Research On 3D Panoramic Viewing System Based On Embedded Technology

The development of intelligent vehicle is accelerating the pace of the automotive industry,from the original mechanical system towards electronic system transformation,a new generation of advanced driver assistance system(Advanced Driver Assistant Systems,ADAS)has been developing rapidly.In the fierce market competition of science and technology,the car panoramic imaging system as the latest science and technology talent shows itself,it is a multi camera environment of a vehicle,driving through the computer(Electronic Control Unit,ECU)for seamless image mosaic,real-time image information to eliminate vision blind by the on-board LCD driver is provided with wide field of view.Driver assistance systems help drivers more intuitive,safe driving.At present,the automobile panoramic image system in the market is generally a bird's eye view from the top view angle.This dissertation presents a 3D panoramic scanning system,which is a system framework for multi camera attached to the exterior of a car for visual inspection of the car The frame can achieve the advanced visualization effect according to the camera image brightness,color and other related properties,and the rendered picture is like a 3D picture taken by a position selection in the 3D space around the car.In this dissertation,auto three-dimensional panoramic look around the view of the system scheme and system architecture scheme is formulated according to the view,a three-dimensional image of the original image is constructed around regulation collected by four cameras attached to the car around a shot in the virtual camera above the visual angle of freedom.Need a lot of cars around the original image information system,in order to reduce the number of system camera with fisheye wide-angle camera,this camera is characterized by a large field of view,the amount of image information can capture more suitable for panoramic image acquisition,image distortion is need shortcomings after pretreatment:fisheye camera calibration and correction of the fisheye image distortion and improve the effect of image noise,clarity and contrast,so as to improve the visual quality of video image.After that,the algorithm of panoramic image stitching based on feature point matching is studied in depth.It includes the search and match of image features,the calibration of the camera,the distortion of the image,the exposure compensation and the image fusion.Finally,the panoramic image is panoramic in three-dimensional space,and loaded with a virtual 3D car.The mouse can be dragged on the screen to achieve the effect of multi angle observation.The system is finally implemented on the embedded mobile platform,carrying the touch liquid crystal display,and the camera interface selects the USB interface.Based on this system,various applications can be developed,including automatic parking,obstacle detection,and even automatic driving.