Proceedings Article | 19 October 2023
KEYWORDS: RGB color model, Minerals, Iron, Principal component analysis, Oxides, Raw materials, Cobalt, Image classification, Geology, Remote sensing
To achieve sustainable development goals, it is crucial to ensure the best practices through the whole mining cycle of Critical Raw Materials (CRM) used in high-tech, low-carbon goods (batteries, wind turbines, electronics). This is the aim of the Secure and Sustainable Supply of Raw Materials for EU Industry (S34I) Horizon Europe (HE) project that began on January 1st, 2023. S34I is investigating new data-driven methods to analyze Earth Observation (EO) data, supporting systematic mineral exploration (and other) activities to increase European autonomy regarding CRM resources. The S34I project is based on satellite data, airborne, Unmanned Aerial Vehicle (UAV), ground-based conventional in-situ techniques/methods, and fieldwork. This study presents the first preliminary results based on Sentinel-1 and Sentinel-2 data using classical image processing techniques (RGB combinations, band ratio, Principal Component Analysis – PCA, and unsupervised classification) for mineral exploration in Spain, either: (i) on land (Aramo area), to gain knowledge on cobalt (Co) deposits (and associate CRMs) by hydrothermal alteration mapping; and (ii) offshore/shallow waters (Ria de Vigo area), to update the knowledge on coastal metallic placers including CRMs such as Ti, Sn, Li, Rare Earth Elements (REEs) and Au. Among the techniques, PCA proved most effective, followed by band ratios, and self-proposed algorithms for detecting placer minerals showed promise, warranting further investigation. In the future, S34I will exploit Copernicus Contributing Missions (CCM) and other EU hyperspectral satellite sensors such as PRISMA and EnMAP. Different Artificial Intelligence (AI) techniques will, in the future, promote advances and innovative methods that mining stakeholders can use to address the challenges faced in different phases of the mining life cycle.
