基于旋转标定板的多相机系统标定方法

Calibration Method for Multicamera System Based on Rotating Calibration Plate

针对现有多相机系统在无公共视场时标定过程繁琐问题,提出了一种基于旋转标定板的多相机系统标定方法。首先通过拟合旋转靶标轨迹实现转轴拟合和建立系统坐标系;然后各相机获取单幅完整的靶标图像,使用PnP算法解算相机与转台系统间的位姿关系;接着将转台转角通过罗德里格斯公式转换为靶标旋转前后的旋转矩阵和平移向量;最后使用计算所得的变换关系实现对多相机系统的标定。实验结果表明,本文方法不受限于相机的视场范围,能够快速、自动地标定多相机系统,在测量精度上也满足大尺寸测量需求。

The calibration of existing multicamera systems with no common fields of view is challenging,as it is extremely cumbersome.To solve this problem,this study proposed a calibration method for multicamera systems based on a rotating calibration plate.First,the axis of rotation was fitted and a calibration coordinate system was established by fitting the trajectory of the rotating target.Second,a single complete calibration image was captured by each camera,and then the pose relationship between the camera and the turntable system was estimated using the PnP algorithm.Third,the angle of the table was converted into a rotation matrix and translation vector of the calibration plate using the Rodrigues rotation formula.Finally,the multicamera system was estimated using these transformation relationships.The experimental results show that the proposed method is not limited by the field of view of the camera,can quickly and automatically calibrate the multicamera system,and also meets the needs of largescale measurement in terms of measurement accuracy.