The Research of Lane Marker Detection Algorithm Based on Inverse Perspective Mapping

Abstract

Abstract. This paper presents a robust and fast algorithm for lane marker detection. It is based on generating an inverse perspective mapping image of the road, computing line abstracted regions and the class points from IPM image, which is followed by a post processing step using rectangular approximation to detect lane marker. A reference database containing over 100 images of natural road scenes was built with corresponding manually labeled ground truth images (available at http://www.lcpc.fr/english/products/). This database enables to evaluate lane marker detection algorithm in a systematic way. The experiment proves that the algorithm proposed can detect the lane marker on the road in various conditions.