7.95 cm respectively. Find the height of the tower. 6.21. The distance between the top and the base of a 150-m-high tower was found to be 5.7 mm. If the distance from the nadir to the base of the tower is 8.8 cm and the ground elevation of the top of the tower is 500 m above mean sea level, how high was the aircraft flying above datum at the moment of exposure? 6.22. Find the displacement of a tree that is 100 m high if the photographic distance from its base to the nadir is 7.6 cm and if the photograph was taken vertically at a flying height of 6000 m above datum. 6.23. The distance, on a vertical photograph, from the nadir to the top of a hill is 9.1 cm and the hill is 200 m higher that datum on the ground. If the photograph was taken from an aircraft flying at 5500 m above datum at the time of exposure, by how much is the top of the hill displaced from its base? 6.24. A photograph was taken with a 305 mm focal length from an aircraft flying at 6500 m above datum. By how much would the summit of a 130-m-tall tower be displaced from its base if its base is at 2 kilometers from the nadir on the ground? 6.25. On a truly vertical photograph, a 300-m-high hill was photographed at 8.5 cm from the photo center. If the photograph was taken at a flying altitude of 3000 m above the base of the hill, what would be the amount of relief displacement of the summit of the hill from its base? 6.26. The radial distance between the nadir and the top of building is measured on a photograph as 6.8 cm. The distance between the top and the bottom of the building is 0.95 cm. What is the height of the building if the photograph was taken from an altitude of 4500 m above datum. 6.27. Two trees with the same height (85 m) are 1800 m and 1200 m from the nadir on the ground. The photograph was taken from a flying altitude of 3000 m above datum with a 152.4 mm focal length. Which tree is displaced more and by how much compared to the other tree?

6.28. Assume a tower, at a distance R from the nadir on the ground, was photographed from the same altitude using simultaneously a 152.4-mm and a 305-mm focal lengths. Which focal length will cause a larger displacement and by how much compared to the other focal length? 6.29. Assume a vertical photograph was taken at 2900 m above mean sea level with a 152.4-mm focal length. What is the relief displacement of the top of a hill that extends 750 m above mean sea level and whose image on the photograph is 9.45 cm from the photo center? 6.30. A vertical photograph was taken at a flying altitude of 2000 m above datum. The summit of a hill, which extends 400 m above datum, was imaged at 8.25 cm from the photo center. By how much and in what direction is the top of the hill displaced from its planimetric position? Where would be the true position of the top of the hill on the photograph? 6.31. The figure below shows a parcel that was imaged on a vertical photograph taken at a flying height of 2000 m above datum with a 152.4-mm focal length. The corners a, b, c, and d of the parcel are respectively 3.12 cm, 9.91 cm, 12.15 cm, and 10.98 cm from the photo center. The corresponding ground points A and C are respectively 420 m and 315 m above datum and B and D are respectively 460 m and 500 m below datum. Find the amount of relief displacement for the four corners and draw the correct positions a', b', c', and d'.

6.32. A vertical photograph was taken at a flying height of 2500 m above mean sea level. A pine tree was identified on the photograph and the distance between its bottom and its top was measured to be 6.25 mm on the photograph. the radial distance from the photo center to the base of the tree was found to be 9.16 cm. If the base of the tree is 1200 m above mean sea level, how tall is this tree? 6.33. A 210-mm focal length camera was used to take a vertical aerial photograph at 2500 m above datum. A tall building was located on the photograph and the distance between its bottom and the photo center was measured to be 7.53 cm and that between its top and the photo center was measured to be 8.15 mm. How tall is the tree? 6.34. Two points a and b were identified on a vertical photograph taken at a flying height of 2000 m above datum. Point a is located on the bisector line of the top right quadrant defined by the +x and +y axes of the photograph and at 9.25 cm from the photo center. Point b is located on the bisector line of the bottom left quadrant defined by the +x and -y axes of the photograph and at 7.65 cm from the photo center. The corresponding ground elevations are hA = 300 m above datum and hB = 150 m below datum. Find the relief displacement of points a and b and, after correcting both points for the displacement, compute the angle formed at the photo center between the two lines radial from a and b.

Part III: Ground coordinates from Vertical and tilted photo measurements
6.35. The images a and b of two ground points A and B were identified on a vertical photograph taken with a 152.4-mm focal length camera. The photocoordinates of points a and b are xa = 4.55 cm, ya = -5.62 cm, xb = 8.25 cm, and yb = -7.75 cm. Ground elevations above datum for points A and B are 400 m and 450 m, respectively and the horizontal distance between A and B is 1000 m. Find the flying height of the aircraft above datum. 6.36. The ground elevations of points A, B, and C are 320 m, 80 m, and 150 m, respectively. The photocoordinates of their images on a vertical photograph, after correction for film shrinkage, are xa = 4.61 cm, yb = 7.15 cm, xb = 5.83 cm, yb = -5.20 cm, xc = -5.73 cm, and yc = -8.45 cm. If the photograph were taken at 2800 m above mean sea level with a 210-mm focal length camera, what would be the horizontal distances of lines AB, AC, and BC? 6.37. Points a, b, and c on a vertical photograph are imaged positions for ground points A, B, and C. The photograph was taken from a flying height of 4500 m above datum with a 152.4-mm focal length camera. The photocoordinates of the three points are xa = -52.25 mm, ya = 22.53 mm, xb = 55.15 mm, yb = 89.75 mm, xc = 59.15 mm, and yc = -81.34 mm. The ground elevations are hA = 200 m, hB = 100 m, and hC = 300 m. Find the ground coordinates for points A, B, and C, the ground horizontal distances of AB, AC, and BC, and the horizontal angles at A, B, and C. 6.38. Images a and b, of two ground points A and B, were identified on a vertical photograph taken at a flying altitude of 2000 m above datum and with a 152.4-mm focal length. The photocoordinates of a and b, measured with respect to the x and y axes determined by the fiducial marks, are xa = -6.25 cm, ya = -5.72 cm, xb = 3.14 cm, and yb = 5.28 cm. What is the horizontal ground distance between A and B if A is 200 m and B is 150 m above datum? 6.39. Three ground features, A, B, and C were imaged at a, b, and c on a vertical photograph taken at 2500 m above datum with a 152.4-mm focal length camera. Ground elevations for points A, B, and C were found to be 450 m, 400 m, and 500 m, respectively. The photocoordinates of a, b, and c, measured with respect to x and y axes defined by the fiducial marks, are xa = 5.17 cm, ya = 4.05 cm, xb = 4.10 cm, yb = -6.25 cm, xc = -7.12 cm, and yc = -6.31 cm. Find the horizontal ground distances of AB, AC, and BC. 6.40. Solve problem 6.32 using equations 6.19 and 6.20. 6.41. Solve problem 6.35 assuming that the photograph is not vertical but tilted at 3o20' from the vertical with a swing angle of 260o. 6.42. Two ground points A and B were imaged at points a and b on an aerial photograph taken with a 152.4-mm focal length lens, a 2o 85' tilt, and a swing angle of 67o. The photocoordinates of points a and b are xa = -5.65 cm, ya = 7.12 cm, xb = -6.15 cm, and yb = -3.56 cm. The ground elevations of points A and B are hA = 300 m and hB = 600 m and the horizontal distance between A and B is 1500 m. Find the flying height above mean sea level from which the photograph was taken.

6.43. An aerial photograph was taken with a 210-mm focal length camera inclined at 2o95' from the vertical and with a swing of 115o. Find the photocoordinates of the nadir and the isocenter with respect to the axes defined by the fiducial marks.

Part IV: Ground coordinates from oblique photograph measurements

6.44. The flying height of an oblique photograph is 3200 m above datum and the focal length is 210 mm. If the distance between the principal point and the apparent horizon, measured along the principal meridian, is 8.95 cm, what is the dip angle and the true depression angle? 6.45. Use the information from problem 6.42 to compute the distances between the principal point and: the true horizon, the nadir, and the isocenter. 6.46. A 210-mm focal length camera was used to take an oblique photograph at a flying height of 2000 m above datum. the average ground elevation of the photograph is 380 m above datum and the distance between the principal point and the apparent horizon, measured along the principal line, is 72.5 mm. Find the true depression angle. 6.47. Solve problem 6.44 if the flying height is 1200 m above datum, the focal length is 152.4 mm, the average ground elevation is 200 m above datum, and the distance between the principal point and the apparent horizon, measured along the principal line, is 67.3 mm.

Lecture 7: Scale and Area Measurement

Ahmed Fahsi, Ph.D. Alabama Agricultural and Mechanical University Department of Plant and Soil Science Center for Hydrology, Soil Climatology and Remote Sensing Normal, Alabama 35762

7.1. Nomenclature and abbreviations

MSL Mean Sea Level, assumed to be 0 meter elevation and is used as reference level for altitude (i.e., elevation) Parameters that are dimensionless and are expressed to nearest 10 units: RF Representative Fraction PS Photo Scale PSR Photo Scale Reciprocal MS Map Scale MSR Map Scale Reciprocal Parameters that are expressed to nearest meter: GD Ground Distance Dat Datum Plane; the use of the notation Dat (instead of DP) is to ovoid confusion between this parameter and the abbreviation (DP) used for Difference Parallax found in subsequent chapters H Altitude of aircraft above MSL H-h Altitude of aircraft above a specific reference datum (e.g., datum plane or average elevation, or point elevation) Parameters that are expressed to nearest 0.1 meter: hD Elevation of Data above MSL hav Average ground elevation above MSL hp Elevation of a ground point P above MSL ho Height of an object or difference in height of two points Parameters that are expressed to nearest 0.1 millimeter: MD Map Distance PD Photo Distance f Camera focal length x Photographic x-coordinate measured with respect to the axes defined by the fiducial marks of the tilted photograph y Photographic y-coordinate measured with respect to the axes defined by the fiducial marks of the tilted photograph x' Photographic x-coordinate after tilt correction y' Photographic x-coordinate after tilt correction Parameters expressed in degrees: Depression angle of an oblique photograph from the true horizon Rotation angle (180o - swing angle) of a tilted photograph s Swing angle Angle of tilt on a vertical photograph

Other abbreviations: A Area PA Polygon Area TA Total Area DCP Dot Counted in a Polygon (or aggregates of polygons of the same category) DT Total of Dots counted in the total area CF Correction Factor AE Area Equivalent per dot Dcm Number of dots per square centimeter DI Number of dots per square inch sin sine cos cosine tan tangent cot cotangent m meter cm centimeter mm millimeter in inch ft foot/feet

7.2. Concepts of scale

A photograph or a map is necessarily smaller than the actual ground areas it represents. To be able to relate or establish a relationship between the photographed or mapped areas and their corresponding actual areas on the ground, the photograph -or the map- must state the ratio or proportion between the entities it represents and their corresponding entities on the ground. In figure 7.1, the ground dimension AB will be imaged as dimension ab on the photograph and the ratio (photo dimensions to ground dimensions) is called photo scale (PS) -or map scale (MS). In this chapter, our emphasis will be focused on the photo scale. Scale is one of the most important information for the usage of an aerial photograph or a map. Quantitative measurements and interpretation of features on a photograph are highly dependent upon this information. Scale is what determines the relationship between the objects imaged on a photograph and their correspondings in the real world (i.e., the ground). It is also this information that determines the amount of detail shown on a photograph or a map. The absence of scale makes it impossible to relate the size of or the distance between objects on a photograph to their actual sizes or distances on the ground.

