APPLICATION OF NEURAL NETWORK FOR FABRIC DETECTION

摘要

Fabric detection is one of the most important contents of textile detection and includes physical testing, wet testing and fastness and color measurement. Fabric evaluation still, for the most part, is done by people, so the inspectors' accuracy is a problem that should be improved for many years. By using the neural networks, that disadvantage can be overcome. Successful applications of neural networks to fabric defects have become well established, and other aspects of network involvement in fabric physical testing are stepping up to play significant roles. In this research, the influences of the number of hidden neurons on the convergence speed and the testing accuracy are investigated. The experimental data shows that the neural networks have strongly capability of self-adaptive recognition and are effective for fabric initial cool/warm feeling testing.