Lane detection based on random hough transform on region of interesting

Abstract

In order to improve computational speed of the lane detection, we proposes a new road edge algorithm. The innovation of this algorithm is that it combined with ideas region of interesting (ROI) and random Hough transform (RHT). In this paper we proposed some constraints on the road model, making the algorithm with a high anti-interference performance. The algorithm consists of the initial road edge detection and the follow-up tracking of road borders. In the initial edge detection we used the Hough transform; in the path tracking we used RHT on the ROI to improve computing speed.