Classification of steel materials by laser-induced breakdown spectroscopy coupled with support vector machines

Long Liang; Tianlong Zhang; Kang Wang; Hongsheng Tang; Xiaofeng Yang; Xiaoqin Zhu; Yixiang Duan; Hua Li

doi:10.1364/AO.53.000544

Applied Optics
Vol. 53,
Issue 4,
pp. 544-552
(2014)
•https://doi.org/10.1364/AO.53.000544

Classification of steel materials by laser-induced breakdown spectroscopy coupled with support vector machines

Long Liang, Tianlong Zhang, Kang Wang, Hongsheng Tang, Xiaofeng Yang, Xiaoqin Zhu, Yixiang Duan, and Hua Li

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Long Liang, Tianlong Zhang, Kang Wang, Hongsheng Tang, Xiaofeng Yang, Xiaoqin Zhu, Yixiang Duan, and Hua Li, "Classification of steel materials by laser-induced breakdown spectroscopy coupled with support vector machines," Appl. Opt. 53, 544-552 (2014)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Related Topics
Table of Contents Category
- Lasers and Laser Optics
Optics & Photonics Topics
?

The topics in this list come from the Optics and Photonics Topics applied to this article.

About this Article
History
- Original Manuscript: October 17, 2013
- Revised Manuscript: December 5, 2013
- Manuscript Accepted: December 9, 2013
- Published: January 23, 2014

Abstract

The feasibility of steel materials classification by support vector machines (SVMs), in combination with laser-induced breakdown spectroscopy (LIBS) technology, was investigated. Multi-classification methods based on SVM, the one-against-all and the one-against-one models, and a combination model, are applied to classify nine types of round steel. Due to the inhomogeneity of steel composition, the data obtained using the one-against-all and one-against-one models were ambiguous and difficult to discriminate; whereas, the combination model, was able to successfully distinguish most of the ambiguous data and control the computation cost within an acceptable range. The studies presented here demonstrate that LIBS–SVM is a useful technique for the identification and discrimination of steel materials, and would be very well-suited for process analysis in the steelmaking industry.

Full Article | PDF Article

More Like This

Classification of steel using laser-induced breakdown spectroscopy combined with deep belief network

Guanghui Chen, Qingdong Zeng, Wenxin Li, Xiangang Chen, Mengtian Yuan, Lin Liu, Honghua Ma, Boyun Wang, Yang Liu, Lianbo Guo, and Huaqing Yu
Opt. Express 30(6) 9428-9440 (2022)

Robust validation of pattern classification methods for laser-induced breakdown spectroscopy

Jeremiah Remus and Kehinde S. Dunsin
Appl. Opt. 51(7) B49-B56 (2012)

Laser-induced breakdown spectroscopy assisted chemometric methods for rice geographic origin classification

Ping Yang, Ran Zhou, Wen Zhang, Shisong Tang, Zhongqi Hao, Xiangyou Li, Yongfeng Lu, and Xiaoyan Zeng
Appl. Opt. 57(28) 8297-8302 (2018)

Previous Article Next Article

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (2)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (5)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (5)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Sample Type	Sample Size	C	S	Si	Mn	P	Cr	Ni	Mo	Ti	Cu
20#	$Φ 20 \times 6 mm$	0.180	0.033	0.209	0.581	0.021	0.016	0.010	—	—	0.012
20Cr	$Φ 20 \times 6 mm$	0.203	0.012	0.201	0.733	0.016	0.754	0.030	—	—	0.102
20CrMnTi	$Φ 30 \times 6 mm$	0.235	0.0065	0.306	0.920	0.014	1.12	0.015	—	0.063	0.017
20CrMo	$Φ 20 \times 6 mm$	0.210	0.0087	0.297	0.891	0.013	1.16	0.013	0.187	—	0.026
20CrNiMo	$Φ 20 \times 6 mm$	0.218	0.0050	0.272	0.793	0.0086	0.405	0.451	0.185	—	0.039
35#	$Φ 20 \times 6 mm$	0.343	0.0026	0.199	0.605	0.014	0.108	0.018	—	0.014	0.034
35CrMo	$Φ 20 \times 6 mm$	0.338	0.013	0.200	0.542	0.014	0.972	0.046	0.175	—	0.062
40Cr	$Φ 20 \times 6 mm$	0.425	0.011	0.277	0.622	0.014	0.939	0.018	—	—	0.016
42CrMo	$Φ 25 \times 6 mm$	0.438	0.0072	0.199	0.581	0.014	0.985	0.022	0.164	—	0.031

