Study on Stereo Matching Algorithm for Lunar Rover Based on Multi-feature

Abstract

In order to navigate and avoid obstacle automatically in an unknown environment, a new matching algorithm for lunar rover based on multi-feature was brought forward in the paper. Firstly, the original images were preprocessed by Gauss filter and CLAHE method to weaken the effect of noise and contrast. Secondly, according to the purposes of stereo matching results and the characteristics of lunar surface, the invariant SIFT keypoints and edge features were extracted. Thirdly, a stereo matching algorithm based on multi-feature was carried out. Epipolar constraint, ordering constraint and uniqueness constraint were employed to enhance the matching speed and precision, and the mismatching points were eliminated by continuity constraint. The combined matching results showed that the algorithm can obtain general geometry of the environment and effectively detect the possible obstacles.