Capture of Vehicle Surroundings Using a Pair of Panoramic Annular Lens Cameras

In this paper we propose a new system to detect obstacles round a vehicle for intelligent driver assistance system. The system consists of a pair of panoramic annular lens (PAL) cameras, which are installed on the top of the vehicle. The panoramic annular lens can project the view of the entire 360 degrees around the optical axis onto an annular plane. Two cameras which are fixed properly can capture all view of the surroundings of the vehicle. They are calibrated by using the spherical perspective projection model to unwrapped the conventional perspective image without distortion. A front panoramic image and a back panoramic image of the vehicle form a vehicle dynamic panoramic map (VDPM). A driver can watch the scene in any direction of the vehicle through the VDPM, which will make the drive safer. Three coordinates systems are established in order to obtain the obstacle position information from the vehicle. We create a virtual binocular vision situation using the video stream from each one panoramic camera. The binocular analysis technique with area-based correlation is used to recognize the obstacle and compute the distance from the vehicle. The experiment result shows the basic feasibility of this system. It is believed that a panoramic annular lens camera will play an important role in the future of intelligent driver assistance systems.