MADDA WALABU UNIVERSITY

• COLLEGE OF ENGINEERING
• DEPARTMENT OF SURVEYING ENGINEERING
• COURSE:- Digital image analysis
• Target Group:- 5th year surveying Engineering
students.

02/11/2023 Digital Image Analysis 24/14/2021 1

 What is DIGITAL IMAGE PROCESSING

 An image may be defined as a two-dimensional

function, f(x,y), where x & y are spatial (plane)
coordinates, and the amplitude of f at any pair of
coordinates(x,y) is called the intensity or gray level
of the image at that point.
 When x,y, and the amplitude values of f are all
finite, discrete quantities, we call the image a digital
image. The field of digital image processing refers
to processing digital images by means of a digital
computer.
02/11/2023 Digital Image Analysis 24/14/2021 2
What is DIGITAL IMAGE PROCESSING

• The use of computer algorithms to perform signal

processing on digital images
• It is a subject that deals with NUMERICAL
TRANSFORMATION of images to produce
desired image (improved representation, colour-wise
or theam-wise)

In context of Remote Sensing (RS), DIA would mean

deriving information from remotely sensed data
using a computer.
02/11/2023 Digital Image Analysis 24/14/2021 3
 Why digital image processing?

• To extract information
• To evaluate a sensor
• To evaluate images statistically
• To assess/improve quality
• To make base data sets (e.g. for GIS)
• To make art

02/11/2023 Digital Image Analysis 24/14/2021 4

 DIGITAL IMAGE ANALYSIS/PROCESSING
Advantages :

•Easy in storage and transmission

•Direct use in GIS
•Enhancements and image manipulations
•Various band combinations for MS sensors
•Efficient to detect tonal variations
•Good for large area mapping

02/11/2023 Digital Image Analysis 24/14/2021 5

 DIGITAL IMAGE ANALYSIS/PROCESSING

Disadvantages :
• Difficult to implement Human Expertise
• Lot of resources required

02/11/2023 Digital Image Analysis 24/14/2021 6

 Digital image processing operations
 Geometric transformation
 Color correction
 Image editing
 Image registration
 Image projection
 Compositing
 Image differencing
 Image segmentation
 Pattern recognition

02/11/2023 Digital Image Analysis 24/14/2021 7

 What is an image? (What’s in an image?)
An image is a two-dimensional picture that has the
same appearance to a subject (a person)
 A digital image is a discreet representation of an
image using ones and zeros (binary) as used in computing
A raster image is same as a digital image made up of
a finite number of pixels
A pixel is the smallest individual element in a digital
image holding quantized values of brightness
 A pixel has both a position and a value consisting of
a quantile (sample)

02/11/2023 Digital Image Analysis 24/14/2021 8

 DIGITAL IMAGE

02/11/2023 Digital Image Analysis 24/14/2021 9

 21 38 6 11 100 125

43 55 12 13 89 94
Red Green Blue

02/11/2023 Digital Image Analysis 24/14/2021 10

 DIGITAL IMAGE

02/11/2023 Digital Image Analysis 24/14/2021 11

 Pixels

02/11/2023 Digital Image Analysis 24/14/2021 12

02/11/2023 Digital Image Analysis 24/14/2021 13
• A photograph could also be represented and displayed in a
digital format by subdividing the image into small equal-sized
and shaped areas, called picture elements or pixels, and
representing the brightness of each area with a numeric value
or digital number.

02/11/2023 Digital Image Analysis 24/14/2021 14

 DN Number / Brightness Values

02/11/2023 Digital Image Analysis 24/14/2021 15

 Digital  Image  Data  Formats

 In  order  to  properly  process  remotely  sensed

 data,  the  analyst  must  know  how  the  data  is
 organized  and  stored  on  digital  tapes  and  how
 the  data  are processed  by  computers  and
 software
 Understanding  existing  digital  data  formats  is
 essential before  the  data  can  be  processed.  
 There  are  many  different  data  formats  used for
 storing  digital  remotely  sensed  data.

