Adiri 2020

Ore Geology Reviews 117 (2020) 103332
Contents lists available at ScienceDirect
Ore Geology Reviews

journal homepage: www.elsevier.com/locate/oregeorev
Recent advances in the use of public domain satellite imagery for mineral T
exploration: A review of Landsat-8 and Sentinel-2 applications
⁎
Zakaria Adiria, , Rachid Lhissoub, Abderrazak El Hartia, Amine Jelloulia, Mohcine Chakouria
a
Team of Remote Sensing and GIS Applied to Geosciences and Environment, Faculty of Sciences and Techniques, Beni Mellal, Morocco
b
Institut National de la Recherche Scientifique, Centre Eau Terre Environnement, 490 rue de la Couronne, G1K 9A9 Québec, QC, Canada
A R T I C LE I N FO A B S T R A C T
Keywords: Thanks to their wide coverage and the valuable spectral information, remote sensing data constitute a popular
Mineral mapping instrument in the mineral exploration toolbox. In recent times, the remote sensing community has witnessed the
Hydrothermal alteration launch of the new and improved Landsat-8 and Sentinel-2 multispectral sensors. The former constitutes the
Remote sensing eighth sensor of the Landsat series launched by the National Aeronautics and Space Administration (NASA),
Landsat
while the latter is linked to the Sentinel-mission launched by the European Space Agency (ESA). The main
Sentinel
Geology
objective of our contribution is to provide a comprehensive review of the use of the Landsat-8 and Sentinel-2
multispectral sensors in mineral exploration. The free and open access to these data and their enhanced spectral
and spatial characteristics (compared to the existing multispectral sensors) has clearly promoted the use of
remotely sensed in mineral exploration. In addition, as illustrated by the case studies presented in this paper,
Landsat-8 and Sentinel-2 data present effective and accurate mapping tools for mineral exploration. Both sensors
identified iron oxides and Al-OH absorption features, in addition to silicate and carbonate minerals. Our review
indicated that Landsat-8 is by far the more popular sensor in mineral exploration applications. Greater uptake of
Sentinel-2 and further case studies will be necessary to better demonstrate its capabilities and potential.
1. Introduction geology, geochemistry, geophysics and laboratory analysis has pro-

vided valuable and detailed information which are useful for miner-
Geology is a discipline that focuses on the study of the composition, alogical and geological studies (Bedell et al., 2009; Amer et al., 2012;
structure and history of Earth processes and their products (van der Liu et al., 2013; Abdelnasser et al., 2018). In addition, remote sensing is
Meer et al., 2012). Geological research relies on a variety of data a relatively quick and cost-effective method permitting the acquisition
sources such as field and laboratory studies, geophysics, geochemistry, and interpretation of spectral signatures over relatively large areas of
aerial photography and remote sensing (Gregory, 1979). The latter interest (Rajesh, 2004; Girija and Mayappan, 2019) and facilitating, for
represents a critical data source for a variety of geological applications, example, the acquisition of high-quality geological information over
such as mineral alteration (Abrams et al., 2015; Agar and Coulter, 2007; remote and otherwise poorly accessible areas (Gregory, 1979; Pour
Bedini, 2011, 2017; Hu et al., 2018; Ibrahima et al., 2018; van Der Meer et al., 2017; Girija and Mayappan, 2019).
et al., 2012), lithological (Rowan and Mars, 2003; Rowan et al., 2006; Excellent discussions of the use of remote sensing in geological
Leverington and Moon, 2012; Pour et al., 2018a,b), or structural studies are provided by Cloutis (2007), Sabins (1999), Bedell et al.
(Dubois, 1999; Sabins, 1999; Rajesh, 2004; Corgne et al., 2010; Hashim (2009), van der Meer et al. (2012, 2014a,b), Abrams et al. (2015),
et al., 2013) mapping. However, remote sensing applications are lim- Bedini (2017), and Girija and Mayappan (2019), with landmark pub-
ited or no use in areas when the bedrock study area is covered by thick lications on this topic by Sabins (1996) and Gupta (2003). The con-
regolith or vegetation, rendering the satellite imagery blind to the sensus from the above studies is that remote sensing offers the capacity
bedrock. for optimizing exploration strategy and approaches, identifying new
Remote sensing is the process of detecting and monitoring the prospects, and reducing exploration costs (Pour and Hashim, 2011;
physical characteristics of an area by measuring its reflected and Amer et al., 2012). Of course, the efficiency of remote sensing as a
emitted radiation at a distance from the targeted area (Lambin, 2001). mineral exploration tool is highly contingent upon the targeted bedrock
The use of satellite data in combination with data derived from field being devoid of thick, laterally continuous regolith cover. The search of
⁎
Corresponding author.
E-mail address: [email protected] (Z. Adiri).
https://doi.org/10.1016/j.oregeorev.2020.103332
Received 15 June 2019; Received in revised form 6 January 2020; Accepted 12 January 2020
Available online 15 January 2020
0169-1368/ © 2020 Elsevier B.V. All rights reserved.
Z. Adiri, et al. Ore Geology Reviews 117 (2020) 103332
Fig. 1. (a) General location of the Bouskour Patte d’Oie deposit, (b) simplified geological map of the deposit area and surrounds, (c) and (d) RGB color composites
(band 4/band 3/band 2) extracted from the associated Landsat-8 and Sentinel-2 images respectively.
mineral deposit involves primarily the mapping, identification and 2. Characterization of mineral alteration assemblages
study of alteration mineral assemblages linked to the targeted mineral
deposit type. These tasks are achieved through the analysis of the Most if not all hydrothermal mineral deposits are structurally con-
spectral signature recorded in the VNIR (Visible-Near Infrared), SWIR trolled, and most are enveloped by characteristic, often zoned, hydro-
(Shortwave Infrared) and TIR (Thermal Infrared) regions of the elec- thermal wall rock alteration assemblages (Funedda et al., 2018; Tuduri
tromagnetic spectrum (Hunt, 1977; Sabins, 1999; Rajesh, 2004; Di et al., 2018; Chauvet, 2019). Most remote sensing studies have been
Tommaso and Rubinstein, 2007; Azizi et al., 2010; Bedini, 2011; Pour undertaken on hydrothermal systems because they commonly enclose
and Hashim, 2011, 2012; van der Meer et al., 2012; Bedini, 2017) with various spectrally active mineral groups, including iron oxides, clays,
the This spectral signature constituting the key mineral identification carbonates and others (van der Meer et al., 2012) (Fig. 2). In fact, every
criterion (Girija and Mayappan, 2019). mineral is characterized by diagnostic absorption features in the elec-
This paper provides a comprehensive review of the use in mineral tromagnetic spectrum. These areas, which can be recorded in the VNIR,
exploration of the latest Landsat-8 and Sentinel-2 multispectral sa- SWIR and TIR spectral regions, serve as a key for mineral identification
tellites. Several case studies and different mapping techniques are re- (Hunt, 1977; Harris et al., 1998, Azizi et al., 2010, Kaufmann et al.,
viewed in this paper with the ultimate goal of assessing their effec- 2010; Bedini, 2011, Pour and Hashim, 2011; van der Meer et al., 2012;
tiveness for mineral mapping. A cornerstone of this review is a case Girija and Mayappan, 2019). Mineral deposits form as a result of phy-
study by Adiri et al. (2018a) of the Bouskour Patte d’Oie copper deposit sicochemical interaction of metal-bearing hydrothermal fluids with the
(Fig. 1). surrounding rock, triggering changes in fluid composition and metal
deposition typically associated with the generation of well-defined
hydrothermal alteration zones comprising diagnostic mineral assem-
blages (Fig. 3) (Kaufmann et al., 2010; Pour and Hashim, 2011).
2
Fig. 2. Schematic overview of the key hydrothermal deposit types and the associated minerals and host rocks (adapted from http://earthsci.org/mineral/mindep/
class_dep/class_dep.html).
Hydrothermal alteration can manifest as potassic (quartz, sericite, to highlight the presence of iron oxide minerals (electronic transitions)
biotite), phyllic (quartz, sericite, pyrite, muscovite, illite), argillic (Hunt, 1977; Gad and Kusky, 2007; Kaufmann et al., 2010; Pour and
(quartz, kaolinite, montmorillonite, illite, muscovite), and propylitic Hashim, 2012). Regarding the SWIR part, operating from 1 to 3 µm, it is
(chlorite, carbonates ± calcite, epidote) mineral assemblages devel- specific for the detection of Al-OH, Mg-OH, CO3, NH4, and SO4 bearing
oped in the wall rocks surrounding a mineral deposit (Rajesh, 2004; minerals (vibrational transitions) (Hunt, 1977; Kaufmann et al., 2010;
Rowan et al., 2006; Kaufmann et al., 2010; Azizi et al., 2010; Mars, Kruse, 2010; Leverington and Moon, 2012; Pour and Hashim, 2012;
2014; Bedini, 2011; Pour and Hashim, 2011). The identification, re- Grebby et al., 2014). For the TIR region, operating from 7 to 14 µm, it is
cognition, and discrimination of hydrothermal alteration present a very able to map silicates and carbonates (Sabins, 1999; Rowan and Mars,
important task in mineral exploration, because of the alteration foot- 2003, Gad and Kusky, 2007, Azizi et al., 2010; Pour and Hashim, 2012;
print is typically much larger than the deposit/ore footprint, and thus van Der Meer et al., 2012). Fig. 2 gives a schematic overview of the key
easier to detect (Kaufmann et al., 2010; Groves et al., 2016; Hamisi hydrothermal deposit types and the associated minerals and host rocks,
et al., 2017). while Table 1 highlights the most abundant and spectrally active mi-
Spectrally, the VNIR region (operating from 0.4 to 1 µm) is intended nerals related to these systems. In the literature, the spectral library
Fig. 3. Schematic cross section of hydrothermal alteration mineral zones (porphyry copper system) (adapted from Lowell and Guilbert (1970), Kaufman et al.
(2010)).
3
Table 1
Key hydrothermal deposit types in relation to the main spectrally active minerals (Source: Livo et al. (2007), Herrington (2011), van Der Meer et al. (2012)).
Hydrothermal deposit type Main spectrally active alteration minerals
Porphyry Cu Biotite, anhydrite, chlorite, sericite, pyrophyllite, zeolite, smectite, carbonate, tourmaline
Carlin-type Illite, dickite, kaolinite
Skarn Garnet, clinopyroxene, wollastonite, actinolite, calcite, chlorite, hematite, illite, fosterite, serpentine-talc, magnetite, calcite
Volcanogenic massive sulfide (VMS) Sericite, chlorite, carbonates, anhydrite, gypsum, amphibole
Epithermal Alunite, pyrophyllite, dickite, kaolinite, diaspore, zunyite, smectite, illite, Sericite, illite, chlorite, carbonate
Sedimentary Exhalative (SEDEX) Pyrite, calcite, dolomite, quartz
Mesothermal Biotite, carbonates, pyrite, pyrrhotite
Intrusion Related Albite, alunite, anhydrite, biotite, epidote, hematite, illite, kaolinite, muscovite
Hot-Springs Kaolinite, illite, montmorillonite, calcite, alunite
Fig. 4. Examples of minerals spectra extracted from the USGS (US Geological Survey) spectral library.
provided by Clark et al. (1999) constitutes a valuable source for mi- 3. Landsat-8 and Sentinel-2 multispectral sensors
neralogical remote sensing studies (Bedini, 2017). Then, Fig. 4 illus-
trates some examples of mineral spectra extracted from this library. 3.1. Landsat-8
From their first launch in 1972 by the NASA, the Landsat satellites
have marked the use of remote sensing imagery for land recording over
4
Fig. 5. Schematic View of Landsat-5, Landsat-7, Landsat-8 and Sentinel-2 Spacecrafts (adapted from Sentinel-2 User Handbook (2015), Landsat 8 Data Users
Handbook (2019), https://landsat.gsfc.nasa.gov/landsat-5/, https://eos.com/landsat-7/, https://landsat.gsfc.nasa.gov/landsat-8/landsat-8-overview/).
Fig. 6. Landsat-5 TM, Landsat-7 ETM+, Landsat-8 OLI/TIRS and Sentinel-2A sensors spectral channels and band-passes (adapted from van Der Meer et al. (2014a),
Wulder et al. (2019)).
nearly 47 years. The first generation (Landsat 1, 2 and 3) operated from reported in Table 2.
1972 to 1985, whereas the second generation (Landsat 4, 5 and 7) Since 2013, Landsat-8 is manifested as the new generation of the
began from 1982 and continues to the present. For Landsat 6, it was series of Landsat satellites, with the goal to ensure the continuity of the
launched in 1993 but it failed to reach orbit (Sabins, 1999; Pour and Landsat mission (Landsat Data Continuity Mission (LDCM)). This sa-
Hashim, 2015). Landsat-5 and Landsat-7 satellites have been widely tellite, which joined Landsat-7 on-orbit, was launched on February 11,
used in various geological remote sensing applications (El Harti et al., 2013, from Vandenberg Air Force Base in California. It is a sun-syn-
2004; Kovalskyy and Roy, 2013; Girija and Mayappan, 2019). As illu- chronous satellite that acquires spectral information in the VNIR, SWIR
strated in Fig. 5, Landsat-5 comprises two instruments, namely the and TIR regions (Roy et al., 2014). Landsat-8 carries two sensors,
Multispectral Scanner (MSS) and the Thematic Mapper (TM). The first namely the Operational Land Imager (OLI) and Thermal Infrared Sensor
one is characterized by four bands, whereas the TM contains seven (TIRS). The former records images with nine spectral bands, while the
bands. For Landsat-7, it carries the Enhanced Thematic Mapper Plus latter collects thermal data with two bands (Table 2) (Fig. 6) (Irons
(ETM+) sensor which consists of eight bands, including a panchro- et al., 2012; Roy et al., 2014; Mohy et al., 2018; Girija and Mayappan,
matic one (Fig. 6). The main characteristics of these sensors are 2019). Landsat-8 data are characterized by significant technological
5
Table 2
Landsat-5 TM, Landsat-7 ETM+, Landsat-8 (OLI/TIRS) and Sentinel-2A sensors characteristics.
Sensor Bands Spectral range (µm) Pixel size (m) Spectral resolution Radiometric resolution (bit)
Landsat-5 TM Band 1 – Blue 0.45 – 0.52 30 7 bands 8

