Stereo Vision in Structured Enviroments by Consistent Semi-Global Matching

Abstract

This paper considers the use of stereo vision in structured environments. Sharp discontinuities and large untextured areas must be aticipated, but complex or natural shapes of objects and fine structures should be handled as well. Additionally, radiometric differnces of input images often occur in practice. Finally, computation time is an issue for handling large or many images in acceptable time. the Semi-Global Matching method is chosen as it fullfills already many of the requirements. Remaining problems in structured environments are carefully analyzed and two novel extentions suggested. Firstly, intensity consistent disparity selection is proposed for handling untextured areas. Secondly, discontinuity preserving interpolation is suggested for filling holes in the disparity images that are cause by some filters. It is shown that the performance of the new method on test images with ground truth is comparable to currently best stereo methods, but the complexity and runtime is much lower.