Improving co-registration accuracy of interferometric SAR images using spectral diversity

The co-registration of complex SAR images is a key problem in SAR interferometry, especially in the repeat-pass case, when the relative acquisition geometry is not known with sufficient accuracy and the approximate terrain elevation changes cannot be considered a priori. Considering that future interferometric SAR systems will operate at higher resolutions the requirements on the accuracy of platform position and attitude will permanently increase. This article presents the application of a new algorithm for the determination of the relative misregistration errors between two interferometric SAR images. The method is based on spectral diversity properties of the complex SAR signals. The phase information of different spectral looks is evaluated giving a misregistration information on a pixel by pixel basis. It is shown that the accuracy of the proposed algorithm is superior to the conventional techniques. It performs best when measuring misregistration errors on a subpixel level. Image contrast as in conventional techniques is not required. Interferometric data of the airborne E-SAR sensor of DLR are used to show the performance of the proposed method