基于优化Gabor滤波器的织物疵点检测算法  

Fabric Flaw Detection Algorium Based on Optimal Gabor Filters

作　　者：邢亚敬[1] 韩润萍[1] 姜玲[1] 

机构地区：[1]北京服装学院工业设计与信息工程学院,北京100029

出　　处：《北京服装学院学报（自然科学版）》2008年第1期45-50,共6页Journal of Beijing Institute of Fashion Technology：Natural Science Edition

基　　金：湖北省数字化纺织装备重点实验室开放课题资助(DTL200707)

摘　　要：提出了一种基于优化Gabor滤波器的织物疵点检测算法.首先介绍了Gabor滤波器的滤波原理,然后提出了Gabor滤波器优化算法,该算法针对一幅标准织物图像,通过计算Gabor滤波器与标准织物图像卷积结果的Fisher Cost函数最小值,得到一组优化Gabor滤波器参数.使用具有这些参数的Gabor滤波器分别对标准织物图像和样本织物图像进行滤波处理,将滤波结果进行比较和融合,得到检测结果.实验证明:该算法能获得良好的检测效果.Presented a new automatic and fast approach for textile flaw detection based on optimal Gabor filters. The theory of Gabor filters was introduced, and the effective method based on the temple image was developed. The temple image was convolved with the Gabor filters in order to obtain the filtered images. By minimizing the fisher cost function of the filtered images, the optimal parameters of Gabor filters were derived. Then the temple and sample image were dealt with the Gabor filters, respectively, with optimal parameters. By comparing and fusing the convolved results, the detection result was derived. Results showed that this adaptable approach could achieve the flaw detection precision.

关 键 词：GABOR滤波器 织物疵点 疵点检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]