中文摘要：

为提高大尺寸数字摄影测量的测量精度，需要对目标点图像坐标进行精确的畸变修正，这依赖于精确的畸变修正模型及可靠的标定方法。通过建立随对焦状态和测量点物距变化的相机畸变模型，实现了任意对焦状态和物距下目标像点畸变量的计算。利用回光反射线建立了用于相机标定的共面垂线阵列测试场，并设计了借助高精度摄影测量系统和后方交会进行标定相机站位调整的方法。利用直线的透视成像依然为直线的原理设计优化目标，精确求解不同对焦状态和站位的相机畸变系数。为了验证模型的正确性和标定结果的准确性，设计了比较实验，实验结果表明，该方法给出的畸变量计算值与观测值间的相对误差不大于0．6％，而传统单一物距畸变模型的计算值同观测值问相对误差最大可达5％。数据说明，该畸变模型和相应标定方法能计算更为精确的畸变量，相对传统畸变模型能够进一步减少系统误差并提高摄影测量精度。

英文摘要：

In order to improve the measurement accuracy ol large scale digital plaotogrammetry measurement,me mstortion of image target coordinates needs to be corrected precisely ,which relies on precise distortion correction model and reliable calibration method. An analytical distortion model varying with focus setting and object distance is established, with which the distortion for the object image points under arbitrary focus condition and object distance can be compu- ted. The retro reflect strings is used to establish the coplanar plumb line array test range for camera calibration. A meth- od is designed, in which the high accuracy photogrammetry system V-STARS is employed to establish a camera orienta- tion field ,within which the resection method is used to help adjusting the camera optic axis. The distortion parameters corresponding to different focus settings and various depths of object plane are computed through an optimal least square analytical method based on the projective property that the projective image of straight lines are still straight lines. Comparative experiments were designed and carried out to testify the correctness of the proposed model and the accuracy of the calibration method. Experiment results indicate that the relative error between the computed distortion and observed distortion is not larger than 0.6% for the proposed method;while that relative error can reach up to a maximum of 5% for the traditional distortion model with single prove that the proposed distortion model and calibration method object distance distortion model. The experiment data can achieve much more accurate distortion magnitude,decrease the systematic error and improve the photogrammetry accuracy compared with traditional distortion model.