基于机器视觉的电子线束产品尺寸检测方法  

A Method for Measuring the Size of Electronic Wiring Harness Products Based on Machine Vision

作　　者：李小宝 陆永华[1] 周利华 冯强 赵东标[1] LI Xiaobao;LU Yonghua;ZHOU Lihua;FENG Qiang;ZHAO Dongbiao(College of Mechanical and Electrical Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;State-Owned Jinjiang Machine Factory,Chengdu 610043,China)

机构地区：[1]南京航空航天大学机电学院,江苏南京210016 [2]国营锦江机械厂,四川成都610043

出　　处：《测试技术学报》2024年第3期289-297,共9页Journal of Test and Measurement Technology

摘　　要：针对电子线束产品尺寸的传统人工检测方式效率低、主观测量误差较大等问题,提出一种基于机器视觉的非接触式检测方法。设计和搭建了视觉检测系统,使用张正友标定法对相机进行标定并采集产品图像信息,采用改进的Canny算法对图像进行边缘检测,提取产品轮廓并对图像进行水平倾斜校正;将校正后的产品轮廓进行区域划分并用最小二乘法对轮廓边缘采样点进行直线拟合;根据轮廓边缘上测量点的图像坐标和相机标定结果计算得到世界坐标系下的产品尺寸;对产品分别进行传统游标卡尺测量和视觉测量试验,结果表明,该方法重复测量误差和绝对测量误差均不超过0.1 mm,平均检测耗时在1 s以内,能够实现电子线束产品尺寸的快速准确检测。Aiming at the problems of traditional manual inspection methods of electronic wire harness product dimensions such as low efficiency and large subjective measurement errors,a non-contact inspection method based on machine vision is proposed.Design and build a visual inspection system,use Zhang Zhengyou calibration method to calibrate the camera and collect product image information,use the improved Canny algorithm to detect the edges of the image,extract the product outline and perform horizontal tilt correction on the image;the corrected product outline is divided and the least squares method is used to perform straight line fitting on the contour edge sampling points;the product size in the world coordinate system is calculated based on the image coordinates of the measurement points on the contour edge and the camera calibration results;traditional vernier caliper measurement and visual measurement tests are performed on the product.The test results show that the repeated measurement error and absolute measurement error of this method do not exceed 0.1 mm,the average detection time is within 1 s,and it can achieve rapid and accurate detection of the dimensions of electronic wire harness products.

关 键 词：机器视觉 相机标定 边缘检测 最小二乘拟合 尺寸检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]