金属圆柱工件缺陷的光电检测  

Optoelectronic inspection of defects for metal cylindrical workpieces

作　　者：张静[1] 叶玉堂[1] 谢煜[1] 刘霖[1] 常永鑫[1,2] 

机构地区：[1]电子科技大学光电信息学院,四川成都610054 [2]中国科学院光电技术研究所,四川成都610209

出　　处：《光学精密工程》2014年第7期1871-1876,共6页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.61205004);粤港招标关键领域重点突破项目(No.W0511159)

摘　　要：针对金属工件外观缺陷检测存在光学照明不均、检测缺陷种类繁多、检测系统识别率不高等问题,研究了检测金属圆柱工件缺陷的方法。分析了局部二元模式(LBP)与局部图像方差强度(LVAR)的基本原理,研究了两者在金属纹理表面缺陷检测中的具体实现方法。采用LBP反应局部图形空间纹理模式,LVAR突出图像强度对比信息,然后用LVAR计算结果作为权重值来调整LBP的局部纹理提取和度量结果,实现了金属圆柱工件的自动缺陷检测。实验中采用步进电机控制工件旋转,配合线阵相机采集圆柱工件的展开图像。实验结果显示,这种方法有效克服了金属材质光照不均的缺点,对大量缺陷种类具有较高的鲁棒性,其检出率高达95.1%,漏检率为0%,满足了工业检测要求。To overcome the shortcomings of metal workpiece defect detection in optical uneven illumination,higher detection defect ranges and lower detection system recognition rate,a defect detection method was proposed.The basic principles of Local Binary Pattern（LBP）and Local Image Variance（LVAR）were analyzed,and their specific methods in the algorithm of metal cylindrical detection were discussed.The LBP was used to reflect local graphics texture pattern and the LVAR to outstand the contrast of image intensity.Then,the weight values calculated from LVAR were used to adjust the extraction and measurement of LBP local texture.Thus,the automatic detection of metal cylindrical workpieces was achieved.In the experiments,the rotation of workpieces was controlled by a stepper and the expanded images of cylindrical workpieces were captured by a linear CCD.The experimental results demonstrate that this method effectively overcome the shortcomings of metal uneven illumination and has high robustness to a large number of defect types.The detection rate hasreached to 99.5% and missing rate to 0%,which meets industrial inspection requirements.

关 键 词：金属圆柱工件 缺陷检测 线阵CCD 局部二元模式 局部图像方差 

分 类 号：TG806[金属学及工艺—公差测量技术] TP391[自动化与计算机技术—计算机应用技术]