星载InSAR复图像多级配准及其误差分析

Multistep Coregistration of Spaceborne InSAR Complex images and Its Error Analysis

复图像的配准是星载干涉合成孔径雷达(InSAR)数据处理中的重要步骤。对于星载InSAR复图像的配准,为了达到高的配准精度并同时保持较高的计算效率,需要采用多级配准的方法,即可先后采用几何配准方法、基于窗口的配准方法和基于频谱差别的配准方法来完成复图像的配准。首先讨论了三种算法的原理和处理方法,其中详细分析了基于频谱差别配准方法的原理。然后分析了各算法的误差,由于基于频谱差别配准影响最终的配准精度,分析了通过去噪来减少其误差。使用上述算法对仿真数据的处理结果验证了算法的有效性。

Coregistration of complex images is an important processing step of spaceborne InSAR data. In order to reach high coregistration precision and at the same time keep fine computing efficiency, the coregistrtion method should be multistep. The multistep method is to use geometrical coregistration ,windowed coregistration and spectral diversity coregistration successively. At first the principle and processing steps of the three coregistration algorithms were studied. In the study the principle of spectral diversity coregistration was detailedly analyzed. Then error of the algorithms was analyzed. Because spectral diversity coregistration can influence the last coregistration precision, reducing its error by denoising was analyzed. The processed results of simulated data can validate the coregistration method.