Interpretation of Hyperspectral Shortwave Infrared Core Scanning Data Using SEM-Based Automated Mineralogy: A Machine Learning Approach

Understanding the mineralogy and geochemistry of subsurface is key when assessing exploring for mineral deposits. To achieve this goal, rapid acquisition accurate interpretation drill core data are essential. Hyperspectral shortwave infrared imaging a non-destructive analytical method widely used in minerals industry to map with diagnostic features samples. In paper, we present an automated interpret hyperspectral on decipher major felsic rock-forming using supervised machine learning techniques processing, masking, extracting mineralogical textural information. This study utilizes co-registered training dataset that integrates quantitative scanning electron microscopy instead spectrum matching spectral library. Our methodology overcomes previous limitations full (i.e., quartz feldspar) caused by need identify minerals; particular, it detects presence considered invisible traditional analysis.