Spectroscopy and Spectral Analysis, Volume. 42, Issue 7, 2255(2022)
Study on Nondestructive Identification of Panax Notoginseng Powder Quality Grade Based on Hyperspectral Imaging Technology
Figures & Tables(10)
Fig. 3. (a) Original spectra; (b) Spectra after SG pretreatment; (c) Spectra after MSC pretreatment; (d) Spectra after SNV pretreatment
Table 1. SVM modeling results of different pretreatment methods
View tableView in ArticleTable 1. SVM modeling results of different pretreatment methods
预处理方式 训练集准确率/% 测试集准确率/% SG 74.609 4 89.062 5 MSC 90.625 0 95.312 5 SNV 90.625 0 96.875 0
Table 2. SVM modeling results based on full spectral data and feature wavelengths
View tableView in ArticleTable 2. SVM modeling results based on full spectral data and feature wavelengths
特征选择方式 波长数 训练集准确率/% 测试集准确率/% 无 478 90.625 0 96.875 0 IRIV 30 84.375 0 92.187 5 VCPA 11 86.328 1 92.968 8 VCPA-IRIV 18 90.234 4 96.875 0
Table 3. GSA-SVM and GS-SVM modeling results based on VCPA-IRIV
View tableView in ArticleTable 3. GSA-SVM and GS-SVM modeling results based on VCPA-IRIV
模型 c g 训练集
分类准确
率/%测试集
分类准确
率/%VCPA-IRIV-GSA-SVM 15.868 0 1.048 0 100 100 VCPA-IRIV-GS-SVM 90.509 7 0.044 1 98.828 1 99.218 8
Tools
Get Citation
Copy Citation Text
Fu-jie ZHANG, Lei SHI, Li-xia LI, Hao-ran ZHAO, Yin-long ZHU. Study on Nondestructive Identification of Panax Notoginseng Powder Quality Grade Based on Hyperspectral Imaging Technology[J]. Spectroscopy and Spectral Analysis, 2022, 42(7): 2255
Paper Information
Category: Orginal Article
Received: Jun. 10, 2021
Accepted: --
Published Online: Nov. 16, 2022
The Author Email: ZHANG Fu-jie (20030031@kust.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20