Iniciar sessão

Navegar para Cima
Instituto de Investigação
em Vulcanologia e Avaliação de Riscos
Última hora:

Artigos em revistas ► internacionais com arbitragem


Referência Bibliográfica

PEREIRA, J., PEREIRA, A.J.S.C., GIL, A., MANTAS, V.M. (2022) - Lithology mapping with satellite images, fieldwork-based spectral data, and machine learning algorithms: the case study of Beiras Group (Central Portugal). CATENA, 220, 106653, doi: 10.1016/j.catena.2022.106653.​​​


​Geological remote sensing has been an invaluable means to obtain data to perform geological mapping objectively and with high accuracy. However, there is a significant gap in geological cartography information at 1:50 000 scale throughout the territory in mainland Portugal. The lack of geological mapping is reflected in the geological resources and land management information. This investigation’s main objective was to assess the viability of using remote sensing, machine learning and geochemical techniques as proof of concept for the Portuguese context using the Beiras Group (mainland Portugal) as a case study area of 341 km2. Multispectral analysis was carried out in two Landsat-8 images of 2015, one in the winter and the other in the summer. Hyperspectral data were obtained using a 400–1000 nm spectroradiometer applied to 23 rock samples collected in two field campaigns – the first in January and the second in April 2019. Spectral differences were found distinguishing the two main lithological units, where the granites (Granito de Coentral and Granito de Vila Nova) had an increasing wavelength spectra shape throughout the whole VNIR measurements. Geochemical data was carried out using X-ray Fluorescence, where the average quantity of major elements such as Na2O [2.15 %] and CaO [0.41 %] was higher in granites than metasediments: 0.38 % and 0.11 %, respectively. The J48 machine learning algorithm was performed using as input Landsat-8 reflectance data which showed a high success rate in both confusion matrixes (83,72 % and 94,08 %).