一种多时相卫星影像密集匹配DSM 配准方法

A registration method of multi-temporal DSM from satellite dense matching

DSM的精确配准是利用不同时相DSM进行变化检测等应用的前提。本文针对多时相卫星影像密集匹配DSM,提出一种DSM配准方法。该方法以刚体变换为转换模型,通过影像匹配方法获取2D同名点对的基础上得到3D同名点对,然后基于粗配准模型参数,采用稀疏ICP的方法,得到精配准模型参数。采用两个时相的卫星影像密集匹配DSM进行实验,实验结果表明,该方法能够有效地实现DSM的精确配准。

Accurate registration of DSM is a prerequisites for applications such as change detection using DSMs of different time periods.A registration method is proposed for multi-temporal DSM from dense matching in this paper.In this method,the rigid body transformation is used as transformation model,and 3D matching points are obtained based on 2D matching points form image matching.Then,based on the coarse registration model parameters,the sparse ICP method is used to obtain the fine registration model parameters.Experiment is conducted on two temporal DSM from satellite images,and experiment result shows that the method can achieve accurate registration of DSM effectively.