基于机器视觉的四方垫片内圆缺陷检测方法  

Inner Circle Defect Detection Method of Square Gasket Based on Machine Vision

作　　者：易焕银 郭新[1] 王贵恩[1] 郝建豹[1] 林子其 邱桂茂 YI Huan-yin;GUO Xin;WANG Gui-en;HAO Jian-bao;LIN Zi-qi;QIU Gui-mao(Guangdong Communication Polytechnic,Guangzhou 5108000,China)

机构地区：[1]广东交通职业技术学院,广东广州510800

出　　处：《广东交通职业技术学院学报》2020年第2期41-42,共2页Journal of Guangdong Communication Polytechnic

基　　金：2018年度广东省普通高校青年创新人才类项目(项目编号:2018GkQNCX005);2019年广东大学生科技创新培育专项资金项目(项目编号:pdjh2019b0748)。

摘　　要：提出了一种基于机器视觉的四方垫片内圆缺陷检测方法。首先利用广义Hough变换定位垫片的中心点位置,然后根据设定阈值对原图进行二值化、标记连通区域并删除最大连通区域外的区域,再分段搜索垫片内圆边沿点,最后用最小二乘法拟合圆后检测各边沿点是否为缺陷位置。算法对任意摆放位置和角度的四方垫片都能成功检测,实时性、准确率都可以满足工业检测的需求。A method for detecting inner circle defects of square gasket based on machine vision is proposed.First of all,locate the center point position of the gasket by using generalized Hough transform.Then,binarize the original image according to the set threshold,label the continuous areas and remove all areas other than the biggest.After that,search the edge points of inner circle of the gasket section by section.At last,fit the circle using least square method and detect whether the edge points are defect position or not.The algorithm can successfully detect the square gasket at any position and angle,and both real-time and accuracy can meet the needs of industrial inspection.

关 键 词：机器视觉 四方垫片 缺陷检测 广义HOUGH变换 最小二乘法 

分 类 号：TP391.76[自动化与计算机技术—计算机应用技术]