Local Deformable Template Matching in Robotic Deburring

Abstract

Burr is a common phenomenon in the manufacture of metal parts, which directly affects the machining accuracy and assembly precision. Based on machine vision, robot deburring is a very effective method. For casting deformation and improving the efficiency of robot deburring, a method of target workpiece contour recognition based on local deformable template matching and threshold segmentation for defining the burr size to adjust the robot deburring speed is proposed. Firstly, compared with the algorithm of linear deformable and shape-based template matching for the same type of workpieces that produce casting deformation, the recognition based on local deformable template matching achieves higher precision. Secondly, after extracting and matching the subpixel edge contours of the standard workpiece and the target workpiece, a method is proposed for defining the burr size by point-to-point distance threshold segmentation to adjust the robot deburring speed. Finally, the experiment of contour matching and robot deburring for three types of workpieces shows that the average of matching degree of local deformable template matching method reaches more than 93%, which is 12.90% and 16.13% higher than the linear deformable and shape-based template matching method. And the average of deburring time for each workpiece after adjusting the robot deburring speed is shortened by about 3.94%, which verifies the effectiveness of the proposed method.