Fast Edge - Based Stereo Matching and Geometric Parameters 1

The reduction of the search region in stereo correspondence can improve the performance of the matching process in terms of execution time and accuracy. For the edge-based stereo matching, we establish the relationship between a search space and parameters such as relative displacement of the edges, the disparity under consideration, the image resolution, the CCD dimension, and the focal length of the stereo system. Then, we propose a novel matching strategy for the edge-based stereo. Afterward, we develop a fast algorithm for the edge-based stereo with a combination of the obtained matching strategy and the multiresolution technique by using the Haar wavelet. Considering the conventional multiresolution techniques using the Haar wavelet, the execution time of our method (including feature extraction and feature matching) are decreased by 37%. Moreover, the failed rate and the error are also decreased. Theoretical investigation and experimental results show that our algorithm has a very good performance; therefore, this new algorithm is very suitable for fast edgebased stereo applications like stereo robot vision. Received December 15, 2002 1