A Convenient Calibration Method for Robot Hand Eye System Without External Calibration Objects

For the hand eye system calibration problem of inspection robot equipped with dual robotic arms and RGB-D camera, a simplified joint calibration method based on deep learning is proposed. It eliminates the need for external systems and calibration boards, resolving the calibration of the dual robotic arms and the extrinsic calibration of the RGB-D camera within the hand eye system of the inspection robot. By utilizing nonlinear optimization, it effectively reduces measurement noise and minimizes the propagation and influence of calibration errors in the coordinate transformation. Experimental results demonstrate that the proposed method achieves an average calibration error of 0.5mm for the dual robotic arms, and a minimum calibration error of 1.2mm for the camera's extrinsic parameters. Consequently, it significantly enhances the hand eye localization accuracy of the inspection robot during task execution. The innovative approach holds promising potential for improving the practical application of inspection robots.