Discrimination of Rock Units in Karst Terrains Using Sentinel-2A Imagery

We explored the potential incorporation of Sentinel-2A imagery for rock unit determination in Croatian karst region dominated by carbonate rocks. The various lithological units are sources both stone aggregates and dimension stone, their spatial distribution is high importance mineral resource management. presented approach included preprocessing processing existing analog data (geological maps), satellite images United States Geological Survey spectral indices, all combination with ground truth data. mapping digital legacy maps using K-means random forest algorithm reduced error geometry geological boundaries from 100 m 300 to below m. possibility discriminating individual based on analysis discriminant function was also examined, providing a tool evaluating resources. Despite challenges posed homogeneity terrain, results this study show that use signature derived can be successfully implemented such terrains enhancement resources exploration.