摘要
在工业现场利用机器视觉技术对分离圆孔进行尺寸测量的应用越来越广泛,且测量精度和速度也不断地提升,但传统的圆孔测量方法对于多管类工件堆积在一起整体测量是难以实现的,常常存在漏检甚至无法提取等问题。为了解决此问题提出了一种基于模板匹配的多圆提取算法,首先在分离管状工件中建立单个工件的模板图像,其次使用该模板与实际待测量的工件组进行匹配,匹配完成后分离成单个的待测工件图像,分别对单个的图像进行边缘提取和圆参数拟合,最后建立坐标关系,根据图像尺寸得到实际的物理尺寸。通过对测量的尺寸和使用高精度三坐标仪得到真实尺寸对比,证明该方法的测量误差在0.01mm以内,可以应用于工业测量中。
The machine vision technology in the field of industrial measurement of the size of the separation hole is more and more widely used,and the accuracy and speed of measurement are continuously improved.However,the traditional method of measuring round holes for multi-tube workpieces stacked together is difficult to achieve.There is often missing or even unable to extract.Intending to solve this problem,it proposed a multi-circle extraction algorithm based on template matching.Firstly,a template image of a single workpiece is set up in a separate tubular workpiece.Secondly,the template is used to match with the actual workpieces which need to be measured.After the matching is completed,A single workpiece image to be extracted.Extracting edge and fitting circle parameter for every single image.And finally establishing the coordinate relationship to get the actual physical size according to the image size.Through comparing the measurement of the size with the true size using of high-precision coordinate,proving that the method of measurement error within 0.01mm can be used in industrial measurement.
作者
周应东
廖晓波
蔡勇
ZHOU Ying-dong;LIAO Xiao-bo;CAI Yong(Manufacturing Process Testing Technology Key Laboratory of Ministry of Education,Southwest University of Science and Tec-hnology,Sichuan Mianyang 621010,China)
出处
《机械设计与制造》
北大核心
2018年第A02期119-122,共4页
Machinery Design & Manufacture
基金
高档数控机床与基础制造装备科技重大专项(2012ZX04007-021-07)
关键词
视觉测量
管类工件
内外径测量
Vision Measurement
Tubular Work Piece
Inside and Outside Diameter Measurement
作者简介
周应东,(1992-),男,四川巴中人,硕士研究生,主要研究方向:机器视觉、嵌入式系统设计;廖晓波,(1982-),男,湖北天门人,博士研究生,讲师,主要研究方向:机器视觉、信号处理。
