一种基于光束平差法的相机标定方法

为解决多目视觉测量系统中，相机标定精度要求高、标定视场范围广等问题，提出一种基于光束平差法 的相机标定方法。利用立体视觉原理重构各个编码点的空间坐标，通过相机成像模型将编码点投影到图像中，并利 用Levenberg-Marquardt 方法最小化编码点的投影点与图像中对应像点之间的投影误差，获得最优的相机参数。实验 结果表明：该方法的最大标定误差为0.052 9 像素，相机标定精度高、应用场景灵活，满足工业应用需求。

A Camera Calibration Method Based on Beam Adjustment Method

In order to solve the problems of high precision and wide field of view in multi-view vision measurement system, a camera calibration method based on beam adjustment method is proposed. The spatial coordinates of each coding point are reconstructed based on the stereo vision principle, and the coding points were projected into the image through the camera imaging model. Then, Levenberg-marquardt method is used to minimize the projection error between the coding points and the corresponding image points in the image, so as to obtain the optimal camera parameters. The experimental results show that the maximum calibration error of the method is 0.052 9 pixels. Besides, the camera calibration accuracy is high, and the application scene is flexible, meeting the needs of industrial applications.