Research On Algorithm Of Vhicle Panoramic System Based On Projection To Variable Curved Surface

With the development of computer vision and image processing technology, more and more advanced driver assistance systems with visual sensor were applied to the car. The technology of monitoring vehicle’s surrounding environment by multiple cameras has become focus in the field. And it is also relatively mature at present. However, the 2D imaging technology still has considerable limitations and is only used in reverse images, and it is unable to help drivers to make a judgment accurately. The 3D imaging technology has realized the 360°seamless display of the vehicle helping the driver acquire a wide range vision of easily. And it can ensure traffic safety effectively.This paper analyzes the fisheye camera model、the imaging principle and calibration technology and proposes a calibration method based on vehicle coordinate system firstly. Secondly, it introduces some techniques of 3D reconstruction, such as the three dimensional surface, projection under virtual view and image mosaic algorithm etc. Further, color correction based on the minimum brightness difference in the public areas is proposed. Finally, 3ds Max is used to establish the 3D projection surface, and the model is imported into Open GL. Then, the reprojection of fisheye image is realized using perspective projection、texture mapping principle. And image fusion is realized using blending technology in the Open GL. The reconstruction of the surrounding environment under virtual viewpoint is realized using viewing transform in the Open GL. Next, the vehicle model is established in 3 ds Max. The algorithm of the paper is designed and realized on the PC based on Open GL and Microsoft Visual Studio.Experiments show that the algorithm in the paper is suitable for using in the vehicle and improves the limitation of inaccurate image of 3D object in the 2D around vehicle system. In addition, the color correction algorithm in the paper is realized simply and easily improving the brightness difference between images and the overall visual effect.