Figure 7.1. Relationship of photographic dimensions to their corresponding ground dimensions. The scale of a photograph is defined as the ratio of the distance measured between any two points on the photograph (or a map) to the distance between the same two points on the ground. In figure 7.2 the photo scale is expressed as oa/OA (or ob/OB or ab/AB). This statement means that one unit of distance on a photograph (or a map) represents a specific number of units of actual distance on the ground. For example, a scale of 1:20000 reads 1 unit measured on the photograph (or a map) represents 20000 times the same unit on the ground. In other words, the horizontal distance on the ground between two well identified points is 20000 times larger than the distance between these same two points properly measured on the photograph. Alternatively, this statement may also read that the distance measured between two well identified points on the photograph (or a map) is 1/20000th of the horizontal distance between the same two points on the ground. A scale of 1:20000 simply means that a distance of 1 mm measured on the photograph (or a map) represents 20000 mm (20 meters) on the ground, or 1 cm on the photograph represents 20000 cm (200 m) on the ground, or 1 inch on the photograph represents 20000 inches (or 1667 feet) on the ground.

Figure 7.2. Scale of an aerial photograph as expressed by the ratio of photo distance to ground distance. Scale determination requires both the ground and the photo distances to be measured along the horizontal plane for accurate representation of the photo scale. Unfortunately, unlike maps, most photographs are subject to distortions (see RSCC Volume 1, Module 6) and, therefore, features on the photograph are not identified in their true horizontal position. A convenient way to achieve a horizontal measurement between two points on a photograph is to select two points that lie at the same elevations and at the same distance from the center of the photograph. Further, because the photograph scale (PS) -or the map scale (MS)- is dimensionless, the distance measured on the photograph (PD) between two points and their corresponding ground distance (GD) must be in the same units.

7.3. Representing Scale

The most common expressions of scale are the ratio or representative fraction (RF), the unit equivalents (UE), the photo scale reciprocal (PSR), and the scale bar. Representative fraction (or ratio) is the fraction of a distance measured between two points on a photograph to the distance measured between the same two points on the ground. It can be expressed as 1/20000 or as 1:20000. For printing convenience, the latter form is the most commonly used. Notation for RF can also be written as PS (photo scale). Unit equivalents, also called equivalent scale, expresses the equivalence of a distance measured between two points in photographic units to the distance between the same two points in ground units. For example, a PS of 1:20000 would be expressed as 1 mm = 20

m (or 1 cm = 200 m or 1 inch = 258 ft), meaning that a distance of 1 mm on a photograph is equivalent to 20 m on the ground (or 1 cm is equivalent to 200 m on the ground). To convert from the representative fraction (RF) to the unit equivalents (UE), we simply multiply both the numerator and the denominator of the former by unity so that all units cancel to obtain the unit equivalents desired. For example, if we desire to obtain unit equivalents of centimeters per kilometer for a RF of 1:20000, we proceed as follows: written as 5 cm = 1 km and reads 5 cm measured on the photograph represents 1

kilometer on the ground. Similarly, to convert from UE to RF, we proceed backward, by multiplying by unity so that all the units cancel as follows: Photo scale reciprocal (PSR) is simply the inverse of the representative fraction. For example, an RF of 1:20000 would correspond to a PSR of 20000. Among the different expressions of the scale, the scale reciprocal (i.e., PSR or MSR) may be the most convenient to work with in constructing equations. However, its interpretation may result in confusion, because large values of PSR -or MSP- represent small scales and vice versa. For example, 24000 scale photographs are larger scale than 36000 scale photographs. Therefore, close attention must be paid when working with different expressions of scale. For those who are not familiar with the notion of scale, the terms "large scale" and "small scale" are often confusing. The term large scale is often intuitively associated with a larger denominator of the RF and vice versa. For example, if asked which photograph would have the "larger" scale, the 1:15000 or the 1:20000, it is likely that the answer would intuitively be the 1:20000, where, in fact, the correct answer is the 1:15000. To understand the notion of large and small scales, one easy way is to compute the values of the RF being compared; the smaller value would correspond to the smaller scale. For example, 1:15000 = 0.000067 and 1:20000 = 0.00005. Since 0.00005 is smaller than 0.000067, the 1:20000 scale is smaller than the 1:15000 scale. Another way of understanding large and small scales is to remember that a large number in scale expression denotes a small scale and vice versa. Another method is to associate scale with the area coverage; large scale will show large feature sizes and vice versa. In other words, ground features are imaged larger on large scale photographs and smaller on small scale photographs.

7.4. Types of scale

On a map, all points are supposed to be at the same scale. However, due to the perspective projection of a photograph, the variation in elevation -thus the flying height above ground features, and the effects of relief and tilt, the scale of an aerial photograph may vary considerably from one point to another. A photograph may only present a uniform scale comparable to that of a map if the camera is perfectly vertical at the time of

exposure and the terrain is perfectly flat. Since these conditions are seldom fulfilled, the photographic scale is usually considered for individual points or the average of a set of points. The first is referred to as point scale and the second is referred to as average scale. Point Scale Point scale is the scale at a point with a specific elevation on the ground. This suggests that every point on a vertical photograph at a different elevation will have a different scale (figure 7.3). Therefore, a photograph taken over a rugged terrain will display a varying range of scales associated with the variations in elevation of the ground. In figure 7.3, scale at point A will be different from that at point B, or C.

Figure 7.3. Scale variations with variation in ground elevation Because point scale represents the scale of a single point at a given elevation, it is the most accurate type of scale. This scale depends on the altitude of the point above mean sea level (MSL), the flying height of the aircraft above the point at the time of exposure, and the focal length of the camera used to take the photograph (figure 7.4). Point scale is

expressed as: where: PSP is the photo scale at point P, f is the focal length of the camera used to take the photograph, the distance between the lens and the focal plane, H is the flying height of the aircraft above MSL, the distance between MSL and the lens, hP is the elevation of point P above MSL, and H-hP = HP is the flying height of the aircraft above point P.

Figure 7.4. Determination of point scale on a vertical aerial photograph. Example 7.1 A 305-mm focal length camera was used to take photographs from 4000 m above MSL. Find the scale of a point that is 800 m above MSL and the scale of a point that is at the MSL. Solution:

By equation 7.1, Average scale Unlike point scale, which is specific to a single point on the ground, average scale may be determined for the entire project area, a set of photographs, a single photograph, a portion of a photograph, or between two points on a photograph. Because the average project scale may not represent the actual scale of each individual photograph or each point on a photograph (figure 7.5), it is referred to as a nominal scale. It is the desired average scale intended for the photo mission. From figure 7.5, this scale may be expressed as:

where: PSav is the average scale of the area considered (project, set of photographs, etc.), hav is the average elevation of the area, and H-hav = Hav is the flying height of the aircraft above the average elevation of the area. Example 7.2 A photograph was taken from a flying height of 4500 m above sea level with a 152.4-mm focal length camera. If the average elevation of the photograph is 580 m above MSL, what is the average scale of the photograph? What would be the scale of a road junction that is at 520 m above MSL? Solution: by equation 7.2, and by equation 7.1,

Figure 7.5. Average scale as compared to point scale. Generally, the average elevation of a photograph (or the entire project area) is commonly considered as the datum plane of the photograph (or the project area) and the scale at this elevation is: where: PSD is the scale at the datum plane,hD is the datum elevation above MSL, andH-hD = HD is the flying height above datum. Although average scale is not as accurate as point scale, it is commonly the most used. Average scale may not be representative for all the points imaged on a photograph (or a project) as the scale may be too small for areas that have elevations higher than the average elevation (such as points D, E, and F in figure 7.5), and too large for areas that have elevations lower than the average elevation (such as points A and B in figure 7.5). In fact, the average scale may not represent any, or very few, point in a particular area of the project.

In figure 7.5, for example, the only point on the ground that will have an exact scale is point C because it lies at the average elevation plane, where the scale is been determined. All the points whose elevations are higher than this point will have a smaller scale than their actual scale and all those whose elevations are lower than this point will have a larger scale than their actual scale. Example 7.3 A 152.4-mm focal length was used to take photographs from 4000 m above MSL over an area that has an average elevation of 560 m above MSL. Find the average scale of the area. Compare this scale to that of a plane that is 300 m below the average elevation and to that of a ridge that is 250 m above the average elevation. Solution: Using equation 7.2, the average scale of the area is:

and using equation 7.1, the scales at the plane and at the ridge are, respectively:

and Therefore, an average scale of 1:22570 would be too large for the plane (whose actual scale is 1:24540) and too small for the ridge (whose actual scale is 1:21330).

7.5. Methods of determining scale of a vertical photograph

Scale from photo and ground measurements Various methods are used to determine the scale of a photograph, depending on the types of information available. The basic and most straightforward technique of expressing scale is the one that uses the ratio of the distance between two points on a photograph (PD) to the distance between the same two points on the ground (GD). From the basic geometry of a vertical aerial photograph (figure 7.6 ), we can see that from similar triangles Lop and LOP, photo scale may be expressed as:

Since

Therefore, Similarly, map scale may be expressed as:

where: PS is the photo scale, MS is the map scale, PD is the photo distance measured between two well identified points on the photograph, MD is the map distance measured between two points on the map, and GD is the ground distance between the same two points on the photograph (or on the map), expressed in the same units.

Figure 7.6. Diagram illustrating the basic geometry of determining photo scale from photographic and ground measurements.

As mentioned earlier, the photo scale reciprocal (PSR), sometimes called photo scale factor or the map scale reciprocal (MSR), sometimes called map scale factor is simply the inverse of the photo scale or the map scale and can be expressed as:

Example 7.4 Assume that the distance between two points was measured to be 83.33 mm on a vertical photograph and 125.00 mm on a map. If the surveying ground distance between the same two points is 3000 m, what are the scales and the scale reciprocals of the photograph and the map? Solution: From equations (7.4) and (7.5), we obtain:

and

and from equations (7.6) and (7.7), we obtain:

Example 7.5 Using the information in example 7.4 express the photo scale and the map scale in unit equivalents.

solution: and Equations 7.4 and 7.5 are convenient when the distance between any well identified points on a photograph -or a map- is accurately measured and their corresponding ground distance is known. Unfortunately, many objects are not always well identified on a photograph and surveying ground distances may not always be available. In addition, accurate measurements of these ground distances are tedious, time consuming, and costly, particularly for long distances and/or when a higher number of measures are needed for accurate average project scale determination. This situation suggests that other alternatives must be considered to determine the photo scale. Scale from focal length and flying height Generally, a photograph is annotated with information, including the date and time of acquisition, the project code, the serial identification of the photograph (i.e., line number and exposure number), the focal length, the flying altitude of the aircraft above MSL or above datum (figure 7.7).

