IEEE Geoscience and Remote Sensing Magazine - March 2020 - 92

imaging capabilities. Such peculiarities permit a systematic
data acquisition capability, which is of great importance for
the monitoring of dynamic processes, such as those associated with natural and/or human-induced hazards.
Since their basic and straightforward exploitation for
digital elevation model (DEM) generation at global scale
(see, for instance, the Shuttle Radar Topography Mission
(SRTM) [2] or the TerraSAR-X/TanDEM-X [3] mission), InSAR approaches have been developed considerably to allow the coherent processing of a sequence of multitemporal
SAR data. Today, differential InSAR (DInSAR) and, more recently, advanced DInSAR (A-DInSAR) processing methods
allow not only detection of possible ground deformations
but also monitoring of possible displacements taking place
in the observed area during the whole temporal acquisition
interval [4], [5].
Early applications of DInSAR concerned the measurement of large displacement fields, that is, those associated
with earthquakes [6] or volcanic activity [7]. Then, A-DInSAR and, specifically, persistent scatterer (PS) interferometry
(PSI) allowed improvement of the 3D positioning of scatterers and the extraction of information associated with possible deformations, thereby extending the application range
of SAR. By providing, for the first time, the ability to achieve
accurate measurements of the deformation of scatterers characterized by a persistency of the response (i.e., of PSs), these
algorithms have contributed to the unprecedented spreading
of SAR in several domains-above all, in the geological, geotechnical, and structural engineering fields [4].
In general, urban areas are characterized by the presence of a multitude of anthropic structures with which,
typically, several PSs are associated [8]. Due to the vertical
development of buildings and the particular side-looking
imaging geometry of SAR, however, they represent a challenging scenario from the data-processing viewpoint. The
radar viewing geometry induces the superpositioning of
the responses of different PSs (layover). SAR tomographic
techniques have contributed to the possibility of overcoming the layover [9], provided the vertical resolution is large
enough and data calibration is accurate.
With improvements of the resolution capabilities of SAR
sensors, particularly the availability of VHR X-band systems, research focused on the imaging and monitoring of
buildings, particularly those subject to subsidence associated with groundwater extraction or slope instabilities, has
been boosted. The literature is, in fact, rich with contributions addressing the application of DInSAR monitoring to
the built environment [10]-[18], including buildings [19]-
[23] and infrastructure [24]-[29].
This article offers an overview of improvements in the
development of a methodology for the risk evaluation of
buildings exposed to landslide hazards. To this end, we
provide a review of the interferometric data-processing approaches for 3D reconstruction and monitoring of buildings able to trade off spatial resolution and measurement
coverage. We also include a description of methods for an
92

application-oriented analysis of the measurements to complement the review. In more detail, we provide a methodology for geometric and kinematic characterization, including
the exploitation of models for dealing with incomplete data
[30] and the generation of tools for analyzing and forecasting the vulnerability of the exposed structures.
LANDSLIDE CONSEQUENCE ANALYSIS
In landslide risk theory, risk is defined as "the probability
of the landslide event multiplied by the consequences,"
where the consequences are "the outcomes or the potential outcomes arising from the occurrence of a landslide"
[31]. Focusing on slow-moving landslides affecting existing buildings, the (qualitative or quantitative) consequence analysis involves identifying and quantifying the
buildings at risk and estimating their vulnerability [32].
In particular, vulnerability is defined as the expected degree of loss to a certain building due to the occurrence of
a landslide of a given intensity; for instance, in quantitative analyses, the degree of loss can be associated with the
ratio of repair cost to monetary value/replacement cost
owing to the damage to the building at risk.
The identification includes
◗◗ the detection of buildings located within and/or on the
boundaries of the landslide-affected areas (this can be
done by overlapping the map of inventoried landslides
with territorial data)
◗◗ their classification based on data collected via image
processing and field surveys on some relevant characteristics (age of construction, structural and foundation typologies, occupancy type, and number of
floors) [33]-[38].
Once identified, the buildings at risk must be quantified (e.g., in terms of monetary value or replacement cost
[39]-[41]). On the other hand, vulnerability can be estimated by taking into account the landslide mechanism, a
related intensity parameter (resulting from the interaction
of the building foundation with the landslide body, that
is, differential displacements) and the described building
characteristics [32], [37]. The results can be expressed by
either vulnerability indices, which relate to a (range of)
intensity parameter values with the expected degree of
loss [36], [42], [43], or vulnerability curves, which show
the relationship between the average damage severity (established based on a classification system) expected to a
given building at the specified risk and intensity parameter value (adapted from [44] and [45]).
Vulnerability curves can be generated by managing socalled fragility curves, which provide the conditional probability for a given building at risk of reaching or exceeding
a certain damage severity level as a function of landslide
intensity, thus quantifying the uncertainties [46]. Based on
the scale of analysis and related purposes [31], [47], fragility
(and vulnerability) curves can be distinguished into heuristic, empirical, analytical, and hybrid, according to the
adopted method [38], [39], [46], [48].
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
