PRODUCT DEFECT DETECTION METHOD AND APPARATUS, ELECTRONIC DEVICE AND STORAGE MEDIUM

摘要

The present application relates to the fields of computer vision, image processing, deep learning, cloud computing, and the like, and discloses a product defect detection method, an apparatus, an electronic device, a storage medium, and a computer program. A specific implementation method is to acquire a multi-channel image of a target product; Input the multi-channel image to the defect detection model, the defect detection model includes a plurality of convolutional branches, fusion modules, and convolutional full branches; Performing feature extraction for each channel in the multi-channel image using a plurality of convolution branches to obtain a plurality of first feature information; Fusing the plurality of first feature information using the fusing module to obtain second feature information; Performing feature extraction on the second feature information using the entire branch of the convolution to obtain third feature information output by the defect detection model; According to the third feature information, defect information of the target product is determined. By implementing the technical solution of the present application, the fusion effect of various features can be reinforced.