Figure 7.7. Some of the information printed on a vertical aerial photograph. This nomenclature is printed on the west side for photographs flown east-west and on the north side for photographs flown north-south. When these parameters are present, the photo scale may be determined using the same geometry of a vertical aerial photograph in figure 7.6. Again, from similar triangle Lop and LOP, we may write: where: PS is the photo scale, f is the focal length of the camera used to take the photograph, and H-h is the flying altitude of the aircraft above the ground, which may be the average elevation of the entire project area, the average elevation of the photograph -or between two points on the photograph, or the elevation of a single point on the ground. This technique is fairly accurate, particularly in determining point scale, when the flying height of the aircraft above MSL and the ground elevation are precisely known. For average scales, the accuracy of this technique depends mainly on the terrain. In addition, with the use of a conventional altimeter on the aircraft, flying height above the ground

may not be recoverable. Further, in rugged terrain, the flying height values may not represent an accurate value for the project area, the photograph, or even a point on the photograph, since the flying height printed on the photographs is only a nominal (usually the average) value for the entire project. This is why the photo -or project- scale is usually referred to as nominal rather than true scale. Only maps are supposed to have a true scale and they can be very useful in determining photo scale as well. Scale from photo and map measurements If the information on the photograph and the ground distance between two points are not available, the photo scale can still be determined if a map of the area is available. By examining equations 7.6 and 7.7, it can be inferred that: GD = PD PSR and GD = MD MSR It follows that: PR PSR = MD MSP and PD and MD are distances measured between the same 2 points that are well identified on the photograph and the map, respectively (figure 7.8). For example, in figure 7.8, if the distance between the same two points, A and B, were measured to be 6.05 cm on a photograph and 4.25 cm on a 7.5-minute USGS quadrangle (PRS = 24,000), then the scale of the photograph (PS) would be 6.05 (4.25 24000) = 1:16860. This method is the most useful and the most practical form of determining a photo scale, simply because of easily obtainable values from a map. A map will always have a scale value or, at least, a bar scale from which a scale value can be determined. Further, distances between two points can easily and more accurately be measured on a map. However, many countries and many regions are not entirely covered with maps or ground surveying distances. If a situation arises where ground surveying distances, a map of the area, and the information on a photograph are not available, a photo scale can still be determined if another photograph of the same area and with known scale is available.

Figure 7.8. Determining photo scale from a map of known scale. Scale from an existing aerial photograph This situation requires that either the scale or the focal length and the flying height above the ground of a second photograph are known. When the scale of a second photograph

(PS2) is known, our photo scale (PS1) may be determined in the same manner as it was done using a map. By measuring the distances between the same two points on photo 1 (PD1) and photo 2 (PD2), the unknown scale of photo 1 (PS1) may be determined from the known scale of photo 2 (PS2) as: PS1 is the scale to be determined for the photograph considered (photo 1), PD1 is the distance measured between two well identified points on photo 1, PD2 is the distance measured between the same two points on the known-scale photograph (photo 2), and PSR2 is the scale reciprocal of photo 2. If the scale of photo 2 is not known but the focal length and the flying height above the ground are available, then, the scale of photo 1 may be determined as follows: From equation 7.8 we can write

Substituting equation 7.11 in equation 7.10, we obtain:

where: f2 is the focal length of the camera used to take photo 2 and H-h2 is the flying height of the aircraft above the ground of photo 2. Procedure of determining scale for a single photograph If the information usually printed on an aerial photograph is missing but a map of the area or ground measurements are available, the average scale of a single photograph may be determined as follows: 1. Select two points, preferably at the average ground elevation of the photograph, and approximately equidistant from and opposite the photo center. The two points must accurately be located on a map -or on the ground if a map is not available. 2. Accurately (to nearest 0.1 mm) measure the distance between the two points on the photograph and the map -or determine the ground horizontal distance (to nearest 1 m) if a map is not available. 3. Compute the photo scale using equation 8.9 if map measurements are available or using equation 8.4 if ground measurements are available. Example 7.6 Two points were clearly identified on an aerial photograph and on a 1:24000-scale USGS topographic map of the same area. The distance on the photograph between the two points was measured to be 8.25 cm and the distance on the map between the same points was measured to be 5.72 cm. Find the scale of the photograph.

Solution: By equation 7.9, we obtain:

If the information on the photograph is available, then the photo scale is determined using equation 7.8. Example 7.7 Suppose a photograph was taken from 3500 m above MSL with a 152.4-mm focal length camera as printed on the photograph. If the average ground elevation of the area covered by the photograph is 830 m above MSL, what is the scale of the photograph? Solution: Using equation 7.8,

Example 7.8 Now, suppose that neither a map nor the information on the photograph are available, but the ground distance between two points well identified points is known to be 1320 m and the distance measured on the photograph between the same two points is measured to be 7.86 cm. Find the scale of the photograph. Solution: By equation 7.4, we obtain, Procedure of determining scale for a series of photographs The average scale for the entire project or a series of photographs may be determined as follows: 1. Determine the average elevation for each flight line using a topographic map. 2. Compute the average altitude for the area considered using the average altitudes determined in step 1 and apply equation 7.8. 3. Apply equation 7.2 if H and f are known, otherwise, use the same steps as in the case of a single photograph by considering the average elevation of a set of photograph instead of one single photograph. Example 7.9 A 305-mm focal length was used to take aerial photographs from 4000 m above MSL. Using a topographic map, the average elevations of the flight lines were found to be 700 m for line 1, 580 m for line 2, 650 m for line 3, and 750 for line 4. Find the average scale of the project. Solution:

First, the average elevation of the entire area is determined as:

Then, using equation 8.2, we obtain the average scale of the project as:

Example 7.10 In example 7.9, the information (H and f) on the photographs was missing, but a number of surveying benchmarks were identified on the photographs and a topographic map of the same area. Two benchmark points were identified on line 1 and line 4 and their coordinates were found to be XA = 515,000 m; YA = 5,182,000 m, XB = 518,000 m; and YB = 5,179,000 m. By properly overlapping (mosaiquing) the photographs and measuring the photo distance between these two points, it was found to be 26.5 cm. Find the average scale of the project. Solution: To be able to find the photo scale, we first need to determine the ground distance between the two points. Since we know the ground coordinates, the distance between the two points may be determined using the Pythagorean theorem:

Then, using equation 7.4, the photo scale can be determined as:

Another alternative is to compare known areas on the ground to their imaged areas on the photograph. The area ratio is directly proportional to the square of the photo scale. This latter may be determined as follows: PS = PD / GD therefore, PS2 = PD2 / GD2 It follows that:

and Example 7.11 A rectangular corn field was identified on an aerial photograph and its length and width were measured to be 10 cm and 6.5 cm, respectively. The ground area of the field is known to be 4 hectares. Find the scale of the photograph. Solution:

The area of the field on the photograph is computed as: Photo area = 10 cm 6.5 cm = 65 cm2 Then, using equation 7.13, we obtain:

Remember that, in order to find PS, both photo area and ground area must be expressed in the same units. There are 100,000,000 cm2 in a hectare (or 6,272,640 in2 in an acre and 2.471 acres in a hectares).

7.6. Determining scale of a tilted photograph

Determination of scale on a tilted photograph is different from that on a vertical photograph. Scale on a tilted photograph is increasingly altered as the distance from the principal point increases. Because of this effect, only point scale is effective and meaningful. The scale of a point on a tilted photograph requires not only the knowledge of the flying height, the focal length of the camera, and the elevation of the point, but the position of the point with respect to the principal line as well as to the axis of tilt because tilt occurs in the direction of the principal line. The axis of tilt is the perpendicular to the line connecting the principal point and the nadir and half way between these two points (see RSCC Volume 1, Module 6). Since the positions of the principal point and the nadir are fixed by the angle of swing (figure 7.10) it is necessary to know the angle of swing.

7.10. Basic geometry of determining scale of a tilted photograph. If a is the image on a tilted photograph of a ground point A that is at an elevation of hA above MSL, its photocoordinates may be determined with respect to the axis defined by

the fiducial marks. In RSCC Volume 1, Module 6 (section 6.7.3.) we saw that the coordinates of a point a, after adjustment of the x and y axes were: x'a = xa cos + ya sin where: xa is the photographic x-coordinate of point a measured with respect to the axes defined by the fiducial marks of the tilted photograph, ya is the photographic y-coordinate of point a measured with respect to the axes defined by the fiducial marks of the tilted photograph, x'a is the photographic x-coordinate of point a after tilt correction, y'a is the photographic y-coordinate of point a after tilt correction, is the rotation angle (180o - swing angle) of a tilted photograph, and is the angle of tilt on a vertical photograph. We also saw that the photo distance ma is the image of the ground distance NA. Since the scale is the ratio of the photo distance to the ground distance, the scale at point a at elevation hA on the ground may be expressed as: (7.14) y'a = -xa sin + ya cos + tan

where: f is the focal length of the camera used to take the photograph, is the angle of tilt (angle between the optical axis and the vertical), and H-hA is the flying altitude above point A. If the tilt is small ( is less than 3o), sec is approximately equal to 1 and equation 7.15 reduces to:

Example 7.12 A 152.4-mm focal length camera, tilted at 3o from the vertical was used to take a photograph from a flying height of 5000 m above datum. Find the photo scale at the principal point, at a point that is 11.63 cm from the principal point on the up side of the tilt and another point at the same distance (11.63 cm) from the principal point on the down side of the tilt. Solution:

7.7. Determining scale of an oblique photograph

As seen in RSCC Volume 1, Module 6, distortions in oblique photographs differ from those in vertical photographs in that the geometry of an oblique photograph is fundamentally different from that of a vertical photograph. It follows that scale alteration on oblique photographs is different in all three directions: x-direction (perpendicular to the principal meridian), y-direction (parallel to the principal meridian), and z-direction (vertical). From figure 7.11, both blocks are cubes on the ground, yet each block has different size, different height, and different orientation on the oblique photograph. Tremendous effort and work was done in the 1950s to understand and formulate the geometry of oblique photographs and to compute scales and heights on these photographs. The fundamental principles may be summarized in the following formulas. Scale perpendicular to the principal meridian Because the plans of the vertical and oblique photographs intersect at the line perpendicular to the principal meridian and passing through the isoline, the scale of the oblique photograph along this line is the same as that of a vertical photograph, and the same formula applies (PS = f /HG, where PS is the photo scale, f is the focal length, and HG is the aircraft altitude above the ground). As the distance of objects increases away from the isoline toward the nadir (i.e, the down side of the tilt), the scale of these objects becomes larger. Likewise, as the distance of objects increases from the isoline toward the horizon, the scale of the objects becomes smaller. This change in scale is proportional to the ratio of the distance between the lens (L) and the photo plane (o) to that between the lens (L) and the ground datum (O) (see figure 7.11a). Therefore, the scale (PS) of a point on the oblique photograph at a distance d, measured from the true horizon in the plane of the oblique (parallel to the principal meridian) can be expressed as:

Where is the depression angle of the oblique from the true horizon and HG is the flying altitude of the aircraft above the ground. When the point lies on the perpendicular to the principal meridian which passes through the principal point, the scale can be determined directly from the following formula:

where f is the focal length of the camera used to take the oblique photograph.

