基于主动视觉的结构光手眼系统自标定方法

Self-calibration Approach for Structured Light Eye-in-hand System Based on Active Vision

为实现结构光手眼系统的标定,提出一种基于主动视觉的结构光手眼系统自标定方法。该标定方法精确控制机器人进行2组平移运动和1组带旋转运动:2组平移运动标定出摄像机内参数、手眼矩阵的旋转部分和光平面的法向量;1组带旋转运动标定出手眼矩阵的平移部分和光平面的深度信息。试验结果表明,该标定方法简单,无需使用特定靶标,特征选取容易,测试结果稳定,定位精度可达到±1.02 mm,对结构光手眼系统在实际工业现场的使用有重要意义。

In order to calibrate the structured light eye-in-hand system, the self-calibration approach based on active vision is proposed. In this calibration method, by precisely controlling the robot, 2 groups of translational motion and 1 group of motion with rotating are conducted. The intrinsic parameters of camera, and the normal vectors of eye-in-hand rotation matrix and the light plane are calibrated by 2 groups of translational motion, and the translational section of eye-in-hand matrix and the depth information of the light plane are calibrated by 1 group motion with rotating. The test results show that this calibration approach is simple, specific target is not requested, feature selection is easy, testing result is stable, the accuracy is±1.02 mm. This provides important significance to structured light eye-in-hand system in industrial fields.