Reflectance Spectroscopy for Accurate and Fast Analysis of Saturated Fatty Acid of Edible Oil Using Spectroscopy-Based 2D Convolution Regression Network

Shi-zhuang WENG; Zhao-jie CHU; Man-qin WANG; Nian WANG

doi:10.3964/j.issn.1000-0593(2022)05-1490-07

Spectroscopy and Spectral Analysis, Volume. 42, Issue 5, 1490(2022)

Reflectance Spectroscopy for Accurate and Fast Analysis of Saturated Fatty Acid of Edible Oil Using Spectroscopy-Based 2D Convolution Regression Network

Shi-zhuang WENG*;... Zhao-jie CHU, Man-qin WANG and Nian WANG |Show fewer author(s)

Author Affiliations

National Engineering Research Center for Agro-Ecological Big Data Analysis & Application, Anhui University, Hefei 230601, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Image of acquisition system of the reflectance spectroscopy on edible oils

Download full size

View in Article

Fig. 2. Structure of full convolution network

Download full size

View in Article

Fig. 3. Spectral input form of two-dimensional matrix

Download full size

View in Article

Fig. 4. Structure of spectroscopy-based two Dimension convolution regression network

Download full size

View in Article

Fig. 5. Reflectance spectra for all kinds of oil samples (a) and mean reflectance spectra of esame oil, soybean oil, corn oil, sunflower oil, rapeseed oil, peanut oil and olive oil (b)

Download full size

View in Article

Table 1. Prediction results of palmitic acid content in edible oils by PLSR after different pretreatment methods
View table
View in Article
Table 1. Prediction results of palmitic acid content in edible oils by PLSR after different pretreatment methods
Methods $R_{T}^{2}$ RMSE_T $R_{P}^{2}$ RMSE_P
CEN 0.944 8 0.824 1 0.866 2 1.598 5
MSC 0.958 2 0.769 3 0.872 5 1.525 3
SNV 0.967 4 0.741 2 0.887 8 1.322 1
STA 0.932 6 0.847 6 0.853 5 1.664 7

Table 2. Prediction results of palmitic acid content in edible oils after SNV denoising
View table
View in Article
Table 2. Prediction results of palmitic acid content in edible oils after SNV denoising
Methods $R_{T}^{2}$ RMSE_T $R_{P}^{2}$ RMSE_P
PLSR 0.967 4 0.741 2 0.887 8 1.322 1
RF 0.989 2 0.569 4 0.899 5 1.485 2
SVR 0.926 0 1.011 6 0.877 7 1.262 1
FCN 0.939 3 0.925 1 0.919 3 0.960 7
S2DCRN 0.991 1 0.185 1 0.987 9 0.510 0

Table 3. Important wavelengths selected by different methods

View table

View in Article

Table 3. Important wavelengths selected by different methods

算法	重要波长数	重要波长/nm
RFrog	64	646, 464, 414, 1647, 532, 546, 612, 580, 1 905, 554, 2 005, 551, 622, 2 006, 1 649, 535, 692, 669, 458, 2 158, 547, 534, 2 399, 460, 1 212, 1 253, 481, 629, 455, 645, 543, 732, 1 482, 1 900, 457, 548, 1 655, 632, 465, 388, 696, 439, 401, 530, 1 507, 584, 1 197, 1 638, 1 653, 2 351, 2 219, 1 964, 618, 2 156, 1 635, 545, 625, 456, 482, 1 576, 2 186, 2 217, 560, 1 645
RFrog-GA	16	532, 534, 692, 464, 629, 551, 646, 612, 1 905, 2 005, 1 253, 2 158, 554, 439, 458, 622
RFrog-SFS	14	388, 401, 414, 439, 455, 456, 457, 458, 460, 464, 465, 481, 482, 530

Table 4. Prediction results of S2DCRN on palmitic acid content in edible oil after important wavelength selection
View table
View in Article
Table 4. Prediction results of S2DCRN on palmitic acid content in edible oil after important wavelength selection
变量选择
方法维数 $R_{T}^{2}$ RMSE_T $R_{P}^{2}$ RMSE_P
RAW 2 151 0.991 1 0.185 1 0.987 9 0.510 0
RFrog 64 0.993 2 0.043 7 0.962 4 0.322 1
RFrog-GA 16 0.989 6 0.067 4 0.947 6 0.376 4
RFrog-SFS 14 0.987 3 0.090 4 0.967 9 0.462 7

Table 5. Prediction results based on S2DCRN for arachidic acid and behenic acid content in edible oils after important wavelength selection

View table

View in Article

Table 5. Prediction results based on S2DCRN for arachidic acid and behenic acid content in edible oils after important wavelength selection

饱和脂肪酸种类	变量选择方法	维数	$R_{T}^{2}$	RMSE_T	$R_{P}^{2}$	RMSE_P
花生酸	-	2 151	0.996 7	0.185 1	0.983 0	0.090 2
花生酸	RFrog-SFS	14	0.973 5	0.102 9	0.950 1	0.152 9
山嵛酸	-	2 151	0.983 2	0.035 2	0.966 2	0.051 8
山嵛酸	RFrog-SFS	14	0.957 6	0.056 9	0.948 6	0.067 1

Tools

Get Citation

Copy Citation Text

Shi-zhuang WENG, Zhao-jie CHU, Man-qin WANG, Nian WANG. Reflectance Spectroscopy for Accurate and Fast Analysis of Saturated Fatty Acid of Edible Oil Using Spectroscopy-Based 2D Convolution Regression Network[J]. Spectroscopy and Spectral Analysis, 2022, 42(5): 1490

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Research Articles

Received: Jan. 6, 2021

Accepted: --

Published Online: Nov. 10, 2022

The Author Email: WENG Shi-zhuang (weng_1989@126.com)

DOI:10.3964/j.issn.1000-0593(2022)05-1490-07

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Prediction results of palmitic acid content in edible oils by PLSR after different pretreatment methods

Table 1. Prediction results of palmitic acid content in edible oils by PLSR after different pretreatment methods

Table 2. Prediction results of palmitic acid content in edible oils after SNV denoising

Table 2. Prediction results of palmitic acid content in edible oils after SNV denoising

Table 3. Important wavelengths selected by different methods

Table 3. Important wavelengths selected by different methods

Table 4. Prediction results of S2DCRN on palmitic acid content in edible oil after important wavelength selection

Table 4. Prediction results of S2DCRN on palmitic acid content in edible oil after important wavelength selection

Table 5. Prediction results based on S2DCRN for arachidic acid and behenic acid content in edible oils after important wavelength selection

Table 5. Prediction results based on S2DCRN for arachidic acid and behenic acid content in edible oils after important wavelength selection

微信扫一扫：分享