IEEE Geoscience and Remote Sensing Magazine - March 2020 - 93

For empirical methods, a thorough knowledge of slowmoving landslide features (i.e., geometry/volume and kinematics, materials, state of maintenance, and value of the
exposed facilities) is necessary [36]. However, the collection
of this relevant information is not always affordable, especially when the analysis involves many buildings within a
municipal territory. In particular, the accessibility of comprehensive databases of damage to facilities and its increase
(if any) over the time associated with the intensity parameter are key issues for consequence analysis purposes. For the
displacement patterns (and rates) experienced by a certain
facility or set of facilities affected by (or close to) slow-moving landslides, very useful information may be derived by
complementing conventional monitoring techniques (i.e.,
inclinometers, GPS, topographic leveling) with VHR SAR
satellite data processed via DInSAR techniques, providing
high coverage on built-up areas, as described in the following sections [35], [49]-[51].
MULTIPASS INTERFEROMETRIC DATA PROCESSING
SATELLITE SYNTHETIC APERTURE RADAR RESOURCES
The success story of satellite HR radar imaging began with
SEASAT in 1978 [52], a system designed mainly for oceanographic applications. After SEASAT and following SIR-A/B/C
[53]-[55], the NASA SRTM in 2000 acquired data that allowed delivery of a DEM (at that time, the most accurate) on
a worldwide scale [2]. SRTM products, provided through a
free open-access data policy, are, typically, used as reference
DEMs to which the interferometric processing of the multitemporal SAR stack is tailored. SRTM, therefore, still plays a
fundamental role in the development and application of the
DInSAR technique, including for landslides.
In the last decade of the 20th century, European, Japanese, and Canadian space agencies started operating
satellite SAR sensors [56]-[58]. The ERS-1, ERS-2, and Radarsat-1-all operating at the C band and classified as HR
according to the European Space Agency (ESA) data warehouse specification [59]-have collected historical interferometric archives that are valuable for the back analysis
often needed in several fields, including for landslide risk.
Figure 1 provides a chart describing the development
of InSAR systems, starting from ERS-1. The colors in this
figure highlight the exploitability of the sensor from interferometric applications. Equal colors and color gradations
for different sensors highlight the possibility of performing a joint interferometric analysis, that is, generating cross
interferograms between acquisitions pertinent to different
sensors in overlapping temporal intervals. Accordingly, a
variation in tone for a given color underlines the impossibility of performing coherent processing among different
sensors, thus driving the data processing toward independent subsets.
With Envisat, a multisensory satellite equipped with a
C band advanced SAR (ASAR) sensor [60], the ESA ensured
continuity of data acquisition with the ERS-2 mission.
MARCH 2020

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

Compared to the ERS missions, Envisat featured enhanced
capability, especially in terms of larger coverage. However,
this has rarely been exploited in the context of landslide
monitoring, mainly due to the consequent and unavoidable spatial resolution loss.
Following its Japanese Earth Resources Satellite (JERS),
which experienced problems in data quality for interferometric applications, the Japan Aerospace Exploration Agency launched the Advanced Land Observing Satellite (ALOS) in
2006. ALOS and it successor ALOS-2 instruments provide
L-band data and improved polarimetric modes [61]; worldwide data availability for interferometric applications, especially outside Asia, is still limited.
The development of DInSAR [5], carried out specifically
with ERS-1, Envisat, and Radarsat-1 data, improved the monitoring of natural hazards, including slope instabilities, and
paved the way for modern,
new-generation, meter-resolution SAR sensors. In the last
VERY USEFUL
12 years, second-generation
INFORMATION MAY BE
SAR sensors have seen considDERIVED BY COMPLEMENTING
erable technological advances.
CONVENTIONAL
In fact, the year 2007 marked an
MONITORING TECHNIQUES
important milestone: together
WITH VHR SAR
with the C band Radarsat-2
SATELLITE DATA.
[62] satellite, the X-band spaceborne SAR sensors onboard
the German TerraSAR-X [63]
and Italian COSMO-SkyMed satellites [64] were launched.
Their main advantage was the ability to provide VHR data
up to meter resolution: improvements in the monitoring of
the built environment and of infrastructure located in areas at risk, such as those affected by slope instabilities, have
been remarkable.
In addition to VHR capabilities, second-generation
SAR satellites also achieved reduced scene revisiting time
[3]. On one hand, this was essential for the monitoring of
emergency situations; on the other hand, it allows contrasting of the effects of temporal decorrelation, which
is more drastic at the X band compared to the C band.
As a matter of fact, the increased revisiting frequencies
of the X band allow improved coverage in sparsely vegetated areas compared with the previous-generation C
band sensor.
The German TanDEM-X mission is aimed at generating a
global DEM of Earth's surface with improved accuracy over
the DEM generated by the SRTM mission. The TanDEM-X
mission involved a bistatic X-band SAR antenna configuration flying in twin mode with a rather narrow orbital tube
between 250 and 500 m. The bistatic acquisition mode
has allowed data collection used to generate a DEM on a
global scale, characterized by a 12-m posting and a height
measurement accuracy of 4 m [3]. Especially for landslides,
a more accurate reference DEM facilitates the processing
of repeat-pass data acquired with large baselines, such as
those of the COSMO-SkyMed.
93

IEEE Geoscience and Remote Sensing Magazine - March 2020

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