空间点的立体视觉传感器标定方法

Calibration method for stereo vision sensor based on spatial points

给出了一种基于随机空间点的立体视觉传感器高精度标定方法。该方法无需求解单个摄像机的内外参数,而是将立体视觉传感器透视变换矩阵中的元素作为未知量,当已知一组由高精度三坐标测量机提供的三维空间点坐标及其对应的图像点坐标时,即可利用奇异值分解法求解出各未知量的最小二乘解,从而避免了使用各类标定模板时由于加工和测量误差引入的标定误差。最后用标定过的立体视觉传感器对一组随机空间点进行三维坐标测量,与坐标测量机给出值进行比较表明,在误差最大的X轴方向,测量误差<0.05 mm,在Y轴和Z轴方向,测量误差<0.01 mm,证明了该方法的有效性。

A high-precision calibration method for stereo vision sensor based on random spatial points is presented. By giving 3D spatial point coordinates by a coordinate measurement machine（CMM） and corresponding image point coordinates, the calibration method is realized by taking the projection matrix elements as unknown quantities and using singular value decompositiofl to get the least-squares solutions,which avoids solving internal and external parameters of each camera and eliminates the calibration error caused by template machining and measurement. A group of random spatial points are measured by the calibrated stereo vision sensor, and measuring results show that, compared with CMM measurement values, the errors are less than 0. 05 mm in the X axis direction, and less than 0.01 mm in the Y and Z axis direction,which indicates that the technique is feasible and effective.