钢管直线度的机器视觉测量系统设计  

Design of machine vision system for steel pipe's straightness measurement

作　　者：李小涛[1] 杜晓晴[1] 李小嘎[1] 田健[1] 

机构地区：[1]光电技术及系统教育部重点实验室重庆大学光电工程学院,重庆400044

出　　处：《光学技术》2011年第3期317-321,共5页Optical Technique

基　　金：中央高校基本科研业务费资助项目(CDJRC10120006)

摘　　要：针对钢管在空间方向直线度的测量,采用机器视觉方法,通过获取钢管上下边缘图像,求得钢管中心轴线坐标,评定直线度误差。在整体设计过程中,照明系统采用背光照明方式和平行光管光源,采用双高斯结构设计了光学成像系统基本结构,并用Zemax对成像系统进行了优化设计和像质分析。根据钢管边缘图像的特征,采用了一种新方法——相邻行灰度差异判定法对钢管边缘进行检测。设计结果表明,背光平行照明有利于获得高对比度的钢管边缘,优化后的成像系统畸变小于0.08%,场曲小于0.05 mm,相邻行灰度差异判定法较传统边缘检测方法效果好、速度快,整个测量系统能够实现钢管直线度的高精度、非接触实时测量。Machine vision is adopted to measure the steel pipe＇s 3-D straightness.Straightness error is evaluated by the steel pipe＇s axis coordinate which is calculated by the steel pipe＇s edge coordinate.During the process of the design system,back light illumination and collimator are used for the illumination system.A structure of double-Gauss lens is adopted to design the basic structure of the optical imaging system.Zemax is used to evaluate image quality and also be used to design the optical system optimally.A new method of image edge detection named judgment of different adjacent row＇s gray is adopted to detect the steel pipe＇s edge.The design results show that back light illumination（collimator） is conducive to have a image of high-contrast steel pipe＇s edge.Distortion of the imaging system after optimization is lower than 0.08%,and field-curve is lower than 0.05 mm.Compared with the traditional edge detection,judgment of different adjacent row＇s gray have better quality and lower computation power than it.The system can be applied to realize non-contact and real-time measurement of the steel pipe＇s straightness with high precision.

关 键 词：机器视觉 钢管直线度 照明系统 光学成像系统 边缘检测 

分 类 号：TN29[电子电信—物理电子学]