Road Obstacle Detection Based on Randomized Hough Transform

Abstract

Road obstacle detection is the core of pedestrian safety. As an effective image object detection method, Hough transform can detect straight lines, circles, ellipses, parabolas, and many other analytical graphics. This paper mainly detects bar deceleration strips and round well covers, owing to obvious linear features. Therefore, randomized Hough transform is applied to detect roadblocks to ensure pedestrian safety, improving the detection efficiency compared to the classical Hough transform in this article.