华南理工大学学报(自然科学版) ›› 2011, Vol. 39 ›› Issue (8): 54-59.doi: 10.3969/j.issn.1000-565X.2011.08.010

• 机械工程 • 上一篇    下一篇

基于机器视觉的平面夹角测量方法

雷经发1,2 王德麾1 袁中凡1†   

  1. 1.四川大学 制造科学与工程学院,四川 成都 610065; 2. 安徽建筑工业学院 机械与电气工程学院,安徽 合肥 230061
  • 收稿日期:2011-01-16 修回日期:2011-05-06 出版日期:2011-08-25 发布日期:2011-07-06
  • 通信作者: 袁中凡(1950-),男,教授,主要从事人机测试工程研究.E-mail: zfyuan26@yahoo.com.cn E-mail:rain78828@163.com
  • 作者简介:雷经发(1978-) ,男,博士生,讲师,主要从事人机测试工程研究.
  • 基金资助:

    四川省应用基础研究项目( 07JY029-083)

A Measurement Method of Plane Angle Based on Machine Vision

Lei Jing-fa1,2  Wang De-huiYuan Zhong-fan1   

  1. 1. School of Manufacturing Science and Engineering,Sichuan University,Chengdu 610065,Sichuan,China; 2. School of Mechanical and Electrical Engineering,Anhui University of Architecture,Hefei 230061,Anhui,China
  • Received:2011-01-16 Revised:2011-05-06 Online:2011-08-25 Published:2011-07-06
  • Contact: 袁中凡(1950-),男,教授,主要从事人机测试工程研究.E-mail: zfyuan26@yahoo.com.cn E-mail:rain78828@163.com
  • About author:雷经发(1978-) ,男,博士生,讲师,主要从事人机测试工程研究.
  • Supported by:

    四川省应用基础研究项目( 07JY029-083)

摘要: 为解决制造、建筑等领域中的平面夹角实时测量问题,提出了一种利用机器视觉的非接触式平面夹角测量方法.该方法首先在两待测平面上分别标记出一个圆及其圆心,测量时仅需使用一台已知其内部参数的数字相机,在空间任意位置进行一次拍照,获得圆心和部分标记圆图像即可; 然后通过最小二乘法拟合出各平面上圆图像的方程及其圆心的图像坐标,即可确定出各平面上的消隐线方程,进而可以得到各平面法线方向上的灭点坐标; 最后,根据相机成像方程,利用这两个灭点坐标和已知的相机内部参数计算出两平面法线夹角值,对其求补角即可得到两平面间的夹角.实验表明,此方法简单可行,测量相对误差小于1%,能够满足实际需求.

关键词: 机器视觉, 角度测量, 二次曲线, 消隐线

Abstract:

In order to overcome the difficulties in the real-time measurement of plane angle in many fields such as manufacture and architecture,a non-contact measurement method is proposed based on machine vision. In the method,first,a circle and its center are marked in each of two planes to be measured,and only one digital camera with known intrinsic parameters is needed to take photos once in any position so as to obtain one image with the centers and all or part of the circles. Then,the equations of the images of the circles and the coordinates of their centers are acquired by means of the least square method,thus achieving the equations of the vanishing lines in the two planes and the coordinates of the vanishing points in the normal directions of the two planes. Finally,according to the imaging equation of the camera,the angle between the normals of the two planes,whose supplementary angle is just the angel between the two planes,is figured out by using the coordinates of the vanishing points and the intrinsic parameters of the camera. Experimental results show that the proposed method is simple and feasible,with a relative measurement error of less than 1%,which can meet practical requirements satisfactorily.

Key words: machine vision, angel measurement, quadratic curve, vanishing line