Extraction and analysis of alteration minerals from GF-5 hyperspectral data: a case study of the quartz-vein type tungsten deposit in Jiaoxi, Tibet

With advancements in geological exploration, hyperspectral remote sensing has become essential for mineral identification in complex geological settings.This study, using the Jiaoxi quartz-vein-type tungsten deposit as an example, addresses the limitations of traditional alteration extraction methods, which depend on manual operation, lack efficiency, and struggle with complex mineral distributions. We innovatively introduced domestic GF-5 hyperspectral remote sensing data and, for the first time, achieved automated extraction of alteration minerals in this mining area. The study employed the spectral angle mapper (SAM) technique and the support vector machine (SVM) algorithm to successfully extract six alteration minerals: paragonite, muscovite, phengite, Fe-Mg chlorite, Fe chlorite, and siderite. Integrating field spectra with geological and hyperspectral data enhanced extraction accuracy and efficiency. Comparative analysis confirms SVM’s superior precision and adaptability in complex geology. This study advances GF-5 hyperspectral applications in prospecting, supporting precise, automated mineral extraction with theoretical and practical value.