Optical Flow Based Moving Object Detection and Tracking System

Moving object detection in digital image sequence involves identification of the presence of an object in consecutive frames where as object tracking is used to monitor the movements with respect to the region of interest. In this project, the motion estimation is obtained using Optical Flow. Optical Flow is the distribution of apparent velocities of movement of brightness patterns in an image. Lucas-Kanade algorithm with Sobel, Horn and Guassian smoothing techniques is used in this work for computation of Optical Flow vectors. Single and multiple object movements with respect to the computed vectors are segmented using thresholding. The extracted movements are tracked using edge and centroid information. Suitable image enhancement techniques are applied to the segmented results to avoid the unwanted information present in the image. Real and virtual image data with static and dynamic environment are used as test sequences to validate the developed algorithms. The tracking performance, in terms of their accuracy and computation time, of the different algorithms with and without image pyramid is analysed and compared in MATLAB & C on Intel Core2 Duo processor on Linux environment.