Research On Three-Dimensional Photogrammetry Technology Based On Machine Vision

In recent years,with the development of information technology,image analysis technology,and computer vision technology,machine vision has been widely used in various fields.Machine vision extracts geometric information from images through image acquisition,processing,and analysis techniques,achieving three-dimensional perception of scenes.Although machine vision technology has wide applications in image measurement,most current applications are only suitable for specific measurement tasks,and the equipment cost is relatively high.In addition,how to solve the problem of measuring dimensions that are difficult to measure using traditional measurement methods in industrial measurement is an important direction of research in the field of machine vision.This paper presents a machine vision-based 3D photogrammetric measurement system using a camera device consisting of two industrial cameras.The system achieves plane area measurement with monocular vision and stereo height measurement with binocular vision.The main contributions of this work are as follows:1.Monocular and binocular camera calibration.Camera calibration is the primary task of image measurement.In this paper,we use a camera calibration method based on a planar checkerboard to perform monocular and binocular camera calibration,and evaluate the accuracy of calibration parameters based on drift error.In the monocular measurement system,camera parameters are utilized to achieve image rectification and three-dimensional reconstruction.In the binocular measurement system,camera parameters and the rotation matrix and translation vector describing the relative position of the left and right cameras are used to achieve image rectification and polar correction of the left and right images.2.A monocular camera is used to obtain depth information and clustering algorithms are introduced for planar area measurement.To address the issue of uncertain Z-axis positioning of the planar area in monocular vision,a constraint method is used to determine the plane depth,considering the entire plane depth as a known value,thus overcoming the problem of monocular vision’s inability to obtain depth information and achieving 3D reconstruction of the target points.In addition,for irregularly shaped targets,K-means clustering is used for image segmentation to extract the target region and calculate its area ratio,achieving accurate area measurement.3.Stereo matching of feature points in binocular vision with constraint information.Traditional feature point matching algorithms detect and match all the information of the left and right images,resulting in high computation cost and high error matching rates.In this paper,a constraint condition is added before feature point matching,which requires the epipolar lines of the left and right images to be parallel.An improved Canny algorithm is used to extract the upper and lower edges of the object to be tested,and the intersection of the epipolar line and the upper and lower edges is used as the feature point.This reduces the computational cost of feature point detection and ensures matching accuracy.4.Visualization design of the measurement system.In this paper,a complete visualized 3D photogrammetry system was designed based on the GUIDE(Graphical User Interface Design Environment)of MATLAB platform.After multiple experiments and tests,the results showed that the error of the single-camera vision measurement system in measuring plane area was within 1%,and the error of the binocular vision measurement system in measuring height was within 1mm.The measurement accuracy can meet the requirements of most industrial measurement scenarios.The machine vision-based 3D photogrammetric measurement system designed in this paper is applicable to scenarios such as leather,textiles,wood,metal sheets,rubber,etc.,where there is a need for area and height/thickness measurements.The experimental results show that the system can achieve accurate planar area measurement and stereoscopic height measurement,and has certain practical value.