02/11/2023 Digital Image Analysis 24/14/2021 16

 Cont.
• Many  commercial  data  suppliers  such  as
EOSAT  and  SPOT,  provide  radiometrically
 corrected  data  in  a  customer  specified  format.
•  There  are  four  major  data  formats  used  by
 government  and  commercial data  suppliers:
1 Band  Interleaved  by  Pixel  (BIP)  Format
2 Band  Interleaved  by  Line  (BIL)  Format
3 Band  Sequential  (BSQ)  Format
4 Run-­Length
‐  Encoding  Format

02/11/2023 Digital Image Analysis 24/14/2021 17

Data media & formats for RS Images

- The data are available on practically all media such as

floppies, DAT’s, CD-ROM’s and also through internet

- But the multispectral data are stored in anyone of the three

formats from those four listed above :

1 BIP Band Interleaved by Pixel

2 BIL Band Interleaved by Line
3 BSQ Band Sequential

02/11/2023 Digital Image Analysis 24/14/2021 18

 Band  Interleaved  By  Pixel  Format  (BIP)

• One  of  the  earliest  digital  formats  used  for

 satellite  data  is  band  interleaved  by  pixel
 (BIP)  format.
•  This  format  treats  pixels  as  the  separate
 storage  unit.
•  Brightness  values  for  each  pixel  are  stored
 one  after  another.
•  It  is  practical  to  use  if  all  bands  in   an  image
 are  to  be  used.  
02/11/2023 Digital Image Analysis 24/14/2021 19
 BIP : In this case the data is stored in scan lines and
in each scan line 1st pixel of first band is followed by
data from other bands, sequentially.
Figure  1  shows  the  logic  of  how  the  data  is
 recorded  to  the  computer  tape  in  sequential
 values  for  a  four-­‐band image  in  BIP  format

Record Scan line Pixel

1 1 1 1 2 2 2 2 ....
1 1 b1 b2 b3 b4 b1 b2 b3 b4
2 2 b1 b2 b3 b4 b1 b2 b3 b4 ...
. . . . . . . . . .

02/11/2023 Digital Image Analysis 24/14/2021 20

 Band  Interleaved  By  Pixel  Format  (BIP)
figure 1
Line 1 Pixel 1 Band 1 Line 1 Pixel 2 Band 1 Line 1 Pixel 3 Band 1

Line 1 Pixel 1 Band 2 Line 1 Pixel 2 Band 2 Line 1 Pixel 3 Band 2

Line 1 Pixel 1 Band 3 Line 1 Pixel 2 Band 3 Line 1 Pixel 3 Band 3

Line 1 Pixel 1 Band 4 Line 1 Pixel 1 Band 4 Line 1 Pixel 3 Band 4

 All  four  bands  are  written  to  the  tape  before  values
 for the  next  pixel  are  represented.  
 Any  given  pixel  located  on  the  tape  contains  values  for
 all  four  bands   written  directly  in  sequence.  
02/11/2023 Digital Image Analysis 24/14/2021 21
 21 38 6 11 100 125

43 55 12 13 89 94

21 6 100 38 11 125

43 12 89 55 13 94

BIP

02/11/2023 Digital Image Analysis 24/14/2021 22

 Band  Interleaved  By  Line  Format  (BIL)
 Just  as  the  BIP  format  treats  each  pixel  of
 data  as  the  separate  unit,  the  band
interleaved  by  line  (BIL)  format  is  stored
 by  lines.
  Figure  2 shows  the  logic  of  how  the  data
 is  recorded  to  the  computer  tape in
 sequential  values  for  a  four  band image  in
 BIL  format

02/11/2023 Digital Image Analysis 24/14/2021 23

BIL: In this case all the pixels of a band and scanline
are out first and the equivalent scanline of second band
follows

Record Scanline Pixel

1 2 3 4 5 6 7 8 ....
1 1 b1 b1 b1 b1 b1 b1 b1 b1 ....
2 1 b2 b2 b2 b2 b2 b2 b2 b2 ....
3 1 b3 b3 b3 b3 b3 b3 b3 b3 ....
4 1 b4 b4 b4 b4 b4 b4 b4 b4 ....
5 2 b1 b1 b1 b1 b1 b1 b1 b1 ....
. . . . . . . . . . .... .
02/11/2023 Digital Image Analysis 24/14/2021 24
 Band  Interleaved  By  Line  Format  (BIL)
Line 1 Band 1 Line 1 Band 2 Line 1 Band 3 Line 1 Band 4
Line 2 Band 1 Line 2 Band 2 Line 2 Band 3 Line 2 Band 4
Line 3 Band 1 Line 3 Band 2 Line 3 Band 3 Line 3 Band 4
Line 4 Band 1 Line 4 Band 2 Line 4 Band 3 Line 4 Band 4

