PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
January 2014
7
interferogram generation, removal of
curved Earth phase trend, adaptive
filtering, phase unwrapping, precise
estimation of interferometric baseline,
generation of a digital elevation
model or surface deformation
image, estimation of interferometric
correlation, and rectification of
interferometric products (e.g., Lu,
2007). Using a single pair of SAR
images as input, a typical InSAR
processing chain outputs two SAR
intensity images, a deformation map
or DEM, and an interferometric
correlation map (e.g., Lu
et al
., 2010):
•
The intensity of a SAR backscat-
tering image depends on the size,
shape, roughness, orientation,
dielectric constant (strongly influ-
enced by moisture content), and
terrain slope of the target. SAR in-
tensity images can be used to char-
acterize land cover types and their
changes, and (with long wave-
length SARs) to reveal structures
buried by vegetation canopy.
•
An InSAR coherence image is
a cross-correlation product de-
rived from two co-registered
complex-valued SAR images, and
depicts changes in backscattering
characteristics at the spectrum
of the radar wavelength. Loss of
InSAR coherence, often referred to
as decorrelation, can result from:
1.) thermal decorrelation caused
by uncorrelated noise sources in
radar instruments, 2.) geometric
decorrelation as a consequence of
imaging a target from very dif-
ferent looking angles, 3.) volume
decorrelation caused by volume
backscattering effects, and 4.)
temporal decorrelation due to envi-
ronmental changes over time (e.g.,
Zebker and Villasenor, 1992; Lu
and Kwoun, 2008). Decorrelation
makes an InSAR image useless
for measuring ground surface
deformation. On the other hand,
the pattern of decorrelation can
be used to characterize surface
modifications caused by flooding,
wildfire, volcanic activity, or earth-
quake shaking.
•
An InSAR deformation image is
derived from phase components
of two overlapping SAR images.
The spatial distribution of surface
deformation along the satellite
line-of-sight (LOS) direction can
be used to constrain numerical
models of subsurface deformation
sources (e.g., Lu, 2007).
•
The ideal SAR configuration for
DEM production is a single-pass
(simultaneous) two-antenna sys-
tem (e.g., Shuttle Radar Topogra-
phy Mission (SRTM)). However,
repeat-pass single-antenna InSAR
also can be used to produce DEMs
to characterize and monitor nat-
ural and man-made hazards that
result in significant changes in to-
pography (e.g., Lu
et al
., 2003).
Pol-InSAR
Most spaceborne SARs such as those
onboard ERS-1, ERS-2, JERS-1, and
RADARSAT-1 (Table 1) are single-po-
larized radars (i.e., radar signals
are both transmitted and received
with either vertical or horizontal
polarization). Sensors of this type
only partially capture the scattering
properties of targets on the surface.
Data from a fully-polarized radar (i.e.,
a system in which signals are both
transmitted and received with both
vertical and horizontal polarizations),
such as those onboard the Japanese
ALOS and Canadian Radarsat-2 sat-
ellites and future SAR satellites cur-
rently being planned, can be related
to the signatures of known elemental
targets, making it possible to infer
the type of scattering that is taking
place (e.g., Lee and Pottier, 2009).
For example, polarization signatures
of the vegetation canopy, the bulk
volume of vegetation, and the ground
surface are different and can be sep-
arated using polarimetric analysis.
An optimization procedure can be
employed to maximize interferometric
coherence between two polarimetric
radar images, thereby reducing the
effect of baseline and temporal decor-
relation in the resulting polarimetric
InSAR (Pol-InSAR) image. Then, us-
ing a coherent target decomposition
approach that separates distinctive
backscattering returns from the cano-
PSInSAR analysis of Newport-Inglewood fault over Los Angeles, CA.
Interferometric SAR (InSAR)
noun
uses two or more SAR images of the same area to extract the land surface
topography and any deformation that might have occurred during the temporal
interval spanned by the images.
