Defect Detection and Localization for Stamped Parts Based on Machine Vision

Stamped parts are workpieces manufactured by applying external force to sheet metal, pipes, profile steel and other materials via press machines and dies, so as to realize plastic deformation or separation of the materials and obtain workpieces with the required shape and size. However, various defects tend to occur during production, which are randomly distributed. The main defects include missing corners, surface damage, missing patterns, broken press heads and so on.To detect the surface defects of stamped parts and improve the speed of screening out defective stamped parts, this paper combines the advantages of machine vision-based defect detection to increase the defect detection rate. Experimental results demonstrate that the software and algorithms applied in this system deliver satisfactory detection efficiency and stability for the defect detection of stamped parts. The innovation of this paper is that different detection methods are proposed for different types of defects. It also provides new ideas for the research of defect detection technology for stamped parts and has certain