Recognizing white line markings for vision-guided vehicle navigation by fuzzy reasoning

Abstract This paper presents a new method for recognizing white line markings for navigation of an autonomous vehicle that operates in out-door environments. A fuzzy-reasoning-based general technique for edge detection is used to classify a pixel in an image into a border region or a uniform region based on luminance differences between this pixel and its neighboring pixels. The idea of this paper is to integrate special knowledge about white line markings into a fuzzy rule base using this general technique. The method is implemented on the THMR-III mobile robot, and experiments show that this method is robust to noise and environment changes.