IEEE Geoscience and Remote Sensing Magazine - March 2020 - 102

Height

distortions and visibility limitations induced by the
side-looking geometry and range-imaging mechanism
used by the sensor. Depending on the local slope and
incidence angle, hillsides facing the sensor undergo a
compression in the range direction. This phenomenon,
sketched in Figure 3 and referred to as foreshortening, is
induced by a reduction of the range variation along the
sloped surface. The effect is easily identifiable in SAR images as an increase of the amplitude values with respect
to those observed if the area were flat. The geometric
compression results in a general loss of spatial and radiometric sensitivity of the ground textures of interest (i.e.,
the built environment), drastically impairing the monitoring capabilities in some cases.
As the topography becomes steeper, a further distortion,
referred to as layover, occurs. This phenomenon implies
that, in the resulting SAR image, the summit of the slope
(B in Figure 3) may anticipate in range the slope foot (A in
Figure 3). The valley beyond and the sloped area can, in
such a case, be one folded over the other, leading to a superposition (interference) of the backscattered echoes from
different elements.
The layover has an impact not only on the imaging of wide,
sloped surfaces, such as hills and mountains, but also on a local scale in the presence of vertical surfaces, such as building facades in urban areas. Although foreshortening and layover over
distributed areas cause problematic distortions that cannot be
easily resolved [121], layover in urban areas can be mitigated:
tomographic processing has the ability to separate the position
of the scatterers and provide the correct 3D reconstruction of
the observed built structure [12], [13].
In a similar way, for the same ranging principle, opposite
hillsides typically undergo an expansion (backslope) with
respect to flat regions in the sensed image. For landslide
monitoring, this effect is generally beneficial, provided
there are no critical situations, such as when the slope
exceeds the radar incidence angle, thus giving rise to a

Off-Nadir

Sl
an
tR

ge
A′

Foreshortening
B′

B
A

Backslope
C

Ground Range

FIGURE 3. The foreshortening effect on the hillsides facing the
observing sensor.

102

shadow. As explained previously, the interferometric phase
relative to landslide areas affected by significant geometric
distortion (large foreshortening, layover, and shadowing)
is generally useless.
A radar visibility map can be derived from orbit information and a DEM [30], [121]-[124]. For instance, in [125],
the main geometric constraints on the feasibility of an
ERS-1 and ERS-2 DInSAR analysis are summarized. This
is useful for a preliminary inspection aimed at identifying areas suitable for DInSAR monitoring and for selecting the best available acquisition geometry (ascending
or descending orbit). In practice, topographic configurations of mountainous areas affected by slope instability
could often limit the measurements to only a single-acquisition geometry.
SPECIFIC ASPECTS OF DEFORMATION
MEASUREMENTS
The principle of interferometry relies on the use of spatial
and temporal phase differences: the estimated superficial
deformation is a differential measurement with respect
to both time and space. This means that no information
is available outside the time interval covered by the SAR
multipass data set, and the temporal reference is generally
set to be the first acquisition. In addition, a spatial reference point is needed; therefore, the generated deformation
time series is referred to the deformation of a reference
point/area. The latter is a coherent target selected within
the radar scene area, whose stability is known or can be
inferred from the local geomorphological/geological characteristics. It has a direct influence on the identification
of stable/unstable areas; for example, if the reference point
is affected by periodic displacement (e.g., due to thermal
dilation), this oscillation may propagate on the time series
of the detected targets.
Because DInSAR phase is 2π unwrapped, the ambiguity
of phase measurements implies the impossibility of tracking correctly and unambiguously a single LOS deformation
exceeding m/4 within one revisiting time interval. Consequently, according to the landslide velocity scale proposed
in [126], spaceborne SAR interferometry is able to assess
displacement rates belonging to the first two classes corresponding to the extremely slow (up to 16 mm/year) to very
slow (from 16 mm/year to 1.6 m/year) classes.
The quantitative analysis and interpretation of 1D-LOS
DInSAR data for landslide analysis applications is closely
related to the value of velocity/displacement along a given
movement direction. To address this issue, Cascini et al.
[30] proposed a procedure that assumes a prevalent translational movement along the steepest slope direction for each
PS. Since the LOS projection on the along-slope direction
can be biased by errors related to the projection operation,
Cascini et al. [30] adopted a threshold on the condition
number of the inversion matrix solving the algebraic system used for the projection operation (i.e., from the ascending, descending, or ascending-descending LOS-to-slope
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

MARCH 2020

IEEE Geoscience and Remote Sensing Magazine - March 2020

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - March 2020