An improved method of corner point extraction used in grating projection phase measurement

摘要

The grating projection phase measurement method can obtain 3D surface profile of an object with a grating projection device and two cameras, which is widely used in reverse engineering, industrial detection, cultural relic digitization and human body measurement due to its simple measuring principle, convenient operation, fast and high-precision data acquisition of 3D point clouds. As one of its calibration processes, binocular vision calibration plays a key role in getting high-precision measurement results, and the accuracy of extracting feature points on the calibration target directly affects the binocular calibration accuracy. In this paper, a sub-pixel corner extraction method based on Shi-Tomasi algorithm was proposed to extract corner points of checkerboard. An experiment is done to compare the extraction effect of the proposed algorithm with the traditional Harris method. The experimental results show that the proposed algorithm can locate the position of corner points more effectively with higher extraction accuracy. The mean square value of the reprojection error is about 0.008 pixels, the rate of corner extraction is 74.6%, and the processing time is shorter, about 0.78 seconds. Therefore, the method used in this paper is reliable and feasible for the feature point extraction in 3D reconstruction of measured objects.