Variable Search Region Base on Disparity Gradient
Longyuan Guo, Guoyun Zhang, Jianhu Wu · Procedia Engineering · 2012
Usually, the correspondence for a point in the first image is obtained by searching on the epipolar line and maximum disparity in the second image. The reduction of the search region can increase the performance of the matching process, in the context of the execution time and the accuracy. According to relationship between disparity gradient and search region, the paper presents a novel strategy to reduce the search redundancy for the area-matching algorithm, which called variable search region based on disparity gradient. Differ from other similarity methods, the paper calculate disparity gradient of all the image pixels and search region of each pixel is obtained according to its disparity gradient. The disparity gradient only is applied to two successive edge points between neighborhood epipolar lines in the others. The experiment result indicates that the search redundancy of area-matching algorithm is eliminated after variable search region is applied. Calculation complexity of original algorithm is decreased. And the result of some texture regions are also improved since search region is reduced.