Sorrentino, Anna (2025) Development of multiscale hyperspectral sensing for integrated remote and proximal characterization of mineralized districts. [Tesi di dottorato]

	Documento PDF Anna_Sorrentino_XXXVII_PON_COMPLETO.pdf Visibile a [TBR] Amministratori dell'archivio Download (83MB) | Richiedi una copia
	Documento PDF Anna_Sorrentino_XXXVII_PON_PARZIALE.pdf Visibile a [TBR] Amministratori dell'archivio Download (56MB) | Richiedi una copia

Abstract

Hyperspectral remote sensing has emerged as a critical tool in Earth Observation, particularly for mineral exploration. It enables the detection and quantitative mapping of surface exposed materials through high-resolution reflectance spectra in the Visible Near-Infrared (VNIR) and Shortwave Infrared (SWIR) ranges, between 400 to 2500 nm. The present PhD thesis aims to evaluate the effectiveness of multi-scale hyperspectral imaging, spanning from satellite to district and laboratory scales, to support mineral exploration and mining activities, bridging the gap between investigations conducted at varying spectral and spatial resolutions. Several deposit types worldwide were chosen and investigated, including porphyry-Cu and Iron Oxide-Copper-Gold deposits in Chile (e.g., Escondida, Rio Blanco-Los Bronces, and Marimaca), high-sulfidation epithermal deposits (e.g., the Allumiere quarry and Tolfa old mine in Italy), oxidized Zn mineralizations, both hypogene (e.g., Beltana, Aroona, and Raphook in Australia) and supergene (e.g., Skorpion and Rosh Pinah in Namibia) deposits, as well as orogenic gold orebodies in Mauritania (e.g., Tasiast and Tijirit). Advanced hyperspectral sensors were employed in this research, including the Italian Space Agency’s PRISMA (“PRecursore IperSpettrale della Missione Applicativa”) and German Aerospace Center’ EnMAP (“Environmental Mapping and Analysis Program”) satellites for regional-scale investigations, Headwall Photonics hyperspectral cameras mounted on tripods for outcrop imaging, and both imaging (Headwall and Hyspex cameras) and point (ASD FieldSpec-4) spectrometers for high-resolution sample-based analyses. Data processing across all the observation scale, involved applying multi-feature extraction workflow based on band ratios and minimum wavelength mapping, systematically utilized to extract diagnostic spectral features and determine the relative abundance and chemical composition of various mineral assemblages. Several alteration and ore-related minerals were identified and mapped, such as phyllosilicates (e.g., white mica, Al- and Fe-smectites, kaolinite, dickite, pyrophyllite, chlorite), sulfates (e.g., alunite, natroalunite, jarosite, gypsum), carbonates (i.e., calcite, dolomite and smithsonite), and Zn-bearing minerals (e.g., scholzite, tarbuttite, fraipontite and willemite). Notably, novel spectral signatures for Zn-clays (i.e., fraipontite) and red willemite were identified, expanding existing spectral libraries for Zn-bearing minerals. The results obtained highlight distinct advantages and limitations of hyperspectral imaging across the various observation scales. PRISMA and EnMAP satellites effectively mapped exposed alteration zones in mineralized areas. However, PRISMA encountered challenges in distinguishing mineral phases with partial overlapping spectral features (e.g., white mica vs. kaolinite) due to its lower signal-to-noise ratio (SNR >100:1 in SWIR-2) and slightly coarser spectral resolution (~15 nm). In contrast, EnMAP excelled in discriminating spectrally similar minerals and mapping alteration zones with high precision, owing to its superior SNR (≥250:1 in SWIR) and slightly finer spectral resolution (~10 nm in SWIR), but faced challenges detecting minerals with absorption features at longer wavelengths (>2400 nm) or present in lower concentrations. At the outcrop scale, tripod-mounted hyperspectral cameras enabled high-resolution mineral mapping, revealing spatial zonation and mineral associations with exceptional detail. However, challenges persisted in identifying quartz and amorphous silica due to overlapping spectral features in the SWIR range. Sample-based analyses provided the highest level of accuracy, allowing the creation of spectral libraries, identification of uncharacterized spectral minerals (e.g., Zn-silicates), and validation of broader datasets. The applied workflows, including band ratio analysis and minimum wavelength mapping, proved highly effective for spectral feature extraction. Although band ratios offered simplicity, they were limited in resolving small spectral shifts (<10 nm) in similar minerals. Conversely, minimum wavelength mapping enabled accurate characterization of spectral parameters such as absorption depth, position, and width, offering a comprehensive approach for mineralogical analysis. In conclusion, the integration of multi-scale spectral data demonstrated the reliability of hyperspectral imaging for mapping host rocks and alteration haloes, even in geologically complex and remote areas. This approach provides cost- and time-efficient exploration workflows, offering significant contributions to mineral exploration and the mining value chain by enhancing target delineation and facilitating decision-making processes.

Downloads

Actions (login required)

Modifica documento