912 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
reduce the error in the measurements from a sample con-
taining 50% tail and 50% of concentrate. Three samples
sub samples were scanned by MinDet at P50 of 43.0 µm.
The average error for those three samples was ± 2.14 µm.
CONCLUSIONS
The development and building of an apparatus to analyse
mineral distribution and particle size in real-time has crucial
implications in mineral processing and separation process
plants. This apparatus integrates slurry sampling and separa-
tion, optical microscopy scanning, and a deep learning algo-
rithm to achieve real-time estimation of modal mineralogy.
The methods builds on existing principles of crystallogra-
phy and physics, such as goniometry and optical micros-
copy, but accelerated with machine learning algorithms.
To test the efficacy of this methodology, a validation
study was conducted on a polymetallic sulfide ore sample
which where the ground truth was measured by ICP-MS
and laboratory screening analysis. These are the gold stan-
dard methodologies in industry used to measure assay
and particle size distribution results. In the validation, the
MinDet apparatus could estimate the P50 and mineral-
ogy with good accuracy and found to be reliable for online
measurement of the poly-metal ore sample in this study.
There are several areas to explore and improve the design
of MinDet which includes expansion of the mineral library
and ruggedising the equipment itself.
Through real-time estimation of modal mineralogy of
process plant slurries, operators and process control algo-
rithms can be more proactive and optimise separation effi-
ciency by adjusting the relevant operation set points. For
example in flotation, these could be reagent selection and
dosage control, optimum grind size determination, airflow
and tank pulp level controls. The flexibility of the MinDet
apparatus to operate between batch and continuous mode
also opens the opportunity to be used for rapid assaying in
commercial labs with a minimum amount of sample prepa-
ration, increasing the number of tests possible.
REFERENCES
Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., &
Süsstrunk, S. (2012). SLIC superpixels compared to
state-of-the-art superpixel methods. IEEE transactions
on pattern analysis and machine intelligence, 34(11),
2274–2282. doi:10.1109/TPAMI.2012.120.
Al-Eshaikh, M.A., Kadachi, A.N., &Sarfraz, M.M. (2016).
Determination of uranium content in phosphate ores
using different measurement techniques. Journal of
King Saud University -Engineering Sciences, 28(1),
41–46. doi: 10.1016/j.jksues.2013.09.007.
Figure 9. Elemental assays calculated from MinDet reading of 100 images (five passes) versus ICP-MS results (Koh et al., 2024)
reduce the error in the measurements from a sample con-
taining 50% tail and 50% of concentrate. Three samples
sub samples were scanned by MinDet at P50 of 43.0 µm.
The average error for those three samples was ± 2.14 µm.
CONCLUSIONS
The development and building of an apparatus to analyse
mineral distribution and particle size in real-time has crucial
implications in mineral processing and separation process
plants. This apparatus integrates slurry sampling and separa-
tion, optical microscopy scanning, and a deep learning algo-
rithm to achieve real-time estimation of modal mineralogy.
The methods builds on existing principles of crystallogra-
phy and physics, such as goniometry and optical micros-
copy, but accelerated with machine learning algorithms.
To test the efficacy of this methodology, a validation
study was conducted on a polymetallic sulfide ore sample
which where the ground truth was measured by ICP-MS
and laboratory screening analysis. These are the gold stan-
dard methodologies in industry used to measure assay
and particle size distribution results. In the validation, the
MinDet apparatus could estimate the P50 and mineral-
ogy with good accuracy and found to be reliable for online
measurement of the poly-metal ore sample in this study.
There are several areas to explore and improve the design
of MinDet which includes expansion of the mineral library
and ruggedising the equipment itself.
Through real-time estimation of modal mineralogy of
process plant slurries, operators and process control algo-
rithms can be more proactive and optimise separation effi-
ciency by adjusting the relevant operation set points. For
example in flotation, these could be reagent selection and
dosage control, optimum grind size determination, airflow
and tank pulp level controls. The flexibility of the MinDet
apparatus to operate between batch and continuous mode
also opens the opportunity to be used for rapid assaying in
commercial labs with a minimum amount of sample prepa-
ration, increasing the number of tests possible.
REFERENCES
Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., &
Süsstrunk, S. (2012). SLIC superpixels compared to
state-of-the-art superpixel methods. IEEE transactions
on pattern analysis and machine intelligence, 34(11),
2274–2282. doi:10.1109/TPAMI.2012.120.
Al-Eshaikh, M.A., Kadachi, A.N., &Sarfraz, M.M. (2016).
Determination of uranium content in phosphate ores
using different measurement techniques. Journal of
King Saud University -Engineering Sciences, 28(1),
41–46. doi: 10.1016/j.jksues.2013.09.007.
Figure 9. Elemental assays calculated from MinDet reading of 100 images (five passes) versus ICP-MS results (Koh et al., 2024)