基于机器视觉的圆柱滚子尺寸检测方法  

Cylindrical roller size detection method based on machine vision

作　　者：郭亚盛 张硕 张爱梅[1] GUO Yasheng;ZHANG Shuo;ZHANG Aimei(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450001,China;Henan Institute of Metrology,Zhengzhou 450001,China)

机构地区：[1]郑州大学机械与动力工程学院,郑州450001 [2]河南省计量科学研究院,郑州450001

出　　处：《现代制造工程》2021年第4期109-113,共5页Modern Manufacturing Engineering

基　　金：河南省科技攻关项目(181200210100)。

摘　　要：以轴承圆柱滚子为研究对象,提出一种基于机器视觉的圆柱滚子尺寸检测方法,利用圆柱滚子的图像实现圆柱滚子直径与长度的同步检测。首先,对相机拍摄的圆柱滚子照片做预处理:使用灰度线性变换提高圆柱滚子与背景的对比度,再使用二值化阈值分割方法实现圆柱滚子图像的获取;然后采用Canny算子检测圆柱滚子边缘,为了获得圆柱滚子的直径与长度值,最后运用最小外接矩形拟合边缘。经过与人工检测结果对比,证明该检测方法检测结果准确。Taking cylindrical roller bearing as the research object,cylindrical roller size detection method based on machine vision was proposed.The image of cylindrical roller was used to realize synchronous detection of the diameter and length of cylindrical roller.Firstly,the cylindrical roller photo taken by the camera was preprocessed:gray linear transformation was used to improve the contrast between cylindrical roller and background,the binarization threshold segmentation method was used to realize the extraction of cylindrical roller image.Then,Canny operator was used to detect the edge,in order to obtain the diameter and length of cylindrical roller.Finally,the minimum enclosing rectangle was used to fit the edge.The comparison with the manual test results shows that the test results are accurate.

关 键 词：圆柱滚子 机器视觉 灰度线性变换 最小外接矩形 尺寸检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TG806[自动化与计算机技术—计算机科学与技术]