Band 2 – Green 0.52 – 0.60 30
Band 3 – Red 0.63 – 0.69 30
Band 4 – NIR 0.76 – 0.90 30
Band 5 – SWIR 1 1.55 – 1.75 30
Band 6 – TIR 10.40 – 12.50 120
Band 7 –SWIR 2 2.08 – 2.35 30
Landsat-7 ETM+ Band 1 – Blue 0.45 – 0.52 30 8 bands 8
Band 2 – Green 0.52 – 0.60 30
Band 3 – Red 0.63 – 0.69 30
Band 4 – NIR 0.77 – 0.90 30
Band 5 – SWIR 1 1.55 – 1.75 30
Band 6 – SWIR 2 2.08 – 2.35 30
Band 7 – TIR 10.40 – 12.50 60
Band 8 – Panchro. 0.52 – 0.90 15
Landsat-8 OLI Band 1 – Coastal 0.43 – 0.45 30 11 bands 16
Band 2 – Blue 0.45 – 0.51 30
Band 3 – Green 0.53 – 0.59 30
Sensor Bands Spectral range (µm) Pixel size (m) Spectral resolution Radiometric resolution (bit)
Band 4 – Red 0.64 – 0.67 30

Band 5 – NIR 0.85 – 0.88 30
Band 6 – SWIR 1 1.57 – 1.65 30
Band 7 – SWIR 2 2.11 – 2.29 30
Band 8 – Panchro. 0.50 – 0.68 15
Band 9 – Cirrus 1.36 – 1.38 30
TIRS Band 10 – TIRS 1 10.60 – 11.19 100 16
Band 11 – TIRS 2 11.50 – 12.51 100
Sentinel-2A Band 1 – Coastal 0.433 – 0.453 60 13 bands 12
Band 2 – Blue 0.457 – 0.522 10
Band 3 – Green 0.542 – 0.577 10
Band 4 – Red 0.650 – 0.680 10
Band 5 – Veg red 0.679 – 0.718 20
Band 6 – Veg red 0.732 – 0.747 20
Band 7 – Veg red 0.773 – 0.793 20
Band 8 – NIR 0.784 – 0.899 10
Band 8A – Veg red 0.855 – 0.885 20
Band 9 – Water v. 0.935 – 0.955 60
Band 10 – Cirrus 1.36 – 1.39 60
Band 11 – SWIR 1.565 – 1.655 20
Band 12 – SWIR 2.10 – 2.28 20
advancement in terms of spectral resolution, signal quantization level, 2015, 2016; Mandanici and Bitelli, 2016; Hu et al., 2018; Ibrahima
Signal-to-Noise ratio, and radiometric resolution (16 bits). This allows et al., 2018). These data are characterized by higher spectral and spatial
the measurement of subtle variability in the surface conditions by a resolutions than those of other multispectral sensors such as ALI,
sharper color contrast (compared to Landsat-7 and ASTER sensors) Landsat-7 (Ge et al., 2018).
(Abrams, 2000; Markham et al., 2004; Irons et al., 2012; Morfitt et al., In June 2015, the ESA launched the Sentinel-2A sensor with the aim
2015; Pour and Hashim, 2015; Safari et al., 2017; Pour et al., 2017; to provide global land coverage, as well as ensure the Earth observation
Pour et al., 2018a,b,c; Yang et al., 2018). In addition, Landsat-8 images data continuity (Drusch et al., 2012; van der Meer et al., 2014a). This
cover 185 km × 180 km in width, available at no cost (Roy et al., 2014; sensor carries a high-resolution multispectral instrument (MSI) with
Pour and Hashim, 2015; Zhang et al., 2016). thirteen spectral bands (Table 2) (Figs. 5 and 6) (Drusch et al., 2012;
Additionally, the spectral ranges of OLI bands were refined to avoid van der Meer et al., 2014a; van der Werff and van der Meer, 2015,
the atmospheric absorption features presents within ETM + bands. This 2016; Hu et al., 2018; Ibrahima et al., 2018; Ge et al., 2018). Sentinel-
refinement appears in band 5 (0.850–0.880 µm) that excludes a water 2A is a sun-synchronous satellite with a scene swath width of 290 km,
vapor absorption feature at 0.825 µm, which is present in band 4 encoded in 12 bits, available free of charge with a revisit time of
(0.780–0.900 µm) of Landsat-7 (Irons et al., 2012; Zhang et al., 2016; 10 days at the equator (van der Meer et al., 2014a; Sentinel-2 User
Girija and Mayappan, 2019). Moreover, two new bands are included: Handbook, 2015; Ibrahima et al., 2018). The Sentinel-2 products are
band 1 for coastal aerosol applications, and band 9 for the detection of useful for various applications, such as land cover change detection,
cirrus clouds (Zhang et al., 2016). geochemical/physical variation, vegetation cover and natural disaster
management. In mineral exploration, the Sentinel-2 data have shown
3.2. Sentinel-2 their potential especially in mapping the iron absorption feature. This is
related to the similar or even higher spectral characteristics than
Under the European Earth Observation program (Copernicus), Landsat series and SPOT in the VNIR part (van der Meer et al., 2014a;
which was established by the European Union (EU) and the European van der Werff and van der Meer, 2015; Ge et al., 2018).
Space Agency (ESA), the Sentinel satellite missions have started. These
missions, from Sentinel 1 to 6, aim to provide data continuity for the 3.3. Data preprocessing
current operational satellite systems (e.g., SPOT, Landsat) (Drusch
et al., 2012; van der Meer et al., 2014a; van der Werff and van der Meer, In remote sensing, a good data interpretation relies on the control of
6
potential signal modification recorded during the acquisition process. 4.1. Band math and developed selective principal component analysis
Hence, the preprocessing step (including radiometric calibration, at-
mospheric and geometric corrections) is critical in order to get a useful Safari et al. (2017) evaluated the potentiality of the Landsat-8 data
scientific product (Chander et al., 2009). for mapping hydrothermal alteration mineral zones associated with
Landsat-8 images contain Digital Numbers (DN) values that must be porphyry copper deposits, at Shahr-e-Babak, Kerman, south of Iran. For
converted to the radiance and then to the ground reflectance. These this purpose, Band math and Developed Selective Principal Component
data are distributed at the L1T (corrected terrain) level, with a Analysis (DSPCA) approaches have been used. Band math is a com-
Universal Transverse Mercator projection and a World Geodetic System monly used mathematical technique that can be performed on several
WGS 84 datum. The preprocessing operation has been successfully multi-band images (with the same spatial dimensions) through using
applied through several correction algorithms. Some authors applied mathematical operations (addition, subtraction, division, and multi-
the FLAASH (Fast Line-of-sight Atmospheric Analysis of Spectral plication). The formulated expression is then calculated for each pixel
Hypercubes) module (Zhang et al., 2016; Pour et al., 2017, 2018a,b,c; and the final result is obtained as one band (https://www.
Yang et al., 2018), whereas other authors used the Internal Average harrisgeospatial.com/docs/BandMath.html) (Fig. 8a).
Relative Reflectance (IARR) method (Pour and Hashim, 2015; Safari Regarding the Principal Component Analysis (PCA), it is a statistical
et al., 2017). In addition, there are other studies that applied the Dark technique that reduces the input data dimensionality and compresses
Object Subtraction (DOS) method (Adiri et al., 2016, 2017, 2018a,b). the contained information in a small number of bands called principal
Basically, FLAASH is a radiative transfer model which is based on the components (PCs) (Gabr et al., 2010; Amer et al., 2012; Girija and
MODTRAN (Moderate Resolution Transmittance) code (Zhang et al., Mayappan, 2019) (Fig. 8b). This reduction allows eliminating data re-
2016). After the calculation of the radiance at sensor, this algorithm dundancy, segregating noise content by suppressing irradiance effects,
performs the atmospheric correction taking into account several para- decorrelating raw bands, and therefore enhancing the targeted in-
meters (atmospheric and aerosol models, water vapor, scene visibi- formation in the image (Singh and Harrison, 1985; Pour and Hashim,
lity…) (Pour et al., 2018a,b). For the IARR correction, it aims to nor- 2011, 2012; Amer et al., 2012; Gupta et al., 2013; Girija and Mayappan,
malize the images to a scene average spectrum. This is done by the 2019). Subsequently, each resulting PC receives a contribution from
calculation of an average spectrum for the entire scene, then dividing it each band of the input image. The quantity of this contribution can be
by the spectrum of each pixel of the image (Kawishwar, 2007; Safari determined by examining the eigenvectors in the resulting statistics.
et al., 2017). Regarding the DOS method, it requires the calculation of The value of the eigenvector in each band will be used to find the PC
the Tope of Atmosphere (TOA) reflectance as a first step. Then, the TOA that contains the targeted spectral information (Zhang et al., 2007;
value of the darkest object in the image is estimated and subtracted Gabr et al., 2010; Amer et al., 2012; Pour and Hashim, 2011, 2012;
from the entire pixels values (Chavez, 1988; Zhang et al., 2010; Zoheir and Emam, 2012; Amer et al., 2012; Safari et al., 2017). These
Atmospheric correction, 2011). In geological studies, bands 1 and 9 are considerations have been proposed by Crosta and Moore (1990) as
