Automatic Road Surface Profiling with Sensors Fusion

This paper addresses the issue of automatically profiling the road surface based on sensor fusion. Road surface profiling in this research paper includes road boundary detection, white line detection and lane division. We propose an approach to perform automatic and robust road surface profiling with the fusion of the following sensors: LADAR (Laser Detection and Ranging), GPS (Global Positioning System), INS (Inertial Navigation System), cameras and an odometer. The LADAR is the most important in our research since we propose a new technique that utilizes laser remission to detect white lines. A prototype system has also been developed for testing with the capability of converting profiling results into video files for easy reference and management. The system is able to work under severe weather and light condition, and profiling all lanes within road boundary at one run. Experimental results on a wide variety of roads have demonstrated the effectiveness of the proposed system.