Brief description

The article is devoted to the problem of performing quality control of textile materials using technical vision systems. The authors developed a methodology and model for multitasking neural network training, and also developed software. The software is designed to recognize moving roll materials of large width with different types of surface design. The authors also formed a database of visual representation of defects in the appearance of textile materials and their classification in accordance with the various technological stages of textile and finishing production, where these defects are formed.