摘要
Abstract
Due to the expansion of production scale and the complexity of processing flow,fabric defects have become a common problem in the textile industry.The traditional textile defect detection methods mainly rely on manual visual inspection,which is prone to problems such as visual fatigue and low efficiency.Therefore,a textile defect detection system based on edge computing and improved convolutional neural network is proposed,which realizes the functions of data collection,defect detection,reporting and early warning of textile.The system com-bines the global receptive field of convolutional neural networks with attention mechanisms to efficiently capture and classify defect features,enabling textile enterprises to perform defect detection calculations and data processing in production environments.The overall accuracy of de-fect detection can reach 97.11%,with defect product detection accuracy of 84.02%,defect free product detection accuracy of 97.98%,and de-fect detection recall rate of 95.28%.The system has excellent application effects in practical scenarios,effectively improving detection efficien-cy,strengthening product quality control,and reducing production costs.It is of great significance for realizing the intelligent development of the textile industry.关键词
边缘计算/卷积神经网络/瑕疵检测/智能纺织/自注意力机制Key words
edge computing/convolutional neural network/defect detection/intelligent textile/self-attention mechanism分类
信息技术与安全科学