not used because the first one is intended for recovering the properties Developed Selective PCA technique, with the aim to reduce data di-
of atmospheric aerosols, while the second one is specific to the study of mension and focus on the bands related to targeted material.
cirrus (Roy et al., 2014). The panchromatic (band 8) is useful for spatial Firstly, Landsat-8 sensor was able to differentiate between altered
resolution enhancement (from 30 m to 15 m). and unaltered rocks in the study area, through highlighting the pre-
Regarding the Sentinel-2A images, they are distributed at the L1C sence of clay and carbonate minerals in the altered type spectral sig-
(TOA reflectance) level with the same projection and the same global natures (through bands 5 and 7). Then, a false RGB color composite
geodetic system as Landsat-8 data. The atmospheric correction can be (band 2, (band 6 – band 7), (band 5 – band 6)) was used to separate
carried out using the Sentinel Application Platform (SNAP) software hydrothermal alteration zone from the background (other components).
provided by the ESA (Ge et al., 2018; Ibrahima et al., 2018; Girija and Afterward, DSPCA technique was calculated (based on the VNIR and
Mayappan, 2019) or the DOS method (Adiri et al., 2018a,b). Similarly, SWIR spectral regions) for mapping Al-OH bearing altered rocks. The
bands 1, 9 and 10 (which are specific to atmospheric aerosols, water resulted PCs were discussed in relation with bands 2, 5, 6 and 7 of
vapor, and cirrus detection respectively), in addition to bands 5 to 7 Landsat-8. These analyses allowed discriminating the rocks with Al-OH
and 8A (which are specific to the vegetation studies) are not used for alteration type, as well as some potential new prospects at the regional
geological applications (Drusch et al., 2012; Hu et al., 2018). Fig. 7 scale, which could be subject to future research investigations.
shows some examples of the library mineral spectra after resampling to
Landsat-8 OLI and Sentinel-2A bandwidths. 4.2. Independent component analysis
Pour et al. (2018a) used Landsat-8 data for geological and miner-
4. Mineral exploration with the Landsat-8 and Sentinel-2 data alogical mapping in inaccessible regions context (Morozumi Range and
Helliwell Hills regions of Northern Victoria Land, Antarctica). This task
As mentioned before, the use of remote sensing in geological and was achieved using the Independent Component Analysis (ICA) algo-
mineralogical mapping has known a high success, especially since the rithm. This latter is based on the data independence assumption, which
launch of the Landsat TM and ETM + sensors. Crosta and Souza Filho allows separating the contained information even if it occupies a small
(2009) provided an overview of the use of these sensors in mineral part of the image (Szu and Buss, 2000; Zhang and Chen, 2002; Fiori,
exploration (Bedell et al., 2009). Among others, Band Ratios (BR) 2003; Du et al., 2006; Chen and Jain, 2014) (Fig. 9). This statistical
technique has been widely applied to these data types in order to map method processes data by rotation in an n-dimensional space (n is the
clays and iron oxides bearing minerals (Goetz and Rowan, 1981; Sabins, number of bands) until the input data are as independent as possible
1999). For example, the ratios 7/5 and 3/1 of Landsat TM has been (Chen and Jain, 2014). In other words, the ICA algorithm tries to find a
used to map hydroxyls and iron oxides minerals respectively (van der linear transformation that makes the result as independent as possible
Meer et al., 2012). As highlighted in the next subsections, several au- (Robila, 2004, Villa et al., 2011). For the ICA and PCA methods, an
thors used Landsat-8 and Sentinel-2 in mineral exploration. For this object is mapped according to the eigenvectors in the statistical results:
purpose, they applied various mapping methods based on different al- it is distinguished by bright pixels if the values of the eigenvectors are
gorithms. These methods and the obtained results were described in the positive in the reflectance bands and negative in the absorption bands
following discussion. of this object. In the opposite case, the latter is mapped by dark pixels
(Pour and Hashim, 2011; Adiri et al., 2018b).
In their study, Pour et al. (2018a,b,c) exploited the IC3, IC4 and IC7
7
Fig. 7. Examples of minerals spectra extracted from the USGS (US Geological Survey) spectral library (a, b) and resampled to Landsat-8 OLI (c, d) and Sentinel-2A (e,
f) bandwidths.
8
Fig. 8. (a) Example of Band math operation and (b) an illustration of Developed Selected Principal Component Analysis (DSPCA) technique.
derived from band 4, band 5 and band 10 of Landsat-8 for dis- absorption features, whereas band 10 can identify the presence of si-
criminating snow/ice, iron oxide/hydroxide, and silicate rocks bearing licate, which generally marks the transition from felsic to mafic units.
minerals respectively. The analysis showed that hematite, jarosite, For ICA method, an RGB color combination of IC5 (band 6 of Landsat-
goethite, and limonite indicate strong absorption features in the VNIR 8), IC6 (band 7 of Landsat-8) and IC7 (band 10 of Landsat-8) respec-
part (bands 4 and 5 of Landsat-8). Snow/ice presents a low reflectance tively was used. Clay and carbonate minerals showed reflectance fea-
in the NIR region (band 5 of Landsat-8), whereas silicates were suc- tures in 1.550–1.750 μm (the equivalent of the Landsat-8 band 6) and
cessfully mapped in the TIR band 10 (Hunt and Ashley, 1979) (Fig. 10). absorption features in 2.10–2.400 μm (the equivalent of the Landsat-8
band 7), in addition to a significant variation in the TIR region
(8.50–11.70 µm) due to the presence of silicate minerals.
4.3. Independent component analysis and continuum removal
Pour et al., (2017) applied ICA and Continuum Removal (CR) 4.4. Band ratios, independent component analysis and the mixture tuned
methods to map the exposed lithological units and associated alteration matched filtering
mineral assemblages in the Antarctic Peninsula region. CR is a function
that isolates characteristic absorption areas in a given spectrum. Basi- Adiri et al. (2018a) evaluated the potentiality of the Landsat-8 and
cally, the continuum corresponds to the background signal related to all Sentinel-2A data in mineralogical mapping at Sidi Flah-Bouskour inlier,
the other components of the image. This signal is removed while Morocco. This was performed through using Band Ratio (BR), ICA and
maintaining just the targeted absorptions (Kruse and Lefkoff, 1999; the Mixture Tuned Matched Filtering (MTMF) techniques. BR is a Band
ENVI Tutorial, 2013). Spectrally, the absorption band feature depends math method which is based on dividing, at least, one band by another
on the chemistry of the mineral content, while the intensity is related to in order to highlight spectral information that cannot be distinguished
the amount of the abundance in the target (Hunt and Ashley, 1979). in the raw data (Rowan and Mars, 2003; Zhang et al., 2007; Gabr et al.,
This is quantified by the depth of diagnostic absorption feature in the 2010; Pour and Hashim, 2011; van Der Meer et al., 2012; Amer et al.,
associated spectrum (Kruse et al., 1993). Accordingly, the CR function 2012). This operation exploits the difference between well-defined re-
is very efficient in highlighting the presence of mineral absorptions flectance and absorption areas in the spectral signature, in order to
(Stevens and Ramirez-Lopez, 2014; Magendran and Sanjeevi, 2014). highlight the targeted object (Safari et al., 2017). Moreover, BR pre-
In the results, a pseudo-color ramp image derived from the CR bands sents the advantage of reducing or eliminating the potential impact of
5, 7 and 10 of Landsat-8 was used. The choice of these bands was ex- topography, through normalization of the abnormal reflectance values
plained by that band 5 is sensitive to the presence of ferric oxides such (recorded in high topography regions) to the other normal values in the
as hematite, goethite, amphiboles, pyroxenes, and phylosilicates. For image (Langford, 2015; Safari et al., 2017; Girija and Mayappan, 2019).
band 7, it allows the detection of Al-OH, Fe, Mg-OH and CO3 BR method is very powerful in distinguishing absorptions related to
Fig. 9. Illustration of the Independent

