Abstract: Based on the binary image from edge detection, an image processing method was suggested which applies a new “row” direction distance transform, taking the obtained distance value as the feature for the further lane detection. Moreover, experiments were conducted to validate that the resulting distance map is superior to and more effective than edge map in dealing with discontinuous lane marks. The results show that the new distance transform can not only provide distance information, but also make discontinuous lane boundaries continuous, and so as to improve the efficiency and robustness of lane detection. 

Key words: lane detection, distance transform, edge detection