Sunghoon Kim, S. Park, and K.-H. Choi (Korea)
Road Segmentation, Car Navigation, Autonomous Vehicle, Video-Based Approach
In this paper, we present a novel algorithm to extract road boundary for autonomous vehicles by analyzing road boundary in a video sequence captured from a moving vehicle. More specifically, the proposed algorithm 1) extracts the edge map of an input image, 2) deletes unnecessary edge components that do not belong to road boundary and merges broken edges of road boundary, and 3) removes shadows that lead false positive of road boundary. In addition, extracted edges from multiple images, i.e., a block of images, can be successfully combined to detect changes of road boundary, making possible to notice that an intersection is located ahead. Experimental results are provided to show the robustness and effectiveness of the proposed algorithm under various road conditions such as shadows and snow, etc.
Important Links:
Go Back
