基于PCA算法的织物疵点图像特征选择  

Feature selection of fabric image defects based on PCA algorithm

作　　者：杨莉国[1] 欧付娜[1] 赵静[1] 武善清[1] 

机构地区：[1]青岛理工大学费县校区机电系,山东临沂273400

出　　处：《数字技术与应用》2011年第4期42-43,共2页Digital Technology & Application

摘　　要：为提高疵点分类的正确率,提出PCA算法对织物疵点图像进行特征选择。首先提取正常和带瑕疵织物图像的灰度共生矩阵、灰度梯度共生矩阵、小波域统计特征共47个特征值,然后采用PCA算法对其进行特征选择,最后利用支持向量机对重新选择后的的特征向量进行分类。实验结果表明利用PCA算法选择的特征缩短了支持向量机的训练建模时间,对常见织物疵点的正确分类率由原来的81.65%提高到93.12%。In order to improve the accuracy of defects classification,a PCA algorithm is proposed to apply feature selection to the fabric defects images. The first step of this article was extracting the texture features of image defects,which are based on the 47 characteristics of the gray-level co-occurrence matrix features,Gray Level Co--occurrence matrix and wavelet features, then use PCA algorithm to select these features of the composition of the feature vector.finally classfieates those new features basede on support vector machine algorithm.The experimental results show that features that seleted by PCA algorithm in this artice can reduce training time and the classification rate of fabric defects ascended from83.45 percent to 93.12 percent.

关 键 词：PCA 特征选择 疵点检测 支持向量机 分类 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]