Elsevier

Advances in Space Research

Volume 67, Issue 3, 1 February 2021, Pages 945-963
Advances in Space Research

Comparison of Landsat OLI, ASTER, and Sentinel 2A data in lithological mapping : A Case study of Rich area (Central High Atlas, Morocco)

https://doi.org/10.1016/j.asr.2020.10.037Get rights and content

Abstract

The eastern part of the Rich area consists of the massive Paleozoic and Meso-Cenozoic cover formations that present the geodynamic development of the study area, where is characterized by various carbonate facies of Jurassic age. The geographical characteristic of the study area leaves the zone difficult to map by conventional methods. The objective of this work focuses on the mapping of the constituent lithological units of the study area using multispectral data of Landsat OLI, ASTER, and Sentinel 2A MSI. The processing of these data is based on a precise methodology that distinguishs and highlights the limits of the different lithological units that have an approximate similarity of spectral signature. Three techniques were used to enhance the image including Principal Component Analysis (PCA), Minimum Noise Fraction (MNF), and Independent Component Analysis (ICA). Lithological mapping was performed using two types of supervised classification : Maximum likelihood classifier (MLC) and Support Vector Machine (SVM).

The results of processing data show the effectiveness of Sentinel 2A data in mapping of lithological units than the ASTER and Landsat OLI data. The classification evaluation of two methods of the Sentinel 2A MSI image showed that the SVM method give a better classification with an overall accuracy of 93,93% and a Kappa coefficient of 0.93, while the MLC method present an overall accuracy of 82,86% and a Kappa coefficient of 0.80. The results of mapping obtained show a good correlation with the geological map of the study area as well as the efficiency of remote sensing in identification of different lithological units in the Central High Atlas.

Introduction

Nowadays, remote sensing plays a very important role in lithological, and mineralogical mapping using multispectral and hyperspectral images (Laoufi et al., 2011, Pour et al., 2019). Lithological mapping employing spatial methods are applicable mainly in arid areas where vegetation is scarce (Othman and Gloaguen, 2014), which is made possible by highlighting the earth's surface and the detection of lithological units by satellite. The procedure of this technique is based on the identification of the Physico-chemical characteristics of the rock surface (Bachri et al., 2019).

The use of multispectral and hyperspectral data in the mapping of lithological units is based on the exploration of the characteristics of the spectral bands (Bentahar et al., 2020). Most of previous studies has used ASTER data because of the properties of its constituent bands (Gad and Kusky, 2007, Othman and Gloaguen, 2014, Pour and Hashim, 2012). Visible near infrared (VNIR) bands are used to show iron-oxide formations, the shortwave infrared (SWIR) bands are useful for the identification of carbonates, hydrates, and hydroxides while the TIR bands are used to show silicates (Gad and Kusky, 2007). Recently the researchers are taking advantage of Sentinel-2A MSI imagery in geological mapping because of its band properties (Ge et al., 2018).

The Landsat Operational Land Imager (OLI) data contain 11 spectral bands, from which 9 spectral bands with a spatial resolution of 30 m, panchromatic band with a spatial resolution of 15 m and two TIR bands with a spatial resolution of 100 m (Table 1). These data can be downloaded for free from the USGS website (https://glovis.usgs.gov/). Advanced Space Heat Emission Reflection Radiometer (ASTER) data consist of 14 bands, 3 VNIR bands have 15 m spatial resolution, 6 SWIR bands have 60 m spatial resolution and 5 TIR bands have 90 m spatial resolution (Table 1). The ASTER images are downloaded free of charge from the NASA website (LPDAAC).

Sentinel 2A MSI data consists of 13 spectral bands with a spatial resolution of 10 to 60 m. Four VNIR bands (B2, B3, B4, and B8) at 10 m, four red-edge bands (B5, B6, B7, and B8a) and two SWIR bands (B11 and B12) at 20 m, and three bands (aerosol, water vapor and cirrus SWIR respectively B1, B9, and B10) at 60 m spatial resolution (Table 1). The image recordings are downloaded free of charge from the USGS website (https://glovis.usgs.gov/).

Central High Atlas is constituted an objective of the different field research thanks to variety of lithologic types of the area, where it consists of the formations of Paleozoic massif which exposed in the Mougueur region (Ouanaimi et al., 2018) and the cover formations of the Meso-Cenozoic age which was controlled by the development of the intracontinental basin (Ait Addi and Chafiki, 2013, Babault et al., 2013, Teixell et al., 2017). The identification of these geological formations allows us to understand the geodynamics of the study area and the High Atlas in general.

