Georgian Electronic Scientific Journal: Computer Science and Telecommunications 2007|No.1(12)
A Study of statistical description of interferometric SAR signal
S. REDADAA, A. BOUALLEG and M. BENSLAMA
ENST de Bretagne, Technopôle Brest-Iroise, CS 83818 –29238 Brest Cedex 3, France
Email : firstname.lastname@example.org
Abstract: Interferometric SAR data are frequently multi-look processed for speckle
reduction and data compression. The statistical behavior of an interferometric signal is
for great interest since it allows predicting its performance, and yet, few papers deal
with the signal nature. This paper deals with the statistical behavior of the phase
difference between two interferometric signals. The probability density function of
interferometric signal is derived stating its dependence on the coherence and the mean
of the phase difference. Besides, the Cramer-Rao lower bound (CRB) is also evaluated.
Key words: Synthetic aperture radar (SAR), interferometry, image, statistics.
The main feature of a SAR system, in front of conventional systems, is the azimuth resolution
increase on the basis of carrying out a recording of complex data, as well as a coherent processing
of such a data. Despite the SAR images complex nature, its phase does not carry information at all.
On the other hand, SAR images are contaminated by speckle, a noise-like signal, although it is a
true electromagnetic measurement -.
The combination of different SAR images leads to multichannel SAR imagery. Any change in
the sensor’s geometry, in the operating frequency, in the employed polarization or the reflectivity
scene will produce a change within the SAR image. In those cases, in which the changes gives rise
to correlated images, the phase difference between the SAR images will contain useful information.
When the SAR images are completely correlated, each of the individual images is contaminated by
speckle, but the phase relation between them will contain useful information free