A novel stereo matching algorithm for binocular micro stereo images

A pixel-based stereo matching algorithm that detects discontinuities from a micro stereo image pair is proposed. A dissimilarity function defined in RGB color space determines corresponding points from left (image source) and right images (image target) along epipolar lines. The dissimilarity function is minimized in order to obtain the best set of correspondences within a disparity range that depends on the geometry of the object to be analyzed. The algorithm matches feature points (edges) and non-feature points separately. Occlusions are simultaneously detected and constraints used in the dynamic programming such as ordering and uniqueness constraints are taken into account during the matching process. Experimental results show that the proposed method is fast and efficient for micro stereo images with simple structure consisting of the matching pattern and the background.