The autonomous detection of sheet metal parts using imageprocessing

Abstract

Robot-assisted bending is being applied to the sheet metal bending gradually; however, almost all of the sheet metal parts are positioned by a mechanical positioning device before fed into the press brake. This process is time-consuming and not intelligible. To address this problem, an approach to detect the position of the sheet metal part automatically is proposed in this paper. A passive vision system is designed to capture the image of the sheet metal parts stacked on the “input” pallet. Then, the position of the top sheet metal part is detected by the proposed method. The proposed method consists of region extraction, edge extraction, and position calculation. First, image mask and mathematical morphology are adopted to extract the region roughly. Then, modified canny edge operator is used to extract the edges, and edges of the top sheet metal part are selected. Finally, the position of the top sheet metal part is obtained by probabilistic Hough line transform and the classification of lines. Experiments in different conditions were carried out. And the experimental results show the feasibility and stability of the proposed method and demonstrate the real time of the image processing. The method can replace the mechanical positioning in the robot-assisted bending and eliminates the positioning time which is cost by mechanical positioning.