基于机器视觉的小型零件圆度测量  

The Roundness Measurement of Small Parts Based on Machine Vision

作　　者：丁哲文 张瑞 朱振伟[1,2] 赵华东 刘丙康[2] DING Zhewen;ZHANG Rui;ZHU Zhenwei;ZHAO Huadong;LIU Bingkang(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450001,China;Henan Intelligent Manufacturing Research Institute,Zhengzhou 450066,China)

机构地区：[1]郑州大学机械与动力工程学院,郑州450001 [2]河南省智能制造研究院,郑州450066

出　　处：《组合机床与自动化加工技术》2024年第4期171-174,180,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：郑州市协同创新重大专项项目(18XTZX12006)。

摘　　要：针对小型零件圆度误差测量效率低、精度不稳定等问题,提出了一种基于机器视觉的小型零件圆度测量方法。通过建立机器视觉检测平台,获得零件图像。经过预处理后,通过Canny算子结合8领域扩张算法划定感兴趣区域,采用多项式插值亚像素边缘检测算法获得边缘坐标;然后,利用改进区域搜索算法确定准圆心,并通过最小区域法的几何结构找出最小区域圆心,从而计算圆度误差。实验结果表明,圆度误差的测量重复精度为8.1μm,能够实现小型零件的非接触测量。Aiming at the problems such as low efficiency and unstable accuracy of roundness error measurement for small parts,a method of roundness measurement for small parts based on machine vision was proposed.Through the establishment of machine vision inspection platform,the parts image is obtained.After pre-processing,the region of interest is delimited by Canny operator combined with 8-domain expansion algorithm,and the edge coordinates are obtained by polynomial interpolation subpixel edge detection algorithm.Then,the improved region search algorithm is used to determine the quasi-center of the circle,and the minimum center of the circle is found by the geometric structure of the minimum region method,and the roundness error is calculated.The experimental results show that the repeatability of roundness error is 8.1μm,which can realize the non-contact measurement of small parts.

关 键 词：机器视觉 圆度误差 边缘检测 最小区域法 小型零件 

分 类 号：TH16[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]