摘要
本文提出了一种相机快速标定方法,使用一个平面标定板,仅拍摄一幅图像,就能完成相机标定。与通用的相机标定方法相比,可以减少拍摄图像及计算相机多个参数初值的麻烦,提高了相机标定的效率,尤其适用于多相机同时标定的情况。在计算相机内参时,使用Levenberg-Marquarat方法优化求解。实验结果表明,本算法速度快且精度较高,可应用于机器视觉研究、工业三维测量等多个领域的摄像机标定。本方法也可以选择拍摄二幅或者任意多幅图像,来参与算法的迭代优化过程,以进一步提高标定精度。
In this paper,a fast camera calibration method is proposed. Using a plane calibration board,only one image can be taken to complete the camera calibration. Compared with the general Zhang Zhengyou camera calibration method,it can save the trouble of taking at least three images to calculate the initial value of camera parameters,and improve the efficiency of camera calibration,especially suitable for multi camera calibration at the same time. In the calculation of camera internal parameters,Levenberg Marquarat method is used to optimize the solution. The experimental results show that the algorithm is fast and accurate,and can be applied in machine vision research,industrial3 D measurement and other fields. This method can also choose to take two or any number of images to participate in the iterative optimization process of the algorithm to further improve the calibration accuracy.
作者
徐杰
XU Jie(D Printing Research Institute,Beijing City University,Beijing 100083,China)
出处
《智能计算机与应用》
2021年第5期155-159,共5页
Intelligent Computer and Applications
