应用Canny算子的织物疵点检测改进算法  

Improved algorithm for fabric defect detection based on Canny operator

作　　者：胡克满[1,2] 罗少龙 胡海燕 HU Keman;LUO Siolong;HU Haiyan(Department of Electronics&Information Engineering,Ningbo Polytechnic,Ningbo,Zhejiang 315800,China;Faculty of Information Technology,Macao University of Science and Technology,Macao 999078,China;Technology and Academia-Industry Cooperation Office,Ningbo Polytechnic,Ningbo,Zhejiang 315800,China)

机构地区：[1]宁波职业技术学院电子信息工程学院,浙江宁波315800 [2]澳门科技大学资讯科技学院,澳门999078 [3]宁波职业技术学院科技与产学合作处,浙江宁波315800

出　　处：《纺织学报》2019年第1期153-158,共6页Journal of Textile Research

基　　金：国家自然科学基金项目(11771226);浙江省教育厅科研项目(Y201738411)

摘　　要：为克服当前Canny算子在织物疵点边缘检测中存在的阈值设定、滤波参数选择等自适应问题,提出一种基于Canny算子的改进算法。通过分析不同种类的织物疵点特征,选择不同参数的高斯滤波器,对织物疵点图像进行滤波处理;采用自适应形式获取图像边缘信息的阈值,避免了因阈值取值过高或过低而无法获得较好织物疵点的边缘信息的问题,同时还可根据不同织物疵点类型选择不同的滤波参数。结果表明,改进后的Canny算法可有效地检测到织物疵点的边缘细节,具有较好的自适应能力,并且提高了算法的有效性。同时对典型的织物疵点进行检测并与传统算法比较,其检测效果更优。In order to solve the self-adaption problem that current Canny operator needs to set threshold and to choose the filtering parameters in the fabric defect edge detection,an improved algorithm based on the original Canny operator was proposed.Firstly different filter parameters were chosen according to the type of the fabric flaw,and then self-adaption was used to obtain the threshold and the parameters of the filter,which avoids the wrong choosing of threshold leading to the lack or redundant edge information,and different filter parameters were chosen according to the type of the fabric flaw.The results showed that the improved Canny algorithm can detect the edge detail of fabric defects,and has good self-adaption capability.Compared to the conventional algorithm,the improved algorithm has better detection results.

关 键 词：CANNY算子 自适应 织物疵点 滤波参数 边缘信息 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]