Research On Image Stitching Technology In Vehicle Panoramic Surround View System

In recent years,the number of private cars owned by Chinese residents has been increasing year by year,and people are paying more and more attention to the problem of traffic safety caused by the increase of vehicles.One of the most common traffic accidents is a scratching event when driving on a city's congested roads at low speeds or parking.The main cause of frequent accidents like this is because the driver has a lot of visual blind spots when driving the vehicle,and the driver cannot fully grasp the scene around the vehicle.The scene around the vehicle.The on-board panoramic surround view system aims to provide the driver with a panoramic aerial view around the vehicle,eliminate visual blind spots to the greatest extent,reduce the occurrence of such accidents,and thus contribute to road traffic safety.This paper studies the algorithms used in the image processing technology of the vehicle panoramic surround view system,and focuses on the image stitching technology.The following aspects are the main research contents of this article.First,the difference between ordinary camera with a fish-eye camera imaging principle be summarized by systematic study,compared to commonly used method of camera calibration,calibration method chosen Zhang Zhengyou fish-eye camera calibration,and common distortion correction algorithms are compared and analyzed Then,the polynomial correction method was used to remove the distortion of the fisheye image by using the calibration results.Secondly,the commonly used top-down transform algorithm is researched,and the direct linear algorithm is used to complete the conversion from the side view to the top view.The necessity of image interpolation and the basic principles of interpolation correction are expounded,and the common image interpolation correction algorithms are compared.The last research is the most important in the process of image processing,which is also the focus of this article: image stitching technology,through the two steps of image registration and image fusion research to achieve panoramic image stitching in the vehicle panoramic surround view system.After comparing and researching the SIFT algorithm and the SURF algorithm,an algorithm improvement scheme is proposed.It is experimentally verified that the improved algorithm can combine the advantages of the two algorithms,and is more suitable for vehicle panoramic surround view systems.The weighted average method is used to fuse the joints,and finally the purpose of the vehicle panoramic view system is to provide the driver with a panoramic view image of the surroundings of the vehicle.After optimization and experimental testing of the overall design of the vehicle panoramic surround view system,the algorithm studied in this paper can achieve the expected goal of displaying the vehicle's panoramic panoramic bird's eye view around the vehicle.