Target detection and tracking in FLIR videos and images

TARGET DETECTION AND TRACKING IN FLIR VIDEOS AND IMAGES(M. Sc.)Oğuzhan DAĞGAZI UNIVERSITYINSTITUTE OF SCIENCE AND TECHNOLOGYSeptember 2011ABSTRACTThere is a need for target acquisition and tracing from FLIR images in the unmanned air vehicles, simulation practices, air, land and sea platforms. To obtain better images through image processing, various modifications and improvements are made on the image. In this study, acquisition and tracing of moving targets in the real-time FLIR videos and photos have been made by using image processing techniques. Basically, the first procedure is turning of the RGB image into the halftone. In the second step, moving targets have been found for the purpose of detecting the area concerned (ROI). Then the edges of the halftone should be determined by the Canny method so as to realize the process of target acquisition. The boundary areas of the target whose edges have been determined are drawn. Afterwards, target image is detected and a rectangular target-tracking marker is placed on the moving target determined so that tracking can be done. In addition, the routes followed by the objects are indicated. Even in the cases in which the pixel values of the target object cannot be distinguished from the pixel values of the background, algorithm value is successfully detected and tracked.Key Words : Flir, Target Tracking, Canny Edge Detection, Boundingbox