A novel extrinsic calibration method for robotic systems with structured light sensors

Abstract

The extrinsic calibration for robotic systems with structured light sensor is fundamental to accomplish the tasks of 3D measurement, reverse engineering and quality inspection. In this paper, we proposed a cuboid calibration target with nicks on the surfaces which is easy to extract the control points. In the calibration process, both least square method and the nonlinear optimization based on maximum likelihood criterion are utilized. The degenerate configuration is discussed for non-singularity solution. Both simulation and real data are utilized for robustness test of the algorithm and the experiment shows that the proposed method has 0.6 mm RMS errors of control points. Compared to the existed methods, the proposed method has the advantages of small error, simple operation and high flexibility.