Fusion of digital map and computer vision data for automotive safety applications

When driving in bad weather conditions on rural roads, it is very helpful for the driver to know which way the road turns within the next 120 meters. A technical system with topographic knowledge can provide this information for the driver. The next step is then an obstacle detection and warning system. In order to decide whether a detected object is on the road and thus of immediate importance it is essential to know the course of the road. Neither map nor camera information alone are sufficient for this task. Digital map systems alone will fail, because without expensive additional sensors the car's position and orientation are not measured accurately, whereas image sensors alone are restricted to visibility up to 50 meters. A novel approach of fusing map and image information is presented. The digital map is used to calculate the shape of the road, an ego motion estimator and the image data are used to determine the vehicle's heading and position. For the covering abstract see E134653.