CAESAR: An Approach Based on Covariance Matrix Decomposition to Improve Multibaseline–Multitemporal Interferometric SAR Processing

Synthetic aperture radar (SAR) tomography has been strongly developed in the last years for the analysis at fine scale of data acquired by high-resolution interferometric SAR sensors as a technique alternative to classical persistent scatterer interferometry and able to resolve also multiple scatterers. SqueeSAR is a recently proposed solution which, in the context of SAR interferometry at the coarse scale analysis stage, allows taking advantage of the multilook operation to filter interferometic stacks by extracting, pixel by pixel, equivalent scattering mechanisms from the set of all available interferometric measurement collected in the data covariance matrix. In this paper, we investigate the possibilities to extend SqueeSAR by allowing the identification of multiple scattering mechanisms from the analysis of the covariance matrix. In particular, we present a new approach, named “Component extrAction and sElection SAR” algorithm, that allows taking advantage of the principal component analysis to filter interferograms relevant to the decorrelating scatterer, i.e., scatterers that may exhibit coherence losses depending on the spatial and temporal baseline distributions, and to detect and separate scattering mechanisms possibly interfering in the same pixel due to layover directly at the interferogram generation stage. The proposed module allows providing options useful for classical interferometric processing to monitor ground deformations at lower resolution (coarse scale), as well as for possibly aiding the data calibration preliminary for the subsequent full-resolution interferometric/tomographic (fine scale) analysis. Results achieved by processing high-resolution Cosmo-SkyMed data, characterized by the favorable features of a large baseline span, are presented to explain the advantages and validate this new interferometric processing solution.