Hyperspectral imaging (HSI) offers a powerful means of resolving mineralogical variability at scales directly relevant to both geotechnical stability and ore management in mining operations. In collaboration with the drone-based monitoring company RockBlast Design SpA, we investigated the use of HySpex VNIR–SWIR drone-based data to characterize mineral distributions in open-pit as well as across several stockpiles in a large-scale, active open-pit copper mine in Chile. The work explores the capacity of spectral methods to support both slope stability monitoring and stockpile quality control in this porphyry copper system, typical for many Chilean copper mines, demonstrating the versatility of HSI for applied mining challenges.

The study was motivated by the need to better understand clay mineral distributions, which play a key role in slope stability, and to explore the potential of hyperspectral imaging for quantifying variability in stockpiled materials. Two main complementary methodologies were applied: 1) the USGS material identification and characterization algorithm (MICA) enables pixel-based classification of clay and other spectrally active minerals, focusing on mineralogical confidence expressed through fit value maps; and 2) Minimum Wavelength Mapping (MWL) mapping subtle shifts in absorption feature positions around the 2200 nm OH band to be tracked as proxies for compositional variation within clays and micas.

For the open-pit dataset, spectral analysis revealed spatial patterns in muscovite chemistry and smectite distributions, with trends towards higher fit values and compositional variation in structurally critical zones. These mineralogical indicators provide valuable proxies for identifying planes of weakness and stratigraphic horizons relevant to geotechnical assessment.

For the stockpile datasets, analysis of three regions of interest demonstrated how HSI can highlight mineralogical heterogeneity across different stockpiles. Segment-based analysis aggregated per-pixel results into dominant mineral classes at 15 m intervals, improving interpretability and enabling comparison across stockpiles. Minerals such as montmorillonite, muscovite (low to medium Al content), and kaolinite mixtures were mapped with varying confidence, revealing subtle but meaningful spatial trends. These results point to a role for HSI in supporting stockpile quality management and blending strategies.

Taken together, the results highlight how hyperspectral imaging can bridge geotechnical monitoring and material management applications. The collected data and analysis demonstrate that integrating spectral methods into general mining workflows provides not only detailed mineralogical maps but also actionable insights for both slope stability and ore handling using already existing tools for spectral geology, thereby supporting service providers operating in the drone-space, such as RockBlast Design SpA.