Research Of Ranging System Based On Binocular Stereo Vision

Computer vision is a technique that uses photosensitive elements instead of human vision to perceive and analyze targets.Binocular stereo vision is an important branch of computer vision,namely two cameras in different positions are panned or rotated to shoot the same scene and the three-dimensional coordinate values of the spatial points are obtained by calculating the parallax of the spatial points in the two images.Binocular stereo vision simulates human eyes for judgment and processing.The advantages of stereo vision is simple and easy operation,It has been widely used in the fields of pose detection and control,non-contact measurement and so on.Firstly,the paper is based on the study of binocular stereo vision,the transformation relationship of four coordinate systems in the camera imaging model is determined and the ranging model of binocular stereo vision is established.Zhang Zhengyou calibration algorithm is adopted and using Matlab calibration toolbox for calibration experiments.In order to calibrate non-coplanar and aligned image planes to be coplanar and line aligned,using OpenCV under Visual Studio 2015 Enterprise to do stereo calibration experiment.The feature point dimensions for the SIFT algorithm are too high.If the dimension is too high,the matching speed will be slow and the storage space will be occupied.As the number of matching points increases,the probability of mismatching may increase.This paper optimizes the dimension of the feature points to make it from the original.The 128-dimensional is reduced to the current 64-dimensional.Experiments show that the running time of the optimized SIFT algorithm is shorter than the original.Using the OpenCV under Visual Studio 2015 Enterprise,the BM algorithm was used to perform stereo matching and ranging experiments.Then,using Visual C# under Visual Studio 2015 Enterprise to design and implement the upper computer software,the upper computer software basically realizes the functions of each part of binocular stereo vision ranging system,which are login function,image acquisition function,camera calibration function,Neighborhood processing function,feature matching function,stereo correction function,stereo matching and ranging function.Finally,a binocular stereo vision ranging system based on S3C2440 hardwareplatform and embedded Linux operating system software platform was developed.The hardware part of the system consists of the microprocessor S3C2440 and its peripheral hardware devices.The software part consists of the embedded operating system Linux and applications.Figure [64] table [6] reference [61].