A Novel Method of Obstacle Detection Using Stereo Vision for Unmanned Systems

Obstacle detection is an important problem in unmanned system technologies, which have been rapidly developing in recent years. In this study, a new approach to obstacle detection is presented. For use in unmanned systems, stereo vision is utilized to detect real-time obstacles on the moving path. In this method, reference points are determined by creating lines parallel to the road. By comparing the reference points with the pixel values of the searched region, the obstacle detection and distance are measured accurately.