为了使三坐标测量机快速准确地识别出被测零件的位姿, 提出了一种基于三坐标测量机平动的单目立体视觉识别方法, 并对该方法的原理、位姿参数的求解和识别过程进行了研究。根据双目立体视觉原理, 以三坐标测量机带动摄像机沿X轴或Y轴平移, 在两个不同位置分别拍摄被测零件的一幅图像; 利用本文提出的基于边缘图像质心偏移的同名像点匹配方法, 实现单摄像机立体视觉测量, 得到被测零件上各特征点在摄像机坐标系中的三维坐标; 由摄像机标定参数, 进一步计算出各特征点在机器坐标系中的三维坐标; 再结合各特征点在零件CAD坐标系中的对应坐标, 求解出被测零件的位姿参数。组建了识别系统, 进行了识别实验, 结果显示, 识别出的实验件位姿的平移量分别为: txl=32.65 mm,tyl=-90.23 mm,tzl=13.27 mm,旋转角分别为AX=38°,AY=4°,AZ=-5°,整个识别过程用时1.818 s。得到的实验数据表明该识别方法是可行的, 可满足实时测量要求。

To recognize the poses of parts to be measared by a Coordinate Measuring Machine (CMM) corrcetly and rapidly, a single-camera stereo vision recognization method based on the translation of CMM was proposed, and its principle, pose parameter solution and recognization procedure were studied. According to the double-camera stereo vision principle, an image of the measured part was capured with a CCD camera driven by the CMM along its X axis or Y axis on two different positions correspondly. Thus, the single-camera stereo vision measurement for the part was realized with the proposed matching method on two images with the same feature point, and the 3D coordinates of each feature point in the camera coordinate system were obtained. Then, by using the camera calibration parameters, 3D coordinates of each feature point in the machine coordinate system were calculated. Finally, combining with the 3D coordinates of each feature point in the CAD coordinate system for the part, the pose parameters of the part were solved. The recognization system was set up, and an experiment was conducted. The results show that the recognized translation pose parameters of the experiment part txl tyl and tzl are 32.65 mm,-90.23 mm, and 13.27 mm,respectively; and the rotation angle pose parameter AX,AY and AZ are 38°, 4° and 5°, respectively. Moreover, the recognization time is 1.818 s. Experiment result shows that the recognization method discussed above is correct and pratical, and meets the measuring requirement in real time.