Laser-camera Compositite Sensing for Road Detection and Tracking

An important feature in most urban roads, and similar environments such as in theme parks, campus sites, industrial estates, science parks and the like is the existence of pavements or curbs on either side defining the road boundaries. These curbs, which are mostly parallel to the road, can be harnessed to extract useful features of the road for implementing autonomous navigation or driver assistance systems. However, vision alone methods for extraction of such curbs or road edge features with accurate depth information is a formidable task as the curb is not conspicuous in the vision image and also it requires the use of stereo images. Further, bad lighting, adverse weather conditions, non-linear lens aberrations, lens glare due to sun and other bright light sources can severely impair the road image quality and thus the operation of vision alone methods. In this paper an alternative and a novel approach involving the fusion of 2D laser range and monochrome vision image data is proposed to improve the robustness and reliability. Experimental results are presented to demonstrate the viability, and effectiveness, of the proposed methodology and its robustness to different road configurations and shadows. Key Words—Intelligent sensors, robot vision systems, laser measurement systems, transportation