Sample Type	Average Rate of Correct Classification	Average Rate of Misclassification	Class (No.) of Misclassified Spectra	Average Rate of Nonclassification
20#	0.85 ( $17 / 20$ )	0.05 ( $1 / 20$ )	35#(1)	0.10 ( $2 / 20$ )
20Cr	0.85 ( $17 / 20$ )	0 ( $0 / 20$ )	——	0.15 ( $3 / 20$ )
20CrMnTi	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
20CrMo	0.55 ( $11 / 20$ )	0.25 ( $5 / 20$ )	35CrMo(5)	0.20 ( $4 / 20$ )
20CrNiMo	0.50 ( $10 / 20$ )	0 (0/20)	——	0. 50 ( $10 / 20$ )
35#	0.75 ( $15 / 20$ )	0.20 ( $4 / 20$ )	20#(4)	0.05 ( $1 / 20$ )
35CrMo	0.95 ( $19 / 20$ )	0 ( $0 / 20$ )	20CrMo(1)	0.05 ( $1 / 20$ )
40Cr	0.60 ( $12 / 20$ )	0 ( $0 / 20$ )	——	0.40 ( $8 / 20$ )
42CrMo	0.60 ( $12 / 20$ )	0.10 ( $2 / 20$ )	35CrMo(2)	0.30 ( $6 / 20$ )
Average	0.7389 ( $133 / 180$ )	0.0667 ( $12 / 180$ )		0.1944 ( $35 / 180$ )

Sample Type	Average Rate of Correct Classification	Average Rate of Misclassification	Class (No.) of Misclassified Spectra	Average Rate of Nonclassification
20#	0.85 ( $17 / 20$ )	0.15 ( $3 / 20$ )	35#(1)	0 ( $0 / 20$ )
20Cr	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
20CrMnTi	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
20CrMo	0.40 ( $8 / 20$ )	0.60 ( $12 / 20$ )	35CrMo (12)	0 ( $0 / 20$ )
20CrNiMo	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
35#	0.70 ( $14 / 20$ )	0.30 ( $6 / 20$ )	20#(4) 20CrNiMo (2)	0 ( $0 / 20$ )
35CrMo	1.00 ( $19 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
40Cr	0.55 ( $11 / 20$ )	0.45 ( $9 / 20$ )	20CrMo (9)	0 ( $0 / 20$ )
42CrMo	0.80 ( $16 / 20$ )	0.15 ( $3 / 20$ )	35CrMo (3)	0.05 ( $1 / 20$ )
Average	0.8111 ( $146 / 180$ )	0.1833 ( $33 / 180$ )		0.0056 ( $1 / 180$ )

Multi-Class Method	Training Time (s/360 spectra)	Testing Time (s/180 spectra)
One-against-all	13.22	4.31
One-against-one	11.76	119.78
Combined model	24.19	12.82

Sample Type	Average Rate of Correct Classification	Average Rate of Misclassification	Class (No.) of Misclassified Spectra	Average Rate of Nonclassification
20#	0.95 ( $19 / 20$ )	0.05 ( $1 / 20$ )	35#(1)	0 ( $0 / 20$ )
20Cr	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
20CrMnTi	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
20CrMo	0.75 ( $15 / 20$ )	0.25 ( $5 / 20$ )	35CrMo (5)	0 ( $0 / 20$ )
20CrNiMo	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
35#	0.80 ( $16 / 20$ )	0.20 ( $4 / 20$ )	20#(4)	0 ( $0 / 20$ )
35CrMo	0.95 ( $19 / 20$ )	0.05 ( $1 / 20$ )	20CrMo (1)	0 ( $0 / 20$ )
40Cr	1.00 ( $20 / 20$ )	0 ( $0 / 20$ )	——	0 ( $0 / 20$ )
42CrMo	0.90 ( $18 / 20$ )	0.10 ( $2 / 20$ )	35CrMo (2)	0 ( $0 / 20$ )
Average	0.9278 ( $167 / 180$ )	0.0722 ( $13 / 180$ )		0 ( $0 / 20$ )

Abstract

Cited By

Figures (2)

Tables (5)

Equations (5)

Applied Optics