大视场摄影测量相机的相对定向

Camera relative orientation in Large-scale photogrammetry

为了提高大视场摄影测量系统的测量精度和稳定性,提出一种基于局部迭代搜索算法的相机相对定向方法,仅需要两幅图像间匹配的像面坐标信息,就可以计算相机之间精确的相对外方位参数。通过闭环多项式解法获得两图像之间本质矩阵的初值;将本质矩阵进行局部迭代,计算正交矩阵和外极线约束误差方程;利用两图像上所有匹配点的像面坐标,通过多次迭代获得本质矩阵的全局最优解;分解本质矩阵获得图像之间的相对外方位参数,并结合Cheirality约束条件确定外方位参数的唯一真实解。通过6m×4m×3m可展开桁架微波天线面形测量场的定向实验,验证了提出方法的测量精度和可靠性。相比于传统的相对定向方法,基于局部迭代搜索算法的相对定向方法具有较高的初始定向精度和稳定性,平移向量误差小于5mm,旋转角度误差小于0.1°,而且消除了对于参照物的依赖,对于提高摄影测量系统的性能具有重要意义。

In order to improve the measurement accuracy and stability of large-scale photogrammetry systems,a new relative orientation method based on local iterative search algorithm is proposed.The method calculates the relative orientation parameters between images of an image pair using only the coordinates of correspondences.The initial estimate of essential matrix between the two images is acquired through the closed-loop polynomial algorithm.The essential matrix is iteratively optimized to obtain the orthogonal matrix and epipolar error equations.All the possible relative exterior orientation parameters between camera positions are calculated by essential matrix decomposition,and the correct solution is determined by Cheirality constraint.The orientation experiment of an expandable truss microwave antenna profile measurement（6 m×4 m×3 m）is carried out to verify the accuracy and reliability of the new method.Compared with the traditional orientation methods,the proposed local iterative search algorithm provides more precise and stable orientation results.Translation error is less than 5 mm,rotation angle error is less than 0.1°.The method eliminates the orientation dependence on reference objects,which will play a key role to improve photogrammetric system performance.