Uso do sensoriamento remoto para identificação de ferro e carbonato em uma região do Domínio Médio Coreaú

Abstract

This study evaluates the potential of remote sensing for the identification of iron deposits and carbonate rocks in the Médio Coreaú Domain, Ceará, using multispectral images from the Sentinel-2 satellite, processed with QGIS 3.34 software in a Geographic Information System (GIS) environment. Digital image processing techniques were applied, including RGB composition, band ratio, normalized difference vegetation index (NDVI), principal component analysis (PCA), and supervised classification. The theoretical framework addressed concepts of remote sensing, electromagnetic radiation, and spectroradiometry, essential for analyzing the spectral reflectance behavior of materials. The spectral response of iron oxides (Fe³⁺ and Fe²⁺) and carbonates, characterized by absorption features in specific regions of the electromagnetic spectrum (VNIR and SWIR), was analyzed to differentiate mineralogical compositions. The results indicated that the Random Forest algorithm provided the best performance in distinguishing mineral classes. The study demonstrated that the integration of orbital data with geoprocessing techniques offers an efficient, accurate, and cost-effective approach for mineral prospecting, contributing to the advancement of exploration practices in the mineral sector of Ceará.