Component Analysis (ICA) method prin-
ciple.
9
Fig. 10. Examples of Independent Component Analysis (ICA) results from Landsat-8 and Sentinel-2A imagery of the Bouskour Patte d’Oie deposit (adapted from Adiri
et al., 2018a).
iron, Al-OH, Mg-OH and CO3 groups (Rowan and Mars, 2003, Pour and diagnostic absorption in bands 5 and 8 (NIR region) of these sensors,
Hashim, 2011). which are typical to the presence of iron. In turn, the abundance of
Concerning the Mixture Tuned Matched Filtering (MTMF), it con- chlorite was relatively mapped via the ratio 7/5 of Landsat-8. This
sists of a sub-pixel classification that highlights the targeted abundance mineral was also discriminated in the IC2 of Landsat-8 and IC5 of
while eliminating the background (Boardman, 1998). This operation Sentinel-2A. In addition, the malachite was distinguished in IC5
compares the spectrum of interest and those contained in the image, (Landsat-8) and IC2 (Sentinel-2A). These minerals were also mapped
then every pixel value is checked in order to avoid every potential er- with the MTMF classification (Fig. 12). The assessment of these results
roneous result from the first process part (Agar and Coulter, 2007; with respect to the ground truth indicated that Landsat-8 and Sentinel-
Shekar, 2008; Mitchell and Glenn, 2009; Pour and Hashim, 2012; Zadeh 2A were able to map the studied alterations in the Patte d’Oie area
et al., 2014). Firstly, MTMF performs a Matched Filtering (MF) process (Adiri et al., 2018a).
to estimates the abundance of the targeted material. Then, a physical
control (also called infeasibility) of the obtained values is done for as- 4.5. Crosta technique
sessing the MF result, in order to reduce the number of the false positive
values that are sometimes found when using the MF only (Fig. 11). Hu et al. (2018) applied the Crosta technique to the Sentinel-2A
Consequently, the obtained MF values differ from 0 to 1 (a value close data for mapping hydrothermal altered minerals in the Duolong area,
to 1 means a better correspondence with the targeted material) (Agar Tibetan Plateau, China. It is a selective principal component method
and Coulter, 2007, Shekar 2008, Pour and Hashim, 2012, Zadeh et al., adapted to the Landsat TM spectral bands, which allows the identifi-
2014). cation of the suitable PCs that contain the targeted spectral information
In the results, the ratios 4/5 of Landsat-8 and 4/8 of Sentinel-2A (Liu et al., 2013). Based on the sign and magnitude of the eigenvectors
allowed discriminating the malachite. The associated spectra showed a values, this technique indicates whether this object is highlighted with
dark or bright pixels (Aydal et al., 2007). Basically, an hydroxyl-bearing
minerals image can be obtained from PC3 or PC4 of 1-4-5-7 bands
(Landsat TM) association, whereas an iron oxides image can be derived
from PC4 of 1-3-4-5 (Landsat TM) association (Ranjbar et al., 2004;
Aydal et al., 2007; van der Meer et al., 2012; Abedi et al. 2013;
Sojdehee et al. 2016; Girija and Mayappan, 2019). Similarly, combi-
nations of the Sentinel-2A bands (2, 8a, 11, 12) with PC3 and bands (2,
4, 8a, 11) with PC4 were successfully used for mapping hydroxyls and
iron minerals respectively in the study area.
4.6. Classification and regression trees
de Lucia Lobo et al. (2018) exploited the Sentinel-2A data for

mapping mining areas in the Brazilian Amazon. Their approach was
based on performing a Classification and Regression Trees (CART)
(Breiman et al., 1984, Breiman, 2001b). This latter is a supervised
classification that consists of the identification and construction of a
binary decision tree using training sample data, in order to achieve a
correct classification (Bittencourt and Clarke, 2003; Thamilselvan and
Sathiaseelan, 2015; Jiang et al., 2010). In other words, CART is a pixel-
Fig. 11. Schematic illustration of the Mixture Tuned Matched Filtering (MTMF) based classification type that exploits the digital number of the selected
approach. training samples for performing a decision tree that classifies each pixel
10
Fig. 12. Examples of Mixture Tuned Matched Filtering (MTMF) results from Landsat-8 and Sentinel-2A imagery of the Bouskour Patte d’Oie deposit (adapted from
Adiri et al., 2018a).
of the image (Bittencourt and Clarke, 2003; de Lucia Lobo et al., 2018). boundary between sedimentary and igneous rocks. In turn, MTMF
Using the Sentinel-2A 3, 4, 6, 8 and 11 bands, the obtained maps gave technique was applied to map the abundances of clay, carbonate, and
satisfactory results while allowing delineating gold, tin, iron, copper, silicate minerals in the study area. The first two mineral types indicated
manganese and nickel mining areas. absorption features at 2.1–2.4 µm (band 7 of the Landsat-8) and re-
flectance at 1.55–1.75 µm (band 6 of the Landsat-8), whereas the sili-
4.7. Independent component analysis and mixture tuned matched filtering cate minerals showed higher emissivity values at 10.30–11.70 µm
(which corresponds to band 10 of Landsat-8). For the Patte d’Oie area,
Adiri et al. (2018b) fused the Sentinel-2A and Landsat-8 OLI data the MNF result extracted from the Landsat-8 data highlights the pre-
with those of the hyperspectral EO-1 Hyperion sensor for mapping sence of yellow and green colors in the three veins, which refer to the
copper mineralization, at the Moroccan Anti Atlas. This task was con- main mineralization types in this area. On the contrary, the Sentinel-2A
ducted in performing the color normalized (CN) spectral sharpening as result shows one color in the three veins, which can be used just for
a fusion method, with the MTMF and ICA as mapping techniques. The presenting the shape of the deposit (Fig. 13). Table 3 summarizes some
results analysis showed that the fused images (Hyperion/Landsat-8 OLI examples of discussed mapping methods with Landsat-8 and Sentinel-
and Hyperion/Sentinel-2A) were able to map the targeted miner- 2A for highlighting mineral features.
alogical alterations better than the Hyperion image alone, especially In other studies, Mia et al. (2019) exploited the Landsat-8 data to
with the ICA method. This was explained by the fact that the fusion explore most active hydrothermal altered area of the Kirishima volcano
process provided new images which are characterized by enhanced complex, Japan. The used methodology was based on color composite,
spatial and spectral resolutions. BR, PCA and Crosta techniques. Abdelnasser et al. (2018) used Landsat-
8 data to delineate the distribution of gold-associated alteration zones
4.8. Band ratio, mixture tuned matched filtering and minimum noise in Atud gold deposit, Egypt. This was achieved through applying PCA
fraction and BR methods. Furthermore, Cardoso-Fernandes et al. (2018) eval-
uated the potential of the Sentinel-2 data in the detection of lithium
Pour and Hashim (2015) relied on BR, MTMF and Minimum Noise bearing pegmatites, at Fregeneda (Spain)-Almendra (Portugal) region.
Fraction (MNF) methods applied to the Landsat-8 data in order to detect This evaluation was done using RGB combinations, BR, PCA and su-
hydrothermal alteration minerals and rock units associated with por- pervised classification techniques. In turn, Soltaninejad et al. (2018)
phyry copper deposits in Sar Cheshmeh copper mining district, Iran. mapped evaporate minerals in Sirjane, Iran using the Landsat-8 data
MNF is a transformation that determines the inherent part (useful in- with the Spectral Angle Mapper (SAM) classification. Gahlan and
formation) of the image, while separating it from the noise (Green et al., Ghrefat (2018) applied PCA, MNF, and BR techniques to the Landsat-8
1988; Boardman and Kruse, 1994; Park, 2006; Azizi et al., 2010; Pour data in order to map gossan zones in the Eastern Arabian Shield, Saudi
and Hashim, 2011). It consists of two cascaded PCA calculations: the Arabia.
first one aims to calculate and rescale the noise in the data (until Zhang et al. (2016) evaluated the integration of ASTER (Advanced
achieving no band-to-band correlations), and then apply a second PCA Thermal Emission and Reflection Radiometer) and OLI for mineral
based on the values derived from the first calculation (Green et al., mapping in the area of Duolong porphyry Cu-Au deposit, China. The
1988; Boardman and Kruse, 1994). Consequently, the resulted MNF obtained results showed that the combination of ASTER
bands fall into two parts: the inherent part (with large eigenvalues) and (VNIR + SWIR) and Landsat-8 (VNIR) data provides more information
the noise-dominated part (with near eigenvalues) (Green et al., 1988; about the mineral content better than the raw data alone. Also, Pour
Girija and Mayappan, 2019). et al. (2018c) used Landsat-8 in the exploration of Zn–Pb Sulfide Mi-
Then, (MNF band-1, MNF band-2, MNF band-3), (4/2, 6/7, 5) and neralization in the Franklinian Basin, North Greenland. Also, Yousefi
(4/2, 6/7, 10) color composites were used for discriminating litholo- et al. (2018) exploited the Landsat-8 data to map gossans developed
gical units and hydrothermally altered rocks, as well as identifying the over porphyry type deposits in Kerman Magmatic Arc, Iran.
11
Fig. 13. Examples of Minimum Noise Fraction (MNF) results from Landsat-8 and Sentinel-2A imagery of the Patte d’Oie deposit area; obtained by calculating the
standard MNF transformation of Landsat-8 and Sentinel-2A imagery.
4.9. Other geological studies that the Sentinel-2 data can be useful for iron oxides detection, by in-
troducing a technique to model the shape of the associated absorption
4.9.1. Lithological and structural mapping methods feature. van der Meer et al. (2014a) proved that the Sentinel-2 data
Certainly, lithological and structural studies are very helpful in the products show good potential for geological applications. Several BRs
characterization and delineation of mineral deposits. Numerous re- have been proposed targeting ferric iron, ferrous iron, laterite, gossan,
search studies showed the efficiency of the Landsat-8 and Sentinel-2 ferrous silicate, and ferric oxides minerals. Moreover, Mwaniki et al.
data in these applications. Yang et al. (2018) exploited the Landsat-8 (2015) compared the Landsat-8 OLI and Landsat-7 ETM+ in lineament
OLI data for mapping lithological units in the Huoshaoyun ore field, mapping, at Kenya. After applying sobel edge detector, they found that
China. Through using BR, PCA and Spectral Matched Filtering (SFM) the best results were obtained from the Landsat-8 OLI. This conclusion
techniques, the obtained maps showed that Landsat-8 gave good re- was also achieved by Adiri et al. (2017) in comparison with the ASTER
sults. The same sensor was useful also for lithological mapping in the sensor. The authors applied an automatic approach based on the filter
Bas Drâa inlier, Morocco (Adiri et al., 2016). For this purpose, the au- and edge detection. Landsat-8 data were also exploited by Amara et al.
thors applied BR, PCA and Support Vector Machine (SVM) techniques. (2019) In order to map hydrothermal alteration and structural features
Also, Pour et al. (2018b) exploited the Landsat-8 data for regional in the Guelma basin, Algeria.
geology mapping in Northern Victoria Land, Antarctica, by using BR as
well as the Normalized Difference Snow Index (NDSI). In addition,
Mohy et al. (2018) performed a lithological mapping in the south- 4.9.2. Machine and deep learning classifications
eastern desert of Egypt based on the Landsat-8 data, while applying the In classification algorithms, the new technological developments in
PCA, ICA and MNF transformations. The same objective was achieved spectral/spatial resolutions of satellite data have created interesting
by Amusuk et al. (2016) in north-central Nigeria, by calculating BRs opportunities and solutions for geological mapping (Ge et al., 2018).
from the Landsat-8 data. For van der Werff and van der Meer (2016), For example, Machine Learning Classifications (MLCs) are powerful
they showed the usability of the Sentinel-2A data for geological and data-driven inference tools that provide an automated approach to re-
mineralogical remote sensing studies, through a BR analysis with the cognize patterns in large data. As a result, they are increasingly used in
Landsat-8 OLI and Landsat-5 TM. geological mapping studies by remote sensing (Cracknell and Reading,
In addition, van der Werff and van der Meer (2015) demonstrated 2014). Most of the published papers on geological mapping using MLCs
have focused on comparing this kind of classification such as Random
Table 3
Examples of mapping methods applied to the Landsat-8 and Sentinel-2A data for highlighting mineral features.
Mineral feature Mapping method Sensor
Clays/Carbonate Band math color composite Landsat-8 (bands 5 and 7)

