A Robust Stochastic Approach to Mineral Hyperspectral Analysis for Geometallurgy
Published in Minerals, 2020
Recommended citation: Alvaro F. Egaña, Felipe A. Santibáñez-Leal, Christian Vidal, Gonzalo Díaz, Sergio Liberman, and Alejandro Ehrenfeld. (2020). "A Robust Stochastic Approach to Mineral Hyperspectral Analysis for Geometallurgy." Minerals, Special Issue: Advanced Spectral Techniques for Mineralogical and Elemental Analysis in Mining and Mineral Processing. Online first. https://www.mdpi.com/2075-163X/10/12/1139
Most mining companies have registered important amounts of drill core composites spectra, by means of different acquisition equipment and following diverse protocols.
These companies have used classic spectrography based on absorption features to perform semi quantitative mineralogy, a method that needs to have ideal laboratory conditions to have normalized spectra to compare.
Besides, the inherent variability of spectral features, due to environmental conditions and the geological context, changes need to be managed.
This work summarizes the research carried out for the characterization of geometallurgical samples, through the use of signal and image processing tools, with the use of hyperspectral data.
The development of hyperspectral image processing techniques, which considers ore samples spectral information as an inherently stochastic processes is formulated and validated. A new segmentation algorithm to analyze ore samples is described.
A novel approach is presented for spectral analysis that uses multi pixel hyperspectral images and statistical analysis techniques. It overcomes the effects of acquisition conditions and geological context variability to estimate critical geological and geometallurgical variables.
A hierarchical regression scheme is used to characterize correlation information among spectra, in order to capture statistically significant properties of different spectrum clusters for geological estimation.
A set of experiments were developed, considering white reference spectra characterization and geometallurgical estimation from a set of mineral samples to show the promising results of the current proposed approach