圆盘类零件几何量机器视觉测量技术研究  

Research on Measurement Technology of Geometrical Quantity of Disc Parts Based on Machine Vision

作　　者：李鹏阳[1] 杨文辉 曹利平 乔卫杰 杨天锋 LI Pengyang;YANG Wenhui;CAO Liping;QIAO Weijie;YANG Tianfeng(School of Mechanical and Instrumental Engineering,Xi’an University of Technology,Xi’an 710048,China;Instituteof Technology,Shaanxi Diesel Engine Heavy Industry Co.,Ltd.,Xingping 713105,China)

机构地区：[1]西安理工大学机械与精密仪器工程学院,西安710048 [2]陕西柴油机重工有限公司工艺研究所,兴平713105

出　　处：《西安工业大学学报》2020年第6期598-604,共7页Journal of Xi’an Technological University

基　　金：陕西省科技创新创业“双导师制”合作科技项目(2019JM-593)。

摘　　要：为提高圆盘类零件人工测量效率、精度,设计了一套基于机器视觉的测量系统,对图像进行灰度处理、平滑滤波、边缘检测和最小二乘法曲线拟合,使用方格图计算像素当量的方法完成了系统的标定,开发了相应的测量软件,进行实验并分析测量误差原因。通过张正友标定法代替原有标定方法和在边缘检测基础上使用灰度矩亚像素边缘定位算法改进实验。结果表明:直径和圆度测量误差达到了0.0291 mm和0.0190 mm,满足实际测量精度要求。In order to improve the efficiency and accuracy of manual measurement used for disc parts,a machine vision-based measurement system is designed to perform grayscale processing,smooth filtering,edge detection and least square curve fitting on the image.The method of calculating the pixel equivalent using the grid graph completed the system calibration,and developed the corresponding measurement software to conduct experiments and analyze the cause of the measurement error.By using the Zhang Zhengyou calibration method instead of the original calibration method,an improved experiment of gray moment sub-pixel edge positioning based on edge detection is performed.The results show that the diameter and roundness measurement errors reach 0.0291 mm and 0.0190 mm,which meets the actual measurement accuracy requirements.

关 键 词：机器视觉 圆盘类零件 几何量 灰度矩 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]