基于机器视觉的疵点检测方法的研究进展  

Research Progress of Defect Detection Method Based on Machine Vision

作　　者：王孟涛 李岳阳 杜帅 WANG Mengtao;LI Yueyang;DU Shuai(Engineering Research Center for Knitting Technology,Ministry of Education,Jiangnan University,Wuxi 214122,China)

机构地区：[1]江南大学教育部针织技术工程研究中心

出　　处：《现代纺织技术》2019年第5期57-61,共5页Advanced Textile Technology

基　　金：国家自然科学基金(61772238);江苏省产学研联合创新资金-前瞻性联合研究项目(BY2016022-21)

摘　　要：机器视觉在织物疵点检测方面具有广泛的应用,对纺织品质量控制具有重要的现实意义。本文通过对近几年相关文献的分析,简要综述关于疵点检测各种方法,并按照对织物图像处理的方法不同可分为:结构方法、统计学方法、频谱分析方法、基于模型的方法和基于学习的方法。着重阐述了各类方法的优缺点,并总结该领域的发展趋势。Machine vision is widely applied in the detection of fabric defects,and carries important practical significance for the quality control of textiles.Through the analysis of relevant literatures in recent years,the various methods of defect detection are summarized,and they can be divided into structural method,statistical method,spectrum analysis method,model-based method and learning-based method in accordance with the different methods of fabric image processing.This paper highlights the advantages and disadvantages of various methods and summarizes the development trends of this field.

关 键 词：图像处理 纺织品 疵点检测 机器视觉 

分 类 号：TS186.3[轻工技术与工程—纺织材料与纺织品设计]