Road Edge Detection And Tracking Algorithm Base On Machine Vision For Intelligent Vehicle

We obtain binary image by using median filtering, Sobel edge detection operator and image segmentation algorithm based on maximum entropy from the road gray image. To acquire the feature parameter of the road edge in the binary image, we propose an improved Hough transformation algorithm. According to the prediction result of the Kalman filtering, we establish the area of interest (AOI) of the road edge and adjust the size of AOI dynamically to track the road edge accurately. The experiment results show that the algorithm approach by this paper is reliable and rubust.