IEEE Geoscience and Remote Sensing Magazine - June 2023 - 15

sensing techniques allows us to monitor fire progress over
time [42] and to perform early fire detection [44]. The use
of the airborne methods in this situation is limited due to
smoke and high temperatures, to which drones equipped
with sensors are sensitive. The wide scope and scalability
of satellite imaging is also important to identify active hot
spots. Unmanned aerial vehicles (UAVs) can be additional
sources of information, but they are supplemental due to
their lack of continuity in time [40].
Among the threats induced by volcanic ash (being a
direct result of an eruption), we can indicate air and water
pollution, climate change [45], and affecting aviation
safety; therefore, monitoring and forecasting the location
of volcanic ash clouds is critically important [46]. Additionally,
falling dust, containing silicate particles, causes
diseases of the respiratory system and contaminates drinking
water [45]. Satellite remote sensing techniques are one
of the most widely used tools for monitoring and locating
volcanic ash clouds [46]. Their advantage over terrestrial
methods is the lack of the need to install and maintain
measuring instruments in hard-to-reach often dangerous
locations [45]. An example of an existing solution is
HOTVOLC, which is a Web geographic information system
based on data from the Spinning Enhanced Visible
and Infrared Imager aboard the Meteosat geostationary
satellite. The system exploits the infrared thermal range
(8-14 μm) to distinguish silicon particles from those of
water and sulfuric acid [45]. Utilizing the thermal infrared
(TIR) band also allows us to determine the height of an ash
cloud based on temperature, which is critical information
for aviation safety [45]. The use of a geostationary satellite
enables monitoring volcanic activity and dust clouds
24 h a day [45], but this monitoring can also be performed
cyclically using the Moderate Resolution Imaging Spectroradiometer,
in which both the visible light bands [47] and
TIR [48] are used.
There exists the correlation of MSI data with various
quantifiable indicators for the assessment of volcanic activity
[49]. End-to-end monitoring solutions commonly
exploit other data modalities, such as SAR, to detect the deformation
of volcanoes that precedes eruptions [50]. Afterward,
the detection of volcanic eruptions may be effectively
performed using classic [48] and deep [51] machine learning
techniques operating on MSI data. An important challenge
concerned with this approach is the rapid darkening
of lava, even though its temperature is still high, which may
result in an incorrect volcano rating status if a satellite image
is captured too late. Extending the analysis to capture
high-temperature information available from short-wave
infrared (SWIR) allows us to classify volcanic activity more
precisely [52]. Using spectral information, the presence of
sulfur dioxide (contained in volcanic eruptions) can also be
identified at an early stage of an eruption [52]. Finally, the
detection of volcanic ash clouds by using the 400-600-nm
bands from HSI can enable us to determine the activity of
a volcano, due to the possibility of separating ash clouds
JUNE 2023 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE
from water clouds [47] (hence, satellite imaging may be
useful even in the presence of clouds). Changes in the accuracy
of volcanic ash cloud detection are also observed
due to changes in meteorological conditions induced by
seasonality during the year [53].
DETECTION AND MONITORING OF
EARTHQUAKES AND LANDSLIDES
Detecting an area affected by an earthquake allows us to
estimate the scale of damage and plan rescue activities [54],
and precise information about the effects of an earthquake
obtained quickly increases the effectiveness of mitigating
their impact [55]. Monitoring also allows for preventive actions
[56]. The extent of the damage caused by an earthquake
often exceeds the capacity of ground equipment, due
to the dangerous and difficult access to some areas and the
risk of further negative phenomena related to infrastructure
damage and the limited operating time of the devices.
The major drawback of using terrestrial tools working at
the microscale is the limited area that can be assessed in
a given time; therefore, macroscale imaging with satellites
can increase the scalability of monitoring solutions. Most
often, the purpose of detecting damage caused by an earthquake
is to identify landslides and affected buildings. In the
case of landslides, the assessment is carried out using tools
that are divided into three categories: 1) analysis of landslide
image features by using optical data, including remote
airborne and satellite sensing data [54], [57]; 2) detection
of surface deformation and deposition due to landslides
by using radar data [58]; and 3) fusion of optical and radar
data [59], [60].
Detection of damaged buildings can be carried out
based on the geometrical features of the five main types
of damage: 1) sloped layers, 2) pancake collapses, 3) debris
heaps, 4) overturn collapses, and 5) overhanging
elements [61]. These elements can be identified using
ground methods, but the use of MSIs and HSIs allows for
executing this procedure in a much shorter time, given
that a sufficiently high spatial resolution is maintained
[62]. Damage to buildings, especially in highly urbanized
areas, must be mapped in a short time to improve (and
even enable) rescue operations. For historical buildings
and networks of tunnels, it is pivotal to efficiently map
the damage. The main characteristics of urban rubble to
be assessed encompass location, volume, weight, and the
type of collapsed building materials, including hazardous
components (e.g., asbestos), which can have a major
impact on further operations [59]. Due to the large spectral
and spatial heterogeneity of urbanized areas, classic
methods may become insufficient due to changes, e.g.,
caused by attempts to remove debris. Thus, capturing subtle
spectral differences for such areas may play a key role
in the urban environment.
Current research concerning the analysis of optical
data acquired in areas affected by earthquakes focuses
on MSIs. The use of 450-510- (blue), 520-580- (green),
15
IEEE Geoscience and Remote Sensing Magazine - June 2023

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - June 2023