Clays DSPCA Landsat-8 (2, 5, 6, and 7)
Hematite/Jarosite/Goethite/Limonite ICA Landsat-8 (4 and 5)
Snow-ice ICA Landsat-8 (5)
Silicates ICA Landsat-8 (10)
Ferric oxide CR color composite Landsat-8 (5)
Al-OH, Fe, Mg-OH and CO3 bearing minerals CR color composite Landsat-8 (7)
Malachite/Chlorite BR, MTMF, ICA Landsat-8 (5), Sentinel-2A (8)
Hydroxyls/Iron Crosta technique Sentinel-2A (2, 8A, 11, 12 with PC3 and 2, 4, 81, 11 with PC4)
Clays/Carbonate/Silicates MTMF Landsat-8 (7 and 10)
Environmental impact of mining areas CART Sentinel-2A (3, 4, 6, 8, 11)
12
Forest (RF) and/or SVM, with more traditional classifiers such as transformation which reduces and enhances the targeted information,
Maximum Likelihood (MLC) (Leverington and Moon, 2012; Yu et al., followed by selecting suitable bands for mapping the object (based on
2012; Cracknell and Reading, 2014; Harvey and Fotopoulos, 2016; Ge the extracted statistics from PCA calculation). These methods can be
et al., 2018; Xie et al., 2018). very powerful in some circumstances like mixture issues, presence of
In the work carried out by Ge et al. (2018), the comparison of Ar- vegetation, noise…
tificial Neural Network (ANN), k-Nearest Neighbor (k-NN), MLC, RF, Furthermore, this study showed that Landsat-8 and Sentinel-2 data
and SVM classifications was performed using Sentinel-2A data for dis- have returned satisfactory results in geological and mineral exploration.
criminating lithological units in the Shibanjing Ophiolite Complex, In addition to the VNIR and SWIR regions, Landsat-8 satellite is char-
China. It was found that the best results were obtained from MLC and acterized by the presence of thermal bands allowing the delineation of
SVM methods. Harvey and Fotopoulos (2016) used Landsat-5 and silicates and carbonates minerals. Also, we note the presence of band-8
geophysical data for discriminating lithological units in the Sudbury (panchromatic) which is useful for enhancing the spectral resolution of
region (Canada). After comparison between k-NN, RF, SVM, and Bayes the images. Landsat-8 data are characterized by a high radiometric
classifications, they found that RF is the best approach. The same re- resolution (16 bits), which allows the measurement of subtle variability
sults were obtained by Cracknell and Reading (2014) using Landsat-7 in surface conditions.
and airborne geophysical data in the Broken Hill area (Australia). For Sentinel-2, the data processing has the advantage of being
Xie et al. (2018) compared the above-cited classifications with the quicker and therefore more efficient and more cost-effective. This
Gradient Tree Boosting (GTB) method, while using seven geophysical sensor has shown its potential in mineral exploration, especially in
parameters from well log data over the Ordos Basin, China. Their mapping the iron absorption feature. This is related to its high spectral
findings indicated that the ensemble-based methods (namely RF and characteristics especially in the VNIR part. In addition, the Sentinel-2
GTB) are better than the other machine learning classifications. En- data are encoded in 12 bits with a high spatial resolution of 10 m. This
semble methods are learning algorithms that construct a set of classi- latter could make a big difference especially in the poorly outcropping
fiers and classify new datasets by taking a weighted vote of their pre- terrains. However, this review showed that most of the authors used
dictions (Xie et al., 2018). The performance of these classifications Landsat-8 data in their studies, which can be explained by the reputa-
depends on the amount of input data, the number of training samples tion of the Landsat satellites in geological and mineralogical mapping.
and how these latter are representatives. The training data have to be Therefore, the effectiveness of Sentinel-2 data needs to be demonstrated
balanced in order to achieve the best performance. According to the in more studies.
study area, a limited number of features may reduce the performance of Undoubtedly, remote sensing is a powerful exploration and geolo-
machine learning classifications, since they are mainly based on sta- gical mapping tool for arid and semi-arid regions, where bedrock ex-
tistical parameters extracted from the training data (Cracknell and posure is common and typically laterally extensive. However, this is not
Reading, 2014). necessarily the case temperate to humid climate zones where mineral
Today, Deep learning is increasingly attracting the interest of re- deposits are often concealed by vegetation and regolith (e.g., thick
mote sensing specialists as an approach to improve mapping applica- soils). These conditions make it difficult to reliably map hydrothermally
tions. Deep learning aims to retrieve geophysical or biochemical altered rocks by way of remote sensing tools. In such conditions, the
quantities rather than detecting or recognizing objects, which presents presence of ore deposits may be marked by areas that lack vegetation
some new challenges for remote sensing data complexity (multimodal, because the high metal concentrations are toxic to plants, or by ab-
multi-sources, prior knowledge concerning the sensors, big data issue) normal spectral characteristics in the vegetation pattern (Sabins, 1999).
(Zhu et al., 2017). For instance, Convolutional Neural Networks (CNNs) In areas of thick regolith cover, which is residual and consists of
approach has proven to be highly effective at detecting features from intensely weathered bedrock, the study of its properties becomes a
satellite images by interleaving convolutional and pooling layers. CNN critical step. From a geochemical viewpoint, the regolith can be studied
was already tested in several remote sensing applications but in few by identifying geochemical pathfinder elements that can be used to
geological studies. For instance, Latifovic et al., (2018) applied CNN vector towards mineralization. In geophysical analysis, the related
using aerial photos, Landsat reflectance, and high-resolution digital maps can be used to distinguish between the regolith responses and
elevation data over five areas within the South Rae geological region of those of the basement (Butt, 2004). In other contexts, the recorded
Northwest Territories, Canada. Their comparison revealed that CNN is spectral signature could sometimes be influenced by the topographical
better than machine-learning classifiers (namely RF). This performance effects. This problem is more obvious in the mountainous areas when
of Deep Learning classifications promotes their use in further geological the components of the same land cover have different irradiance and
and mineralogical studies, especially while exploiting the effectiveness thus different spectral behavior. Accordingly, the correction of this ef-
of Landsat-8 and Sentinel-2 data. fect is critical in order to avoid any interpretation errors in the used
satellite images (Proy et al., 1989; Chu and Yingcheng, 1992).
4.10. Discussion In summary, the reflectance spectra of a covered mineralized area
may differ from adjacent non-covered areas. These spectral differences
After analyzing the discussed studies, it turned out that some pro- are variable according to the type and content of the cover. The com-
cessing methods (e.g., Band Math, Minimum Noise Fraction) can be bination of multisource data such as remote sensing (in the presence of
used as a first mapping step (for example for the separation between cover, the use of hyperspectral data is recommended), geochemistry,
altered and unaltered rocks). In order to map specific alterations, we geophysics, aerial photography and field observations remains very
must use more developed methods that allow mapping more accurately important in mineral and geological mapping (Sabins, 1999).
the target. Mixture Tuned Matched Filtering (MTMF), Independent
Component Analysis (ICA), and Developed Selective Principal 5. Conclusion
Component Analysis (DSPCA) present some processing methods that
have given satisfactory results in mineral exploration task. This can be In conclusion, Landsat-8 and Sentinel-2 multispectral data have
related to the efficiency of these algorithms: MTMF highlights the been successfully used in mineral mapping of hydrothermal alteration
target, eliminates the background, and controls the resulted pixel in zones associated with a variety of mineral deposits types. Commonly
order to avoid every potential erroneous value. Concerning the ICA, it is processing methods include, for example, Band Ratios, Principal
a transformation that separates the available information even if it Component Analysis, Developed Selective Principal Component
occupies a small part of the image, which makes the result as in- Analysis, Independent Component Analysis, Minimum Noise Fraction,
dependent as possible. For the DSPCA, it is based on a PCA Mixture Tuned Matched Filtering, and Crosta. In some applications, the
13
spatial/spectral characteristics of the multispectral data are combined Hydrothermal alteration mapping and structural features in the Guelma basin
with other sensors, such as Hyperion or ASTER, using fusion and in- (Northeastern Algeria): contribution of Landsat-8 data. Arab. J. Geosci. 12, 94.
Amer, R., Kusky, T., El Mezayen, A., 2012. Remote sensing detection of gold related
tegration techniques. Ultimately, the combination of these multi-source alteration zones in um Rus area, central eastern desert of Egypt. Adv. Space Res. 49,
data provides a more thorough way of characterizing alteration mi- 121–134.
nerals, as neither multispectral nor hyperspectral data on their own Amusuk, D.J., Hashim, M., Pour, A.B., Musa, S.I., 2016. Utilization of Landsat-8 data for
