An Adaptive Method for Detecting Lane Boundary in Night Scene

Lane detection plays a key role in the vision-based driver assistance system and is used for vehicle navigation, lateral control, collision prevention, or lane departure warning system. In this paper, we present an adaptive method for detecting lane marking based on the intensity of road images in night scene which is the cause of numerous accidents. First, a region of interest (ROI) image is extracted from the original image and converted to its grayscale image in which the value of each pixel is the maximum value of R, G and B channel of ROI image. After that, we find the maximum intensity on each row of grayscale image. Finally, the lane boundary is detected by Hough transform. Experiment results indicate that the proposed approach was robust and accurate in night scene.