Metallic workpiece surface defect recognition method and device based on machine vision

Abstract

The invention relates to a metallic workpiece surface defect recognition method and device based on machine vision. The method comprises the steps that a surface image of a metallic workpiece is acquired, and the surface image of the metallic workpiece is converted into a grayscale image with edge features through image preprocessing; image features of the grayscale image are enhanced through image enhancement processing; feature image data of the grayscale image obtained after enhancement processing is extracted through image segmentation processing; and the feature image data is input into apre-trained metallic workpiece defect recognition model, and a defect recognition result of the metallic workpiece is output. The device comprises a manipulator, a lifting stock bin, image collectionequipment and a central processing module, wherein the central processing module is used for performing defect recognition on the surface image of the metallic workpiece, generating a control instruction through the defect recognition result and controlling the manipulator to clamp and place the metallic workpiece according to the control instruction. The metallic workpiece surface defect recognition method and device provide better help for optimizing the recognition result of a surface defect of the metallic workpiece.

Description

technical field [0001] The invention relates to the technical field of workpiece defect identification, in particular to a method and device for identifying surface defects of metal workpieces based on machine vision. Background technique [0002] In machinery manufacturing enterprises, most of the production enterprises use manual inspection to detect the surface quality of metal processing workpieces. It cannot be detected and controlled in time. The method of manual inspection is based on human subjective judgment, so it has the problem of low accuracy. When defective parts are not screened out and mixed with normal parts, it takes a lot of labor costs to separate them again. In addition, the manual inspection method not only has high input costs, but also has low efficiency. Therefore, production enterprises urgently need a method with high accuracy, high efficiency, low cost, and capable of separating defective parts from normal parts. [0003] In view of the above p...

AI Technical Summary

Problems solved by technology

[0004] The above method can realize the quality inspection of columnar products, but the surface shape of the machined metal workpiece is more complicated, and the surface of the metal workpiece is affected by interference factors such as light and workpiece surface burrs after processing, so that the defect features are not obvious and not obvious. protrude

If the above-mentioned defect detection method is directly applied to the defect detection on the surface of metal workpieces, there will be the following problems: the gray scale analysis method in this method is limited by the gray scale, and when the gray scale of a small number of different types of objects is very different from each other, it can However, in the face of metal workpieces with complex surface shapes, there are no obvious grayscale differences or large grayscale value range overlaps in the images, so it is difficult to obtain accurate segmentation results, and the images obtained after segmentation cannot It shows the details of the image very well, so that the effect of identifying the surface defects of the metal workpiece is not good.

Images