基于标准杆的双目视觉传感器标定技术研究  

Research on Calibration of Stereo Vision Sensor Based on Standard Bar

作　　者：李丽娟[1] 李宁[1] 林雪竹[1,2] 

机构地区：[1]长春理工大学光电工程学院,长春130022 [2]长春理工大学机电工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2015年第5期34-38,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：国家863重点项目(2013XX0406)

摘　　要：目前制造业对大范围精密测量技术要求越来越高,为了克服传统双目立体视觉标定过程中测量场难实现、标定精度低等缺点,提出了一种较灵活的基于一维标准杆的双目视觉传感器的标定方法。该方法首先利用正方形模板和带有四个特征点的标准杆实现主点和畸变的标定,然后在测量场多次摆放标准杆,由两摄像机拍摄标准杆图像,利用标准杆上对应特征点之间的约束以及基本矩阵估计双目视觉传感器其余参数,最后光束法平差进行整体优化,消除累积误差。仿真实验和真实实验结果均表明该方法适用于大视场双目视觉传感器的标定,操作简单、标定效率高。The more accurate of large scope precise measurement is increasingly required in manufacturing industry,this paper presents a more flexible method based on one-dimensional standard bar for calibrating stereo vision sensor,which overcomes difficulty and low precision in factory in traditional binocular stereo vision sensor. Firstly,a square template and a standard bar with four points are used to calibrate principal point and aberration,then the standard bar is randomly placed in measurement space for several times and the images of the points on the standard bar are acquired by the cameras,the other parameters are estimated based on the foundational matrix and the constraint of the known distance between the points on the target,lastly,the bundle adjustment method is applied to optimize all the parameters of the cameras and calibration points to reduce the accumulated error. The results of simulation and physical experiment show that the proposed calibration method is efficient and easy to operate,and this method is suitable for the calibration of stereo vision sensor with large field of view.

关 键 词：双目视觉传感器 标定 标准杆 光束法平差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]