轮胎帘线缺陷特征提取的图像处理算法研究  

Study on Image Processing Algorithm of Tire Cord Defect Feature Extraction

作　　者：徐培[1] 马铁军[1,2] 林丽红 XU Pei;MA Tie-jun;LIN Li-hong(School of Mechanical ＆ Automotive Engineering,South China University of Technology,Guangdong Guangzhou 510640,China;Guangzhou SCUT Bestry Technology Corporation,Guangdong Guangzhou 510530,China)

机构地区：[1]华南理工大学机械与汽车工程学院,广东广州510640 [2]广州华工百川科技股份有限公司,广东广州510530

出　　处：《机械设计与制造》2018年第12期27-31,共5页Machinery Design & Manufacture

基　　金：国家自然科学基金项目(51278201)

摘　　要：根据轮胎X光图像在缺陷区域表现出不同的灰度分布特点,研究了一种通过分析统计像素点多种特征的缺陷检测方法。使用形态学二值细化算法对原始图像细化处理,对帘线缺陷分类统计局部缺陷特征和曲线缺陷特征。跟踪并记录所有细化曲线的像素点,根据像素邻域的目标灰度统计特征判断局部缺陷;根据曲线的斜率和曲率直方图及弦弧距-弦长比曲线判断曲线缺陷。分析对比无缺陷轮胎X光图和缺陷轮胎X光图,结果表明该方法区分度高,可同时完成轮胎帘线断线、杂质、交叉和弯曲多种缺陷的自动化检测。According to the different gray level distribution characteristics of the tire X-ray image in the defect area, a defect detection method based on the analysis of the statistical features of pixelswas studied.The original image was thinned based on morphological two valued thinning algorithm. The cord defects were classified as local defects and curve defects which had different statistical characteristics. Tracking and recording all pixel points of the thinning curve, localizeddefectswere judged according to the statistical characteristics of the target gray value of the pixel neighborhood; then cord bending was judged according tothe curve slope and curvature and chord arc distance-chord length ratio. By analyzing and comparing the normal tire and the defective tire, the results show that the method hadhigh degree of discrimination and can be used to detect tire cord breakage, impurity, cross and bend.

关 键 词：轮胎帘线 缺陷识别 弦弧距-弦长比 斜率 曲率 

分 类 号：TH16[机械工程—机械制造及自动化] TP391.4[自动化与计算机技术—计算机应用技术]