基于组合特征和GA_SVM的织物疵点分类  

Classification of Fabric Defects Based on Combination Features and GA_SVM

作　　者：王春妍 郭宇 张晓丽 WANG Chunyan;GUO Yu;ZHANG Xiaoli(School of Intelligent Manufacturing Industry,Shanxi University of Electronic Science and Technology,Linfen 041000,China)

机构地区：[1]山西电子科技学院智能制造产业学院,山西临汾041000

出　　处：《纺织科技进展》2024年第10期20-23,共4页Progress in Textile Science & Technology

基　　金：山西省高等学校科技创新项目(2023L467)。

摘　　要：对织物疵点进行检测分割后,需要对疵点进行分类,采用组合特征和遗传算法优化参数的支持向量机模型对织物疵点进行分类识别算法。提取其几何特征和纹理特征,降低样本数据维度,构建特征向量,利用GA对SVM分类器参数进行优化,获取GA_SVM分类器最佳参数,对几种常见的疵点类型实现分类评定。对不同类型的织物图像进行检测,准确率达91%以上,可根据不同类型的疵点特性进行后续的修补或改进工艺流程,降低疵点产生的概率。After detecting and segmenting fabric defects,it is necessary to classify them and propose a support vector machine model that uses combination features and genetic algorithm optimization parameters to classify and recognize fabric defects.Its geometric and texture features were extracted and the feature vectors were constructed,and the genetic algorithm was used to optimize the classifer parameters of support vector machine,while the optimal parameter of GA_SVM classifier was obtained,achieving classification evaluation for several common defect types.The experiment detected images of different types of fabrics with an accuracy rate of over 91%,and subsequent repairs or improvement of the process flow can be carried out based on the characteristics of different types of defects,reducing the probability of defect occurrence.

关 键 词：疵点检测 组合特征 遗传算法(GA) 支持向量机(SVM) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]