This work focuses on mapping of the lithological units using Landsat OLI, ASTER, and Sentinel 2A MSI multispectral data in the eastern part of Rich High Atlas. The application of remote sensing data provides information from inaccessible areas to understand the development of geological formations in the Central High Atlas. This work consists of an update of the geological map of study area and identification of lithological units that are identified in the new works (Ouanaimi et al., 2018, Teixell et al., 2017). The processing of multispectral data is mainly based on image enhancement by using various techniques to highlight the lithological units. In this work, the identification of lithological units has been carried out using Principal Component Analysis (PCA), Independent Component Analysis (ICA), and Minimum Noise Fraction (MNF). This step allows the collection of lithological data based on the spectral signature of the rock surface. Then, the mapping of the lithological units based on the application of classifications algorithm techniques including supervised classification MLC and SVM.

Section snippets

Geological settings

The study section is located in the eastern part of the Rich area. It presents the boundary between the Central and Eastern High Atlas (Fig. 1). The Central High Atlas belongs to the great atlasic chain which is considered a typically intracontinental chain representing the foreland of the Mediterranean alpine system of the Rifaine chain (Babault et al., 2013). The lithological formations in the study area (Fig. 2) consist of Palaeozoic Massif masked by Mesozoic and Cenozoic cover (Benammi et

Methodology

The processing methodology of multispectral data used in this work is summarized in the flowchart (Fig. 3). After pre-processing the multispectral images, the image processing process is based on image enhancement using Principal Component Analysis (PCA), Independent Component Analysis (ICA), and Minimum Noise Fraction (MNF). Mapping of lithological units was performed using two techniques of supervised classification : Maximum Likelihood Classification (MLC) and Support Vector Machine SVM

Principal component analysis (PCA)

Principal Component Analysis PCA is a processing technique of multispectral data widely used in lithological and structural lineaments mapping (Bentahar et al., 2020). This technique is a multivariate statistical method. based on noise elimination and highlighting of targeted information to clarify the relationship between variables and the phenomenon (Adiri et al., 2017, Gasmi et al., 2016).

Minimum Noise Fraction transformation (MNF)

Minimum Noise Fraction Transformation (MNF) is similar to Principal Component Analysis (PCA). It is

Spectral characterization of lithological units

The spectral signatures extracted from the three types of multispectral images (Fig. 4) show the continental formations have maximum reflectance, while the Triassic basalts show low reflectance in the spectral signature diagram of the Landsat OLI and Sentinel 2A images. The pleozoic sandstone and pelite show low reflectance according to the ASTER diagram. The maximum reflectances are 0.68 µm, 0.65 µm, and 1.58 µm recorded for Landsat OLI, ASTER, and Sentinel 2A data, respectively. The minimum

Discussion

Lithological mapping using multispectral or hyperspectral images are very profitable for geologists. In this work the lithological mapping was done using multispectral Landsat OLI, ASTER, and Sentinel 2A data. The image processing is based on image enhancement to highlight all the lithological units that make up the region. The PCA shows that color composite RGB of the Landsat OLI (PC2, PC3, PC5) gives good results. allowing us to differentiate between the different lithological units: satin

Conclusion

The mapping of lithological units using satellite data allows us to visualize inaccessible areas by conventional methods. The present work focuses on the mapping of lithological units in the Rich region using three types multispectral data including Landsat OLI, ASTER, and Sentinel 2A. The processing of these data is mainly based on enhancing and increasing the visualization of the lithological units for this purpose we used PCA, MNF, and ICA. These three techniques allow us to differentiate

Funding

The research did not receive specific funding.

Declarations

Author's contribution statement:

Ibtissame Bentahar : treated the satellite images, Analyzed and interpreted the data, she is the main author of the article, she wrote the article.

Mohammed Raji : Analyzed and interpreted of the results, he followed the methodology applied in the image processing as well as supervised the writing of the article.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to thank reviewers for their very helpful comments and constructive reviews of this manuscript.

References (31)

Cited by (22)

  • Effective delineation of rare metal-bearing granites from remote sensing data using machine learning methods: A case study from the Umm Naggat Area, Central Eastern Desert, Egypt

    2022, Ore Geology Reviews
    Citation Excerpt :

    Consequently, the objective of ICA is to identify a group of uncorrelated elements that are as independent of one another as possible (uncorrelatedness is weaker than independence) (Yang and Cheng, 2015). Regarding several previous studies (Bentahar and Raji, 2021, Chen and Zhang, 1999, Kumar et al., 2020, Pour et al., 2019, Shebl et al., 2021b), the current study implements ICA for better lithological discrimination. The minimum noise fraction (MNF) approach refers to an orthogonal rotation that tends to result in components organized in increasing rank of random noise as opposed to decreasing rank of variance in principal components (Assiri, 2016, Green et al., 1988, Khan et al., 2007, Kruse et al., 2003).

