基于加权局部熵的织物疵点检测  

Detection of Fabric Defect Based on Weighted Local Entropy

作　　者：赵训明[1] 邓中民[1] 冉友廷[1] 

机构地区：[1]武汉纺织大学

出　　处：《棉纺织技术》2012年第10期20-23,共4页Cotton Textile Technology

基　　金：湖北省重大项目(C2010036)

摘　　要：探讨基于加权局部熵的织物疵点检测方法的效果。针对图像的局部熵在织物疵点检测中的缺陷,在局部熵的基础上提出了加权局部熵的检测方法,并对织物缺纬、断经、破洞疵点进行了检测。结果表明:基于加权局部熵的织物疵点检测方法识别率在90%以上。指出:相比单一局部熵的方法,此方法在缺纬、断经、油污、破洞四类常见疵点的检测上更方便、更精确。Effect of fabric defect detection method based on weighted local entropy was discussed. According to defect of image local entropy in fabric defect detection, detection method of weighted local entropy was put forwards. Fabric missing picks, broken ends and hole defects were detected. The result shows that the fabric defects recognition rate based on weighted local entropy can be reached above 90%. It is pointed out that compared to single local entropy method, the new method is more convenient and accurate in detecting four kinds defects i.e. missing pick, broken ends, oil and hole defects.

关 键 词：疵点检测 加权局部熵 MATLAB 图像处理 缺纬 断经 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP274[自动化与计算机技术—计算机科学与技术]