lithological mapping of basement rocks of plateau state North Central Nigeria. The
constitute a thorough mineral exploration study. Either way, the best International Archives of the Photogrammetry, Remote Sensing and Spatial
method for testing the reliability and optimizing the obtained results is Information Sciences. Kuala Lumpur, Malaysia.
through integrated field and laboratory studies. Atmospheric correction, 2011. Available from < http://www.gisagmaps.com/atco-
guide/ > .
In geological remote sensing, spectral resolution influences the Aydal, D., Ardal, E., Dumanlilar, O., 2007. Application of the Crosta technique for al-
quality and amount of the extracted information. Landsat-8 and teration mapping of granitoidic rocks using ETM+ data: case study from eastern
Sentinel-2 are able to highlight clay, hydroxyl groups, and iron-bearing Tauride belt (SE Turkey). Int. J. Remote Sens. 28, 3895–3913.
Azizi, H., Tarverdi, M.A., Akbarpour, A., 2010. Extraction of hydrothermal alterations
minerals. Both sensors are characterized by good spatial resolution
from ASTER SWIR data from east Zanjan, northern Iran. Adv. Space Res. 46, 99–109.
(15 m for Landsat-8 and 10 m for Sentinel-2) and have the ability to Bedell, R., Crósta, A.P., Grunsky, E., 2009. Remote sensing and spectral geology. Rev.
map silicate and carbonate minerals (with the Landsat-8 TIRS). In Econ. Geol. 16.
summary, the freely available Landsat-8 and Sentinel-2 data have Bedini, E., 2011. Mineral mapping in the kap simpson complex, central eastgreenland,
using hymap and ASTER remote sensing data. Adv. Space Res. 47, 60–73.
clearly demonstrated their capabilities and potential with respect to Bittencourt, H.R., Clarke, R.T., 2003. Use of Classification and Regression Trees (CART) to
mineral exploration, and, thus, will ensure data continuity for the Classify Remotely-Sensed Digital Images. In: Proceedings in the IGARSS IEEE
geological remote sensing community beyond the Landsat-5, Landsat-7 International Geoscience and Remote Sensing Symposium. Toulouse, France.
Bedini, E., 2017. The use of hyperspectral remote sensing for mineral exploration: a re-
and ASTER missions. view. J. Hyperspectral Remote Sens. 7, 189–211.
Boardman, J.W., 1998. Leveraging the high dimensionality of AVIRIS data for improved
Declaration of Competing Interest sub-pixel target unmixing and rejection of false positives mixture tuned matched
filtering. In: AVIRIS Proceeding. JPL Publication, pp. 97–101.
Boardman, J.W., Kruse, F.A., 1994. Automated spectral analysis: a geological example
The authors declare that they have no known competing financial using AVIRIS data, North Grapevine Mountains, Nevada. In: ERIM (Ed.), Proc. 10th
interests or personal relationships that could have appeared to influ- Thematic Conference on Geological Remote Sensing, San Antonio.
Breiman, L., 2001b. In: Random Forests. Machine Learning. Springer Link, pp. 5–32.
ence the work reported in this paper. Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J., 1984. Classification and Regression
Trees. Wadsworth Inc.
Acknowledgements Butt, C.R.M., 2004. Understanding the regolith in tropical and sub-tropical terrains: the
key to exploration under cover. Predictive Mineral Discovery Under Cover: SEG
Conference & Exhibition, Perth, Western Australia, 74–78.
This review is based on peer-reviewed journal papers. We apologize Cardoso-Fernandes, J., Lima, A., Teodoro, A.C., 2018. Potential of Sentinel-2 data in the
to those authors whose papers have not been included. We thank the detection of lithium (Li)-bearing pegmatites: a study case. Proc. SPIE 10790, Earth
Faculty of Sciences and Techniques, Beni Mellal, Morocco. Thanks are Resources and Environmental Remote Sensing/GIS Applications IX, 107900T.
Chander, G., Markham, B.L., Helder, D.L., 2009. Summary of current radiometric cali-
due also to the anonymous reviewers for their valuable comments. We bration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors. Remote Sens.
also gratefully acknowledge the helpful comments and recommenda- Environ. 113, 893–903.
tions of Dr. Oliver Kreuzer, from the James Cook University Economic Chauvet, A., 2019. Structural control of mineral deposits. Theory and reality. Minerals
256.
Geology Research Centre, Douglas, Queensland, Australia. Chavez Jr., P.S., 1988. An improved dark-object subtraction technique for atmospheric
scattering correction of multispectral data. Remote Sens. Environ. 24, 459–479.
Appendix A. Supplementary data Chen, Y.W., Jain, L.C., 2014. Independent component analysis and its application to
classification of high-resolution remote sensing images. In: Subspace Methods for
Pattern Recognition 57 in Intelligent Environment, Studies in Computational
Supplementary data to this article can be found online at https:// Intelligence. Berlin: Springer-Verlag, 552.
doi.org/10.1016/j.oregeorev.2020.103332. Chu, L., Yingcheng, Y., 1992. ISPRS Archives, Volume XXIX Part B7, Technical
Commission VII: Interpretation of Photographic and Remote Sensing Data.,
Washington, D.C., USA. 182–188.
References Clark, R.N., Gregg, A.S., Andrea, A.G., Trude, V.V.K., Wendy, M.C., 1999. The U. S.
Geological Survey, Digital Spectral Library: Version 1: 0.2 to 3.0.
Cloutis, E.A., 2007. Review Article Hyperspectral geological remote sensing: evaluation of
Abdelnasser, A., Kumral, M., Zoheir, B., Karaman, M., Weihed, P., 2018. REE geochemical
analytical techniques. Int. J. Remote Sens. 17 (12), 2215–2242.
characteristics and satellite-based mapping of hydrothermal alteration in Atud gold
Corgne, S., Magagi, R., Yergeau, M., Sylla, D., 2010. An integrated approach to hydro-
deposit, Egypt. J. African Earth Sci. 145, 317–330.
geological lineament mapping of a semi-arid region of West Africa using Radarsat-1
Abedi, M., Gholami, A., Norouzi, G.-H., 2013. A stable downward continuation of air-
and GIS. Remote Sens. Environ. 114, 1863–1875.
borne magnetic data: a case study for mineral prospectivity mapping in Central Iran.
Cracknell, M.J., Reading, A.M., 2014. Geological mapping using remote sensing data: a
Comput. Geosci-UK 52, 269–280.
comparison of five machine learning algorithms, their response to variations in the
Abrams, M., 2000. The Advanced Spaceborne Thermal Emission and Reflection
spatial distribution of training data and the use of explicit spatial information.
Radiometer (ASTER): data products for the high spatial resolution imager on NASA‘s
Comput. Geosci. 63, 22–33.
Terra platform. Int. J. Rem. Sens. 21, 847–859.
Crosta, A., Moore, J., 1990. Enhancement of Landsat Thematic Mapper imagery for re-
Abrams, M., Tsu, H., Hulley, G., Iwao, K., Pieri, D., Cudahy, T., Kargel, J., 2015. The
sidual soil mapping in SW Minas Gerais State, Brazil- A prospecting case history in
Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) after
greenstone belt terrain. In Thematic conference on remote sensing for exploration
fifteen years: review of global products. Int. J. Appl. Earth Obs. Geoinf. 38, 292–301.
geology- methods, integration, solutions, 7 th, Calgary, Canada, 1173–1187.
Adiri, Z., El Harti, A., Jellouli, A., Maacha, L., Bachaoui, E., 2016. Lithological mapping
Crosta, A.P., Souza Filho, C.R., 2009. Mineral exploration with landsat thematic mapper
using Landsat 8 OLI and Terra ASTER multispectral data in the Bas Drâa inlier,
(TM)/enhanced thematic maper plus (ETM+): A review of the fundamentals, char-
Moroccan Anti Atlas. J. Appl. Remote Sens. 10 (1), 016005.
acteristics, data processing and case studies. Rev Econ. Geol. 16, 59–82.
Adiri, Z., El Harti, A., Jellouli, A., Lhissou, R., Maacha, L., Azmi, M., Zouhair, M.,
de Lucia Lobo, F., Souza-Filho, W.M., de Moraes Novo, E.M.L., Carlos, F.M., Barbosa,
Bachaoui, E., 2017. Comparison of Landsat-8, ASTER and Sentinel 1 satellite remote
C.C.F., 2018. Mapping Mining Areas in the Brazilian Amazon Using MSI/Sentinel-2
sensing data in automatic lineaments extraction: a case study of Sidi Flah-Bouskour
Imagery (2017). Remote Sens. 10, 1178.
inlier, Moroccan Anti Atlas. Adv. Space Res. 60, 2355–2367.
Di Tommaso, I., Rubinstein, N., 2007. Hydrothermal alteration mapping using ASTER
Adiri, Z., El Harti, A., Jellouli, A., Maacha, L., Azmi, M., Zouhair, M., Bachaoui, E., 2018b.
data in theInfiernillo porphyry deposit, Argentina. Ore Geol. Rev. 32, 275–290.
Mapping copper mineralization using EO-1 Hyperion data fusion with Landsat 8 OLI
Drusch, M., Del Bello, U., Carlier, S., Colin, O., Fernandez, V., Gascon, F., Hoersch, B.,
and Sentinel-2A in Moroccan Anti-Atlas. Geocarto. Int.
Isola, C., Laberinti, P., Martimort, P., Meygret, A., Spoto, F., Sy, O., Marchese, F.,
Adiri, Z., El Harti, A., Jellouli, A., Maacha, L., Azmi, M., Zouhair, M., Bachaoui, E., 2018a.
Bargellini, P., 2012. Sentinel-2: ESA's optical high-resolution mission for GMES op-
Mineralogical mapping using Landsat-8 OLI, Terra ASTER and Sentinel-2A multi-
erational services. Remote Sens. Environ. 120, 25–36.
spectral data in Sidi Flah-Bouskour inlier, Moroccan Anti-Atlas. J. Spat. Sci.
Du, Q., Kopriva, I., Szu, H., 2006. Independent-component analysis for hyperspectral
Agar, B., Coulter, D., 2007. Remote sensing for mineral exploration–a decade perspective
remote sensing imagery classification. Opt. Eng. 45 (1), 017008.
1997–2007. In: Milkereit, B. (Ed.), Proceedings of Exploration 07: Fifth Decennial
Dubois, J., 1999. Identification des Linéaments dans les images satellitaires par ajuste-
International Conference on Mineral Exploration, pp. 109–136.
ment et suivi des segments. Project application submitted to school top technology
Amara, B.N., Aissa, D.E., Maouche, S., Braham, M., Machane, D., Guessoum, N., 2019.
university of Quebec, Canada.
14
