You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.


Experiments of 3D SAR Tomography Techniques with P-Band Polarimetric Data

Fabrizio Lombardini(1) and Matteo Pardini(1)

(1) University of Pisa, Via G. Caruso 16, 56122 Pisa, Italy


SAR systems operating with long (decimetric to metric) radio wavelengths posses the capability of penetrating deep into volumes, making them an ideal sensor e.g. for forest and ice investigations. However, since the 3D scene scattering properties are projected onto the 2D azimuth-range plane, ambiguities arise in the inversion of physical and geometrical parameters, causing several drawbacks depending on the application. The 3D SAR Tomography (Tomo-SAR) [1]-[6] and Polarimetric SAR Tomography (PolTomo-SAR) [7] techniques are very promising advances in this field. Tomo-SAR is a multibaseline (MB) extension of conventional cross-track SAR interferometry, employing many passes over the same area. Differently from conventional interferometry, which can only furnish a measure of the terrain topography, Tomo-SAR can resolve multiple scatterers at different heights in each given range-azimuth cell, and produce a continuous profiling along the height dimension. In fact, the Tomo-SAR technique exploits the synthesized aperture along the height-ground range plane to get full 3D imaging through elevation beamforming [2], [4]. Given the correspondence between the scatterer height and the spatial frequency in the MB data vector components [2], this imaging process can also be interpreted as a spatial spectral analysis obtained by sweeping a proper spatial pass band filter along the height dimension. As a consequence, Tomo-SAR can add more features, e.g., for biomass estimation, forest classification, tree and building height estimation, and other geophysical parameter extraction problems; interest is growing also for ice thickness monitoring. This technique has also been extended to PolTomo-SAR, which jointly exploits MB SAR data acquired with different polarization channels in order to retrieve at the same time the height information of the multiple scatterers and a set of normalized complex coefficients describing the polarimetric scattering mechanism [7]. In this work, we experiment and extend different tomographic techniques with fully polarimetric P-band airborne MB data. Data have been acquired in the framework of the ESA BIOSAR project by the DLR’s E-SAR system over a (frozen) forest site in Remningstorp, Sweden, in a two months temporal span. Firstly, we investigate the behaviour of the advanced adaptive beamforming and superresolution tomographic techniques. The adaptive beamforming algorithm (Capon) is based on spatial leakage rejection, offering an improvement in terms of both sidelobe level and resolution, at the cost of possible radiometric non-linearities (i.e., self cancellation phenomena of the scattering components of interest) [4]. Superresolution tomography exploits model parameter fitting of multiple scatterers compact in height (MUSIC, multiple signal classification) [5]. The analysis is focused on typical problems which may limit future 3D SAR missions. In particular, the gains of adaptive beamforming and superresolution tomography are evaluated with respect to classical Fourier-based elevation focusing in terms of resolution with a small overall baseline, and in terms of sidelobes with irregular baseline distribution. The self-cancellation phenomenon affecting adaptive beamforming will be analyzed and discussed, in presence of different emulated miscalibration levels in the data. Also, data temporal decorrelation issues in the tomographic image formation are analyzed by checking height focusing on specific areas. Statistical accuracy of the height estimates is evaluated, in particular by comparing the tomo-extracted sub-canopy digital terrain model (DTM) with a LIDAR DTM over the same scene. Furthermore, the possibility is investigated of extending the tomographic principle based on swept pass-band filtering to a new formulation where a proper spatial filter furnishes a MB dataset with undesired height components cancelled for specific applications. This layer-canceller is a fixed multiband (i.e. pass- and stop-band) filter; its possible use could be a pre-processing for applying in a two layer scenario the band synthesis for range resolution enhancement [8], and other interferometric techniques matched to single scatterer scenarios. Moreover, PolTomo-SAR processing [7] is applied to extract the polarimetric scattering mechanism of the imaged P-band scatterers. In particular, effects of adaptive beamforming radiometric non-linearity are investigated in terms of possible relative unbalances of the different channel filters [7], distorting the estimated polarimetric scattering mechanism. The comparison is carried out with linear, but lower resolution, PolTomo Fourier beamforming. Finally, an improved radiometrically stable adaptive beamforming method for PolTomo-SAR is presented and tested. This method is based on a regularization approach in the adaptive filters design in order to reduce their unbalances, at the cost of increased computational burden, and to enhance the estimation accuracy of the polarimetric coefficients.

ACKNOWLEDGEMENTS The authors would like to thank the Microwaves and Radar Institute team of the DLR for processing and providing the SAR data set that it acquired in the framework of the ESA project BIOSAR, and Prof. Fabio Rocca and Stefano Tebaldini from PoliMi for useful and stimulating discussions.

REFERENCES [1] R.N. Treuhaft, B.E. Law, G.P. Asner, S. Hensley, “Vegetation Profiles Estimates from Multialtitude, Multifrequency Radar Interferometric and Polarimetric Data,” Proc. of IEEE 2000 Int. Geosci. and Remote Sensing Symp. (IGARSS), pp. 126-128, Hawaii, 2000. [2] A. Reigber, A. Moreira, “First Demonstration of Airborne SAR Tomography using Multibaseline L-band Data,” IEEE Trans. on Geosci. and Remote Sensing, vol. 38 (5), pp. 2142-2152, 2000. [3] G. Fornaro, F. Serafino, and F. Soldovieri, “Three Dimensional Focusing With Multipass SAR Data,” IEEE Trans. Geosci. Remote Sens., vol. 41(3), pp. 507-517, 2003. [4] F. Lombardini, A. Reigber, “Adaptive Spectral Estimators for Multibaseline SAR Tomography with Airborne L-Band Data,” Proc. of IEEE 2003 Int. Geosci. and Remote Sensing Symp. (IGARSS), Tolouse, France, 2003. [5] F. Lombardini, J. Ender, L. Rößing, M. Galletto, L. Verrazzani, “Experiments of Interferometric Layover Solution with the Three-antenna Airborne AER-II SAR System,” Proc. IEEE Int. Geosci. and Remote Sensing Symp., Anchorage, Alaska, Sept. 2004. [6] S. Tebaldini, F. Rocca, “SAR Tomography over Decorrelating Targets for Forestry Applications,” Proc. of 2008 European SAR Conference (EUSAR), Friedrichshafen, Germany, 2008. [7] S. Sauer, L. Ferro-Famil, A. Reigber, E. Pottier, “Multi-Aspect POL-InSAR 3D Urban Scene Reconstruction at L-band,” Proc. of 2008 European SAR Conference (EUSAR), Friedrichshafen, Germany, 2008. [8] C. Prati, F. Rocca, “Improving slant-range resolution with multiple SAR surveys,” IEEE Trans. on Aerospace and Electronic Systems, Vol. 29(1), pp. 135-144, 1993.


Workshop presentation

Full paper


  Higher level                 Last modified: 07.05.06