基于人工神经网络的织物疵点检测  

Detection of Fabric Defect Based on Artificial Neural Network

作　　者：沈咏军 朱桂英[2] 

机构地区：[1]浙江广厦建设职业技术学院建筑工程系,浙江金华322100 [2]浙江理工大学信息电子学院,杭州310018

出　　处：《丝绸》2007年第6期38-41,共4页Journal of Silk

摘　　要：根据疵点的特征对常见织物疵点进行了简单的划分。采用直方图均衡化、二值化、中值滤波、腐蚀和膨胀等方法对织物图像进行一系列的预处理,对织物疵点的特征参数进行提取,利用人工BP神经网络来判别疵点的类别并进行分类。结果表明,利用BP神经网络识别织物疵点并进行分级是行之有效的。The common fabric defects were classified by according to their characteristic.The-defect images are pretreated by series of methods, such as histogram transformation, two-value transformation, middle-value wave flitting, eroding and dilating. Then some characteristic parameters of each defect were acquired. After that, artificial BP neural network was used to identify the classification of fabric defects. The results show that it is effective to identify and classify the fabric defects by using BP neural network.

关 键 词：人工神经网络 织物疵点 分类 质量等级评定 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]