| Popis: |
The past two decades have seen continuing progress in the field of deformation analysis based on Interferometric Synthetic Aperture Radar (InSAR) data. In particular, refined methods for processing distributed scatterers (DS) have been developed. A general idea is that in a pre-processing step DS can be transformed to quasi PS and that the quasi PS together with PS can be analyzed with an arbitrary PS algorithm. This can be achieved by grouping statistically similar pixels, estimating the coherence matrix based on the group of pixels and extracting the signal of the DS from the coherence matrix. In this study different alternatives for these three steps are investigated with help of comprehensive simulations. In particular, a powerful regularization method is introduced that has not been used in the InSAR context before. It improves considerable the results for those estimators of DS signal that make use of the inverse of the coherence matrix. |
