A Lane Marking Detection Algorithm Based on Improved Hough Transformation

This paper proposes a suitable for DSP's lane marking detection algorithm based on improved Hough transform,the algorithm first computes gradient of the edge image from the edge extraction which is just the preclass process of the classical Hough transform,and then use the computed gradient results to determine the parameter of edge orientation which should be searched from 360 degree in the next process of Hough transform, further more the algorithm to reduce its computation divided the parameter space into many little blocks in the precondition of no affect lane marking detection results.Experimental results showed that:the method not only can accurately detect the lane marking,but also can run in DM648DSP hardware platform in real-time,and get a wonderful lane marking detection results.