Figure 7.11. Diagram illustrating the geometry of determining scale of an oblique photograph. Scale parallel to the principal meridian The scale along the principal meridian (y-scale), applicable to distances measured away from or toward the horizon point, is determined as:

Where d is the distance from the true horizon to one point on the principal meridian and d' is the distance to the other point, between which the scale is to be computed. This formula can be used to determine the scale between any two points that are along the same line converging toward and passing through the horizon point (z). Scaling horizontal distances obliquely On an oblique photograph, any distance that is neither perpendicular to the principal meridian (x-scale) nor along a line passing through the horizon point (y-scale) must be determined before computing the scale. Determination of this distance may be done by measuring the two components (parallel and perpendicular to the principal meridian) at right angles to one another and computing the actual distance as the hypotenuse of the right triangle thus formed using the Pythagorean theorem. For example, in figure 7.11 (b), the diagonal distance from point X' to Y across the cubic block would be determined by measuring the distance from X to X' by the formula for scale perpendicular to the principal meridian, measuring the distance X to Y by the

formula for scale parallel to that meridian, and computing X'Y as the square root of the sum of the squares of the two distances:

It should be noted that even though the triangle formed by points X', X, and Y does not appear "right" when viewed in perspective, the three points form a right triangle in actuality. Scaling vertical distances On an oblique photograph, vertical distances or heights (z-scale) may be determined by the formula:

where d is the distance from the base of the object to the horizon point (z) and e is the distance from the top of the object to the nadir (n). Measurement from the principal parallel Because on an oblique photograph, the principal point is easier to identify than the horizon point, it is often more convenient to measure distances from the parallel passing through the principal point (principal parallel) than from the horizon point or the nadir. If y is the distance along the converging line from the point to the principal parallel, then:

For d, y is measured along the ray that passes through the principal point and the horizon point (z) and for e, y is measured along the ray that passes through the nadir. Also, these formulas are for the far side of the principal parallel toward the horizon. For the near side of the principal parallel, the signs preceding y are reversed. Because a high oblique photograph is a highly tilted photograph along the principal meridian, also called principal line, the scale of such photograph is constant along any line that lies perpendicular to the principal meridian but varies from point to point along any other line. Since the principal line represents the x-axis of the photograph, the scale of a point along a constant-scale line at a distance x from the principal point is designated as PSx along the line. Conversely, the scale of a point that falls on a line perpendicular to the principal line varies from one point to another and it is designated as PSy. If dz is the distance measured parallel to the principal line from the true horizon (z) to the point whose scale is being computed, then the x-scale at that point is expressed as:

and the y-scale of a point is expressed as:

where: dz is the distance between the true horizon and the point considered, measured along a line parallel to the principal line, is the true depression angle, HG is the flying height of the aircraft above the ground, and f is the focal length of the camera used to take the photograph. Example 7.13 A 152.4-mm focal length camera was used to take a photograph from a flying height of 3500 m above datum. The photo distance between the true horizon and the principal point was measured to be 9.87 cm. Find the scale along the constant-scale line passing through the principal point and the scale along another line that lies half way between the principal point and the true horizon. Solution: From equation 6.47 (RSCC Volume 1, Module 6), we obtain:

By equation 7.22, the x-scale at the principal point is:

and the x-scale halfway between the principal point ant the true horizon along the principal line is:

Example 7.14 Using the information in example 7.3, find the y-scale at the principal point. Solution:

7.8. Factors affecting scale

By examining equations 7.1 and 7.2, we can see that the scale of a vertical photograph varies directly as the focal length of the camera and inversely as the flying height of the aircraft above the ground. This latter is itself affected by the tilt of the camera axis from the vertical (i.e., plumb line). Effect of focal length on photo scale Equations 7.1 and 7.2 suggest that for a given flying height, a larger focal length will produce a larger photo scale. This is because as the focal length increases, the proportion of the focal length (f) to the flying height (H-h) (or the proportion of the photo distance,

PD, to the ground distance, GD) increases and vice versa. In figure 8.13, the ground distance (GD) is photographed as PD1 on photo 1, taken with a 152.4-mm focal length camera and as PD2 photo 2, taken with a 305-mm focal length camera. Because f2 is twice as long as f1, distance PD2 on photo 2 will be twice as long as distance PD1 on photo 1 and the scale will be twice as large on photo 2 than on photo 1, provided the flying height above the ground is constant. Therefore, photo distance, and hence photo scale, varies directly as the focal length. Conversely, the ground area covered by a photograph varies inversely with the camera focal length. Thus, because f2 = 2 f1, the area covered by photo 2 is one fourth of the area covered by photo 1. Note the trade-off between the detail allowed by the scale of a photograph and the area covered by the photograph. The larger the scale, the greater the detail recorded on the photograph, but the smaller the area covered by the photograph. This trade-off should always be taken into consideration when acquiring aerial photographs or planning for a photo mission. Scale and area coverage may be controlled using two techniques: varying the flying height above the ground or varying the focal length as it will be discussed in the following sections. The most common focal lengths used in aerial photography are the 152.4 mm (6 in), the 210 mm (8.25 in), and the 305 mm (12 in).

Figure 7.13. Effect of the focal length on the photo scale and ground coverage. Example 7.15 Two cameras equipped with a 152.4-mm and a 305-mm focal lengths were mounted on an aircraft to take photographs from 3000 m above the ground. a. What will be the scale of the photographs taken with the two different focal lengths? b. If the format of the photographs is 23 by 23 cm (9 by 9 inches), how much area will be covered by one photograph (photo 1) taken with the 152.4 mm and by one photograph (photo 2) taken with the 305 mm focal length? c. What will be the length on photo 1 and photo 2 between two points that are 1350 m apart on the ground? Solution: a. By equation 7.8, we obtain:

b. c. Using equation 7.6, we can write:

In example 7.15, notice that because f2 is twice as long as f1, PD2 is twice as long as PD1, PS2 is twice as large as PS1, but Area2 is one fourth of Area1. Thus, photo distance and photo scale vary directly as the focal length but the area varies inversely as the focal length. Effect of flying height on photo scale Equation 7.8 also indicates that, for a given focal length, higher flying altitudes will produce smaller scales. This is because as the flying height above the ground (H-h) increases, the proportion of the focal length (f) to the flying height (H-h) (or photo distance, PD, to ground distance, GD) decreases. Figure 7.14 illustrates two photographs, P1 and P2, taken from two different flying heights, H1-h and H2-h, respectively. Because P2 is taken from a higher flying altitude than P1, the image of the same distance on the ground (GD) will appear smaller on P2 than on P1. In fact, for a given focal length, f, if H2-h is twice as high as H1-h, the distance, PD2, between two well identified points on photo 2 will be half the distance, PD1, between the same two points on photo 1. Likewise, the scale of photo 2, PS2, will be half the scale of photo 1, PS1 (or PSR2 = 2 PSR1). Conversely, the area covered by photo 2 will be four times larger than that covered by photo 1. Notice that the flying height has an inverse effect on photo scale and the amount of ground area covered by a photograph compared to the focal length. The same trade-off depicted with the focal length applies to the flying height.

Figure 7.14. Effect of flying height on the photo scale and the ground coverage. Example 7.16 An area was photographed from two different flying heights using the same focal length of 152.4 mm. If photo 1 (P1) and photo 2 (P2) were taken from 2000 m and 4000 m above ground, respectively, find the scale of these two photos and compare the areas covered by each photo if their format is 23 by 23 cm. Solution: From equation 7.8, we obtain:

and

and Notice that PS1 is twice as large as PS2 but Area1 is only 1/4 Area2 because the flying height (4000 m) used to take photo 2 is twice as high as the flying height (2000 m) used to take photo 1. Effect of Topography on scale Varying terrain is a major source of scale variation in an aerial photograph. Difference in ground elevation causes the flying height above the terrain to vary from one photograph to another or from one point to another in a single photograph. Focal length (f) and flying height (H) above MSL remain constant on an aerial photo mission, but the flying height (H-h) above ground features varies as the ground elevation changes (figure 8.15). Because flying height above the terrain decreases with higher ground elevations, objects at higher altitudes (e.g., ridges, plateau, summit of the mountains) will have larger scale than objects at lower altitudes (e.g., planes, valleys, canyons). This variation in scale is more pronounced when flying at low altitude above the ground. For example, if the terrain varies from 0 to 1000 m above MSL and the flying height of the aircraft is 2000 m above MSL, the photo scale will be twice as large at 1000 m as at 0 m (i.e., MSL). Example 7.17 If a 152.4-mm focal length camera were mounted on an aircraft flying at 4000 m above MSL and taking aerial photographs over an area whose elevation varies from 500 m to 1200 m above MSL, what would be the variation in the photo scale? Solution: Again, using equation 7.8, we can determine the scale at 500 m and at 1200 m elevations as follows:

Consequently, over varying terrain, a given photograph will have as many scales as there are ground elevations in the area photographed. The average scale of a photograph may be very different from the other photographs of the same project. Therefore, a nominal project scale of 1:15840 may be the average of individual photo scales ranging from 1:13565 to 1:20634 or other intervals. To minimize the effect of topography, we usually use the average elevation of the area considered. Because scale is not an exact representation of ground dimensions, it is usually expressed to nearest 10 units. For example, a ratio of 1:15837 would be expressed as 1:15840 and sometimes even as 1:15800.

Figure 7.15. Effect of topography on photo scale; higher ground elevations (B) have larger scale than lower ground elevation (A) in a single photograph. Usually, over a varying terrain, the scale of a series of photographs is determined at the average ground elevation of the considered area and it is referred to as average or nominal scale. In example 7.17, for example, the average scale for the entire area would be:

Effect of tilt on the photo scale Another variation in photographic scale is caused by the tilt of the optical axis of the camera from the vertical at the time of exposure. As the angle between the camera axis and the vertical increases, the distance from the camera to the ground increases and the proportion of the focal length to this distance (or the photo distance to the ground distance) decreases (figure 7.16). Hence, the photo scale decreases.

