Texture surface defect detection method and system

Abstract

The invention belongs to the field of image processing, and discloses a texture surface defect detection method and system. The method comprises an offline training stage and an online detection stageI; in the offline training stage, multi-channel texture priori of an input texture image is extracted through a priori extraction step; a texture background of the input image is accurately reconstructed through a texture reconstruction step under the guidance of the extracted priori, more accurate texture features are encoded in a hidden space of a texture reconstruction module, and defects are inhibited from being reconstructed in the texture background; and finally, through pixel-level adversarial learning, the texture reconstruction precision is further improved. During detection, the defects can be detected only by subtracting the reconstructed texture background image from the input image. The method has high detection precision for defects with different sizes and different contrastratios on different texture surfaces.

Description

technical field [0001] The invention belongs to the field of image processing, and relates to a texture surface defect detection method and system, and more specifically, to a texture surface defect detection method and system based on a priori guided feature coding confrontation network. Background technique [0002] In the field of industrial manufacturing, the quality of raw materials varies, and the manufacturing process is complex. Surface defects often occur on the surface of products, such as steel, wood, textiles, ceramic tiles, and new display devices such as TFT-LCD, OLED, etc. Since the surface of various products often exhibits texture features, surface defects refer to localized areas that are different from the surrounding texture structure and pattern, or localized areas with irregular brightness changes. These texture surface defects will directly reduce product quality and affect user experience. In order to improve the production quality, all types of surf...

AI Technical Summary

Problems solved by technology

[0006] In view of the above defects or improvement needs of the prior art, the present invention provides a texture surface defect detection method and system, the purpose of which is to use the powerful data modeling capability of deep learning to conduct industrial Product surface defect detection, which solves the problem of low detection accuracy caused by different types of texture surface defect detection algorithms, large-scale changes, extremely low contrast, irregular brightness changes, changeable shapes, and lack of samples. Can greatly improve the detection accuracy

Images