Single camera stereo vision coordinate measurement in parts pose recognization on CMM

Abstract

In order to recognize parts’ pose on Coordinate Measuring Machine (CMM) correctly and fast, based on the translation of CMM, A single camera stereo vision measurement method for feature points’ 3D coordinate on the measured parts is proposed. According to the double cameras stereo vision principle, a image of the part to be measured is captured by A CCD camera, which is driven by CMM along its X or Y axis, on two different position correspondly. Thus, the part’s single camera stereo vision measurement is realized with the proposed image matching method, which is based on the centroid offset of image edge, on two images of the same feature point on the part, and each feature point’s 3D coordinate in the camera coordinate system can be obtained. The measuring system is set up, and the experiment is conducted. The feature point’s coordinate measuring time is 1.818s, and the difference value, which is between feature points’ 3D coordinate calculated with the experiment result and that measured by CMM in the machine coordinate system, is less than 0.3mm. This measuring result can meet parts’ pose real-time recognization requirement on the intelligent CMM, and also show that the method proposed in this paper is feasible.