A novel lane changing algorithm with efficient method of lane detection

This paper presents an efficient method of lane detection and apply the detection results in vehicle auto lane departure. The inverse perspective mapping image is taken for line detection. During image preprocessing, PPHT method and a third degree Bezier spline are applied for line fitting step. Meanwhile, linear clustering is taken to reduce the excessive lines. Not only the lines in the current driving lane can be detected, but also lines of neighboring lanes when vehicle is changing lanes can be detected. Besides, an algorithm of trajectory planning is introduced as a prerequisite for the implementation of lane changing. Fields results have shown the effectiveness of the proposed system.