Figure 7.16. Effect of tilt on the scale of an aerial photograph. More specifically, on a tilted photograph, the scale is smaller on the up side and larger on the down side of the tilted photograph. As shown in figure 7.16, the same length of ground distance (GD) is photographed smaller (d1) on the up side and larger (d2) on the down side of the photograph. Fortunately, this change in scale is regular throughout the tilted photograph, which makes its correction possible. To minimize errors due to tilt, it is recommended to compute the average photo scale by measuring the photo distance between two points that are at the same distance and opposite from the photo center. On a vertical photograph, variation in scale due to tilt is usually small that it can be neglected.

7.9. Factors affected by scale

One of the most factors affected by scale is the ground surface coverage. The area covered by a photograph is a function of the scale of the photograph. Large scale photographs cover small areas on the ground and vice versa. The amount of ground surface covered by an aerial photograph dictates the amount of detail depicted on the

photograph. Usually, more detail can be depicted on large scale than small scale photographs. With larger scales, objects appear bigger and easier to identify and interpret, but the area covered is smaller. It follows that it would take more photographs to cover the same area with large scale than with small scale photographs. This also implies that the cost involved with large scale photographs would be higher. Therefore, there is always a trade-off between the amount of detail desired and the cost involved, which must not be ignored when planning for a photo mission.

7.10. Determining areas from aerial photographs

Areas of land surfaces can be determined directly from aerial photographs. This requires, however, the delineation of the boundaries of the units whose areas are to be determined. The area of many regularly shaped surfaces can be calculated by using the appropriate formulas shown in Appendix, Table 1. Areas of non regular shaped surfaces require more advanced calculus and techniques discussed later in this chapter. A variety of methods for determining land areas of any unit shape from aerial photographs or maps have been developed. Instruments and techniques for determining areas range from simple to complex, archaic to high technology, cheap to very expensive. The principal methods and devices used for area determination are: integration, coordinates, weighing, transects, dot grid, and planimeters. Whether the surface is regular or irregular, mathematical area measurements basically require knowledge of the linear magnitude of two dimensions. However, beside accurate mathematical formulas, several practical techniques and procedures are available to estimate surface areas. Area determination of the earth's surface usually considers the area of the orthogonal projection of the irregular land surface to a datum-level plane. Area determination for portions of the earth's surface can be expressed as the square of any of the linear units. The most common area units are the international metric system units: square centimeters (cm2), square meters (m2), hectares (ha), and square kilometers (km2). However, in various countries, other units have become standard through common usage. For example, in the United States, the acre (43,560 square feet or 10 square chains) is the most common area unit (see Appendix, Table 2 for converting factors).

7.10.1. Area by coordinates

If we can establish an xy-axis system on a plane (e.g., aerial photos), the area of a polygon of any shape can be computed. Suppose we take the east-west direction of a photograph as the x-axis, the north-south direction as a y-axis, and the xy origin as (0,0). If a polygon of any shape is bounded by straight segments, we can determine the x and y coordinates of the vertices by measuring the distance between the origin and the orthogonal projection of that vertex to the x-axis (for x coordinates of the vertex) and to the y axis (for y coordinates of the vertex) (figure 7.18).

Figure 7.18. Area by coordinates; xi and yi are the photo coordinates of vertices. If the coordinates are designated as x1, x2, x3, ....., xn and the y coordinates are y1, y2, y3, ...., yn, then the area of the polygon whose vertices are (y1, y1), (y2, y2), (y3, y3), ..., (yn, yn) is the absolute value of:

if the ground coordinates of vertices are known the ground area may be directly determined by substituting these coordinates in equation 7.25. Otherwise, the scale of the photograph may be used to determine the surface area on the ground as follows: after determining the polygon area using the photo coordinates, this surface is then multiplied by the square of the photo scale reciprocal: Photo area PSR2 = Ground area Example 7.19 The photo coordinates of the 5 vertices of a polygon delineated on a 1:12,000 vertical aerial photograph were measured to be x1 = 5.12 cm, y1 = 1.54 cm, x2 = 5.61 cm, y2 = 4.37 cm, x3 = 7.13 cm, y3 = 6.20 cm, x4 = 9.83, y4 = 6.81 cm, x5 = 8.08 cm, y5 = 3.44 cm. Find the area of the polygon. Solution: Using equation 7.25, the photographic area of the polygon is: (|[(5.12)(4.37) + (5.61)(6.20) + (7.13)(6.81) + (9.83)(3.44) + (8.08)(1.54)] [(5.61)(1.54) + (7.13)(4.37) + (9.83)(6.20) + (8.08)(6.81) + (5.12)(3.44)]|) = |151.97 - 173.38| = 21.41 cm2 The ground area of the polygon is then: 21.41 cm2 (12000)2 = 308304 104 cm2 or 30.83 hectares Example 7.20 Using the same photograph in example 7.19, another polygon was delineated and the ground coordinates of its 5 vertices were found to be X1 = 514852, Y1 = 5178256, X2 = 515495, Y2 = 5180687, X3 = 518168, Y3 = 5181112, X4 = 518738, Y4 = 5179680, X5 = 517766, Y5 = 5178184. Find the ground area of the polygon. Solution: Using equation 7.25, the ground area of the polygon can directly be computed as: |[(514852)(5180687) + (515495)(5181112) + (518168)(5179680) + (518738)(5178184) + (517766)(5178256)] - [(515495)(5178256) + (518168)(5180687) + (518738)(5181112) + (517766)(5179680) + (514852)(5178184)]| = 851.50 ha If the polygon is delineated by arcs instead of straight segments, the above formula can still be used to estimate the area of the polygon. In this case, the boundaries of the polygon should be divided into small arcs that could approach a straight segment (figure 7.19). The line segments shown in figure 7.19 are purposely drawn large to better

illustrate the concept of this technique. Usually, the curved boundary is divided into small arcs that approach straight segments. The shorter the arcs, the more accurate the area estimation.

Figure 7.19. Diagram illustrating area estimation by coordinates when the surface boundaries are curves instead of straight lines. However, a more accurate method to determine the area of a "non-straight-line-boundarypolygon" is the use of integration, planimeters or computers. All of these techniques are based on the coordinate method described above to determine the area of a polygon.

7.10.2 Area by integration

This technique supposes that the plane surface is bounded by known functions (figure 7.20). In this figure, the area of the shaded surface may be determined by using the area bounded by the X-axis and ordinates at x = a and x = b or by using the area bounded by the Y-axis and ordinates y = c and y = d. The area of this plane surface can then be calculated using the definite integral:

Where dx and dy are the small portions of the curve that approach a straight segment. This technique is rarely used in practical photography area measurement and, therefore, will not be discussed further.

Figure 7.20. Estimating polygon area by integration.

8.10.3. area by weighing

A surface area of a polygon on plane sheet of uniform thickness can be estimated using the weighing procedure. First, the area and weight of the entire plane sheet are determined; the weight of the map of the photograph paper will be directly proportional to its surface area. Then, the total photo surface and weight are used as the basis to determine the ratio of area to unit weight. For example, if the plane sheet is a 23 cm by 23 cm (9 inches by 9 inches) aerial photograph, its area is 23 23 = 529 cm2. The photo

area is then converted to ground surface by multiplying the value of the photo area (529 cm2 in this case) by the square value of the photo scale reciprocal (i.e., PSR2): Ground Area = Photo Area (PSR)2 For example, if a 23 cm by 23 cm aerial photo weighs 20 grams, then the ratio of area to unit weight is 529 cm2 20 g = 26.45 cm2/g. In other words, 1 g of paper would be equivalent to an area of 26.45 cm2 on the photo or 26.45 cm2 PSR2 on the ground. Areas of polygons of any configuration can then be cut out and weighed. Each gram will be equivalent to 26.45 cm2 PSR2 ground area. Suppose we have interpreted and delineated the units of interest on a set of photographs. The units can then be cut out and sorted by category of information (e.g., all the polygons that represent bare soil are cut and put in the same pile, all the polygons that represent forested areas are cut out and put in a separate pile, etc.). Then, by weighing the piles of paper for each category, the total area of each category can be determined and the percentage area may be compared to the entire project area. The total project area will also serve for checking against the total weights of the individual categories. Example 7.21 suppose we interpreted an aerial photograph at a nominal scale of 1:20,000 for a land use project. If the aerial photograph is 23 cm by 23 cm and weighs 20 grams, find the ratio of area to the unit weight and the area of a lake that was cut out and found to weigh 4 grams. Solution: 20 g is equivalent to an area of 23 cm 23 cm = 529 cm2 on the photograph or 529 (20,000)2 cm2 = 2116 108 cm2 on the ground = 2116 104 m2 on the ground (because 1 m2 = 10,000 cm2) = 2116 ha on the ground (because 1 ha = 10,000 m2). If 20 grams = 2116 ha, therefore, 1 gram = 105.8 ha, which represents the ratio of area to the unit weight. If the polygon corresponding to the lake weighs 4 grams, its surface is: 4 105.8 = 423.2 ha or 20 % ([423.2 2116] 100) of the entire area covered by the photo. The same procedure can be applied for the rest of the categories interpreted on the photograph or a set of photographs or a map. This simple technique was developed by the U.S. Soil Conservation Service and has been found to be both efficient and accurate. The accuracy is of course dependent, among others, upon the homogeneity of the thickness of the paper, the precision of cutting the polygons, and the scale used to weigh the paper.

7.10.4. Area by transects

The transect method is based on the assumption that linear measurements vary directly as do areal measurements, provided the sampling technique is unbiased and the number of samples is sufficient. Like the weighing method, the transect method is basically a technique for proportioning the total acreage of a known area among the constituent categories of unknown areas. The transect method uses regularly spaced lines that are drawn directly on the area of interest or on a transparent overlay that can be laid on top of the area (figure 7.21). Using a graduated scale, the distance of all the segments passing through each unit are measured to the nearest desired unit (e.g., 0.1 mm) and added up for each category constituting the area of interest. The area of each category is then

determined by relating the total measure of a given area to the total linear distance covering the area:

where a is the area of the category to be determined, A is the total area of all the categories (i.e., the project area), l is the sum of all the segments (transects) passing through the category, and L is the sum of all the transects inside the project area considered. For example, in figure 7.21, L = Li = L1 + L2 + ... +L11, l = li = l1 + l2 + l3 + l4, A = x y, and a = (A l) L.

Figure 7.21. Determining polygon area from total area using the transect technique. For example, if 15 equally spaced and parallel lines were overlaid parallel to the side of a 23 cm by 23 cm photograph, the total transect length would be 23 cm x 15 = 345 cm. If wetlands areas were crossed for a total of 146 cm, this category would occupy an area of 146 345 or 42.32 percent of the total area of the 23 cm by 23 cm photograph. On a

1:15840 scale photograph, this percentage would be equivalent to: 0.4232 (0.23 0.23) m2 158402 = 561.7 104 m2 or 561.7 ha. The number of and spacing between transects used depend upon the size and arrangement of the polygons to be measured; the higher the number, the more accurate the measurements. Generally, at least 10 transects should be used and a higher number for accuracy, depending on the size of the area to be measured. The transect lines may randomly be laid across the area to be measured. However, for regularly shaped units (e.g., rectangular or linear), the transects lines should be laid across the short dimensions to obtain the best sample (figure 7.22). The transect lines must not be laid parallel to the boundaries of long polygons.

