%0 Journal Article
%T 基于高斯牛顿法的DEM匹配算法<br>DEM Co-registration Algorithm Based on Gauss-Newton Method
%A 张同刚
%A 王昆仑
%A 金国清&lt
%A br&gt
%A ZHANG Tonggang
%A WANG Kunlun
%A JIN Guoqing
%J 西南交通大学学报
%D 2017
%R 10.3969/j.issn.0258-2724.2017.03.020
%X 为提升DEM（digital elevation model）匹配效率，建立了一种基于高斯牛顿法的快速DEM匹配算法.该算法采用高斯牛顿法替代最小二乘法来进行DEM匹配模型的目标方程求解，加速了目标方程求解的迭代过程.新算法匹配过程中，匹配参数沿梯度最大方向逼近目标值，迭代次数大幅度减少，具有更稳定的迭代收敛性，显著提高了算法的执行效率.通过多组模拟试验对新算法进行了测试，并与具有代表性的最近点迭代算法进行了比较.结果表明：新算法对匹配参数的收敛速率平均提高了42.1%，完成匹配所需的总时间平均减少了74.9%.<br>： To improve the efficiency of DEM (digital elevation model) co-registration, a fast algorithm based on Gauss-Newton method was proposed. This algorithm uses Gauss-Newton method instead of the least squares method, to solve the objective equation of the DEM co-registration model, and greatly accelerates the iterative convergence. During the iterations of the new algorithm, matching parameters approach the target values by following the direction of maximal gradient, which significantly reduces the number of iterations. Moreover, the iterative convergence is more stable and the algorithm operation efficiency is greatly enhanced. The new algorithm was tested with several groups of simulated datasets, and compared with the representative iterative closest points (ICP) algorithm. The experimental results show that the average convergence rate of the proposed algorithm is improved by 42.1%, and the computation time for matching is reduced by about 74.9%
%K DEM匹配
%K 算法
%K 高斯牛顿法
%K 迭代收敛性
%K 执行效率
%K &lt
%K br&gt
%K DEM co-registration
%K algorithm
%K Gauss-Newton method
%K convergence rate
%K performance efficiency
%U http://manu19.magtech.com.cn/Jweb_xnjd/CN/abstract/abstract12442.shtml