Cloth defect detection method based on adversarial neural network, terminal and storage medium

Abstract

The invention discloses a cloth defect detection method based on an adversarial neural network, a terminal and a storage medium. The method comprises the following steps: 1, training a GAN model according to a positive sample of cloth, and outputting the GAN model after the training is completed; 2, sending a detection image sample of the cloth to the GAN model, and reconstructing the detection image sample by the GAN model to generate a good reconstructed image sample; and performing structural similarity comparison on the detection image sample and the reconstructed image sample, and judgingwhether the detection image sample has defects or not according to the difference degree of comparison. According to the invention, the GAN model is trained and generated through the positive sampleof the cloth, and the test image sample of the cloth is reconstructed through the GAN model to generate a reconstructed image sample, and the structural similarity between the test image sample and the detection image sample is obtained through comparison to judge whether defects exist or not, so that various defects of the cloth can be detected.

Description

technical field [0001] The invention relates to the technical field of cloth defect detection, in particular to a cloth defect detection method based on an adversarial neural network, a terminal and a storage medium. Background technique [0002] With the rapid development of the economy, my country's manufacturing industry is also developing rapidly, and products can be manufactured on a large scale every day for market demand. The increase in the number and types of products has led to an increase in people's requirements for product quality, and products with beautiful appearance are more likely to be favored by consumers. The surface quality of the product will affect the commercial value of the product, and the appearance defect will directly cause the depreciation of the commercial value of the product. The surface quality of the product has an important impact on the direct use or deep processing of the product. In the process of product production, due to the influe...

AI Technical Summary

Problems solved by technology

[0006] The existing deep learning detection algorithms mainly include: Faster Rcnn, SSD, Yolov3, etc. These algorithms have good detection effects for general scenarios, but the effect is not good in the cloth industry. The main reason is that there are too many types of cloth and the shapes of defects are different. And for the complex and ever-changing background of plaid pattern and printing, there is no good solution so far

Deep learning has requirements for data samples. It is time-consuming and laborious to collect hundreds of cloth defects. In the final cloth manufacturing process, processes such as shaping, mercerizing, and liquid ammonia are very fast. Deep learning has high computational complexity. It is useful for online detection of cloth. very challenging

Images