多传感器的天文标定与空间配准

Astronomical calibration and spatial registration of multi-sensor

为了实现共平台的多个空间观测成像系统的空间配准,对多个相机进行天文标定和姿态测量,提出一种基于星图模拟的星图识别和姿态解算算法。首先,建立共平台的多传感器天文观测模型,并产生模拟星图,将星点之间的特征转换为图像特征;其次,根据四边形对角线的共线不变性特征,对相机观测星图和模拟星图进行特征匹配,提取相机内参数的初始值;然后,利用星点坐标矩阵奇异值不变性,自动匹配剩余星点并求解姿态矩阵的初始值;最后,对相机内参数和姿态矩阵进行非线性优化求解。实验结果表明,在未精确标定光学系统内外参数的情况下,星点正确识别率大于97.4%,相机像元角分辨率为3.9"×3.4",最后计算得出的投影误差低于1个像素,满足多传感器空间观测系统空间配准的精度和鲁棒性要求。

In order to achieve the spatial registration of multi-sensor on the same platform for astronomical observation,a camera calibration and attitude measurement algorithm was proposed based on the star image simulation which includes four steps.Firstly,according to star observation model and star catalogue,the simulated star image was produced in order to transform constellation features into image features.Then,based on the invariant collinearity of quadrilateral diagonal,image features between the observed star map and the simulated star image were matched,consequently,the initial values of the camera parameters could be extracted.Through using the singular value decomposition method,all the remaining control points were matched automatically and the attitudes of cameras were calculated.Finally,a complicated nonlinear optimization on camera intrinsic parameters and attitude matrix was required for this algorithm.The experimental result indicates that the re-projection error between different cameras is better than 1pixel,and the correct recognition rate is more than 97.4%when the calibrations of optical system parameters are imprecise.This method can meet the requirements of accuracy and robust in registration of multi-sensor astronomical observation system.