Rapid Determination of TBARS Contents in Tan Mutton Using Hyperspectral Imaging

Nai-yun FAN; Gui-shan LIU; Jing-jing ZHANG; Rui-rui YUAN; You-rui SUN; Yue LI

doi:10.3964/j.issn.1000-0593(2022)03-0713-06

Journals >Spectroscopy and Spectral Analysis >Volume 42 >Issue 3 >Page 713 > Article

Spectroscopy and Spectral Analysis
Vol. 42, Issue 3, 713 (2022)

Rapid Determination of TBARS Contents in Tan Mutton Using Hyperspectral Imaging

Nai-yun FAN^*, Gui-shan LIU^*;, Jing-jing ZHANG, Rui-rui YUAN, You-rui SUN, and Yue LI

Author Affiliations

School of Food & Wine, Ningxia University, Yinchuan 750021, China

show less

DOI: 10.3964/j.issn.1000-0593(2022)03-0713-06 Cite this Article

Nai-yun FAN, Gui-shan LIU, Jing-jing ZHANG, Rui-rui YUAN, You-rui SUN, Yue LI. Rapid Determination of TBARS Contents in Tan Mutton Using Hyperspectral Imaging[J]. Spectroscopy and Spectral Analysis, 2022, 42(3): 713 Copy Citation Text

show less

Fig. 1. Linear fitting diagrams of measured and predicted values of TBARS contents
(a): Raw-PLSR; (b): SG-PLSR; (c): De-trending-PLSR; (d): SG+De-trending-PLSR

Download full size | View in the Article

Fig. 2. Two-dimensional correlation spectra and its slice spectra
(a): Synchronous contour map; (b): Slice spectra

Download full size | View in the Article

Fig. 3. Characteristic wavelength extracted by variable selection algorithm
(a): VCPA algorithm; (b): SPA algorithm

Download full size | View in the Article

Data set	Samples	TBARS contents/(mg·kg^-1)
Data set	Samples	Max	Min	Mean	Standard deviation
Calibration set	135	1.47	0.16	0.67	0.30
Prediction set	45	1.37	0.17	0.61	0.35

Table 1. Statistics of measured TBARS contents in Tan mutton

View in the Article

Pretreatment method	No. LV	Calibration set		Cross-validation		Prediction set
Pretreatment method	No. LV	$R_{C}^{2}$	RMSEC/(mg·kg^-1)	$R_{CV}^{2}$	RMSECV/(mg·kg^-1)	$R_{P}^{2}$	RMSEP/(mg·kg^-1)
Raw	17	0.830	0.123	0.722	0.159	0.821	0.157
SG	20	0.829	0.123	0.706	0.164	0.805	0.189
de-trending	20	0.874	0.106	0.770	0.144	0.853	0.139
SG+de-trending	20	0.850	0.115	0.749	0.150	0.827	0.243

Table 2. PLSR modeling results of TBARS content with different pretreatment methods

View in the Article

Wavelength/nm	579	699	756	867
Synchronous
579	+	-	-	+
699		+	+	-
756			+	-
867				+

Table 3. Signs of cross-peaks generated from two-dimensional correlation analysis

View in the Article

Wavelength extraction method	Variable number	Characteristic wavelength/nm
FS+VCPA	7	636, 641, 646, 660, 675, 694, 915
FS+CARS	16	415, 713, 727, 732, 751, 756, 785, 862, 895, 915, 924, 939, 958, 972, 977, 982
FS+SPA	20	401, 449, 458, 468, 492, 516, 559, 593, 603, 631, 675, 723, 795, 819, 843, 852, 943, 953, 972, 987
2DCOS+VCPA	8	631, 641, 646, 651, 660, 675, 689, 761
2DCOS+CARS	24	588, 593, 598, 603, 612, 622, 631, 636, 646, 655, 675, 684, 689, 703, 713, 727, 742, 751, 775, 780, 790, 799, 809, 819
2DCOS+SPA	14	593, 598, 603, 612, 622, 631, 641, 651, 660, 675, 689, 785, 809, 823

Table 4. The results of extracting characteristic wavelengths

View in the Article

Extraction method	No. LV	Calibration set		Cross-validation		Prediction set
Extraction method	No. LV	$R_{C}^{2}$	RMSEC/(mg·kg^-1)	$R_{CV}^{2}$	RMSECV/(mg·kg^-1)	$R_{P}^{2}$	RMSEP/(mg·kg^-1)
FS	20	0.874	0.106	0.770	0.144	0.853	0.139
FS+VCPA	7	0.794	0.135	0.770	0.143	0.858	0.139
FS+CARS	15	0.821	0.126	0.771	0.143	0.806	0.148
FS+SPA	19	0.792	0.136	0.717	0.159	0.842	0.140
2DCOS	20	0.850	0.115	0.751	0.150	0.854	0.139
2DCOS+VCPA	8	0.795	0.135	0.769	0.143	0.839	0.146
2DCOS+CARS	19	0.857	0.113	0.790	0.137	0.862	0.132
2DCOS+SPA	14	0.815	0.128	0.771	0.143	0.845	0.143

Table 5. Comparison of PLSR models based on different wavelength extraction methods

Download Citation

Save the article for my favorites

Paper Information