基于小波提升格式的织物疵点检测  

Fabric defect detection based on wavelet lifting scheme

作　　者：管声启[1] 石秀华[1] 

机构地区：[1]西北工业大学航海学院,西安710072

出　　处：《计算机工程与应用》2008年第25期219-221,228,共4页Computer Engineering and Applications

基　　金：西安市科技攻关资助项目(No.GG04039)

摘　　要：为了满足织物疵点检测快速而准确的要求,提出了一种基于小波提升格式疵点检测的新方法。首先根据已知滤波器,通过提高消失矩阶次构造与织物纹理相匹配的小波。在此基础上,对构造小波的滤波器进行提升和对偶提升,来获得不同的提升算子和对偶提升算子,从而实现小波的提升分解。最后提取小波分解后的高频子图细节特征,通过与正常织物高频子图细节特征相比较,从而实现疵点检测。实验证明了该方法是可行有效的,检测准确率达到92.5%以上。To fabric defect detection of rapidity and accuracy,a new method for defect detection based on wavelet lifting scheme is presented.Firstly,according to known the set of finite filters,the new wavelet with matching fabric texture properties is constructed by improving Vanishing Moments.Secondly,the set of construction wavelet filters are followed by lifting and dual lifting,and different lifting operators and dual lifting operators are obtained to wavelet decomposition.Lastly,the detail signal after wavelet decomposition is extracted,and it is compared with the detail signal of normal fabric to detect defect.The experimental result confirms that the proposed method is validity and feasibility,and the detection accuracy rate is over 92.5%.

关 键 词：提升格式 小波构造 疵点检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]