基于PCNN和SVM的帘子布疵点识别算法

Algorithm of cord defect detection based on improved PCNN and SVM

为了使用机器对帘子布疵点进行有效的检测和分类,根据帘子布灰度分布的特点,分析和讨论了改进型PCNN模型,利用改进型PCNN对帘子布疵点特征值进行提取,然后将支持向量机作为分类器最终实现疵点图像的识别.实验结果表明,该方法的疵点识别率较高,在90%以上,是一种简单有效的识别方法.

In order to detect and classify the cord defection effectively by use of the machines and according to the re-quirements of the cord fabric defect detection,the improvement of PCNN is analyzed and discussed.An im-proved PCNN is used to extract the eigenvalues of the cord fabric.The SVM is finally used to classify different kinds of the Cord Fabric.The experiment shows that the algorithm has higher rate of recognition rate（more than 90%）,which is a simple and effective recognition algorithm.