针对工业视觉测量中的摄像机标定精度及效率低的问题, 提出一种高精度CCD相机分区域标定方法。该方法首先利用三坐标测量机(CMM)带动圆形发光二极管(LED)光靶标按梯形台状做精确空间移动, 结合最小二乘椭圆拟合算法求解光靶标的像面位置, 并与CMM三维坐标形成精确的空间标定点集。同时, 将像平面按圆环形对称结构分割成N个区域, 并结合改进的Tsai算法分别对每个子区域进行相机参数的标定。标定实验结果表明: 经过分区域标定, 相机采集点的总误差比单区域标定法降低了17%(N=8), 单点平均误差降低了20%左右。算法可实现自动精确标定点采集, 操作过程简单, 基本满足中等精度的工业测量要求。所提出的相机标定法可应用于工业视觉测量, 特别是大工件测量领域。

A reliable sub-regional calibration method for CCD cameras was proposed to improve the accuracy and efficiency of industrial vision measurement. With the proposed method, a round light target was moved with the Coordinate Measuring Machine (CMM) according to a trapezoidal shape. The images of light target were processed with least square ellipse fitting algorithm to obtain the center positions. By 3D coordinates of CMM, a set of accurate calibration data were obtained. Furthermore, the image plane was separated into several sections according to the circle symmetrical structure and camera parameters in different regions were calculated with the sub-regional theory and improved Tsai method. Comparing with the sub-regional calibration method to single calibration one, the actual calibration experiments show that the total error is decreased by 17% when the image is separated 8 regions and the average error of single point is decreased about 20%. Experiments demonstrate that the data collection process is automatic and simple, and the camera calibration accuracy meets the medium precision of industrial measurement.It indicates that the method proposed is suitable for the industrial vision measurement, especially for the larger work piece measurement.