Environmental Perception for Intelligent Vehicles Using Catadioptric Stereo Vision Systems

This paper presents advantages of catadioptric camera systems for robotic applications using the example of autonomous vehicles and driver assistance systems. We propose two different catadioptric stereo camera systems to improve environmental perception. Both stereo camera systems consist of at least one catadioptric camera, which combines a lens with a hyperbolic mirror and provides a panoramic view of the surrounding. This allows a 360◦ field of view of the vehicle’s surrounding and ultimately allows e.g. object detection around the vehicle or the detection of lane markings next to the vehicle. We discuss the advantages of both camera systems for stereo vision and 3D reconstruction. Furthermore, we show results for depth estimation with sparse features in a simulated environment as well as on a real data set.