基于小波静态分解的离散小疵点检测  

Small fabric defect detection based on wavelet static decomposition

作　　者：管声启[1,2] 石秀华[2] 王筠[1] 

机构地区：[1]西安工程大学机电工程学院,西安710048 [2]西北工业大学航海学院,西安710072

出　　处：《天津工业大学学报》2010年第5期73-76,共4页Journal of Tiangong University

基　　金：西安工程大学博士科研基金项目(BS1005)

摘　　要：在分析织物疵点特点的基础上,提出了一种新的检测方法.首先根据正常织物小波分解后的能量来确定小波基和小波分解层数.接着对测试图像进行小波静态分解,然后提取子图中的子窗口标准差作为提取的特征;在此基础上,通过与正常子图中的特征相比较来判断小疵点是否存在.最后通过提取特征构造的标准差图确定小疵点的位置.实验表明,该检测方法是有效的.On the basis of analysis of the defect characteristicsof fabrics,a new method for small defect detection was presented.Firstly,the normal fabric energy of wavelet decomposition was used to determine the wavelet func-tion and level of wavelet decomposition.The test image was decomposed statically by wavelet,and then stan-dard deviation of sub-window in the sub-image was extracted as the extracted features,which was compared with the normal sub-image features to determine the existence of defects.Finally,the location of defect was determined by the standard deviation image.Experimental results show that this method is effective.

关 键 词：小波分解 特征提取 标准差图 疵点检测 

分 类 号：TS101.97[轻工技术与工程—纺织工程] TP391.41[轻工技术与工程—纺织科学与工程]