Polarimetric Ground and Volume Decomposition based on the PolInSAR Two-Layer model

Abstract

This work introduces a methodology to decompose the ground and volume scattering contributions using Polarimetric Interferometric SAR (PolInSAR) acquisitions. The introduction of interferometric information allows to overcome the limitation of conventional polarimetric model-based decompositions and to obtain full-rank polarimetric coherency matrices for the ground and volume scattering contributions. This is achieved by establishing a link between polarimetric and interferometric measurements in the context of a two-layer model. Assuming that these two dimensions of the data are separable, the polarimetric response of the two layers may be obtained as a function of their interferometric coherences. The proposed approach is applicable to single as well as multibaseline PolInSAR data. Accordingly, the technique is applied and evaluated with real single and multibaseline PolInSAR data at P- and L-band acquired by the DLR’s F-SAR airborne sensor over tropical forest from the AfriSAR 2016 campaign. Experimental results demonstrate the effectiveness of the proposed decomposition, which provides a uniform high-entropy volume component, while the retrieved ground layer preserves topographic features consistent with the expected Polarimetric Orientation Angle (POA). The analysis confirms the capability of the method to retrieve full-rank component matrices without imposing polarimetric constraints on the scattering mechanisms.