El Harti, A., Bannari, A., Bachaoui, E., Aarab, E., Girouard, G., El Ghmari, A., 2004. visualization and analysis of imaging spectrometer data. Remote Sens. Environ. 44,
Synergie entre la télédétection multispectrale et les données de terrain pour la con- 145–163.
ception d’un nouveau modèle géodynamique d’ouverture du bassin paléozoïque des Kruse, F.A., 2010. Mineral mapping using spectroscopy: from field measurements to
Jebilet centrales (Maroc). C. R. Geosc. 336, 1311–1320. airborne and satellite-based imaging spectrometry. in Proceedings, ASARS
ENVI Tutorial, 2013. Selected Hyperspectral Mapping Methods. Available from < http:// Symposium, February 23–25 2010, Boulder, Colorado (Invited Paper).
www.harrisgeospatial.com/portals/0/pdfs/envi/Mapping_Methods.pdf > . Lambin, E.F., 2001. Remote sensing and geographic information systems analysis. Int.
Fiori, S., 2003. Overview of independent component analysis technique with an appli- Encyclopedia Soc. Behav. Sci. 13150–13155.
cation to synthetic aperture radar _SAR_ imagery processing. Neural Networks 16, Landsat 8 (L8) Data Users Handbook, (2019). LSDS-1574 version 4.0.
453–467. Langford, R.L., 2015. Temporal merging of remote sensing data to enhance spectral re-
Funedda, A., Naitza, S., Buttau, C., Cocco, F., Dini, A., 2018. Structural controls of ore golith, lithological and alteration patterns for regional mineral exploration. Ore Geol.
mineralization in a polydeformed basement: field examples from the Variscan Baccu Rev. 68, 14–29.
Locci Shear Zone (SE Sardinia, Italy). Minerals 8 (10), 456. Latifovic, R., Pouliot, D., Campbell, J., 2018. Assessment of Convolution Neural Networks
Gabr, S., Ghulam, A., Kusky, T., 2010. Detecting areas of high-potential gold miner- for Surficial Geology Mapping in the South Rae Geological Region, Northwest
alization using ASTER data. Ore Geol. Rev. 38, 59–69. Territories, Canada. Remote Sens. 10 (2), 307.
Gad, S., Kusky, T., 2007. ASTER spectral ratioing for lithological mapping in the Leverington, D.W., Moon, W.M., 2012. Landsat-TM-based discrimination of lithological
Arabian–nubian shield, the Neoproterozoic Wadi Kid area, Sinai, Egypt. Gondwana units associated with the purtuniq ophiolite, Quebec, Canada. Remote Sens. 4,
Res. 11, 326–335. 1208–1231.
Gahlan, H., Ghrefat, H., 2018. Detection of Gossan zones in arid regions using landsat 8 Liu, L., Zhou, J., Jiang, D., Zhuang, D., Mansaray, L.R., Zhang, B., 2013. Targeting mineral
OLI data: implication for mineral exploration in the Eastern Arabian Shield, Saudi resources with remote sensing and field data in the Xiemisitai area, West Junggar,
Arabia. Nat. Resour. Res. 27, 109–124. Xinjiang, China. Remote Sens. 5, 3156–3171.
Ge, W., Cheng, Q., Tang, Y., Jing, H., Gao, C., 2018. Lithological classification using Livo, K.E., Kruse, F.A., Clark, R.N., Kokaly, R.F., Shanks, W.C., 2007. Hydrothermally
sentinel-2A data in the Shibanjing Ophiolite complex in Inner Mongolia, China. altered rock and hot-spring deposits at yellowstone national park—characterized
Remote Sens. 10, 638. using airborne visible-and infrared-spectroscopy data. Chapter O of Integrated
Girija, R.R., Mayappan, S., 2019. Mapping of mineral resources and lithological units: a Geoscience Studies in the Greater Yellowstone Area— Volcanic, Tectonic, and
review of remote sensing techniques. Int. J Image Data Fusion. Hydrothermal Processes in the Yellowstone Geoecosystem.
Goetz, A.F.H., Rowan, L.C., 1981. Geol. Remote Sens. Sci. 211, 781–791. Lowell, J.D., Guilbert, J.M., 1970. Lateral and vertical alteration-mineralization zoning in
Grebby, S., Cunningham, D., Tansey, K., Naden, J., 2014. The impact of vegetation on porphyry ore deposits. Econ. Geol. 65, 373–408.
lithological mapping using airborne multispectral data: a case study for the north Magendran, T., Sanjeevi, S., 2014. Hyperion image analysis and linear spectral unmixing
troodos region, cyprus. Remote Sens. 6, 0860–10887. to evaluatethe grades of iron ores in parts of Noamundi, Eastern India. Int. J. Appl.
Green, A.A., Berman, M., Switzer, P., Craig, M/D., 1988. A transformation for ordering Earth Obs. Geoinf. 26, 413–426.
multispectral data in terms of image quality with implications for noise removal. IEEE Mandanici, E., Bitelli, G., 2016. Preliminary comparison of Sentinel-2 and Landsat 8
T. Geosci. Remote Sens. 26, 65–74. imagery for a combined use. Remote Sens. 8 (12), 1014.
Gregory, A.F., 1979. Remote sensing in the search for metallic ores: a review of current Markham, B.L., Storey, J.C., Williams, D.L., Irons, J.R., 2004. Landsat sensor perfor-
practice and future potential. Geophysics and geochemistry in the search for metallic mance: History and current status. IEEE T. Geosci. Remote Sens. 42, 2691–2694.
ores; Geological Survey of Canada. Econ. Geol. Report 31, 511–526. Mars, J.C. 2014. Hydrothermal Alteration Maps of the Central and Southern Basin and
Groves, D.I., Goldfarb, R.J., Santosh, M., 2016. The conjunction of factors that lead to Range Province of the United States Compiled From Advanced Spaceborne Thermal
formation of giant gold provinces and deposits in non-arc settings. Geosci. Front. 7 Emission and Reflection Radiometer (ASTER) Data (ver 1.1, April 8, 2014: U.S.
(3), 303–314. Geological Survey Open-File Report 2013–1139, 6 p., 13 plates, scale 1:1,300,000).
Gupta, R.P., 2003. Remote Sensing Geology. Springer-Verlag, Berlin, Heidelberg, Mia, Md.B., Fujimitsu, Y., Nishijima, J., 2019. Exploration of hydrothermal alteration and
Germany. monitoring of thermal activity using multi-source satellite images: a case study of the
Gupta, R.P., Tiwari, R.K., Saini, V., Srivastava, N., 2013. A simplified approach for in- recently active Kirishima volcano complex on Kyushu Island, Japan. Geothermics 79,
terpreting principal component images. Adv. Remote Sens. 2, 111–119. 26–45.
Hamisi, J., MacKenzie, D., Pitcairn, L., Blakmore, H., Zack, T., Craw, D., 2017. The hy- Mitchell, J.J., Glenn, N.F., 2009. Subpixel abundance estimates in mixture-tuned matched
drothermal footprint around the Reefton orogenic gold deposit, New Zealand. J. Geol. filtering classifications of leafy spurge (Euphorbia esula L.). Int. J. Remote Sens. 23.
Geophys. 60 (2), 59–72. Mohy, H., Abou El-Magd, I., Basta, F., Amasha, A., 2018. Utilization of full-polarimetric
Harris, J.R., Rencz, A.N., Ballantyne, B., Sheridon, C., 1998. Mapping altered rocks using SAR data (RADARSAT-2), ASTER and Landsat 8 data in geological mapping of the
landsat tm and lithogeochemical data: sulphurets-brucejack lake district, British west Gebal Elba area, Halayeb district, South Eastern Desert, Egypt. J. Indian Society
Columbia, Canada. Photogramm. Eng. Remote Sens. 4, 309–322. Remote Sens.
Harvey, A., Fotopoulos, G., 2016. Geological mapping using machine learning algorithms. Morfitt, R., Barsi, J., Levy, R., Markham, B., Micijevic, E., Ong, L., Scaramuzza, P., van der
International Archives of the Photogrammetry, Remote Sensing 41, 423–430. Werff, K., 2015. Landsat-8 Operational Land Imager (OLI) Radiometric Performance
Hashim, M., Ahmad, S., Md Johari, M.A., Pour, A.B., 2013. Automatic lineament ex- On-Orbit. Remote Sens. 7 (2), 2208–2237.
traction in a heavily vegetated region using Landsat Enhanced Thematic Mapper Mwaniki, M.W., Moeller, M.S., Schellmann, G., 2015. A comparison of Landsat 8 (OLI)
(ETM+) imagery. Adv. Space Res. 51, 874–890. and Landsat 7 (ETM+) in mapping geology and visualising lineaments: A case study
Herrington, R., 2011. Geological features and genetic models of mineral deposits. SME of central region Kenya. The International Archives of the Photogrammetry, Remote
Mining Engineering HandbooK, Chapter 3.1, 83–104. Sensing and Spatial Information Sciences, Volume XL-7/W3, 36th International
Hu, B., Xu, Y., Wan, B., Wu, X., Yi, G., 2018. Hydrothermally altered mineral mapping Symposium on Remote Sensing of Environment, Berlin, Germany.
using synthetic application of Sentinel-2A MSI, ASTER and Hyperion data in the Park, S., 2006. Mapping and Spatial Characterization of Nonnative Grasses in the Big
Duolong area, Tibetan Plateau, China. Ore Geol. Rev. 101, 384–397. Island, Hawaii using Hyperspectral Imagery. ASPRS Annual Conference Reno,
Hunt, G.R., 1977. Spectral signatures of particulate minerals in the visible and near in- Nevada.
frared. Geophysics 42, 501–513. Pour, A.B., Hashim, M., 2011. Identification of hydrothermal alteration minerals for ex-
Hunt, G.R., Ashley, R.P., 1979. Spectra of altered rocks in the visible and near infrared. ploring of porphyry copper deposit using ASTER data, SE Iran. J. Asian Earth Sci. 42,
Econ. Geol. 74 (7), 1613–1629. 1309–1323.
Ibrahima, E., Barnabéa, P., Ramanaidoub, E., Pirarda, E., 2018. Mapping mineral Pour, A.B., Hashim, M., 2012. Identifying areas of high economic-potential copper mi-
chemistry of a lateritic outcrop in new Caledonia through generalized regression neralization using ASTER data in the urumieh–dokhtar volcanic belt, Iran. Adv. Space
using Sentinel-2 and field reflectance spectra. Int. J. Appl. Earth Obs. Geoinf. 73, Res. 49, 753–769.
653–665. Pour, A.B., Hashim, M., 2015. Hydrothermal alteration mapping from Landsat-8 data, Sar
Irons, J.R., Dwyer, J.L., Barsi, J.A., 2012. The next Landsat satellite: the Landsat data Cheshmeh copper mining district, south-eastern Islamic Republic of Iran. J. Taibah
continuity mission. Remote Sens. Environ. 122, 11–21. Univ. Sci. 9, 155–166.
Jiang, L., Wang, W., Yang, X., Xie, N., Cheng, Y., 2010. Classification methods of remote Pour, A.B., Hashim, M., Hong, J.K., Park, Y., 2017. Lithological and alteration mineral
sensing image based on decision tree technologies. In: 4th Conference on Computer mapping in poorly exposed lithologies using Landsat-8 and ASTER satellite data:
