基于深度学习的织物疵点检测研究进展  

Progress in fabric defect detection based on deep learning

作　　者：贺智明[1] 彭亚楠 HE Zhiming;PENG Yanan(School of Information Engineering,Jiangxi University of Science and Technology,Ganzhou,Jiangxi 341000,China)

机构地区：[1]江西理工大学信息工程学院

出　　处：《毛纺科技》2019年第8期83-88,共6页Wool Textile Journal

摘　　要：针对深度学习在图像处理、目标检测等领域中的应用,综述了几种常用的织物疵点检测方法,主要分为基于结构的方法、基于频谱的方法、基于统计的方法、基于模型的方法和基于学习的方法,概括这些方法的原理并比较分析其优缺点。着重阐述基于深度学习的织物疵点检测方法和发展状况,分析其未来研究方向,为相关研究提供学术参考。In recent years,deep learning has achieved great success in image processing,target detection and other fields,providing a new method for fabric defect detection.The commonly used fabric defect detection methods were summarized,which are mainly divided into structure-based method,spectrum-based analysis method,statics-based method,model-based method and learning-based method.The principles of these methods were summarized and compared,and their advantages and disadvantages were analyzed.Then,the main methods and development status of fabric defect detection based on deep learning technology in recent years were described,and the future research direction in this field was analyzed,providing valuable academic reference for relevant researchers.

关 键 词：织物疵点 图像处理 目标检测 深度学习 

分 类 号：TS101.9[轻工技术与工程—纺织工程]