Tuesday, October 09, 2012

A Data-Adaptive Compressed Sensing Approach to Polarimetric SAR Tomography of Forested Areas

Esteban Aguilera sent me the following:

Hi Igor,
We’ve recently put together the final version of the following short paper: A Data-Adaptive Compressed Sensing Approach to Polarimetric SAR Tomography of Forested Areas.
Anyone can get a hold of it at: http://elib.dlr.de/75995/1/06302171.pdf
The accompanying presentation slides can be found at: http://elib.dlr.de/75795/2/presentation.pdf
It is basically a MMV problem where physical validity as well as a specific subspace of every row is enforced (necessary for subsequent physical analysis of 3D radar images). If you think it’s of interest, I wouldn’t mind your featuring it on Nuit-Blanche.

Abstract—Super-resolution imaging via compressed sensing (CS)-based spectral estimators has been recently introduced to synthetic aperture radar (SAR) tomography. In the case of partial scatterers, the mainstream has so far been twofold, in that the tomographic reconstruction is conducted by either directly working with multiple looks and/or polarimetric channels or by exploiting the corresponding single-channel second-order statistics. In this letter, we unify these two methodologies in the context of covariance fitting. In essence, we exploit the fact that both vertical structures and the unknown polarimetric signatures can be approximated in a low-dimensional subspace. For this purpose, we make use of a wavelet basis in order to sparsely represent vertical structures. Additionally, we synthesize a data-adaptive orthonormal basis that spans the space of polarimetric signatures. Finally, we validate this approach by using fully polarimetric L-band data acquired by the E-SAR sensor of the German Aerospace Center (DLR)

Thanks Esteban  for the heads-up.

No comments: