基于机器视觉的线缆表面缺陷快速检测算法研究  

Research on Fast Detection Algorithm of Cable Surface Defect Based on Machine Vision

作　　者：王海芳 焦龙 乔湘洋 祁超飞 WANG Hai-fang;JIAO Long;QIAO Xiang-yang;QI Chao-fei(School of Control Engineering,Northeastern University at Qinhuangdao,Qinhuangdao Hebei 066004,China)

机构地区：[1]东北大学秦皇岛分校控制工程学院

出　　处：《组合机床与自动化加工技术》2020年第2期119-122,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：秦皇岛科技支撑项目(201501B011);秦皇岛市大学生科技创新创业专项资金项目(2018-79,121)

摘　　要：利用机器视觉技术检测线缆表面缺陷时,检测时间长、漏检率高。为此,提出一种基于机器视觉的线缆表面缺陷快速检测算法。通过引入CV-Kmeans区域分类算法建立自适应滤波窗口改进高斯滤波算法,在此基础上建立自适应模板,然后计算原图像与模板的Pearson(皮尔逊)相关系数快速判断图像是否含有缺陷。对含有缺陷的图像进行模板与原图差分,最后对差分所得到的图像用自适应阈值分割法提取缺陷。实验表明,算法可有效识别缺陷并减少检测时间,漏检率为3.22%,满足线缆生产需求。When the machine vision technology is used to detect the surface defects of cables,the detection time is long and the rate of leakage is high.Therefore,proposes a fast detection algorithm for cable surface defects based on machine vision.By introducing the cv-kmeans regional classification algorithm to establish the adaptive filtering window and the improved Gaussian filtering algorithm,an adaptive template is established on this basis.Then,the Pearson correlation coefficient between the original image and the template is calculated to quickly determine whether the image contains defects.The image with defects was differentiated between the template and the original image,and finally the image obtained by the difference was extracted by adaptive threshold segmentation method.Experiments show that this algorithm can effectively identify defects and reduce the detection time,with a miss detection rate of 3.22%,meeting the needs of cable production.

关 键 词：机器视觉 表面缺陷检测 自适应模板 Pearson相关系数 

分 类 号：TH16[机械工程—机械制造及自动化] TG65[金属学及工艺—金属切削加工及机床]