Research On Binocular Vision Ranging System

Binocular vision ranging is to extract scenery information by simulating human eyes.Compared with traditional ranging methods,binocular vision ranging has the advantages of high accuracy and non-contact,and can restore scenery information more comprehensively and concretely.Through studying the principle of binocular vision ranging,this paper studies the process of binocular camera calibration,stereo correction and stereo matching,and improves the traditional stereo matching algorithm based on census transform.According to the ranging principle,a binocular vision ranging platform is built,which is programmed by Visual Studio 2013 and OpenCv visual library version 2.4.10.This paper uses Zhang Zhengyou's camera calibration method to calibrate.The black-and-white grid corner information in the self-made calibration board is extracted,and sub-pixel corner detection is used to further precise corner coordinate information.On the premise of satisfying the accuracy requirement,the camera parameters obtained by calibration are analyzed,and the stereo correction of the images captured by binocular cameras is carried out by using the obtained parameters,so that the images are aligned in the imaging plane coordinate system in line,so as to reduce the complexity of stereo matching.In the process of stereo matching,because the accuracy of traditional census algorithm depends heavily on the pixel value of the central point,and the matching effect in disparity discontinuous regions is not ideal,this paper proposes improvements in the process of matching cost calculation and cost aggregation.By comparing the average value of the pixels in the neighborhood of the central point with the average value of the central point,when the difference is large,it shows that the central point is disturbed by noise.At this time,the average value of the pixels in the neighborhood is used to replace the central point.In the process of aggregation of matching costs,bilateral filtering is used to enhance the matching effect for disparity discontinuous regions,taking into account the location information and gray value information of the pixels in the neighborhood.The experimental results show that the improved census algorithm has better anti-noise ability,and the matching effect in disparity discontinuous region is strengthened.Finally,a number of groups of image information are collected for ranging experiments,and the disparity map obtained by stereo matching is used to obtain the three-dimensional distance information of objects in space based on the principle of geometric change.The experimental results show that it has high ranging accuracy within a certain distance.