• Each  line  is  represented  in  all  four  bands  before  the  next
 line  is  recorded.
•  Like  the  BIP  format,  it  is  a  useful  to  use  if  all  bands  of  the
 imagery  are  to  be  used  in  the  analysis.  
• If  some  bands  are  not  of  interest,  the  format  is  inefficient ,if the
data are on tape,  since  it  is  necessary  to  read  serially  past
 unwanted  data.
02/11/2023 Digital Image Analysis 24/14/2021 25
 21 38

6 11
21 38 6 11 100 125
100 125
43 55 12 13 89 94
43 55

12 13

89 94
02/11/2023 Digital Image Analysis 24/14/2021 BIL 26
 Band  Sequential  Format
 The  band  sequential  format  requires  that  all  data  for  a single
 band  covering  the  entire  scene  be  written  as  one  file.
 Thus,  if  an  analyst  wanted  to   extract  the  area  in  the  center
 of  a  scene  in  four  bands,  it  would  be  necessary  to  read  into
 this  location  in  four  separate  files  to  extract  the  desired
 information.
 Many  researchers  like  this  format  because  it  is  not
 necessary  to  read  serially  past  unwanted  information  if
 certain  bands  are  of  no  value,  especially  when  the  data
 are  on  a  number  of  different  tapes.
 Random-access  optical  disk  technology,  however, makes  this
 serial  argument  obsolete.

02/11/2023 Digital Image Analysis 24/14/2021 27

BSQ: In this format the entire image of 1st band is given
and is then followed by entire image of 2nd band and so on.

Record Scanline Pixel

1 2 3 4 5 6 7 8 9...
1 1 b1 b1 b1 b1 b1 b1 b1 . ....
2 2 b1 b1 b1 b1 b1 b1 b1 . ....
. . . . . . . . . . ...
. . b2 b2 b2 b2 b2 b2 b2 . ....
. . . . . . . . . . . .. .

02/11/2023 Digital Image Analysis 24/14/2021 28

 21 38

43 55

21 38 6 11 100 125 6 11

43 55 12 13 89 94 12 13

100 125

89 94

BSQ
02/11/2023 Digital Image Analysis 24/14/2021 29
 Statistical description of images
 image histogram
 individual pixel values
univariate descriptive statistics
statistics derived from a single variable
multivariate statistics
statistics derived from multiple variables

Image histogram:-The frequency of occurrence

of individual or binned brightness values in an
image
02/11/2023 Digital Image Analysis 24/14/2021 30
 Why do we care about histograms?
 Histograms provide a lot of information on images even
without looking at the images themselves such as
presence or absence of features, distribution etc.
 Histograms help evaluate images statistically e.g. normal,
skewed, bimodal distribution
 Histograms are used in individual image enhancements
 Histograms are used in image classification
 Histograms are used in image segmentation
 Histograms help matching of images across time or space

02/11/2023 Digital Image Analysis 24/14/2021 31

 Histograms of image

02/11/2023 Digital Image Analysis 24/14/2021 32

 Image Statistics
Why extract initial statistics about digital image data?

 Representation (or summary) of data in each spectral

band.

 Assist in image enhancements.

 Selection of bands with maximum information content.

 Eliminate bands with redundant/unnecessary

information.

 Use in other image processing functions (e.g., image

classification.
02/11/2023 Digital Image Analysis 24/14/2021 33
 Minimum
Shows the lowest value of the data. For example, the
satellite has minimum and maximum value which is 0
and 255 respectively. Once you open the properties that
will gives the minimum value information about the data. In
the image, minimum value starts from 0.

02/11/2023 Digital Image Analysis 24/14/2021 34

 Maximum
Shows the highest value of the data. For example, the
DN values of satellite has a maximum range about 255
which is for pure soil. The maximum DN value information
presents in the properties of data.

02/11/2023 Digital Image Analysis 24/14/2021 35

 Mean
Average of all brightness values in the image. If the
histogram shows the DN/pixel values from 0 to 84 then it
will gives the average value. Sum of DN values/ No. of
Pixels.

02/11/2023 Digital Image Analysis 24/14/2021 36

 Median

Value midway (or at midpoint) in the

distribution of brightness values.

02/11/2023 Digital Image Analysis 24/14/2021 37

 Mode

The value that occurs most frequently in a

distribution (usually highest point on
distribution curve of histogram)
Typical of most remotely sensed
spectral image data.

Minimum
Minimum Maximum Maximum

Range Range

02/11/2023 Digital Image Analysis 24/14/2021 38