半导体光电, 2014, 35 (3): 555, 网络出版: 2014-06-24
智能坐标测量中零件位姿的单目立体识别
Single Camera Stereo Recognition of Parts’ Pose in Intelligent Coordinate Measurement
位姿识别 单目立体视觉 图像匹配 智能坐标测量 pose recognition single camera stereo vision image matching intelligent coordinate measurement
摘要
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, 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, 经过同名像点的匹配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。进行了识别实验, 实验数据表明该识别方法是可行的, 满足实时测量要求。
Abstract
In order to realize correct and fast recognition of parts’ pose on Coordinate Measuring Machine (CMM), a single camera stereo vision recognition method based on the translation of CMM was proposed, and its principles, solution and recognition procedure of the pose parameters were studied. According to the double-camera stereo vision principles, an image of the part to be measured is captured by the CCD camera, which is driven by CMM along its x or y axis, on two different positions correspondingly. Thus, the part’s single camera stereo vision measurement is realized with the image matching on two images of the same feature point to be measured on the part, and each feature point’s 3D coordinates in the camera coordinate system can be obtained. Then, using the camera calibration parameters, each feature point’s 3D coordinates in the machine coordinate system can be calculated. At last, combining with each feature point’s 3D coordinates in the part CAD coordinate system, the part’s pose parameters can be solved. The recognition system was set up, and experiments were conducted. Experimental results show that the recognition method is correct and practical, which can meet the real-time measuring requirement.
王春梅, 黄风山. 智能坐标测量中零件位姿的单目立体识别[J]. 半导体光电, 2014, 35(3): 555. WANG Chunmei, HUANG Fengshan. Single Camera Stereo Recognition of Parts’ Pose in Intelligent Coordinate Measurement[J]. Semiconductor Optoelectronics, 2014, 35(3): 555.