A Stereo Matching Algorithm Based on Fuzzy Identification
基于模糊判别的立体匹配算法

Stereo vision has long been one of the central research problems in computer vision, and stereo matching is the most important and difficulty issue of stereo vision. There are some limits for existing approaches to recover precise and dense disparity map. Feature-based stereo can produce more precise matching but only sparse disparity map. On the other hand, Area-based approaches can provide dense disparity map but less precise matching. In the situation of image synthesis for IBR, we need not only precise matching but also a dense disparity map. Thinking of the uncertain and fuzzy characteristic during matching, we introduce the widely used fuzzy set theory to the field of stereo matching, and propose an algorithm based on fuzzy identification. The algorithm uses the information of gradient magnitudes, angles of orientation and gray value information of nearby points as the identification facts. Experiments with real and synthetic images have been performed, they show that this algorithm is effective and it is of great value to use.