基于非成像模型的摄像机校准方法

作为摄影测量关键部件的摄像机，其校准精确度直接影响了整个测量系统的测量不确定度．通过对常规摄像机校准技术的研究，发现基于内参数成像模型的校准方法会因为内参数综合优化的相关性干扰和非线性求解的不确定性使校准的畸变参数不稳定，且无法修正由镜头透镜组制造中的曲率变化非连续平滑等造成的非模型化畸变．结合航空摄影测量光学实验室方法与垂线法，以摄像机的成像原理为依据提出基于非成像模型的校准方法，以实现逐像元（甚至亚像素级）畸变的细化修正．实验表明，该校准方法的角度测量精度为±5″，系统不确定度为10×10-6，并可适用于所有的相机和镜头．

Camera Calibration Method Independent of Imaging Model

The camera is the main component for photography measurement, and its calibration accuracy determines the uncertainty of the measuring system. Through normal camera calibration research, it can be concluded that the distortion parameters calibrated by the calibration methods based on the intrinsic parameter imaging model will be unstable because of the correlation interferences of intrinsic parameter optimization and the uncertainty of nonlinear solution, and the distortion caused by the discontinuous smoothing of curvature change in camera lens group manufac- turing is independent of the imaging model. From the imaging principle of the camera, a new calibration method inde- pendent of the imaging model is proposed, combined with the aerophotogrammetry optical laboratory method and the vertical method, to realize the refinement correction of each pixel （even sub-pixel）. The experimental results show that by using the proposed calibration method, the angle measurement accuracy is ±5″, the indeterminacy of the measuring system is 10 x 10-6, and the method is applicable to all cameras and lens.