Generating Disparity Maps from Stereo Images Using NCC, SSD, and SAD Algorithms

On the MATLAB platform, stereo images are processed using algorithms such as Normalized Cross-Correlation (NCC), Sum of Squared Differences (SSD), and Sum of Absolute Differences (SAD) to compute disparity maps. These algorithms evaluate similarity between images and calculate disparity values based on the similarity metrics. By combining these algorithms, more accurate disparity maps can be obtained, enabling 3D reconstruction and depth information extraction. In MATLAB implementation, NCC measures normalized correlation between image patches using sliding window techniques, SSD computes pixel-wise squared differences for matching cost calculation, and SAD utilizes absolute difference sums for efficient disparity estimation. Key functions like normxcorr2() for NCC, block matching with sum-squared differences for SSD, and absolute difference operations for SAD are typically employed. The combination approach allows leveraging NCC's illumination invariance, SSD's sensitivity to pixel variations, and SAD's computational efficiency for robust stereo matching.