Image processing and visual control method for arc welding robot

The methods to improve the robustness and adaptability of arc welding robots are presented. Image processing algorithm for feature extraction from structured light images of weld seam is discussed in the case of a large amount of strong reflection, arc light and splash disturbance. The processes consist of defining target region, selecting adaptive threshold and extracting feature points. Then a novel hybrid visual servo control method for arc welding robot is provided, in which the position-based and image-based visual servo method are integrated to increase the robustness of weld seam tracking. The welding experiment is well conducted to verify the methods.