The Coordinate Measurement Technology Of Long Distance Moving Object Based On Stereo Vision

Stereo vision technology is one of the non-contact,high-precision,low-operation measurement technology,has been widely used in various industries.When measuring long-distance moving targets,the traditional short-baseline stereo vision measurement system is limited by the length of the baseline,and the measurement accuracy is low.Compared with the short-baseline system,the long-baseline system has higher measurement accuracy and a wider application range.This paper applies the stereo vision technology to the measurement of long-distance moving targets.Based on the actual conditions of the experimental site,combined with the characteristics of the target to be measured,such as the distance,size,movement mode,etc.,a set of long baseline stereo vision system is built,which can quickly and accurately Calculate the distance of the target to be measured,and then locate its three-dimensional coordinates.The main work completed in this paper is as follows:(1)According to the characteristics of the target to be tested,select appropriate hardware parameters and build a stereo distance measurement system for long-distance measurement;(2)Research the camera mathematical model and calibration principle,and improve Zhang Zhengyou's calibration method according to the characteristics of the hardware system.Build a stereo reference system,calibrate the internal and external parameters of the camera step by step,use re-projection error to filter high-quality calibration images to obtain accurate internal parameters,use least squares fitting to extract the stereo reference frame feature points,complete the external parameter calibration,and use The camera parameters perform distortion correction and stereo correction on the binocular image to obtain a line-aligned image without distortion;(3)In order to solve the problem of moving target detection under dynamic background interference,a variety of target detection algorithms are tested,and finally an improved five-frame difference method is used to obtain complete moving target information.Thenaccording to the characteristics of the moving target,use Hough transform circle fitting and centroid extraction to detect the target contour,locate the feature points,obtain the target parallax information,and achieve the purpose of distance measurement and coordinate positioning;(4)Using the method of this paper,measure multiple sets of dynamic targets at different distances,compare with the actual distance of the target,check the measurement accuracy and stability of the method of this paper,and analyze the cause of the error based on the experimental results;(5)In the environment of Windows7 system and Visual Studio2015,combined with OpenCV3.2 computer vision library,the software design and MFC interface design are implemented in C language to realize the calibration and correction functions of the long baseline stereo vision system.The experimental results show that the measurement error of this system increases with the distance of the target to be measured.When measuring the target at 500 m,the measurement error is within 4.7%,which can be used for a variety of dynamic target measurements,with high measurement accuracy and good robustness.