自动插件机元件管脚视觉检测算法的研究  

Research of Visual Inspection Algorithm for Component Pins of Auto Insertion Machine

作　　者：韩浩宇 宋建 

机构地区：[1]华南理工大学聚合物新型成型装备国家工程研究中心聚合物成型加工工程教育部重点实验室,广东广州510640

出　　处：《机械设计与制造》2018年第3期109-111,115,共4页Machinery Design & Manufacture

基　　金：国家自然科学基金项目(11572129);广东省科技计划项目(2014B090921006)

摘　　要：针对自动插件机的应用需求,提出一种元件管脚视觉检测算法,可对管脚完整度及成型跨距进行检测。首先,对图像进行二值化处理,用连通区域面积特征及最小包围圆对管脚粗定位;其次,通过形态学处理构建ROI,运用Canny检测算子对目标边缘进行像素级定位,之后再根据曲线拟合原理对目标边缘进行亚像素级定位;最后,用权重函数将目标边缘进行圆拟合,获得管脚的精确位置信息。实验结果表明算法具有较高的稳定性,且算法耗时少,能满足自动机插件机的实际应用要求。According to the application requirement of auto insertion machine,we construct a visual inspection algorithm for component pins,which can inspect the integrity and center distance of pins. Firstly,using the area of connected region and the smallest enclosing circle after threshold the image to obtain the rough localization of pins. Secondly,creating the ROI through morphology and using the theory of fitting curve,sub-pixel target edges are obtained after Canny edge detection.Furthermore,we fit a circle robustly to the target edges using weight function to obtain the precise location of pins. The results of experiments demonstrate that the algorithm has high stability,less time consuming,andthe requirements of practical application in autoinsertion machine can be met by the algorithm.

关 键 词：自动插件机 管脚 视觉检测 亚像素级 权重函数 圆拟合 

分 类 号：TH16[机械工程—机械制造及自动化]