|
Abstract:
|
This paper investigates the polarimetric dependence
of the interferometric complex correlation and proposes amethodology
for cancelling individual scattering mechanisms, in terms of
the complex correlation coefficient phase, under the assumption of
the random volume over ground model. This allows the estimation
of the ground topography on forested and vegetated areas. The
first part of the analysis considers the separation of the volume
from the ground (including the double-bounce scattering mechanism).
This process identifies the polarization states, without
constraining them to be equal in both polarimetric acquisitions,
which allow to cancel either the volume scattering contribution or
the ground contribution. In order to have access to the interferometric
phase of the remaining or isolated scattering mechanism,
the polarimetric phase contribution of this scattering mechanism
has to be removed in a second step. In the case of forested areas,
the previous methodology is considered from two different point of
views. For the estimation of the underlying ground topography, the
cancellation of the volume scattering contribution makes possible
to access the interferometric phase associated to the ground contribution.
In addition, the interferometric information associated to
the volume scattering contribution is estimated based on the cancellation
of the ground contribution. The proposed techniques are
analyzed on the basis of simulated and experimental polarimetric
interferometric synthetic aperture radar data, demonstrating that
the ground topography, as well as the height associated to the
volume contribution, are asymptotically nonbiased and dependent
on the shape of the particles of the random volume. In case of
spheres (η = 0),
the ground-to-volume ratio presents large values
favoring the accurate estimation of the topographic phase. For
the case of dipole like particles (η = 0.5), the ground-to-volume
ration decreases producing a coherence |ρ| in the order of 0.1,
making necessary a large speckle filtering to obtain a reliable
estimation of the topographic phase. |
