改进的基于最大熵和Gabor滤波器的织物疵点检测  

Improved method of fabric defects detection based on maximum entropy and Gabor filter

作　　者：赵波[1] 郑力新[2] 

机构地区：[1]华侨大学计算机科学与技术学院,福建厦门362021 [2]华侨大学信息科学与工程学院,福建厦门362021

出　　处：《微型机与应用》2011年第13期29-31,35,共4页Microcomputer & Its Applications

基　　金：福建省高等学校新世纪优秀人才支持计划(07FJRC01);福建省产业计划开发项目(25201071)

摘　　要：提出了一种改进的基于最大熵原理和Gabor滤波技术的织物疵点检测方法。采用多通道Gabor滤波算法,取模值特征为输出,利用最大熵分割模值图像,再进行图像融合,最后计算轮廓的周长和面积去除孤立点得到最终检测结果。利用OpenCV算法库,选取了四种具有代表的织物疵点图片进行验证,实验结果表明,该方法降低了计算复杂度、检测速度快、检测效果好、无须事先学习,适用于不同疵点类型的各种检测。An improved design method based on maximum entropy and Gabor filter for fabric defect detection is proposed. This method emploies the multi-channel Gabor filter algorithm and takes the amplitude as a output characteristics, uses maximum entropy to segment the amplitude image and then fused the sub-images. The result is achiered by calculating perimeter and area of the contours of removal of isolated points. The experiment selects four types of typical fabric defect images and OpenCV library is used. The experiment result shows that the algorithm can reduce the computational complexity, without prior learning, and also characterized by high testing speed and testing effect, it can be applied to a variety of different typos of defect detection.

关 键 词：疵点检测 GABOR滤波器 最大熵 OPENCV 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]