1522 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
the geometallurgical questions. Additionally, there are 3
sources of spectra for including into a spectral library that
can be used for processing of hyperspectral data: 1) exem-
plar spectral library signature (i.e., USGS spectral library
(Kokaly 2017)), 2) site specific spectra (commonly referred
to as ground-truth), and lastly 3) in-scene spectra that are
based on pixel coordinates and/or geographic coordinates.
Figure 7 shows the complexity in processing of hyperspec-
tral data for geometallurgy and includes the spectral library
into a systematic hyperspectral data processing method-
ology. Figure 7 also incorporates mapping compositional
change in mineralogy (i.e., wavelength mapping of white
mica (Meyer, Kokaly et al. 2022)) and classification (gen-
eralized from McDowell and Kruse (2016)) into the auto-
mated spectral approach.
ANALYSIS AND INTERPRETATION
Analysis and interpretation of hyperspectral data processing
should be examined in the context of the geometallurgi-
cal objectives. Mapping minerals or increased mineralogi-
cal understanding are primary geometallurgical objectives
that affect mineral processing or metallurgical extraction.
If analysis of the results from processing the data didn’t
answer geometallurgical objectives, these data can be repro-
cessed with new or different algorithms. Separately, these
results can be readily into 3D block models that are based
on geographic information systems (GIS) or 3D modeling
packages (Pears and Chalke 2016).
INTEGRATION AND MODELING
Due to the 2D nature of the results from hyperspectral
imaging, these data can be integrated into GIS software
or spatial modeling software for integration and modeling
with other data. The specific methodology for integrated
interpretation doesn’t need to be determined at the out-
set of the collection of hyperspectral data with mineral
processing, metallurgical, geological, geochemical, and/or
geophysical data according to the geometallurgical objec-
tive. Moreover, since hyperspectral data is geophysical data,
integrated geometallurgical modeling of hyperspectral data
with geochemical, mineral chemical, mineral processing,
and metallurgical data follows a similar approach to the
integrated interpretation of geological and geophysical data.
This methodology is 1) Interpret your data, 2) Develop a
starting model and establish a geometallurgical framework
for modeling, 3) Implement quantitative modeling and
inversion, 4) Model validation and update geometallurgical
inversion techniques (Pears and Chalke 2016).
CONCLUSION—GEOMETALLURGICAL
OBJECTIVE
Imaging spectroscopy (aka hyperspectral imaging) can read-
ily help address geometallurgical questions in the mining
life-cycle. By understanding that materials (geological, min-
eral processing, metallurgical, waste) have spectroscopic
signatures that can be measured in the in-situ or in the lab-
oratory rapidly, imaging spectroscopy has the opportunity
Figure 6. Automated Spectral Processing (aka ENVI
Hourglass Method)(Boardman and Kruse 1994)
Figure 7. Generalized automated spectral processing
methodology
the geometallurgical questions. Additionally, there are 3
sources of spectra for including into a spectral library that
can be used for processing of hyperspectral data: 1) exem-
plar spectral library signature (i.e., USGS spectral library
(Kokaly 2017)), 2) site specific spectra (commonly referred
to as ground-truth), and lastly 3) in-scene spectra that are
based on pixel coordinates and/or geographic coordinates.
Figure 7 shows the complexity in processing of hyperspec-
tral data for geometallurgy and includes the spectral library
into a systematic hyperspectral data processing method-
ology. Figure 7 also incorporates mapping compositional
change in mineralogy (i.e., wavelength mapping of white
mica (Meyer, Kokaly et al. 2022)) and classification (gen-
eralized from McDowell and Kruse (2016)) into the auto-
mated spectral approach.
ANALYSIS AND INTERPRETATION
Analysis and interpretation of hyperspectral data processing
should be examined in the context of the geometallurgi-
cal objectives. Mapping minerals or increased mineralogi-
cal understanding are primary geometallurgical objectives
that affect mineral processing or metallurgical extraction.
If analysis of the results from processing the data didn’t
answer geometallurgical objectives, these data can be repro-
cessed with new or different algorithms. Separately, these
results can be readily into 3D block models that are based
on geographic information systems (GIS) or 3D modeling
packages (Pears and Chalke 2016).
INTEGRATION AND MODELING
Due to the 2D nature of the results from hyperspectral
imaging, these data can be integrated into GIS software
or spatial modeling software for integration and modeling
with other data. The specific methodology for integrated
interpretation doesn’t need to be determined at the out-
set of the collection of hyperspectral data with mineral
processing, metallurgical, geological, geochemical, and/or
geophysical data according to the geometallurgical objec-
tive. Moreover, since hyperspectral data is geophysical data,
integrated geometallurgical modeling of hyperspectral data
with geochemical, mineral chemical, mineral processing,
and metallurgical data follows a similar approach to the
integrated interpretation of geological and geophysical data.
This methodology is 1) Interpret your data, 2) Develop a
starting model and establish a geometallurgical framework
for modeling, 3) Implement quantitative modeling and
inversion, 4) Model validation and update geometallurgical
inversion techniques (Pears and Chalke 2016).
CONCLUSION—GEOMETALLURGICAL
OBJECTIVE
Imaging spectroscopy (aka hyperspectral imaging) can read-
ily help address geometallurgical questions in the mining
life-cycle. By understanding that materials (geological, min-
eral processing, metallurgical, waste) have spectroscopic
signatures that can be measured in the in-situ or in the lab-
oratory rapidly, imaging spectroscopy has the opportunity
Figure 6. Automated Spectral Processing (aka ENVI
Hourglass Method)(Boardman and Kruse 1994)
Figure 7. Generalized automated spectral processing
methodology