Fig ure 7.2 2. Dia gra

m illu stra tin g the pro per ori ent atio n of the tra nse cts in res pec t to the ori ent atio n of the pol yg ons ; the tra nse cts mu st not par alle l any lon g stra igh t

pol yg on bo un dar y. One of the procedures of using the transect method may be as follows: Starting about half a centimeter from an appropriate boundary (so that the transect lines are not parallel to the sides of long polygons as described in the previous paragraph) of the area to be measured, transect lines are drawn 1 cm apart until the opposite boundary of the area (figure 7.23). The same process may be performed by overlaying a transparency on which the transects lines are already drawn. Next, the line segments intercepting the boundaries of the polygons are measured and the transect line distances are summarized for each polygon and each category (Table 7.1). The sum of proportions of all categories can then be verified to make sure they total one hundred percent (or 1.00) and the sum of the areas of individual categories totals the total area of the entire project area considered.

Ta ble 7.1 . Co ntr ol tabl e for are a det er mi nati on of ind ivi dua l cat ego ries

usi ng the tra nse ct tec hni que .

Pol Tot Pro Are Cor Adj ygo al port a of recti uste n dist ion a on d type anc to cate fact area (1) e in total gor or (6) a dist y cate anc (4) (5) gor e in y the (2) enti re area (3)

A B C . . Z

liA liB liC . . liZ

Equ Equ (4) liA / atio atio + n n (5) Li 7.27 7.28 = = LiB / = = = Li = = = = = = = = l / Li . . liZ / Li
iC

EL 100 PAT %

TA

Figure 7.26. Illustration of (a) Mechanical planimeter, (b) Electronic planimeter, and (c) schematic representati on of a mechanical planimeter. Verification of the paper plane Measurements from paper may be subject to errors introduced from shrinkage or dilatation. When making measurements from a grided map, errors may also be introduced from defective griding. In either case, these errors must be known prior to area measurement. The dimensions of the grid on a map are usually known and errors due to

paper deformation or defective griding may be determined by measuring the dimensions of the grid on the map and comparing them to the actual dimensions. For example, if a 100 mm by 100 mm grid were measured to be 100.2 mm and 100.3 mm, its area is 10050.06 mm2. Therefore, a correction factor of -50 mm2 per 10000 mm2 (or -5/1000) should be applied to the measured surfaces. Determining areas using a planimeter Before making measurement with a planimeter, the following issues must be taken into account. 1. The paper should be taped down on a horizontal table to ovoid displacement of the paper or the sliding of the planimeter. 2. The drawing paper should be smooth and free from cuts and wrinkles as these anomalies disturb the planimeter's maneuver. 3. The pivot point, P, should be placed so that the polar arm, PB, is at a right angle with the tracing arm, BT, when T is placed in the middle of the polygon to be measured (figure 7.27). 4. The base circle must be outside the polygon being measured and all the planimeter's components must be on the drawing paper. If the polygon is too big, it may be arbitrarily divided into smaller surfaces, which are measured separately and added up. Although it is possible to measure large areas with the base circle inside the polygon being measured, this method is more complicated than dividing the polygon into smaller areas so that the base circle is outside the area measured. 5. In a given tracing direction, the angle PBT must not exceed 180o during the rotation of the wheel. This situation is shown in figure 7.28, which exhibits an incompatible positions of PBT and PB'T'. 6. The roller edge (i.e., measuring wheel) should not be touched at any time as this may introduce errors in measurements and/or cause it to rust or corrode, which will adversely affect the measuring accuracy.

Figure 7.27. Proper positioning of the planimeter for area measurement. The polar arm, PB, should be perpendicular to the tracing arm, BT, when the tracing point is approximately at the center of the polygon to be measured.

Figure 7.28. Diagram showing an improper positioning of the polar and the tracer arms because the position, from B to B', of these two components exceeds 180o. Starting at point A (chosen by the operator), an initial reading is taken on the recorder, R, then the tracing point T is moved progressively and carefully around the perimeter until it returns to the starting point A and a second reading is taken on the recorder. As the tracing point T is moved around the perimeter of the polygon, the wheel, W, combine sliding, positive and negative rotations to give the algebraic sum of the rotation, n, which is registered on the recorder. The difference between the initial and the final reading is the actual planimeter reading, n. For convenience, it is usually preferable to set the starting point to zero so that the final reading on the recorder is the planimeter reading, n. For accurate measurements, it is recommended that the planimetric reading be reiterated at least twice. The area, PA, of the polygon is obtained by multiplying the difference between the final and the initial readings by the scale coefficient, k: PA = n k (7.40) The value of n is a function of the radius of the wheel and the polar arm of the planimeter and the scale factor, k, is a function of the scale of the plane containing the polygon being measured. The scale factor also varies with the polar arm of the planimeter and is usually

provided with the planimeter. A table for unit area may be presented as indicated in table 7.3. Table 7.3. Table unit for area in metric scale for two tracer arm lengths at different scales.

Tracer Arm Length

149.5

116.2

Scale 1:100 1:200 1:300 1:400 1:500 1:600 1:1000 1:1500 1:2000 1:3000 1:5000 1:6000 1:10000 1:20000 1:24000 1:25000 1:30000 1:50000 Example 7.27

Unit Area 0.1 m2 0.4 m2 0.9 m2 1.6 m2 2.5 m2 3.6 m2 10 m2 22.5 m2 40 m2 90 m2 250 m2 360 m2 1000 m2 4000 m2 5760 m2 6250 m2 9000 m2 25000 m2

Unit Area 0.08 m2 0.32 m2 0.72 m2 1.28 m2 2 m2 2.88 m2 8 m2 18 m2 32 m2 72 m2 200 m2 288 m2 800 m2 3200 m2 4608 m2 5000 m2 7200 m2 20000 m2

Assume in figure 7.27, the scale is 1:24,000. After tracing the perimeter of the polygon the difference between the final and the initial readings of the planimeter was found to be 1265. What is the ground area of the polygon? Solution: From the table 7.3, if the scale coefficient is 5760 m2 (tracer arm length = 149.5) the area is: 1265 5760 m2 = 7286400 m2 or 728.64 ha if the scale coefficient is 4608 m2 (tracer arm length = 149.5) the area is: 1265 4608 m2 = 5829120 m2 or 582.91 ha REFERENCES American Society of Photogrammetry. 1966. Manual of Photogrammetry, 3rd edition. Menasha, Wisconsin. George Banta Co., Chapters 2 and 10. Anonymous. 1988. How to Use Aerial Photographs in Natural Resource Applications. United States Department of Agriculture, Forest Service, Nationwide Forestry Applications Program, November, 1988. Chapters 7, 8, and 9. Avery, E. T. and L. G. Berlin. 1985. Interpretation of Aerial Photographs. Fourth Edition, Burgess Publishing Company, Minneapolis, Minnesota, Chapters 3 and 4. Axelson, Hans. 1956. Effect of photo scale on the use of aerial photographs in Swedish forestry. Norrlands Skogsvrdsfrbunds Tidskrift 3: 255-292. Barrett, J. P. and J. S. Philbrook. 1970. Dot grid area estimates: precision by repeated trials. Journal of Forestry 68: 149-151, illus. Hager, R. K., and W. Nagy. 1955. A method of determining scale of vertical aerial photos. Journal of Forestry 53: 588-589. Howard, J. A. 1970. Aerial photo-ecology. American Elsvier Publishing Company, Inc., New York. N. Y., Chapter 7. Husch, B., C. I. Miller, and T. W. Beers. 1972. Forest Mensuration. Second Edition, John Wiley & Sons, New York. Chapter 6. Landis, G. H. and H. A. Meyer. 1954. The accuracy of Scale Determination on Aerial Photographs. Journal of Forestry 52: 863. Lane, Benjamin B., JR. 1950. Scales of oblique photographs. Photogrammetric Engineering 16: 409-414 Lillesand, Thomas M. and R. W. Kiefer. 1987. Remote Sensing and Image Interpretation. Second Edition, John Wiley & Sons, New York, N.Y., Chapter 2. Moffitt, Francis H. and E. M. Mikhail. 1980. Photogrammetry. Third Edition. Harper & Row, Publishers, New York, Chapter 7. Paine, D. P. 1981. Aerial Photography and Image Interpretation for Resource Management. John Wiley & Sons, New York, N.Y. Chapters 4 and 5. Pope, Robert B. 1957. The effect of photo scale on the accuracy of forestry measurements. Photogrammetric Engineering 23: 869-873. Spurr, S. H. 1960. Photogrammetry and Photo-interpretation. Second Edition. The Ronald Press Company, New York, Chapters 5, 6, and 10.

Trorey, Lyle G. 1950. Handbook of Aerial Mapping and Photogrammetry. Cambridge University Press, London. 178 pp. Ulliman, J. J. 1995. Aerial Photo Interpretation of Renewable Natural Resources: Course Notes and Workbook, 170 p. Department of Forest Resources, University of Idaho, Moscow, Idaho 83844. Wolf, Paul R. 1983. Elements of Photogrammetry with Air Photo Interpretation and Remote Sensing. Second Edition. McGraw-Hill Book Company. New York, N.Y., Chapters 6. Practice Problems Part I: Scale of a vertical photograph Vertical photographs 8.1. Define the followings: Photo scale; Point scale; Average scale; Nominal scale 8.2. How do you express scale? 8.3. What are the major factors that affect scale? Explain how you can reduce the effect of these factors on photo scale. 8.4. A vertical photograph was taken at 3000 m above a soccer filed that is 350 m above mean sea level. If the focal length of the camera used to take the photograph is 152.4 mm, what is the scale of a point on the photograph whose ground elevation is 500 m above mean sea level? Express the scale in representative fraction (RF) and unit equivalents (UE). 8.5. The average elevation of a vertical photograph is 300 m above mean sea level and its scale is 1 mm = 15 m. How high above mean sea level was the aircraft flying when taking the photograph with a 210-mm focal length camera? 8.6. The average ground elevation between two points is 280 m and the distance, measured on a vertical photograph, between these two points is 8.65 cm. The same 2 points were identified on a 1:24000 scale map and the distance was measured to be 2.73 cm. If the photograph was taken with a 152.4-mm focal length camera, find the scale of the photograph and the flying height above mean sea level at which the photograph was taken. 8.7. The distance between two points, measured on a vertical photograph, is 5.21 cm. The distance between these same two points, measured on a 1:50000 scale map, was found to be 1.43 cm. The average ground elevation between the two points is 300 m above mean sea level. Find the flying height at which the photograph was taken if the focal length used is 152.4 mm. 8.8. Vertical photographs was taken with a 152.4-mm focal length camera over an area whose elevation ranges between 280 m and 520 m. The smallest scale requested for the project is 1:10000. How high above mean sea level must the aircraft fly to meet this scale requirement and what will be the largest scale in the project? 8.9. A vertical photograph was taken at a flying height of 3500 m above datum. The distance between two points was measured to be 6.25 cm on the photograph and the corresponding distance on the ground is 1 kilometer. What was the focal length used to take the photograph?

