实Gabor滤波器的设计及其应用  

Design of Real Gabor Filters and Their Applications

作　　者：陈泽虹[1] 冯晓霞[1] 吴崇茂[1] 

机构地区：[1]闽南师范大学数学与统计学院,福建漳州363000

出　　处：《北华大学学报（自然科学版）》2013年第5期604-609,共6页Journal of Beihua University（Natural Science）

基　　金：福建省自然科学基金项目(2010J01018)

摘　　要：针对常见疵点的特征,以及Gabor滤波器对织物图像的检测效果,考虑计算的复杂度和检测的适用范围,设计了4个椭圆形的实Gabor滤波器,分别分布在水平和垂直方向,同一方向上2个不同尺度的滤波器相切.基于这些滤波器,提出了一种织物疵点检测算法.特别地,在图像融合的环节,给出了一种新方法.实验表明,新算法能够准确地检测出多种织物疵点.最后,从算法步骤、计算复杂度、实验效果三个方面对新算法与文献算法进行了比较.结果表明,新算法简单,计算复杂度较低,检测效果较好.Based on the features of common defects, detection effect of fabric images by Gabor filters, computational complexity and application scope, four ellipse-shaped real Gabor filters are designed, which distributed at horizontal and vertical directions, two filters in different scales are tangent respectively in every direction. Furthermore,a new algorithm based on those filters is proposed to detect fabric defects, especially, a new method is presented in the step of fusing subimages. The experimental results show that the new algorithm can detect many kinds of fabric defects accurately. Finally, the new algorithm are compared with Zou Chao＇ s one in three aspects：algorithm steps, computational complexity and experimental effect, the results indicate that the new algorithm has lower complexity and better effect.

关 键 词：实Gabor滤波器 织物疵点检测 滤波 图像融合 二值化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]