A new elliptic contour extraction method for reference hole detection in robotic drilling

Abstract

In robotic drilling of aircraft structures, reference holes are pre-drilled on aircraft structures and then detected by vision systems in the drilling process to compensate for the relative positioning errors between the robot tool center point and the workpiece, thus achieving improved position accuracy of drilled holes. In this paper, a novel elliptic contour extraction method based on salient region detection and optimization with snakes model is proposed for reference hole detection. Firstly, salient region detection is used for segmenting the region of reference hole from the background, and the resultant image of this operation is used for contours retrieving. Secondly, the initial contour of the reference hole is obtained from the retrieved contours using the voting method. Then the initial contour of the reference hole is further refined with the snakes model through energy minimizing of the snake. Finally, the elliptical parameters of the reference hole are computed by fitting an ellipse to the evolving result of the snake. The robustness and accuracy of reference hole detection with respect to noise and environmental disturbance are enhanced significantly through saliency estimation and optimization with the snakes model. Experimental results reveal that the proposed method can be applied to detect reference holes accurately and robustly in the jamming environment of aircraft assembly.