8.10. On a 7.5-minute U.S.G.S topographic map, the distance between 2 points was measured to be 5.15 cm. On a photograph, taken with a 152.4-mm focal length camera, the distance between these same two points measures 8.65 cm. Find the flying height above datum at which the photograph was taken. 8.11. The distance between two points is 12.83 cm on a vertical photograph and 3.32 cm on a map, whose scale is 1:50000. What is the corresponding distance between these same two points on the ground and what is the scale of the photograph expressed in RF and UE? 8.12. Two points, A and B, are 450 m and 585 m above mean sea level, respectively. The average scale between these two points on a photograph taken with a 152.4-mm focal length camera is 1 mm = 12 m. What would be the scale of a point on the photograph whose elevation is 300 m above mean sea level? 8.13. A 152.4-mm focal length camera was used to take a vertical photograph. A 1.5 km road section was imaged as 4.78 cm on the photograph. What is the flying height above datum at which the photograph was taken? 8.14. The ground distance of a straight boundary of a corn field is 850 m. The field is nearly level and lies beside the principal point of the photo. The distance on the photo of this boundary is 5.15 cm. a) What is the scale of the photograph? b) Express this scale in unit equivalents in centimeter per kilometer and inch per mile. c) If the focal length of the camera used to take the photograph is 305 mm, how high was the aircraft above the ground when taking the photograph? 8.15. A vertical aerial photograph was taken from 4500 m above MSL using a 305-mm focal length. What would be the variation in the photo scale if the ground elevation of the photograph varies between 270 m and 730 m? Find the average scale of the photograph and the scale of the top of a hill that is 580 m above MSL. 8.16. The image of a redwood log was identified on a photograph and measured to be 1.52 cm. The log was also measured on the ground and was found to be 182.5 m. How high was the aircraft above the log when taking the photograph if the focal length used is 152.4 mm? 8.17. A photograph was taken with a 152.4-mm focal length camera from 3800 m above MSL. Four points, a, b, c, and d were identified on the photograph and their ground elevations were found to be A = 460 m, B = 510 m, C = 395, and D = 615. Find the scale at each point and the average scale of the photograph. Conclude. 8.18. Two points A and B whose elevations are hA = 335 m and hB = 516 m above mean sea level are imaged at points a and b on a vertical photograph. The photocoordinates of a and b are xa = -5.12 cm, ya = -6.54 cm, xb = -5.47 cm, and yb = 2.73 cm. The photograph was taken with a 152.4-mm focal length camera and the ground distance between A and B is 2150 m. Find the flying height from which the photograph was taken. 8.19. A road section that is 1350 m long on the ground begins at point A, whose elevation is hA = 385 m, and ends at point B, whose elevation is hB= 575 m above mean sea level. These two points are imaged at points a and b on a vertical photo that was taken with a 152.4-mm focal length. The photocoordinates of points a and b are xa = -19.85 mm, ya =

73.36 mm, xb = -27.37 mm, and yb = -98.16 mm. What is the flying height the photograph was taken from? 8.20. On a 1:24,000 topographic map, the distance between two road intersections was measured to be 51.2 mm. The two road intersections were identified on a vertical aerial photograph and the distance between these two points was measured to be 122.3 mm. Find the scale of the photograph and the ground distance between the two road intersections. 8.21. A 152.4-mm focal length camera was used to take a vertical aerial photograph from 3260 m above MSL. What is the scale of the top of a hill that is 620 m above MSL? 8.22. In your aerial photography laboratory, find three photographs at different scales and compute the average scale of each photograph using the topographic maps (measure at least three distances on each photograph to accurately determine the average scale of these photos). 8.23. A soccer field that is 120 m long measures 15.3 mm on a vertical photograph. What is the scale of the photograph? How high above the ground was the aircraft flying when taking the photograph if a 152.4-mm focal length camera were used. 8.24. A rectangular corn field was measured on a vertical aerial photograph (photo 1) to be 5.26 cm long. This same field was identified on a 1:10,000 photograph (photo 2) on which the same length was measured to be 8.72 cm. Find the scale of photo 1. 8.25. Two road intersections were identified on a vertical aerial photograph (photo 1) and were measured to be 4.73 cm apart. The same distance was measured on a second photograph (photo 2) that was taken with a 210-mm focal length camera from 3280 m above the ground. Find the scale of photo 1 and the ground distance between the two road intersections. 8.26. Two points, a and b, are 6.35 cm apart on photo 1 and 8.32 cm on photo 2. Photo 2 was taken with a 152.4-mm focal length camera from 4260 m above MSL. If a is 820 m and b is 680 m above MSL, what is the average scale of photo 1? If photo 1 were taken with a 210- mm focal length, from what altitude above the ground was this photo taken? 8.27. A football filed was measured to be 0.3528 cm2 on a vertical aerial photograph. The ground dimensions of the field are 50 m by 100 m. Find the scale of the photograph. 8.28. A ten-hectare cotton filed was delineated on a vertical aerial photograph and its photographic area was measured to be 3.24 cm2. Find the scale of the photograph and the altitude above the ground from which the photograph was taken if the a 152.4-mm focal length was used. Tilted photographs 8.29. How does tilt affect the scale of a photograph? 8.30. What can you do to minimize the effect of tilt on aerial photograph? 8.31. A 3o tilt aerial photo was taken from 3800 m above the ground using a 210-mm focal length. What is the scale of the photo at the principal point, at 8.32 cm from the principal point on the up side of the tilt, and at 6.85 cm from the principal point on the down side of the tilt? 8.32. Repeat problem 8.31 if the tilt of the photograph was 2o45'.

8.33. A tilted aerial photo was taken from 3680 m above the ground using a 152.4-mm focal length. The scale of the photo was found to be 1:24890 at 10.43 cm from the principal point on the up side of the tilt and 1:23530 at 9.16 cm from the principal point on the down side of the tilt. By how much was the camera tilted from the vertical when taking the photo? 8.34. In problem 8.33, determine the scale at the same points (10.43 cm and 9.16 cm from the principal point) if the tilt were 1o56'. Conclude. 8.35. Repeat problem 8.33 if the photograph were taken with a focal length of 210 mm instead of 152.4 mm. Conclude. 8.36. Repeat problem 8.33 if the photo were taken from 2800 m above the ground. Conclude. Oblique photographs 8.37. A 152.4-mm focal length was used to take an oblique aerial photograph from 4260 m above the ground. Find the photo scale at the isoline (i.e., the line passing through the isocenter, which is a point halfway between the principal point and the nadir). How would the scales at the principal point and at the nadir compare to this scale (i.e., at the isocenter)? 8.38. An oblique photo was taken from 3500 m above the ground with a depression angle of = 33o and a focal length of 152.4 mm. Find the scale of the photo along the principal line (i.e., the line perpendicular to the principal meridian and passing through the principal point). 8.39. The scale, at the principal line, of an oblique aerial photograph is 1:15,000. If the photograph were taken with a 210-mm focal length camera inclined 66o30' from the vertical, how high above the ground was the aircraft flying when taking the photograph? 8.40. A 152.4-mm focal length camera, inclined 31o40' from the true horizon, was used to take an oblique photograph from 4260 m above the ground. Two points were identified on the photograph and their distances from the true horizon were measured to be 12.73 cm and 19.37 cm. Find the average scale between these two points. 8.41. The average scale between two points that are 21.16 cm and 13.93 cm from the true horizon on an oblique photograph is 1:14560. If the photograph were taken with a 210mm focal length camera inclined 57o25' from the vertical, how high above the ground was the aircraft flying when taking the photograph? 8.42. In problem 8.41, if the two points were 19.75 cm and 12.14 cm from the true horizon, the scale of the photograph were 1:13600, the focal length were 152.4 mm, and the height of the aircraft above the ground were 1560 m above the ground, by how much was the camera inclined from the vertical when taking the photograph? Part II: Area measurements 8.43. What affects the accuracy of area measurements? Explain. 8.44. How does slope affect area measurements on aerial photographs?

8.45. What are the different techniques for determining area on aerial photographs? Briefly explain the principle of each technique and state its advantages and disadvantages. 8.46. A 5 sided polygon was delineated on a 1:15840 vertical photograph. Taking the photo center as the coordinate origin, the photo coordinates of the polygon vertices are: x1 = 23.8 mm, y1 = 42.6 mm, x2 = 32.4 mm, y2 = 61.2 mm, x3 = 63.7 mm, y3 = 53.6 mm, x4 = 63.2 mm, y4 = 29.8 mm, x5 = 32.1 mm, and y5 = 18.1 mm. Find the area of the polygon using the coordinate method. 8.47. A forester desires to estimate the total area of forested land on a U.S. Geological Survey 7.5- minute map covering 18292 hectares. The forested areas were delineated on the map and 40 equally spaced transect lines were drawn horizontally (i.e., east-west) on the map, which is 32.5 cm wide. The total length of all the transects passing through the forested areas was measured to be 730 cm. Find the total area of the forested land on the 7.5-minute map. 8.48. Find a set of three photographs in your laboratory and, using the appropriate procedure you have learned through your photo-interpretation class, determine the effective area on the central photo of your set. Next, using the photo-interpretation techniques, interpret and delineate the different classes inside the effective area. Finally, determine the area of each class category using the transect, the dot grid, and the planimeter methods. Use tables 8.1 and 8.2 as models for your calculation. Compare the three techniques 8.49. Lay down a dot grid transparency on top of a section of a U.S.G.S 7.5-minute map and determine the area, in hectares, of the section. 8.50. In Problem 8.49, if the area of the section were 259 hectares, how accurate was your area determination of the section in problem 8.49? Use table 8.2 to correct for the discrepancies. 8.51. A 10-dot-per cm2 dot grid is overlaid on top of a polygon, which is delineated on a 1:15840 photograph. If 127 dots are counted within the polygon, what is its area in hectares? 8.52. Assume the total area of the rectangle bellow is 460 hectares. Use a dot grid to determine the area of polygons A, B, C, D, and E by the total known area technique.

8.53. Repeat problem 8.52 using the transect technique. 8.54. In problem 8.52, assume the total area of the rectangle is unknown, but the average scale of the map is 1:20,000. Use a 10-dot-per-cm2 dot grid to determine the area of the five polygons. 8.55. It is desired to determine the number of hectares of wetlands in a project area containing 1000 23 cm by 23 cm vertical aerial photographs at 1:15800. The effective area of each photograph was determined and cut out to make a mosaic of the project area. Each effective area was found to represent 30 percent of the total area of the 23 cm by 23 cm photograph. The wetlands were delineated and were found to represent 23% of the entire area covered by the mosaic formed by the effective areas. If one photograph weighs 16 grams, find the total area of the wetland areas. 8.56. One section of a U.S.G.S. 7.5-minute map was identified on a 1:15,000 photograph. Three fifth of the section was interpreted to be forest land. Find the area of this forest land if the 23 cm by 23 cm photograph weighs 16 grams. 8.57. Starting one half cm from the edge of the rectangle bellow, properly draw 10 equally spaced transect lines and determine the area of polygons A, B, and C if the scale of the map is 1:12,000.

