Classification of Ores Using Laser-Induced Breakdown Spectroscopy Combined with PCA-PSO-SVM

Dapeng Wen; Xiyin Liang; Maogen Su; Fuchun Yang; Tianchen Zhang; Ruilin Chen; Meng Wu

doi:10.3788/LOP202158.2314006

Laser & Optoelectronics Progress, Volume. 58, Issue 23, 2314006(2021)

Classification of Ores Using Laser-Induced Breakdown Spectroscopy Combined with PCA-PSO-SVM

Dapeng Wen¹, Xiyin Liang^1、*, Maogen Su², Fuchun Yang², Tianchen Zhang¹, Ruilin Chen¹, and Meng Wu¹

¹Engineering Research Center of Gansu Province for Intelligent Information Technology and Application, College of Physics and Electronic Engineering, Northwest Normal University, Lanzhou, Gansu 730070, China

²Key Laboratory of Atomic and Molecular Physics & Functional Material of Gansu Province, College of Physics and Electronic Engineering, Northwest Normal University, Lanzhou, Gansu 730070, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(9)

Fig. 1. Flow of PSO-SVM classification model

Download full size

Fig. 2. LIBS spectra of 12 types of ore samples

Download full size

Fig. 3. LIBS spectra of U ore. (a) Original spectrum; (b) spectrum after preprocessing

Download full size

Fig. 4. PCA analysis results of LIBS data for 12 types of ores. (a) Each principal component score and the cumulative score of principal components; (b) three-dimensional scatter plot of the first three principal components

Download full size

Fig. 5. PCA-PSO-SVM recognition results on 12 types of ores. (a) PSO optimizing results; (b) PCA-PSO-SVM recognition results

Download full size

Fig. 6. Comparison of number of errors in ore classification by three classification models

Download full size

Table 1. Ore category information in the dataset

View table

Table 1. Ore category information in the dataset

Ore category label	Ore name	Ore category label	Ore name	Ore category label	Ore name
Ore 1	U ore	Ore 5	Au-Cu ore	Ore 9	Hematite ore
Ore 2	Zn-Pb-Ag sulfide ore	Ore 6	Mn ore	Ore 10	Anomalous ferruginous ore
Ore 3	Ni-Cu ore	Ore 7	Gold oxide ore	Ore 11	Silver copper gold ore
Ore 4	Sn ore	Ore 8	Zinc sulfide ore	Ore 12	Skarn tungsten magnetite ore

Table 2. PSO-SVM recognition results with different input variables
View table
Table 2. PSO-SVM recognition results with different input variables
Numbers of variables Cumulative interpretation rate /% Average accuracy rate /%
3 94.63 41.46
5 96.16 76.88
10 97.64 98.54
15 98.03 99.79
25 98.23 99.92
50 98.47 99.77

Table 3. Recognition results of three classification models
View table
Table 3. Recognition results of three classification models
Model Mean error Average classification time /s Average classification accuracy /%
PCA-LDA 7.00 1.60 98.54
PCA-PSO-BP 5.58 3.03 98.84
PCA-PSO-SVM 0.46 10.76 99.90

Tools

Get Citation

Copy Citation Text

Dapeng Wen, Xiyin Liang, Maogen Su, Fuchun Yang, Tianchen Zhang, Ruilin Chen, Meng Wu. Classification of Ores Using Laser-Induced Breakdown Spectroscopy Combined with PCA-PSO-SVM[J]. Laser & Optoelectronics Progress, 2021, 58(23): 2314006

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites