为了提高光笔式三坐标视觉测量的正确性和有效性, 对其多点透视模型进行了研究。以经典的“N点透视问题”理论为依据, 将世界坐标系的原点巧妙地构建在光笔笔尖处, 使坐标测量变为传统的摄像机外部参数标定问题, 采用坐标变换方法, 建立了光笔式三坐标视觉测量系统的多点透视数学模型, 得出光笔上点光源的个数应大于等于3。结合光笔参数的自标定, 采用十点透视模型, 开发出光笔式三坐标视觉测量实验系统, 在三坐标测量机上进行了比对测量实验。比对测量结果表明: 在300mm 测量范围内, 实验测量系统和三坐标测量机在x、y、z轴方向上的测量数据最大差值分别为0.13、0.14和0.2mm。

In order to improve the correctness and validity for three coordinate vision measurement using a lignt-pen, the perspective-n-point model is studied. According to the classical "perspective-n-point problem" theory, by positioning the origin of the global coordinate system on the pin-point of the light-pen, the process of measuring coordinate is converted into calibrating the cameras external parameters. With the coordinate transforming method, the perspective-n-point model for three coordinate vision measurement system is established, and the conclusion is drawn that the number of point light source must be more than three. Combining with the parameters selfcalibration, an experimental measuring system with ten point light sources on the light-pen is developed, and comparative tests were performed on the Coordinate Measuring Machine(CMM). The results show that the maximum difference between the measuring data of the experimental system and CMM is 0.13mm, 0.14mm and 0.2mm in 300mm measuring range along the x, y, z axises respectively.