Outcrop sensing for the exploration of REEs and lithium

Non-invasive technologies are key to ensure the sustainable exploration and mining of critical materials such as lithium and rare earth elements (REE), which are needed to fuel battery, magnet, and photovoltaic technologies for the transition towards low-carbon economies (Simandl, 2014; Heredia et al., 2020). Innovative sensing methods such as hyperspectral imaging allow the remote identification and high spatial resolution mapping of characteristic light-material interactions (“spectral fingerprints”) indicative of specific minerals and raw materials. These maps can be produced rapidly to support geological investigations during both exploration and mining. Accurate positioning of the spectral information in 3D space allows creating digital twins of outcrops and mines and integrating the spectral mapping results with other data such as from drill-cores or geochemical sampling. However, the low concentration and generally subtle spectral features associated with Li-bearing minerals and REEs present challenges for spectral mapping at the outcrop scale and, as a result, past approaches tended to use indirect methods based on, for example, mineral associations (Cardoso-Fernandes et al., 2020).

Direct mapping of REE and Li-minerals abundances is possible using high-resolution, high-quality spectral data and appropriate correction techniques. Sensor miniaturisation and development of new platforms now allow rapid data collection across large areas. These advances, combined with novel radiometric and geometric correction workflows (e.g., Thiele et al., 2021), recently allowed the direct detection of REE- and Li-bearing minerals at outcrop scale, using close-range ground-based (Boesche et al., 2015; Booysen et al., 2021) and drone-borne platforms (Booysen et al., 2020).

Two case studies of open pit mines, one from Siilinjärvi, Finland, (Fig. 1), the second from Uis, Namibia (Fig. 2) demonstrate direct 3D mapping of REE and Li mineralization. First, we captured ground-based photographs that were used for structure-from-motion, multi-view-stereo (SfM-MVS) photogrammetry to create digital 3D models/point clouds. We also acquired terrestrial hyperspectral data using a Specim AisaFENIX hyperspectral line scanner (push-broom scanner). The sensor covers the electromagnetic spectrum from 350 to 2500 nm over 384 spatial pixels and creates a datacube by steady rotation on a tripod. The resulting panoramic scan covers the important spectral features for both REE (578 nm, 740 nm and 799 nm for neodymium) and Li-bearing minerals (1593 nm and 1839 nm for cookeite, and 1538 nm, 1766 nm and 1850 nm for montebrasite). However, the results are subject to immense geometric distortion as well as radiometric influences that complicate spectral mapping, spatial localization, and scaling. We applied the processing workflow of Thiele et al. (2021) to perform accurate re-projection in original 3D space, as well as important radiometric and topographic corrections by combining the hyperspectral scans with the photogrammetric 3D point clouds. We subsequently mapped the occurrence of REEs and Li-bearing minerals based on their subtle characteristic spectral absorptions, and validated the occurrences by field sampling and mineralogical and elemental analysis.

Innovative remote sensing approaches allow us to rapidly create accurate digital twins of mine faces and natural geological outcrops. The 3D point clouds provide information on geological structure and can be populated with an arbitrary number of point-specific, high-dimensional spectral attributes to create so-called “hyperclouds”. Accurate processing and correction of the contained spectral information enables us to directly map the subtle spectral features indicative of Li- and REE-bearing minerals. Efficient remote mapping, even in complex terrains, is highly beneficial not only for initial targeting, but also for monitoring ore extraction, thus facilitating both exploration and optimized extraction.

We thank AfriTin and Yara Oy for support during field work, access to the mine site, and geological information and geochemical data.