and Computing Technologies in Agriculture (CCTA), Nanchang, China, pp. 353–358. North-eastern Graham Land, Antarctic Peninsula. Ore Geol. Rev.
Kaufmann, H., Segl, K., Itzerott, S., Bach, H., Wagner, A., Hill, J., Heim, B., Oppermann, Pour, A.B., Park, T.-Y.S., Park, Y., Hong, J.K., Zoheir, B., Pradhan, B., Ayoobi, I., Hashim,
K., Heldens, W., Stein, E., Müller, A., van der Linden, S., Leitão, P.J., Rabe, A., M., 2018c. Application of multi-sensor satellite data for exploration of Zn–Pb Sulfide
Hostert, P., 2010. Hyperspectral algorithms report in the frame of EnMAP prepara- mineralization in the franklinian basin, North Greenland. Remote Sens. 10, 1186.
tion activities. Sci. Tech. Rep STR10/08. Pour, A.B., Park, Y., Park, T.-Y.S., Hong, J.K., Hashim, M., Woo, J., Ayoobi, I., 2018b.
Kawishwar, P., 2007. Atmospheric Correction Models for Retrievals of Calibrated Spectral Regional geology mapping using satellite-based remote sensing approach in Northern
Profiles from Hyperion EO-1 Data. Thesis submitted to the International Institute for Victoria Land, Antarctica. Polar Sci. 16, 23–46.
Geo-information Science and Earth Observation. Pour, A.B., Park, Y., Park, T.-Y.S., Hong, J.K., Hashim, M., Woo, J., Ayoobi, I., 2018a.
Kovalskyy, V., Roy, D.P., 2013. The global availability of Landsat 5 TM and Landsat 7 Evaluation of ICA and CEM algorithms with Landsat-8/ASTER data for geological
ETM+ land surface observations and implications for global 30 m Landsat data mapping in inaccessible regions. Geocarto Int.
product generation. Remote Sens. Environ. 130, 280–293. Proy, C., Tanré, D., Deschamps, P.Y., 1989. Evaluation of Topographic Effects in
Kruse, F.A., Lefkoff, A.B., 1999. Analysis of Spectral Data of Manmade Materials, Military Remotely Sensed Data. Remote Sens. Environ. 30, 21–32.
Targets, and Background Using an Expert System Based Approach. In Proceedings, Rajesh, H.M., 2004. Application of remote sensing and GIS in mineral resource mapping –
ISSSR’99, ISPRS, 339–350. an overview. J. Miner Petrol. Sci. 99, 83–103.
Kruse, F.A., Lefkoff, A.B., Boardman, J.W., Heidebrecht, K.B., Shapiro, A.T., Barloon, P.J., Ranjbar, H., Honarmand, M., Moezifar, Z., 2004. Application of the Crosta technique for
Goetz, A.F.H., 1993. The spectral image processing system (SIPS)-interactive porphyry copper alteration mapping, using ETM+ data in the southern part of the
15
Iranian volcanic sedimentary belt. J. Asian Earth Sci. 24, 237–243. van der Meer, F., Hecker, C., van Ruitenbeek, F., van der Werff, H., de Wijkerslooth, C.,
Robila, S., 2004. Independent Component Analysis. In: Advanced Image Processing Wechsler, C., 2014b. Geologic remote sensing for geothermal exploration: a review.
Techniques for Remotely Sensed Hyperspectral Data Book, Part II. Berlin: Springer, Int. J. Appl. Earth Obs. Geoinf. 33, 255–269.
109–132. van der Meer, F.D., van derWerff, H.M.A., van Ruitenbeek, F.J.A., 2014a. Potential of
Rowan, L.C., Mars, J.C., 2003. Lithologic mapping in the mountain pass, california area ESA's Sentinel-2 for geological applications. Remote Sens. Environ. 148, 124–133.
using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) van der Werff, H., van der Meer, F., 2015. Sentinel-2 for mapping iron absorption feature
data. Remote Sens. Environ. 84, 350–366. parameters. Remote Sens. 7, 12635–12653.
Rowan, L.C., Schmidt, R.G., Mars, J.C., 2006. Distribution of hydrothermally altered van der Werff, H., van der Meer, F., 2016. Sentinel-2A MSI and Landsat 8 OLI provide
rocks in the Reko Diq, Pakistan mineralized area based on spectral analysis of ASTER data continuity for geological remote sensing. Remote Sens. 8, 883.
data. Remote Sens. Environ. 104, 74–87. Villa, A.J.A., Chanussot, B.J., Jutten, C., 2011. Hyperspectral image classi_cation with
Roy, D.P., Wulder, M.A., Loveland, T.R., Woodcock, C.E., Allen, R.G., Anderson, M.C., independent component discriminant analysis. IEEE T. Geosci. Remote Sens. 49,
Helder, D., Irons, J.R., Johnson, D.M., Kennedy, R., Scambos, T.A., Schaaf, C.B., 4865–4876.
Schott, J.R., Sheng, Y., Vermote, E.F., Belward, A.S., Bindschadler, R., Cohen, W.B., Wulder, M.A., Loveland, T.R., Roy, D.P., et al., 2019. Current status of Landsat program,
Gao, F., Hipple, J.D., Hostert, P., Huntington, J., Justice, C.O., Kilic, A., Kovalskyy, science, and applications. Remote Sens. Environ. 225, 127–147.
V., Lee, Z.P., Lymburner, L., Masek, J.G., McCorkel, J., Shuai, Y., Trezza, R., Xie, Y., Zhu, C., Zhou, W., Li, Z., Liu, X., Tu, M., 2018. Evaluation of machine learning
Vogelmann, J., Wynne, R.H., Zhu, Z., 2014. Landsat-8: science and product vision for methods for formation lithology identification: a comparison of tuning processes and
terrestrial global change research. Remote Sens. Environ. 145, 154–172. model performances. J. Petrol. Sci. Eng. 160, 182–193.
Sabins, F.F., 1996. Remote Sensing: Principles and Interpretation. Worth Publishers, New Yang, M., Ren, G., Han, L., Yi, H., Gao, T., 2018. Detection of Pb–Zn mineralization zones
York. in west Kunlun using Landsat 8 and ASTER remote sensing data. J. Appl. Remote
Sabins, F.F., 1999. Remote sensing for mineral exploration. Ore Geol. Rev. 14, 157–183. Sens. 12(2), 026018.
Safari, M., Maghsoudi, A., Pour, A.B., 2017. Application of Landsat-8 and ASTER satellite Yousefi, S.J., Ranjbar, H., Alirezaei, S., Dargahi, S., Lentz, D.R., 2018. Comparison of
remote sensing data for porphyry copper exploration: a case study from Shahr-e- hydrothermal alteration patterns associated with porphyry Cu deposits hosted by
Babak, Kerman, south of Iran. Geocarto Int. granitoids and intermediate-mafic volcanic rocks, Kerman Magmatic Arc, Iran: ap-
Sentinel-2 User Handbook, 2015. European Spatial Agency (ESA), Issue 1, Rev 2. plication of geological, mineralogical and remote sensing data. J. African Earth Sci.
Reference GMESS1OP-EOPG-TN-13-0001. 142, 112–123.
Shekar, A. 2008. Mineral abundance mapping using hyperion dataset in parts of Udaipur Yu, L., Porwal, A., Holden, E.J., Dentith, M.C., 2012. Towards automatic lithological
(Rajasthan) and Keonjhar (Orissa), India. Thesis presented in the Indian institut of classification from remote sensing data using support vector machines. Comput.
Remote Sensing. Geosci. 45, 229–239.
Singh, A., Harrison, A., 1985. Standardized principal components. Int. J. Remote Sens. 6 Zadeh, M.H., Tangestani, M.H., Roldan, F.V., Yusta, I., 2014. Sub-pixel mineral mapping
(6), 883–896. of a porphyry copper belt using EO-1 hyperion data. Adv. Space Res. 53, 440–451.
Sojdehee, M., Rasa, I., Nezafati, N., Abedini, M.V., 2016. Application of spectral analysis Zhang, X., Chen, C.H., 2002. New independent component analysis method using high
to discriminate hydrothermal alteration zones at Daralu copper deposit, SE Iran. order statistics with application to remote sensing images. Opt. Eng. 41 (7),
Arabian J. Geosci. 9 (1), 41. 1717–1728.
Soltaninejad, A., Ranjbar, H., Honarmand, M., Dargahi, S., 2018. Evaporite mineral Zhang, Z., He, G., Wang, X., 2010. A practical DOS model-based atmospheric correction
mapping and determining their source rocks using remote sensing data in Sirjan algorithm. Int. J. Remote Sens. 31, 2837–2852.
playa, Kerman, Iran. Carbonate Eavaporite 33, 255–274. Zhang, X., Pazner, M., Duke, N., 2007. Lithologic and mineral information extraction for
Stevens, A., Ramirez–Lopez, L. 2014. An introduction to the prospectr package. gold exploration using ASTER data in the south chocolate mountains (California).
Szu, H., Buss, J., 2000. ICA neural net to refine remote sensing with multiple labels. Proc. ISPRS J. Photogramm. Remote Sens. 62, 271–282.
SPIE 4056, 32–49. Zhang, T., Yi, G., Li, H., Wang, Z., Tang, J., Zhong, K., Li, Y., Qin, W., Bie, X., 2016.
Thamilselvan, P., Sathiaseelan, J.G.R., 2015. A comparative study of data mining algo- Integrating data of ASTER and Landsat-8 OLI (AO) for hydrothermal alteration mi-
rithms for image classification. I.J. Educ. Manage. Eng. 2, 1–9. neral mapping in Duolong Porphyry Cu-Au Deposit, Tibetan Plateau, China. Remote
Tuduri, J., Chauvet, A., Barbanson, L., Labriki, M., Dubois, M., Trapy, P.H., et al., 2018. Sens. 8, 890.
Structural control, magmatic-hydrothermal evolution and formation of hornfels- Zhu, X.X., Tuia, D., Mou, L., Xia, G., Zhang, L., Xu, F., et al., 2017. Deep learning in
hosted, intrusion-related gold deposits: insight from the Thaghassa deposit in Eastern remote sensing: a comprehensive review and list of resources. IEEE Geosci. Remote
Anti-Atlas, Morocco. Ore Geol. Rev. 97, 171–198. Sens. Mag. 5, 8–36.
van Der Meer, F.D., van der Werff, H.M.A., van Ruitenbeek, F.J.A., Hecker, C.A., Bakker, Zoheir, B., Emam, A., 2012. Integrating geologic and satellite imagery data for high-
W.H., Noomen, M.F., van der Meijde, M., E. Carranza, E.J.M., de Smeth, J.B., Woldai, resolution mapping and gold exploration targets in the South Eastern Desert, Egypt. J.
T., 2012. Multi- and hyperspectral geologic remote sensing: A review. Int. J. Appl. African Earth Sci. 66–67, 22–34.
Earth Obs. Geoinf. 14, 112–128.
16

Adiri 2020

Uploaded by

Copyright:

Available Formats

Adiri 2020

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Adiri 2020

Uploaded by

Copyright:

Available Formats

Ore Geology Reviews 117 (2020) 103332

Contents lists available at ScienceDirect

Ore Geology Reviews

1. Introduction geology, geochemistry, geophysics and laboratory analysis has pro-

Landsat-5 TM Band 1 – Blue 0.45 – 0.52 30 7 bands 8

Band 4 – Red 0.64 – 0.67 30

Fig. 9. Illustration of the Independent

4.6. Classiﬁcation and regression trees

de Lucia Lobo et al. (2018) exploited the Sentinel-2A data for

Clays/Carbonate Band math color composite Landsat-8 (bands 5 and 7)

You might also like