I. Introduction
In the process of strip steel production, a perfect quality inspection system can effectively guarantee the yield of products [1]. However, the uncertainty of the production environment and the constraints of the production process will cause some defects on the surface of the strip that affect the performance of the product. The occurrence of defects will affect the safe use of strip steel. Under the general trend of increasingly strict requirements on product quality, accurate and efficient detection of product surface defects has become an urgent problem to be solved in the industry, which has caused research on different forms of defect detection methods in the industry [2].