Applying self-organizing maps to characterize hyperspectral drill core data from three ore prospects in Northern Finland

Hyperspectral (HS) data offers a fast, non-invasive and non-destructive way of obtaining information about the composition of rocks. In the Hyperspectral Lapland (HypeLAP) -project we are using HS imaging and point spectral data to map the mineralogical and mineral chemical variations of a total of 500 m of drill core data. These data, acquired in the visible-near infrared-short-wave infrared (VNIR-SWIR; 400-2,500 nm) and long-wave infrared (LWIR; 8,000-12,000 nm) wavelength regions, represent two gold prospects (Ruosselka and Hirvilavanmaa) and one Cu-Zn volcanogenic massive sulphide (VMS) prospect (Pahtavuoma) in northern Finland. In this paper, we present preliminary results from the Hirvilavanmaa study area, where we mapped the mineralogy of a single drill core box using the self-organizing maps (SOM) multivariate data analysis technology. In the process, point spectral data, Mineral Liberation Analyzer (MLA) data and X-Ray Fluorescence (XRF) data were used to guide the analysis of the HS image data. The results suggest that HS data has the potential to offer valuable information for mineral exploration activities in the Hirvilavanmaa study area. In the next steps, we will focus on analyzing LWIR wavelength region data from the three study areas to obtain a holistic view of their mineralogy and mineral chemistry. The HypeLAP-project is financed through the European Regional Development Fund (Sustainable growth and jobs 2014-2020 -programme).