Industrial defect intelligent identification method for negative electrode material graphite box plate

Abstract

The invention relates to the technical field of defect detection, and discloses an intelligent identification method for industrial defects of a negative electrode material graphite box plate, which comprises the following steps of: acquiring a surface image through high-resolution imaging equipment, highlighting defect characteristics by utilizing an edge enhancement algorithm, inputting the preprocessed image into a deep learning model for classification and positioning, and obtaining a deep learning model; after a defect distribution diagram is generated, grading judgment and automatic generation of a quality analysis report are realized. According to the method, the defect detection precision and efficiency can be remarkably improved, the generalization ability is high, the method is suitable for detecting various types of defects, meanwhile, the system development and maintenance workload is reduced, a scientific basis is provided for production management, and the production efficiency and the product quality are effectively improved.

Images