Stereo-Based Dense Disparity Estimation for Lunar Surface

Abstract

A stereo-based dense disparity estimation algorithm is proposed to build high quality dense disparity map of lunar surface under special illumination, weak texture and occlusion condition. To avoidance the serious shadow effect effectively, intrinsic image of stereo images are preprocessed. A color similarity probability based belief propagation algorithm (BP) is proposed to solve the depth discontinuous problem of occlusion and obtain an initial dense disparity map. Mean-Shift segmentation algorithm and adaptive threshold optimization are utilized to improve the precision of initial dense disparity map of weak and high similarity texture region in stereo images. Experimental results of the images of standard test library and the stereo images of simulated lunar surface validate that our algorithm is robust to build high quality stereo dense disparity map for Lunar surface.