Stereo Matching Algorithm for Autonomous Positioning of Underground Mine Robots

Abstract

Binocular stereo vision theory comes from the principle of parallax. Two cameras capture different images of the same target from different angles at the same time, or capture different images of the same target from different angles with the same camera at different times to obtain a pair of visual image pairs. According to the principle of parallax, stereo matching and three-dimensional reconstruction of objects are carried out. Different stereo matching algorithms have different advantages and disadvantages. Through several of these algorithm simulation and data processing, and by comparing the results of the related parameters, it is concluded that the performance of several matching algorithms, and with the demand of robot positioning under mine, a summary of these algorithms were analyzed, and improved method is put forward.