8.58. Repeat problem 8.57 using a dot grid. Compare the results. 8.59. Assuming the scale of the diagram in problem 8.57 is 1:14500. Use a known scale dot grid and the corresponding calculation formulae to determine the area of the three polygons.

Lecture 8: Stereoscopy
Prof. John E. Estes Michael J. Lawless Remote Sensing Research Unit Department of Geography University of California Santa Barbara, California 93106-4060 Telephone: (805) 893-3845 Fax: (805) 893-3703

Stereoscopy
Stereoscopy is the science and art that deals with the use of binocular vision for the observation of overlapping photographs or other perspective views and the method by which such views are produced. Essentially most of us with normal eyesight have stereoscopic vision (i.e. The ability to see and appreciate depth of field through the perception of parallax.)

Monocular Depth Perception

There are clues to the depth of field of objects which can be perceived/appreciated monocularily. Monocular or one eye depth perception deals with in-born cues that humans instinctively employ when viewing objects. These are things we do we dont normally think about. Such cues or clues to the distance we think an object is at include the fact that close objects typically appear: 1. Larger; 2. Brighter; 3. To be seen in more detail; and, 4. To obscure the view of objects that are more distant. There are limits, however, to monocular depth perception. Put a coin on end on a table and view it with only one eye. You see only the edge of the coin. The same can be said for looking straight on at other objects. A chimney viewed from directly above can appear to be as an o. Is it a well; or is it a smokestack?

Binocular Vision
With binocular vision each eye sees a different picture and the brain fuses the two images into one. {Talk about the dominant eye, and experiment with glasses that turn the world up-side-down.} Stereoscopic vision is not an all or nothing affair. Individuals, depending on a number of factors (e.g. interpupillary distance, near or far sightedness, etc.) have stereoscopic vision to varying degrees. It is best if interpreters who do a lot of work with stereo have nearly equal vision in both eyes. Don't worry if you don't have equal eye strength. Very few interpreters work only with stereo, or with stereo all the time. The people who do this type of work are the photogrammetric engineers who work all day on stereo plotting equipment.

Depth Perception
The normal interpupillary distance in humans is 2.5 to 2.6 inches. If we could increase this distance we would increase our perception of depth. Stereo pairs greatly stretch this normal eye base (interpupillary distance) and give up the exaggerated 3-D photographic effect we perceive when viewing the stereo pairs.

Figure 8-1. Stereo Eyes

Figure 8-2. Stereo Photos Now, normally when viewing objects greater than 1,500 feet it 2,000 feet away the special ability to perceive depth is essentially lost. Look, say at a far off line of trees on a flat plane. It is very difficult to tell if they are really in a line or staggered. Basically what your eye brain mechanism will do in this case is employ other clues to help determine the depth of the trees (again, these are essentially the monocular cues described above). With normal eyes from an aerial view we would see a tree like this; while with a stereo pair it would appear like this. From this you can begin to see the difference between unaided stereo viewing and viewing objects and features in the environment with the aid of stereo pairs. In normal unaided stereo the lines of sight converge, and it is this convergence that gives us depth perception. Basically, then when our eyes focus on a near by object they also converge so that the lines of sight from each eye intersect at the object. However, in stereo photo vision our lines of sight are essentially parallel or focused at infinity. Now,

because of a process called accommodation we have a problem. Accommodation relates to the change in focus of the eye with distance. Basically, convergence and accommodation go together. As the eye focuses on an object they also turn (or move) so that the lines of sight intersect at the object. The issue (or problem if you will) with stereo vision is that to get stereo we must maintain parallel lines of sight while focusing our eyes at close range. This is not a normal situation and may cause strain on the eyes.

Figure 8-3. Sausage Exercise The "Sausage Exercise" can be helpful in developing the ability to see stereo. Essentially in this exercise you focus your eyes on a distant object and then slowly bring your forefingers into the line of vision. The farther apart your fingers and the larger the sausage when it forms the more nearly parallel are your lines of sight.

Proper Use of Stereoscopes

The following are some guidelines that will help you use your stereoscopes properly. They are important and should be kept in mind when performing stereo based interpretations: Be sure that the photos are properly aligned, preferably with the shadows toward the viewer. (Having the shadows away from the viewer can cause terrain reversal or "false stereo".) 1. Be sure to keep the eye base and the long axis of the stereoscope parallel to the flight line. 2. Try to maintain an even glare free illumination of the images and make yourself comfortable. 3. Keep the lenses of your stereoscope clean. 4. Do Not Attempt Stereoviewing For Long Periods in the Beginning.

Problems/Issues Involved in Viewing Stereo

1. People with eyes of unequal strength may have difficulty seeing stereo. If you wear glasses it is best that they be worn when stereo viewing. 2. Poor photographic illumination, misalaigned prints or uncomfortable viewing positions may result in eye fatigue. 3. Illness or severe emotional distress may create sensations of dizziness when using strereoscopes. 4. Pseudo stereo can be caused by: 1. Erroneous reversal of prints; 2. Viewing photos with shadows falling away rather than towards the interpreter. 5. Objects which change position from one photo to another cannot be viewed in stereo. 6. In areas of high vertical relief, scale differences in adjacent photos may make it difficult to obtain a 3-D effect. 7. Dark shadows or clouds may prohibit stereoscopic study by obscuring details of the scene on one print or the other. 8. Individuals who have difficulty with stereoscopic viewing, should not attempt unaided eye stereo viewing.

Lecture 9: Height Measurement

Prof. John E. Estes Michael J. Lawless Remote Sensing Research Unit Department of Geography University of California Santa Barbara, California 93106-4060 Telephone: (805) 893-3845 Fax: (805) 893-3703

Height Measurement
There are a number of methods for measuring the heights of objects using aerial photography. I will briefly discuss three that I feel are important for photo interpreters to be familiar with. Two deal with the measurement of height from a single photo; while the third deals with the determination of height by measuring parallax differences. The types of parallax being measured here are: 1. Absolute Parallax; and, 2. Differential Parallax. Remember that according to the American Society of Photogrammetry and Remote Sensing's Manual of Photogrammetry, 3rd.ed.: Parallax: The apparent displacement of the position of a body, with respect to a reference point or system, caused by a shift in the point of observation. Absolute Parallax: Considering a pair of aerial photographs of equal principle distance, the absolute parallax is the difference of a point is the algebraic difference of the distances of the two images from their respective photo nadirs measured in a horizontal plane and parallel to the air base. Differential Parallax: The difference in the absolute stereoscopic parallaxes of two points imaged on a pair of photographs. This is usually employed in the determination of the differences in the elevation of objects.

Single photo methods of height determination

Shadow Height Method:
Basically, if the shadow cast by an object can be measured and the sun angle causing the shadow is known or can be derived (from latitude, date and time) then the height of the object can be calculated using simple trigonometry, as follows: h = Ls x tan Where: tan = the tangent of the sun angle from the ground surface Ls = Length of the shadow. Here we assume that the shadow on which the ground falls is level and that the object is vertical. The object's top must be sharply defined so that it creates a distinct image.

There are many sources of error here, shadows are not always distinct, and the calculation of the sun angle is an involved process.

Displacement Method:
In this method of height determination from a single aerial photo we: 1. Accept the principle point as the photo nadir (were assuming a true vertical photo); 2. Must precisely know or be able to determine the altitude from which the photo was acquired. 3. Both the top and the bottom of the object to be measured should be clearly visible. 4. The degree of image displacement must be great enough to be accurately measured with available equipment. If the above conditions are met, the formula for the displacement method of height determination from a single aerial aerial photo can be written as: Ho = Ha x D R Where: Ho = Height of the object; Ha = Altitude above the surface where the photo is taken; D = Length of the displaced image;

Figure 9-1. Displacement Method R = Radial distance from the photo nadir to the top of the object. Basically, its important that you know that these techniques exist. That they can be accomplished in a practical fashion if necessary. If you ever have to do this you can always go back and cookbook it. Parallax Height Measurement This is the most used method of measuring heights on air photos. There are many forms of the parallax equations. Avery and Berlin give one; Paine in his book lists three: 1) for mountainous terrain; 2) for level terrain; and, 3) the short cut equation. What I will give here is the basic form of the equation: Ho = Ha x dP Pb + dP Where: Ho = The height of the object of interest; Ha = Platform height or altitude above datum; dP = Differential Parallax; and Pb = Absolute Parallax.

So if the altitude of the aircraft above datum is: 1. known or can be calculated; and, 2. if, from the available stereo pairs, we can calculate the differential and the absolute parallax; then, 3. We can ascertain the heights of objects in the photos. Important things to remember here include: Ha, the height of the aircraft should be in the same units as the objects height. Feet are typically employed. dP and Pb, should also be in the same units. Yet, here you would typically use hundredths of inches. you need accurate measuring devices to get accurate measurements.

Now for a small trick. If we can assume that: 1. Photo tilt is less than 3; 2. Both negatives or positive transparencies of the stereo pair were taken from the same flying height; 3. Both nadirs and principle points are at essentially the same ground elevation; and, 4. The base of the objects to be measured are, essentially, at the same elevation as that of the principle point. Then, we can substitute the average photobase of the stereo pairs being used can be substituted for Pb (absolute parallax). Let's say, as Avery and Berlin do in their book (5th. ed.) on Pg. 78 and 79; that we are going to measure the height of the Washington Monument from a stereo pair. The nominal photo scale we have is 1:4,800. We have gone in and corrected this in the area of the monument monument to 1:4,600 at the base of the Monument. The camera focal length was 12 inches. So the flying height was what? 4,600 feet. The average photobase (P) of the stereopair is calculated to be: 4.40 inches. Absolute stereo parallax at the base; and at the top of the monument is measured parallel to the line of flight with an engineers scale. The difference is: 2.06 in. - 1.46 in. This gives a dP of 0.60 in. So, 0.60 inches is the differential parallax of the displaced images. Substituting these values into a form of our formula: Ho = [H] dP Pb + dP Ho = 4,600 Ft. 0.60 in. 4.40 in. - 0.60 in. The actual height of the Washington Monument is 555 feet. This is a very accurate measurement for this type of exercise. For example if we had used the nominal photo scale of 1: 4,800; instead of the corrected 1: 4,600 scale we would have gotten a height of 576 feet. A 21 foot error as opposed to the 3 foot error we did get. It just goes to show that the more time you put in setting up the problem and the more precise the instruments are the better ( up to a point) the measurements that you can achieve. Well that's essentially it for stereo and height measurement. = 552 feet