GRSG 35th Conference 2024 Abstract
Title: ESA Satellites and beyond: a multi sensor approach for lithium pegmatite exploration
Author: Olivia Rhind
Organisation: Viridien Satellite Mapping
This study showcases the improvement of traditional remote sensing methods for mineral exploration by employing advanced processing and interpretation techniques. ESA’s Sentinel-2 optical imagery is a critical sensor for remote sensing geological mapping, enabling 1:25,000 scale lithological and structural interpretation. Exploration workflows are further supported by ASTER Bare Earth+ mosaics, and where applicable, high-resolution commercial multispectral data and derivatives. In addition to multispectral datasets, recent advances in hyperspectral satellite technology and processing techniques have made such datasets more accessible for regional-scale analysis.
In order to meet the growing demand for critical minerals such as lithium, there is an increasing need to develop innovative techniques and methodologies for exploration. Lithium-bearing pegmatites are an attractive exploration target due to their often-high grade and suitability to conventional mining methods. This study presents a multi-sensor approach to target Li-bearing pegmatites in the Cape Cross-Uis belt and the Nainais-Kahero belt in northwest Namibia. The study also reviews this approach in the more challenging remote sensing conditions of the Pilbara Craton in Western Australia, where pegmatites are not exposed at surface and regolith mapping is necessary to identify potential spectral signatures of Li-bearing minerals. Multispectral and hyperspectral imagery is used to visualize lithological changes, map structural lineaments, and identify potential pegmatite occurrences, particularly for lithium-bearing mineralization.
The ASTER Bare Earth+ mosaic is derived from processing the entire archive of ASTER satellite imagery acquired over both study areas, resulting in an enhanced representation of the Earth’s surface with optimal signal to noise ratio and minimum vegetation, cloud cover and cloud shadow. This allows the most representative multispectral signatures to be derived via processing specifically tailored to the Li-bearing granitic pegmatite system. In the Cape Cross-Uis belt WorldView-3 high-resolution imagery, with a resolution of 3.7 m in the short-wave infrared (SWIR) and 1.2 m in the visible near infrared (VNIR) bands is processed to highlight the pegmatites that are indistinguishable in ASTER-derived outputs due to the lower spatial resolution of ASTER imagery. Hyperspectral EMIT, PRISMA, and EnMAP data were processed over both study areas to support lithological and surface cover discrimination, in addition to distinguishing the distribution and relative abundance of Li-bearing minerals such as petalite, spodumene and cookeite.
A semi-automated ML-driven lineament extraction technique was applied to all case study regions to support subsequent expert structural interpretation. The lineament extraction model is fine-tuned for the task of lineament extraction using tiles of RGB composites derived from the GLO-30 global elevation model. Expert structural interpretations, using Sentinel-2 and Copernicus 30m DEM, served as training labels. This enabled the contextualization of the relationships between granite and pegmatite occurrence, the structural framework, tectonic evolution, and the alteration of the associated igneous bodies. The relative prospectivity of the study areas and the effectiveness of this approach in both regions can then be assessed, and where successful, exploration targets can be identified and prioritized.