Monocular stereo vision based method for validating path accuracy of industrial robots

The evaluation of industrial robot path accuracy is significant for improving its dynamic operating performance. In this paper, a monocular stereo vision (MSV) based method is proposed for measuring the path error of robot end-effectors. Firstly, the structure parameters of the MSV system are analyzed to determine appropriate imaging parameters of the equivalent virtual binocular vision (VBV). Secondly, to save cost and measurement time, an all-in-one high-accuracy artifact, which integrates path representation and camera calibration, is designed. And accordingly, an improved decoding method is proposed to distinguish different tags on the artifact. After calibrating depth of field (DoF) dependent lens distortion, the path accuracy can be assessed with high-accuracy by the difference between path measured by the vision and the nominal one. Experimental results show that the MSV system with an average measurement accuracy of 0.012 mm can evaluate the path performance of the robot, and the maximum circular path error at 1.5 m/min is 0.252 mm.