  • Integrated Landsat multispectral imagery and airborne geophysical data for enhanced litho-structural mapping: A case study of Adamawa Massif, north-eastern Nigeria

    2022, Journal of African Earth Sciences
    Citation Excerpt :

    Interpretation of airborne aeromagnetic, gravity, radiometric, and remotely sensed Landsat data constrained by detailed field investigations have proven to be highly effective in unraveling the lithological and structural relationships in crystalline basement complex terrains (e.g., Martelet et al., 2006; Schetselaar et al., 2008; Metelka et al., 2011; Fagbohun et al., 2017; Aboelkhair et al., 2020; Okeyode et al., 2021). Spaceborne multispectral scanners on Landsat systems with 4–7 spectral channels such as Enhanced Thematic Mapper Plus (ETM+) are notable for their efficacy indiscriminating lithological, structural, and geomorphologic features in a rugged and inaccessible terrains for only a fraction of time at an inexpensive rate (Mandeng et al., 2018; Grochowski et al., 2019; Abdelkareem et al., 2021; Bentahar and Raji, 2021) owing to their synoptic coverage, spectral, and spatial mid-range resolutions (Sabins, 1999). Their robustness in lithological mapping is demonstrated in arid and semi-arid areas devoid of dense vegetation with clear rock and soil exposures (Grebby et al., 2011; Harris et al., 2011).

  • Stacked vector multi-source lithologic classification utilizing Machine Learning Algorithms: Data potentiality and dimensionality monitoring

    2021, Remote Sensing Applications: Society and Environment
    Citation Excerpt :

    The classifier performance may vary depending on the implemented datasets, thus a wide range of datasets are implemented for mapping rock units using several classifiers. Most of the results highlighted SVM (Bachri et al., 2019; Bentahar and Raji, 2021; Kumar et al., 2021; Manap and San, 2018), MLC (Fatima et al., 2013; Ge et al., 2018; Hadigheh and Ranjbar, 2013; Jellouli et al., 2016; Mehr et al., 2013), and ANN (Grebby et al., 2011; He et al., 2015). Thus, we decided to intensively investigate the performance of these classifiers over a wide range of the utilized data sets.

  • Contribution of remote sensing and GIS to mapping groundwater vulnerability in arid zone: Case from Amour Mountains- Algerian Saharan Atlas

    2021, Journal of African Earth Sciences
    Citation Excerpt :

    Furthermore, spatially explicit information is necessary, the scale of which must be sufficiently detailed to obtain reliable and accurate cartography. Remote sensing (RS) is an ideal approach for detailed mapping of terrain features and surface structure and, thank to the several platforms, such as Copernicus Land Monitoring Service or U.S. Geological Survey Landsat series of Earth Observation for accessing to free satellite imagery, this is an efficient tool for land monitoring (Zumsprekel and Prinz, 2000, Zoheir and Emam, 2012, Pournamdary et al., 2014, Gabr et al., 2015, Eldosouky et al., 2017, Bentaher and Raji, 2020). Multispectral data is able to provide information on lithology or rock composition based on spectral reflectance (Ridwan et al., 2018).

  • Lineament mapping in the Ikniouen area (Eastern Anti-Atlas, Morocco) using Landsat-8 Oli and SRTM data

    2021, Remote Sensing Applications: Society and Environment
    Citation Excerpt :

    In the central High Atlas, a major orientation of the lineaments ENE-WSW, WNW-ESE, NE-SW, and NW-SE has been pointed out on the basis of lineament extraction from Landsat ETM satellite image processing (El Moujahid et al., 2016), while (Nouayti et al. 2017),had demonstrated that the direction of the lineaments was NE-SW to E-W, on the flip side in the Rich area (Central High Atlas, Morocco). Researchers have also shown that the majority of the lineaments had NE-SW and E-W directions based on the image processing of Landsat-8 OLI, Sentinel-2A, ASTER, and ASTER-GDEM data,(Abdelouhed et al., 2021; Bentahar and Raji 2021),carried out a study following the same approach of extracting lineaments in the Central High Atlas, at the level of the Anmiter-Asfalou-Imini district; the authors found three major dominant directions NE-SW, E-W, and ENE-WSW. It is currently accepted that the High Atlas chain was built mainly by the post-Cretaceous inversion of ancient Mesozoic normal faults and decays, themselves inherited from the Paleozoic and Proterozoic basement.

View all citing articles on Scopus
View full text