Mineralogical classification and concentration estimation in mining with app using hyper-spectral imaging and machine learning

Abstract

This study presents an innovative method for identifying minerals by combining the capabilities of hyperspectral imaging with machine learning. Although hyperspectral images are challenging to process due to their extensive dimensions and substantial size, our solution effectively tackles this complexity by providing a user-friendly machine learning tool specifically tailored for hyperspectral data. This self-developed tool simplifies the process of constructing datasets and enhances machine learning processes for identifying mineral species and estimating their concentrations. The interface is designed to be easy to use, allowing non-experts to effectively identify minerals without needing professional expertise. This is further enhanced by the integration of machine learning capabilities. Our instrument is positioned as an innovative solution that greatly enhances geological surveys in mining regions, leading to useful outcomes for mineral-related research and industrial applications.