基于双目视觉一维标定算法的研究

Research on One-dimensional Calibration Algorithm Based on Binocular Vision

相机标定在视觉测量、三维重建中起着至关重要的作用。传统的二维标定在大视场或者户外条件下受光照和自身遮挡的影响,很难对相机进行高精度的标定。提出了一种利用定距的一维标定点来标定双目相机内、外参数的方法。该方法首先拍摄一组标定点在空间任意摆放4个及以上位姿的图片,利用归一化8点算法得出基本矩阵,由基本矩阵恢复内、外参数的初值。为了提高一维标定结果的精度,考虑到相机镜头畸变的影响,选用了“Weng”的镜头畸变模型。真实实验的结果证明该方法标定的结果与一维标定传统的“Zhang”模型相比较具有更高的精度。

This paper proposes a method for calibrating the internal and external parameters of a binocular camera by using a fixed one-dimensional calibration point.This method firstly takes a set of pictures where the calibration points are randomly placed in space with 4 or more poses,and uses a normalized 8-point algorithm to obtain a basic matrix.The basic matrix restores the initial values of internal and external parameters.In order to improve the accuracy of the one-dimensional calibration results,and considering the effects of camera lens distortion,a“Weng”lens distortion model is selected in this paper.The results of real experiments prove that the calibration results obtained by this method have higher accuracy than the traditional″Zhang″model of one-dimensional calibration.