Abstract:Aiming at the limitations of traditional manual measurement of plate dimension, such as low dimensional accuracy, large workload, and easy damage to the plate surface, a vision measurement system for plate dimension is designed based on binocular vision technology. The checkerboard image is collected by binocular camera, camera calibration and image correction are carried out by MATLAB, left and right images are taken, and feature points are stereo matched by semi global matching (SGM) algorithm to reconstruct the three-dimensional point cloud model of the target. In order to improve the accuracy of target feature point coordinate acquisition, a sub- pixel detection method based on HARRIS is proposed. The Region Growing algorithm combined with expansion and corrosion operations is used to extract the surface contour of the plate, and the three-dimensional coordinates of each point on the plate contour are calculated according to the principle of triangulation to realize the dimension measurement of the plate, and point cloud reconstruction is performed to enhance the three-dimensional display effect. The practice results show that the sub-pixel detection method has advantages in corner extraction, and realizes high-precision dimension measurement in actual plate measurement applications, meeting the industrial measurement requirements.
