基于匹配点变形估计的布标缺陷检测算法  

Label defect detection algorithm based on matching point deformation estimation

作　　者：宋强 高国刚 李文峰 冯东海 贡军民 SONG Qiang;GAO Guogang;LI Wenfeng;FENG Donghai;GONG Junmin(Faculty of Engineering,China University of Petroleum-Beijing at Karamay,Karamay,Xinjiang 834000,China;Zhongheng Yongchuang(Beijing)Technology Co.,Ltd.,Beijing 100089,China;Engineering Technology Branch,Xinjiang Petroleum Administrative Bureau,Karamay,Xinjiang 834000,China)

机构地区：[1]中国石油大学(北京)克拉玛依校区工学院,新疆克拉玛依834000 [2]中恒永创(北京)科技有限公司,北京100089 [3]新疆石油管理局工程技术公司,新疆克拉玛依834000

出　　处：《毛纺科技》2021年第10期73-76,共4页Wool Textile Journal

基　　金：新疆维吾尔自治区高校科研计划项目(XJEDU2017M047)。

摘　　要：基于机器视觉和图像处理的全自动检测技术是布标缺陷检测的发展趋势。布标产品刚度较低,在检测过程中受张力作用,容易产生大的不均匀变形,导致模板匹配检测算法的准确率不高。针对该问题,文章运用多匹配点检测思路,利用光流法分析计算布标模板图像与待检测布标图像对应匹配点的相对运动矢量,该运动矢量反映待检测布标受张力作用引起的形变。通过逆变换可消除待检测布标的受力变形,提高模板匹配检测算法的准确率。该研究成果对于拓展光流分析应用领域、提高全自动布标缺陷检测准确率、提升服装行业检测自动化程度,具有重要理论价值与工程意义。Automatic detection technology based on machine vision and image processing is the development trend of label defect detection.However,due to the low stiffness of the label,large uneven deformation is easy to occur due to the tension in the detection process,which leads to the low accuracy of template matching detection algorithm.Aiming at this problem,the idea of multi matching point detection was used in this paper,optical flow analysis was used to calculate the relative motion vector of the matching point between the label template image and the label image to be detected.Through the inverse transformation,the stress deformation of the label to be detected can be eliminated,and the accuracy of the template matching detection algorithm can be improved.The research results of this paper have important theoretical value and engineering significance for expanding the application field of optical flow analysis,improving the detection accuracy of automatic label inspection,and improving the detection automation of clothing industry.

关 键 词：布标 缺陷检测 机器视觉 光流分析 

分 类 号：TS107[轻工技术与工程—纺织工程]