Spectroscopy and Spectral Analysis, Volume. 42, Issue 1, 176(2022)
Combination of Spectral and Textural Informations of Hyperspectral Imaging for Predictions of Soluble Protein and GSH Contents in Mutton
Figures & Tables(11)
Fig. 1. Statistical analysis of soluble protein and GSH contents in different parts
Fig. 3. Characteristic wavelengths selected by CARS algorithm
A: soluble protein content; B: GSH content
Fig. 4. Selection of characteristic wavelengths using iVISSA-IRIV algorithm for soluble protein concent
Fig. 5. Selection of characteristic wavelengths using iVISSA-IRIV algorithm for GSH content
Fig. 6. The first three principal component images of mutton samples
Fig. 7. Comparison of model results
(a): Comparison chart for soluble protein content in mutton;(b): Comparison chart for GSH content in mutton
Fig. 8. Visualizations of spatial distributions of soluble protein and GSH contents in mutton
Table 1. Statistical of soluble protein and GSH contents of mutton of sample sets
View tableView in ArticleTable 1. Statistical of soluble protein and GSH contents of mutton of sample sets
样本集 校正集 预测集 样本量 测量范围 平均值 标准差 样本量 测量范围 平均值 标准差 可溶性蛋白 128 19.96~110.87 70.55 18.29 43 38.81~93.55 71.77 15.83 GSH 128 17.34~172.75 46.43 24.51 43 17.81~169.95 44.08 26.38
Table 2. PLSR models for mutton indicators by different pretreatment methods
View tableView in ArticleTable 2. PLSR models for mutton indicators by different pretreatment methods
建模
对象预处理
方法LVs 校正集 预测集 Rc RMSEC Rp RMSEP 可溶性
蛋白Raw 17 0.875 7 8.830 0 0.854 7 9.827 3 SG 16 0.857 8 9.398 0 0.814 9 10.584 5 SNV 11 0.828 2 10.248 1 0.854 2 9.726 8 OSC 13 0.848 6 9.674 2 0.846 8 9.919 2 Detrend 10 0.830 1 10.195 9 0.881 4 9.074 5 GSH Raw 17 0.810 0 14.375 0 0.780 5 13.564 0 SG 16 0.751 7 16.168 4 0.732 2 18.155 8 SNV 17 0.804 8 14.049 6 0.826 5 12.355 8 OSC 11 0.862 1 12.423 1 0.702 0 20.074 0 Detrend 14 0.816 0 14.170 7 0.745 8 16.678 1
Table 3. Prediction results for soluble protein and GSH content in mutton using different characteristic wavelengths and models
View tableView in ArticleTable 3. Prediction results for soluble protein and GSH content in mutton using different characteristic wavelengths and models
建模对象 模型 特征波长数 特征波长 校正集 预测集 Rc RMSEC Rp RMSEP 可溶性蛋白 MLR 125 Raw 0.872 1 10.214 2 0.832 1 10.987 3 48 CARS 0.918 4 9.210 9 0.860 4 10.822 0 31 iVISSA-IRIV 0.904 2 9.020 5 0.881 6 9.454 6 LS-SVM 125 Raw 0.891 8 8.971 3 0.867 0 8.423 8 48 CARS 0.929 6 6.806 1 0.885 2 8.640 1 31 iVISSA-IRIV 0.914 6 7.422 5 0.881 8 8.527 2 GSH MLR 125 Raw 0.798 6 15.864 5 0.773 4 20.062 3 19 CARS 0.841 0 14.440 1 0.795 5 24.017 4 29 iVISSA-IRIV 0.844 6 14.998 5 0.870 5 15.517 7 LS-SVM 125 Raw 0.777 2 15.430 6 0.724 2 46.348 1 19 CARS 0.855 8 12.689 3 0.729 7 27.272 9 29 iVISSA-IRIV 0.936 4 9.102 8 0.708 2 30.307 2
Tools
Get Citation
Copy Citation Text
Lu QIAO, Song-lei WANG, Jian-hong GUO, Xiao-guang HE. Combination of Spectral and Textural Informations of Hyperspectral Imaging for Predictions of Soluble Protein and GSH Contents in Mutton[J]. Spectroscopy and Spectral Analysis, 2022, 42(1): 176
Paper Information
Category: Research Articles
Received: Dec. 23, 2020
Accepted: --
Published Online: Mar. 31, 2022
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20