Fig. 1. Robot eye-in-hand hand-eye calibration system

Fig. 2. Abnormal sample detection method based on Z-scores

Fig. 3. Hand-eye matrix optimization algorithm based on LM

Fig. 4. Hand-eye calibration platform of blade repair robot

Fig. 5. Sample scaling example

Fig. 6. Z-score scatter plot of dataset. (a) First inversion; (b) Second inversion; (c) Third inversion

Fig. 7. Example of experimental steps. (a) TCP calibration; (b) Touching characteristic points; (c) Taking images; (d) Obtain the coordinates of feature points

Fig. 8. Comparison diagram of coordinate fluctuation from experimental results. (a) Method A；(b) Method B；(c) Ours

Fig. 9. Histogram of Euclidean distance comparison

Fig. 10. Robot grinding repair operation. (a) Robotic inspection; (b) Installation of sanding tools; (c) Grinding action; (d) Depth cameras provide information