基于机器视觉的圆形精密工件尺寸检测  

Dimension measurement of circular precision components based on machine vision

作　　者：于佳昊 刘嘉承 王连锴[1] 杨展 宋德[1] Yu Jiahao;Liu Jiacheng;Wang Liankai;Yang Zhan;Song De(School of Physics,Changchun University of Science and Technology,Changchun 130013,China)

机构地区：[1]长春理工大学物理学院,长春130013

出　　处：《现代计算机》2024年第20期41-44,共4页Modern Computer

基　　金：国家自然科学基金资助项目(U2031113)。

摘　　要：针对圆形精密工件尺寸测量困难、检测成本较高等问题提出一种基于区域灰度模型的快速、精确的检测算法。对基于机器视觉的圆形精密尺寸检测方法进行了研究。首先,通过形态学处理完成图像的预处理;然后,在区域灰度模型算法中引入OTSU自适应阈值,提高检测精度;之后,通过Canny算子定义的圆环完成缺陷处理;最后,采用最小二乘法拟合圆形直径长度。该方法运算速度快、检测精度高,最大测量误差小于0.002 mm,平均测量误差精度小于0.0007 mm。满足工件检测精度要求。A fast and accurate detection algorithm based on regional grayscale model is proposed to address the difficulties in measuring the dimensions and the high cost of detection for circular precision components.Research on Machine Vision-Based Circular Precision Dimension Measurement Methods.Firstly,preprocessing of the image is accomplished through morphological operations;Then,in the Partial-area-effect model algorithm,OTSU adaptive thresholding is introduced to improve detection accu-racy;Afterwards,defect handling is accomplished through a circular ring defined by the Canny operator;Finally,the diameter length of the circular shape is fitted using the least squares method.The method boasts high computational speed and accuracy,with a maximum measurement error of less than 0.002 mm and an average measurement error precision of less than 0.0007 mm,

关 键 词：精密检测 机器视觉 区域灰度模型 

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]