Application of Hyperspectral Image to Detect the Content of Total Nitrogen in Fish Meat Volatile Base

Jin-ping ZOU; Shuai ZHANG; Wen-tao DONG; Hai-liang ZHANG

doi:10.3964/j.issn.1000-0593(2021)08-2586-05

Spectroscopy and Spectral Analysis, Volume. 41, Issue 8, 2586(2021)

Application of Hyperspectral Image to Detect the Content of Total Nitrogen in Fish Meat Volatile Base

Jin-ping ZOU^1、*, Shuai ZHANG^2、2;, Wen-tao DONG^2、2;, and Hai-liang ZHANG^{2、2; *;}

Author Affiliations

¹1. Jiangxi Biotech Vocational College, Nanchang 330013, China

²2. School of Electrical and Automation Engineering, East China JiaoTong University, Nanchang 330013, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Imaging system schematic diagram of hyperspectral

Download full size

View in Article

Fig. 2. The reflectance spectra of salmon fish fillets at different TVB-N values

Download full size

View in Article

Fig. 3. Predicted and measured TVB-N values for both the PLS (a) and LS-SVM (b) models under the whole spectral wavelengths range

Download full size

View in Article

Fig. 4. Position of selected effective wavelengths by SPA

Download full size

View in Article

Fig. 5. Visualization of the predicted TVB-N content for 4 fish samples in prediction set

Download full size

View in Article

Fig. 6. Main steps of detection of salmon fish fillets TVB-N content using HIS

Download full size

View in Article

Table 1. Reference values of TVB-N content (mg N/100 g) measured by a traditional method
View table
View in Article
Table 1. Reference values of TVB-N content (mg N/100 g) measured by a traditional method
TVB-N Calibration Prediction
Minimum 7.84 8.01
Maximum 16.49 15.98
Mean 12.42 12.13
Standard deviation 5.42 5.28

Table 2. Prediction results of LS-SVM and PLS based on full spectral parameters
View table
View in Article
Table 2. Prediction results of LS-SVM and PLS based on full spectral parameters
Model
type Number of
wavelength LVs Calibration Prediction
$R_{c}^{2}$ RMSEC/% $R_{P}^{2}$ RMSEP/%
LS-SVM 167 / 0.919 2.292 0.918 2.312
PLS 167 10 0.905 2.742 0.907 2.751

Table 3. Effective wavelengths selected by SPA
View table
View in Article
Table 3. Effective wavelengths selected by SPA
算法数量选择波长/nm
SPA 8 956, 1 013, 1 152, 1 210, 1 286,
1 301, 1 397, 1 464

Table 4. SPA-LS-SVM and SPA-PLS models for predicting TVB-N in salmon fillets by using the only 8 important wavelengths
View table
View in Article
Table 4. SPA-LS-SVM and SPA-PLS models for predicting TVB-N in salmon fillets by using the only 8 important wavelengths
Model
type Number of
wavelength LVs Calibration Prediction
$R_{c}^{2}$ RMSEC/% $R_{p}^{2}$ RMSEP/%
SPA-LS-SVM 8 / 0.904 2.802 0.903 2.761
SPA-PLS 8 5 0.903 2.794 0.901 2.801

Tools

Get Citation

Copy Citation Text

Jin-ping ZOU, Shuai ZHANG, Wen-tao DONG, Hai-liang ZHANG. Application of Hyperspectral Image to Detect the Content of Total Nitrogen in Fish Meat Volatile Base[J]. Spectroscopy and Spectral Analysis, 2021, 41(8): 2586

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Research Articles

Received: Aug. 7, 2020

Accepted: --

Published Online: Sep. 8, 2021

The Author Email: ZOU Jin-ping (2356255615@qq.com)

DOI:10.3964/j.issn.1000-0593(2021)08-2586-05

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Reference values of TVB-N content (mg N/100 g) measured by a traditional method

Table 1. Reference values of TVB-N content (mg N/100 g) measured by a traditional method

Table 2. Prediction results of LS-SVM and PLS based on full spectral parameters

Table 2. Prediction results of LS-SVM and PLS based on full spectral parameters

Table 3. Effective wavelengths selected by SPA

Table 3. Effective wavelengths selected by SPA

Table 4. SPA-LS-SVM and SPA-PLS models for predicting TVB-N in salmon fillets by using the only 8 important wavelengths

Table 4. SPA-LS-SVM and SPA-PLS models for predicting TVB-N in salmon fillets by using the only 8 important wavelengths