摘要： Hydrothermal alteration due to geothermal fluids often introduces mineral alteration and weathering that poses significant natural hazards around volcanoes. Hydrothermal alteration can be mapped remotely using satellite and airborne derive images. In this study, we explored the capacity of available multispectral satellites, high-resolution airborne hyperspectral and LiDAR imagery to provide an improved geological mapping and classification capability for volcanic terrains. Image classification experiments using a Random Forest approach trained using ground class data to classify 15 ground cover types show that Sentinel-2 and Landsat 8 OLI+TIR can provide a geological map with Overall (OA) and Kappa Accuracies (KA) of 69% and 66% respectively. Classification accuracy was dramatically improved when high-resolution airborne datasets were included. The use of full-spectrum AisaFENIX hyperspectral images improved accuracies to OA = 84% and KA = 82%. The maximum image classification accuracy is reached (OA = 87, KA = 85%) when all input features were combined.