基于遗传算法的织物疵点特征选择

Feature selection of fabric defects based on genetic algorithm

为提高疵点分类的正确率,提出应用遗传算法对织物的疵点进行特征选择。首先提取机织物疵点图像,基于直方图、灰度共生矩阵、灰度差分统计、小波差分统计等描述纹理特征,采用遗传算法对这些特征组成的特征向量进行特征选择,再用支持向量机(SVM)分别对原特征向量和选择的特征子向量进行分类。实验结果显示,织物疵点的平均识别率从原来的89%提高到95%,说明该算法对织物疵点特征选择是有效的。

In order to improve the accuracy of defects classification, a genetic algorithm is proposed to apply feature selection to the fabric defects. The first step of this method was extracting the texture features of image defects, which are based on the characteristics of the histogram, the gray-level co-occurrence matrix （GLCM） features, gray-scale statistical difference, gray statistical difference in wavelet domain, etc. Then the genetic algorithm （GA） was applied to select these features of the composition of the feature vector. Finally, support vector machine （SVM） was used to classify the original feature vector and new feature vector, respectively. The results showed that the average recognition rate of fabric defects ascended from 89 percent to 95 percent, demonstrating that the method is valid for feature selection of fabric defects.