Research And Implementation Of Panoramic View Parking System

With the increasing of car number, the traffic burden is gradually worsen, the problems that hard to park and lack of parking space are becoming more and more common. Though the speed of parking is low, accidents still happen frequently, because it has many blind areas during the parking process. In this paper, we propose a solution of a panoramic vision system which can offer a panoramic image around the car to the driver, eliminate driver's blind spots and improve the safety of parking.In this paper, we have done some research about the algorithms of panoramic vision system, include: studied the imaging model of the fish-eye camera and corrected the fish-eye camera distortion with Zhang Zhengyou calibration method; Transformed perspective to top-view with DLT algorithm; Stitched the panoramic image with the image mosaic algorithm based on mark point and eliminated the stitching gap with the algorithm of weighted average fusion; we propose a lookup table method to improve the real-time capability of the system for some specific characters it has.We also have research image interpolation as well as two-dimension graphic engine and proposed an image interpolation method based on graphic engine, which greatly reduces the time of stitching the panoramic image. Offering a 3D virtual video is one trend of panoramic view parking system. We researched texture mapping and GPU programming, then Mapped the panoramic image to a hemisphere model, which reduces the distortion of top-view, improves the quality of panoramic image and provides drivers with multi-angle of view. And we optimized the startup of the system, which reduces the startup time and improves user experience.