基于线结构光扫描的烟支表面缺陷检测方法研究  

Study on cigarette surface defect detection method based on linear structured light scanning

作　　者：钟琳[1] 徐海[1] 米芳芳[1] 赵德华 詹妮 丁凡 牛燕丽[1] ZHONG Lin;XU Hai;MI Fangfang;ZHAO Dehua;ZHAN Ni;DING Fan;NIU Yanli(China Tobacco Hubei Industrial Co.,Ltd.,Wuhan 430040,China)

机构地区：[1]湖北中烟工业有限责任公司,武汉430040

出　　处：《包装与食品机械》2024年第2期82-87,共6页Packaging and Food Machinery

基　　金：湖北中烟工业有限责任公司科研项目(2022JSGY4WH2B041)。

摘　　要：在烟支表面缺陷检测时,为解决传统二维图像方法无法获得深度信息的问题,提出一种基于线结构光传感器旋转扫描的检测方法。烟支在两同向辊轮的旋转带动下实现自转,通过线结构光视觉传感器获取物体轴向轮廓线,使用随机采样一致性拟合轮廓线,通过计算原始点到拟合轮廓的距离判断缺陷点。选取某品牌烟支在检测过程中得到的原始数据,其中多组样本的Kappa值均达到0.97,重复性达到99.3%。研究为烟支缺陷检测提供参考。In order to solve the problem of traditional two-dimensional image methods being unable to obtain depth information during surface detection of cigarettes,a new detection method based on rotary scanning of linear structured light sensor was proposed.The cigarette is rotated by the rotation of two rolling wheels in the same direction.The linear structured light vision sensor was used to obtain the axial contour of the object,the random sampling consistency was used to fit the contour,and the distance between the original point and the fitted contour was calculated to judge the defect point.The original data obtained from a certain brand cigarette during the detection process were selected,and the Kappa value of multiple groups of samples reached 0.97,with repeatability reaching 99.3%.The study provides reference for the subsequent detection of cigarette defects.

关 键 词：线结构光视觉传感器 烟支 随机采样一致性 直线拟合 

分 类 号：TS47[农业科学—烟草工业]