Forest temporal decorrelation: 3D analyses and processing in the Diff-Tomo framework

Abstract

Techniques based on coherent combination of complex (amplitude and phase) SAR data, in particular Pol-InSAR and 3D multibaseline SAR Tomography (Tomo-SAR), can extract rich information on complex scenarios with multiple scatterers mapped in the SAR cell. However, forest scenarios are characterized by a temporal decorrelating volume canopy scatterer, and a set of related open problems exists, in particular for Tomo-SAR techniques to be applied to space monitoring of biomass. Multipass/multibaseline 4D Differential Tomography (Diff-Tomo) is a promising advancement in this field, furnishing space-time signatures of multiple scatterer dynamics in the SAR cell, in particular with urban applications. In this paper, to deal with the forest decorrelation issues in Tomo-SAR, experimental advances are presented of the original extension of Diff-Tomo methods for analyzing vegetated scenes, to extract jointly geometric and dynamic information of forest layers. The Diff-Tomo enabled functionalities are separation in the height dimension of different temporal coherence levels that are mixed (undiscriminated) in classical total coherence analyses ("Tomography of coherences"), and 3D Tomography robust to temporal decorrelation. Extended airborne P-band multipolarimetric results and phenomenological investigations are shown.

4D quick Diff-Tomo analyses of short-term decorrelation of distributed scatterers

“…Diff-Tomo capacity to identify such spectral signatures has been confirmed with non-parametric 4D analysis of BioSAR-1 ESA campaign P-band forest data with 2 months time span, acquired by the DLR's E-SAR airborne platform, over a boreal region in Sweden [11,15]. A very large area (500 hectares) P-band separation of height-varying coherence times [11] has been also obtained with a semiparametric [15] (AR model) 4D+ Diff-Tomo processor, which can complement previous dedicated radar-tower ESA campaigns [3] and in particular the ongoing AfriScat and BorealScat.…”

“…Extending the setup in [11,15], two small scale (short range) experiments have been carried out, termed Partial coherence InSAR-tomo Stratigraphic short-term Anal-yses of trees Scattering (PisaScat). These use a special groundbased miniradar (ack.…”

“…tandem) satellites systems, such as the investigated SAOCOM-CS, even with the planned use of 3D correlative (track-pair only) imaging processing [5,10], can still be sensibly influenced by short-term temporal decorrelation of the typical windblown vegetation, depending on their formation-flying configuration time lag design. In addition to other emerging long-term non-stationarity issues related to seasonal and weather effects [8], it is thus important to expand characterizations of forest temporal decorrelation, especially at the short-time scale, also accounting for the height-varying nature of the decorrelation phenomena [3,6,9,11]. In fact, blurring origins are local, and the classical (ensemble) coherence metric considering the overall behaviour of the height-distributed scattering is not enough in the Tomo/Pol-InSAR context.…”

“…While initially usefully applied to subsidence monitoring in garbled urban areas (see e.g. [14]), this has actually the potential to extract dynamic information in layered media, including separating along the height dimension different temporal decorrelation mechanisms that are mixed (undiscriminated) in the classical coherence analysis [9,11,15]. This separation is based on the concept of space-time signatures of long-term temporal decorrelation [9], which are scattering component distributions in the domain of spatial (height) and temporal frequency of harmonics in which a signal from a scatterer with temporal decorrelation can be decomposed.…”

“…A very large area (500 hectares) P-band separation of height-varying coherence times [11] has been also obtained with a semiparametric [15] (AR model) 4D+ Diff-Tomo processor, which can complement previous dedicated radar-tower ESA campaigns [3] and in particular the ongoing AfriScat and BorealScat.…”