Optical Flow Computation from Two Frames in AVI Video Using MATLAB

This MATLAB program extracts two consecutive frames from an AVI video file and performs optical flow analysis to track motion patterns between frames. The implementation includes three distinct optical flow computation methods: the gradient-based Horn-Schunck (HS) algorithm that computes dense flow fields through global energy minimization, the point-tracking Lucas-Kanade (LK) method that assumes constant flow in local neighborhoods using least squares calculation, and block matching technique that identifies motion vectors by comparing rectangular blocks between frames using similarity measures like sum of absolute differences (SAD) or mean squared error (MSE). These algorithms enable comprehensive analysis of object trajectories and velocity variations within the video sequence. The program provides a straightforward yet effective approach for video motion analysis, featuring frame extraction using VideoReader function, image preprocessing with rgb2gray conversion, and customizable parameters for algorithm optimization. This implementation serves as a practical tool for understanding motion dynamics in computer vision applications.