Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 May 21;14(1):11544.
doi: 10.1038/s41598-024-62166-0.

Optimizing multi-spectral ore sorting incorporating wavelength selection utilizing neighborhood component analysis for effective arsenic mineral detection

Natsuo Okada  1 , Hiromasa Nozaki  2 , Shinichiro Nakamura  2 , Elsa Pansilvania Andre Manjate  3 , Angesom Gebretsadik  2   4 , Yoko Ohtomo  5 , Takahiko Arima  5 , Youhei Kawamura  5
Affiliations

Optimizing multi-spectral ore sorting incorporating wavelength selection utilizing neighborhood component analysis for effective arsenic mineral detection

Natsuo Okada et al. Sci Rep. .

Abstract

Arsenic contamination not only complicates mineral processing but also poses environmental and health risks. To address these challenges, this research investigates the feasibility of utilizing Hyperspectral imaging combined with machine learning techniques for the identification of arsenic-containing minerals in copper ore samples, with a focus on practical application in sorting and processing operations. Through experimentation with various copper sulfide ores, Neighborhood Component Analysis (NCA) was employed to select essential wavelength bands from Hyperspectral data, subsequently used as inputs for machine learning algorithms to identify arsenic concentrations. Results demonstrate that by selecting a subset of informative bands using NCA, accurate mineral identification can be achieved with a significantly reduced the size of dataset, enabling efficient processing and analysis. Comparison with other wavelength selection methods highlights the superiority of NCA in optimizing classification accuracy. Specifically, the identification accuracy showed 91.9% or more when utilizing 8 or more bands selected by NCA and was comparable to hyperspectral data analysis with 204 bands. The findings suggest potential for cost-effective implementation of multispectral cameras in mineral processing operations. Future research directions include refining machine learning algorithms, exploring broader applications across diverse ore types, and integrating hyperspectral imaging with emerging sensor technologies for enhanced mineral processing capabilities.

Keywords: Machine learning; Mineral processing; Neighborhood component analysis; Sensor-based ore sorting; System; Wavelength selection.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
A summary of a flowchart of the research.
Figure 2
Figure 2
An overview of the NCA algorithm, which combines the Leave-One-Out and neighborhood methods to indicate the importance of the input bands.
Figure 3
Figure 3
The sample is illuminated from two directions by a halogen light source, and the wavelength of the reflected light from the sample is acquired with a HS camera. The left figure shows the actual experimental scene, and the right figure shows the schematic diagram of the experiment.
Figure 4
Figure 4
Pre-processing method for HS data, where H is height, W is width, and N is the number of bands. The data were cut out along the band direction to have height h and width w.
Figure 5
Figure 5
Samples images of 14 arsenic samples diluted by quartz.
Figure 6
Figure 6
Distinctive protrusions appear at 700 nm and 800 nm; the spectral intensity tends to increase in the near-infrared region after 800 nm.
Figure 7
Figure 7
Shows each band weighted by NCA. The vertical axis shows feature weight and the horizontal axis shows the number of bands.
Figure 8
Figure 8
Machine learning results (Validation accuracy).
Figure 9
Figure 9
Machine learning results (Test accuracy).
Figure 10
Figure 10
Wavelength selection method comparison of classification accuracy (NCA, MRMR, Chi2, ReliefF, ANOVA, and Kruskal–Wallis).
Figure 11
Figure 11
Confusion Matrix of NCA results using 1 band.
Figure 12
Figure 12
Confusion Matrix of NCA results using 2 bands.
Figure 13
Figure 13
Confusion Matrix of NCA results using 3 bands.
Figure 14
Figure 14
Confusion Matrix of NCA results using 4 bands.
Figure 15
Figure 15
Confusion Matrix of NCA results using 5 bands.
Figure 16
Figure 16
Confusion Matrix of NCA results using 6 bands.
Figure 17
Figure 17
Confusion Matrix of NCA results using 7 bands.
Figure 18
Figure 18
Confusion Matrix of NCA results using 8 bands.
Figure 19
Figure 19
Confusion Matrix of NCA results using 9 bands.
Figure 20
Figure 20
Confusion Matrix of NCA results using 10 bands.
See this image and copyright information in PMC

References

    1. Elshkaki A, Graedel T, Ciacci L, Change BR. Copper demand, supply, and associated energy use to 2050. Glob. Environ. Chang. 2016;39:305–315. doi: 10.1016/j.gloenvcha.2016.06.006. - DOI
    1. Pell R, et al. Towards sustainable extraction of technology materials through integrated approaches. Nat. Rev. Earth Environ. 2021;2(10):665–679. doi: 10.1038/s43017-021-00211-6. - DOI
    1. Ferreccio C, Sancha AM. Arsenic exposure and its impact on health in Chile. J. Health Popul. Nutr. 2006;24:164–175. - PubMed
    1. Mohammed Abdul KS, Jayasinghe SS, Chandana EPS, Jayasumana C, De Silva PMCS. Arsenic and human health effects: A review. Environ. Toxicol. Pharmacol. 2015;40(3):828–846. doi: 10.1016/J.ETAP.2015.09.016. - DOI - PubMed
    1. Okada N, Maekawa Y, Owada N, Haga K, Shibayama A, Kawamura Y. Automated identification of mineral types and grain size using hyperspectral imaging and deep learning for mineral processing. Minerals. 2020 doi: 10.3390/min10090809. - DOI