Abstract: In order to solve the real-time road marking recognition problem in the intelligent vehicle vision navigation system, a novel pedestrian crossing and stop line detection algorithm based on oriented-edge match is proposed. It builds top-view image through Inverse Perspective Mapping(IPM) transformation by a mapping table to accelerate IPM process. It abstracts vertical oriented-edge and horizontal oriented-edge. The pedestrian crossing and stop line can be detected by matching adjacent up and down edges to bands. Experimental results illustrate that the algorithm is effective for real-time analysis and robust under difficult urban road circumstances.
Inverse Perspective Mapping(IPM),
Region of Interest(ROI),
random sample consensus