Rapid Detection of Tocopherol Equivalent Antioxidant Capacity in Tan Mutton Based on the Fusion of Hyperspectral Imaging and Spectral
Information
YUAN Jiang-tao1, GUO Jia-jun1, SUN You-rui1, LIU Gui-shan1*, LI Yue1, WU Di1, JING Yi-xuan2
1. College of Food Science and Engineering, Ningxia University, Yinchuan 750021, China
2. College of Physics and Electronic Engineering, Ningxia University, Yinchuan 750021, China
Abstract：Trolox-Equivalent Antioxidant Capacity (TEAC) is one of the endogenous antioxidant indexes of muscle, which can be used to determine the antioxidant activity of hydrophilic compounds and free radical scavenging ability. The visible near/infrared (Vis/NIR) hyperspectral imaging technology was used to explore the feasibility of rapid detection of the TEAC in Tan mutton, a quantitative prediction model for TEAC based on spectral information fusion of image texture features (TFS) was built. The samples from different parts were randomly split into calibration set and prediction set according to the ratio of 3∶1. The spectral reflectance images were collected in the range of 400~1 000 nm, and the regions of interest (ROI) were selected to obtain raw spectral data. Four algorithms, including Median Filtering (MF), Baseline, Savitzky-Golay (S-G) and multiplicative scatter correction (MSC), were used to correct the scattering and interference information in the original spectrum, and the Partial Least Squares Regression (PLSR) model was established to correlate spectral data with TEAC values. Representative characteristic spectra of TEAC concentrations were extracted using Interval random frog (IRF), Variable combination population analysis (VCPA), Competitive adaptive reweighted sampling (CARS), and Iteratively variable subset optimization (IVSO) algorithms. The meat's main texture features were extracted sequentially by using the Gray level co-occurrence matrix (GLCM) algorithm. Based on the characteristic spectrum and spectral fusion information, the Back-propagation artificial neural network (BP-ANN) and Least-squares support vector machines (LSSVM) model were established to predict and compare the TEAC content in Tan mutton. The results showed that (1) The PLSR model established by the preprocessed spectra of Baseline was the best with Rc of 0.912 1, RMSEC of 0.963 5, Rp of 0.868 3, RMSEP of 1.277 0; (2) The 71，9，22 and 39 characteristic bands based on the original spectral were extracted by IRF，VCPA，CARS and IVSO methods, respectively, accounting for 56.8%，7.2%，17.6% and 31.2% of the total bands; (3) Compared with model effects of BP-ANN and LSSVM models in feature variables extraction based on multiple algorithms, the optimal prediction model for TEAC content was Baseline-IVSO-LSSVM (Rc=0.913 2, RMSEC=0.962 0, Rp=0.864 6, RMSEP=1.288 3); (4) The fusion model IVSO-TF1-BP-ANN showed better results (Rp=0.891 6) with improving by 0.028 6, compared with model based on the characteristic wavelength.
Key words：Mutton; Hyperspectral imaging; Tocopherol equivalent antioxidant capacity; Fusion of spectra and texture feature
袁江涛，郭佳俊，孙有瑞，刘贵珊，李 月，吴 迪，景怡萱. 滩羊肉TEAC含量的高光谱快速检测技术[J]. 光谱学与光谱分析, 2024, 44(02): 588-593.
YUAN Jiang-tao, GUO Jia-jun, SUN You-rui, LIU Gui-shan, LI Yue, WU Di, JING Yi-xuan. Rapid Detection of Tocopherol Equivalent Antioxidant Capacity in Tan Mutton Based on the Fusion of Hyperspectral Imaging and Spectral
Information. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(02): 588-593.
[1] Silva L de O, Garrett R, Monteiro M L G, et al. Food Chemistry, 2021, 362: 130159.
[2] Feng X, Tjia J Y Y, Zhou Y G, et al. LWT—Food Science and Technology, 2020, 118: 108737.
[3] Zahir S A D M, Omar A F, Jamlos M F, et al. Sensors and Actuators A: Physical, 2022, 338: 113468.
[4] Wang C X, Wang S L, He X G, et al. Meat Science, 2020, 169: 108194.
[5] Yuan R R, Liu G S, He J G, et al. Journal of Food Science, 2020, 85（5）: 1403.
[6] Anticona M, Blesa J, Lopez-Malo D, et al. Food Bioscience, 2021, 42: 101185.
[7] Weng S Z, Guo B Q, Du Y H, et al. Food Analysis Methods, 2021, 14: 834.
[8] Kumar Y, Chandrakant Karne S. Trends in Food Science & Technology, 2017, 62: 59.
[9] Yang X Y, Liu G S, He J G, et al. Journal of Food Science, 2021, 86: 1201.
[10] Bonah E, Huang X Y, Aheto J H, et al. Infrared Physics & Technology, 2020, 107: 103327.
[11] Yang Y, She X T, Cao X Q, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2022, 277: 121249.
[12] Wang Y J, Li M H, Li L Q, et al. Food Chemistry, 2021, 345: 128816.
[13] Dong F J, Hao J, Luo R M, et al. Computers and Electronics in Agriculture, 2022, 198: 107027.
[14] Liu C Y, Yu T. Neural Computing and Applications, 2020, 32: 1639.
[15] FAN Nai-yun, LIU Gui-shan, ZHANG Jing-jing, et al(樊奈昀，刘贵珊，张晶晶，等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2021, 42(3): 713.
[16] Hasan M M, Chaudhry M M A, Erkinbaev C, et al. Meat Science, 2022, 188: 108774.
[17] Cheng W W, Sun D W, Pu H B, et al. LWT-Food Science and Technology, 2016, 72: 322.
[18] Dixit Y, Casado M P, Cama-Moncunill R, et al. Analysis Methods, 2016, 8: 4134.
[19] Jiang H Z, Yoon S-C, Zhuang H, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2019, 213: 118.
[20] Wu L G, He J G, Liu G S, et al. Postharvest Biology and Technology, 2016, 112: 134.
