Retrieving unrecognized objects from HSV into jpeg videoat various light resolutions
T. Arumuga Maria Devi
Assistant Professor, Dept. of CITE HOD, SMIEEE, Dept. of CITE P G Scholar, Dept. of CITECentre for Information Technology and EngineeringManonmaniam Sundaranar University, Tirunelveli
This paper deals mainly with the performance study andanalysis of image retrieval techniques for retrieving unrecognized objectsfrom an image using Hyper spectral camera at low light resolution. Sincethe identification of moving object in a camera is not possible in a lowlight environment as the object has low reflectance due to lack of lights.Using Hyper spectral data cubes, each object can be identified on thebasis of object luminosity. Moving object can be identified by identifyingthe variation in frame value. The main work identified are that efficientretrieval of unrecognized objects in an image will be made possible usingHyper spectral analysis and various other methods such as Estimation of Reflectance, Feature and mean shift tracker, Traced feature located onimage, Band pass filter (Background removal) etc. These methods usedabove to retrieve unrecognized object from a low light resolution arefound to be more efficient in comparison with the other image retrievaltechniques.
Anomaly suspect, mean shift algorithms,spectral detection, .
he process of recovering unrecognized objects froman image in low light is a trivial task which finds itsneed in recognizing objects from a distant location.Since there is a need in retrieving unrecognized objects fromthe image, some form of object extraction method from animage is necessary. The application of detecting objects froman image is as follows. Here, we focus on the problem of tracking objects through challenging conditions, such astracking objects at low light where the presence of the objectis difficult to identify. For example, an object which is fastlymoving on a plane surface in an abrupt weather condition isnormally difficult to identify. A new framework thatincorporates emission theory to estimate object reflectanceand the mean shift algorithm to simultaneously track theobject based on its reflectance spectra is proposed. Thecombination of spectral detection and motion predictionenables the tracker to be robust against abrupt motions, andfacilitate fast convergence of the mean shift tracker. Videoimages are moving pictures which are sampled at frequentintervals usually, 25 frames per second and stored as sequenceof frames. A problem, however, is that digital video data ratesare very large, typically in the range of 150
Megabits/second.Data rates of this magnitude would
consume a lot of thebandwidth in transmission, storage
and computing resourcesin the typical personal computer.
Hence, to overcome theseissues, Video Compression
standards have been developedand intensive research is
going on to derive effectivetechniques to eliminate picture redundancy, allowing videoinformation to be transmitted and stored in a compact andefficient manner.A video image consists of a time-orderedsequence of frames of still images as in figure 1. Generally,two types of image frames are defined: Intra-frames (I-frames)and Inter-frames (P- frames). I-frames are treated asindependent key images and P-frames are treated as Predictedframes.An obvious solution to video compression would bepredictive coding of P-frames based on previous frames andcompression is made by coding the residual error.Temporalredundancy removal is included in P-frame coding, whereas I-frame coding performs only spatial redundancy removal. II.
TECHNIQUEThe problem laid in the past decades in identifying theunrecognized objects from a low light resolution. If the imageis created from a hyper spectral camera, the problem still laidin identifying what actually the object was, since the hyperspectral image detects only the presence of an object, not whatan object actually is. Various reflectance  methods wereused in order to obtain the specific property of the image. Butsince the above methods does not specify what the objectproperty was, there should be a method in order to specifywhat the object in an image actually was. Since the imagetaken from a hyper spectral camera suffers from lowresolution, we could not identify what actually the particularobject was, even though it detects the presence of an object.There is a need for image applications in the detection of objects from a distant location. Normally, the image would besuch that the presence of an object could not be detected fromit. But, from a hyper spectral camera, the object, if it was onthat location, could be captured in the hyper spectral camera.Also, an image taken from a hyper spectral camera suffersfrom low resolution and thus does not show the exactproperties of an image. Since the identification of movingobject in a camera is not possible from distant location, toovercome this problem we can use Hyper spectral camera toidentify the object.. Thus, the problem areas are such thatthere should be a methodology in identifying an object from alow light resolution. That
is, it should detect the points from ahyper spectral image which are the points that specify theparticular objects in the image by reflectance mechanisms of the object. The next problem is such that if an
object is fastlymoving on a plane
it is not necessary that the objectwill be present on every frame. The points that resembles theobject in the hyper spectral image should be able to be used inretrieving the objects by using background removal. Relatedto the implementation of transcoding, the work is as follows
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 10, October 2011152http://sites.google.com/site/ijcsis/ISSN 1947-5500