Detection Tracking and Classification of Object in Forest Surveillance Video Footage

Object tracking is to find the state of the target object in the image sequence. It plays an important role in numerous applications like surveillance, recognition, motion analysis and user interface. Many improvements are made in recent years but the challenging problem is the illumination variation, camera motion, occlusion, shape variation and pose variation. A new proposed approach is provided for efficient object tracking using SDC classifier. This approach first split the given video into key frames, uses SDC classifier to split the frame into positive template and negative template and trained in the classifier after that occlusion in the frame is converted into 8*8 blocks. The SGM classifier uses batches for handling occlusions. When a new video is given as input the video is converted into frames and sent to the SDC classifier the classifier will test the input frame with the trained frame and produce the result. Dynamic appearance changes are also effectively handled by this system. In this we consider animal video to track the animal in the given video. This proposed method efficiently tracks the animals in the given frame when compared with other methods.