Infrared and Laser Engineering, Volume. 54, Issue 6, 20250003(2025)
Origin identification of Astragalus membranaceus based on LIBS-NIR spectral information fusion
Figures & Tables(12)
Fig. 2. (a) LIBS experimental setup diagram; (b) NIR experimental setup diagram
Fig. 3. (a) LIBS classification confusion matrix; (b) NIR classification confusion matrix
Fig. 4. (a) Low-level fusion policy flow; (b) Mid-level fusion policy flow
Fig. 5. Classification confusion matrix for low-level fused spectra
Fig. 6. (a) LIBS characteristic variable diagram; (b) NIR characteristic variable diagram
Table 1. Classification results based on LIBS and NIR spectra
View tableView in ArticleTable 1. Classification results based on LIBS and NIR spectra
Spectra Model Train set Test set M-P M-R M-F1 ACC AUC M-P M-R M-F1 ACC AUC LIBS LG 92.70 92.57 92.49 94.57 98.17% 85.38% 84.00% 83.45% 84.00% 95.67% SVM 96.03% 96.00% 95.99% 96.00% 100.00% 80.87% 80.00% 79.32% 80.00% 96.02% LDA 92.37% 92.00% 92.06% 92.00% 98.03% 88.44% 88.00% 87.67% 88.00% 98.09% NIR LG 84.27% 84.57% 84.13% 84.57% 96.73% 79.65% 78.67% 78.43% 78.67% 94.13% SVM 89.80% 89.71% 89.68% 89.71% 98.93% 77.44% 77.33% 76.92% 77.33% 94.04% LDA 89.76% 89.71% 89.58% 89.71% 99.01% 79.28% 78.67% 78.57% 78.67% 96.29%
Table 2. Spectral classification results based on different characteristic variable screening methods
View tableView in ArticleTable 2. Spectral classification results based on different characteristic variable screening methods
Spectra Variable Number Train set Test set M-P M-R M-F1 ACC AUC M-P M-R M-F1 ACC AUC LIBS PCA 31 94.66% 94.29% 94.32% 94.29% 99.27% 88.21% 85.33% 84.99% 85.33% 98.04% SPA 23 90.00% 89.71% 89.71% 89.71% 98.60% 83.07% 78.67% 79.26% 78.67% 94.44% Selection 51 91.38% 91.43% 91.40% 91.43% 98.60% 89.21% 88.00% 88.04% 88.00% 98.84% NIR PCA 13 87.97% 88.00% 87.82% 88.00% 97.62% 76.02% 74.67% 73.97% 74.67% 93.56% SPA 19 84.38% 84.57% 84.44% 84.57% 96.90% 77.30% 76.00% 75.96% 76.00% 94.18%
Table 3. Optimal classification results of single spectrum and fusion spectrum
View tableView in ArticleTable 3. Optimal classification results of single spectrum and fusion spectrum
Spectra Variable Train set Test set M-P M-R M-F1 ACC AUC M-P M-R M-F1 ACC AUC NIR 509 89.76% 89.71% 89.58% 89.71% 99.01% 79.28% 78.67% 78.57% 78.67% 96.29% LIBS 24477 92.37% 92.00% 92.06% 92.00% 98.03% 88.44% 88.00% 87.67% 88.00% 98.09% Low-level 24986 93.88% 93.71% 93.70% 93.71% 99.07% 93.00% 92.00% 91.92% 92.00% 99.64% Mid-level 70 96.72% 96.57% 96.59% 96.57% 99.81% 96.31% 96.00% 95.99% 96.00% 99.98%
Tools
Get Citation
Copy Citation Text
Yuqiang LIU, Shengqun SHI, Mengsheng ZHANG, Yebin XU, Lianbo GUO. Origin identification of Astragalus membranaceus based on LIBS-NIR spectral information fusion[J]. Infrared and Laser Engineering, 2025, 54(6): 20250003
Paper Information
Category: Advanced interdisciplinary studies
Received: Jan. 3, 2025
Accepted: --
Published Online: Jul. 1, 2025
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20