Differential SAR Tomography (Diff-Tomo) has emerged as a powerful crossing of Differential Interferometry and 3D Tomography, producing 4D (3D+Time) imaging capabilities extensively applied to urban monitoring. More recently, through Diff-Tomo, identifying spatial (height)-temporal spectra of multiple height-distributed decorrelating (forest) scatterers, an advanced decorrelation-robust Tomography has been obtained, together with separation of decorrelation mechanisms in forest layers. After the related first Diff- Tomo processor, full model-based Generalized-MUSIC, a semi-parametric Generalized-Capon processor has been conceived and tested for the non-stationary distributed scenarios. In this paper, the feasibility of exploiting the Diff- Tomo framework for decoupling the interfering canopy scatterers in the retrieval of surface deformation in vegetated areas is analysed, expanding first trials in [1]. This potential is tested applying the Generalized-MUSIC/-Capon methods for this challenging application, the former being expected to be more accurate when correctly matched and the latter more flexible. Diff-Tomo subcanopy deformation (subsidence or deep landslide) monitoring is explored by simulated analyses of an L-band scenario; system tradeoffs in resorting to higher or lower carrier frequency are also hinted. Comparison with advanced interferometry methods is discussed, and real forest data Diff-Tomo tests of deformation monitoring, currently at P-band, are presented.
Gen-Capon and Gen-MUSIC Diff-Tomo for Non-Stationary Distributed Media: Exploration of Potential for Subcanopy Subsidence Monitoring
F. Lombardini
Primo
;
2020-01-01
Abstract
Differential SAR Tomography (Diff-Tomo) has emerged as a powerful crossing of Differential Interferometry and 3D Tomography, producing 4D (3D+Time) imaging capabilities extensively applied to urban monitoring. More recently, through Diff-Tomo, identifying spatial (height)-temporal spectra of multiple height-distributed decorrelating (forest) scatterers, an advanced decorrelation-robust Tomography has been obtained, together with separation of decorrelation mechanisms in forest layers. After the related first Diff- Tomo processor, full model-based Generalized-MUSIC, a semi-parametric Generalized-Capon processor has been conceived and tested for the non-stationary distributed scenarios. In this paper, the feasibility of exploiting the Diff- Tomo framework for decoupling the interfering canopy scatterers in the retrieval of surface deformation in vegetated areas is analysed, expanding first trials in [1]. This potential is tested applying the Generalized-MUSIC/-Capon methods for this challenging application, the former being expected to be more accurate when correctly matched and the latter more flexible. Diff-Tomo subcanopy deformation (subsidence or deep landslide) monitoring is explored by simulated analyses of an L-band scenario; system tradeoffs in resorting to higher or lower carrier frequency are also hinted. Comparison with advanced interferometry methods is discussed, and real forest data Diff-Tomo tests of deformation monitoring, currently at P-band, are presented.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
