Identification of Adulterated Edible Oils Based on 3D Fluorescence
Spectroscopy Combined With 2D-LDA
JIANG Hai-yang1, 3, CUI Yao-yao2, JIA Yan-guo1*, CHEN Zhi-peng3
1. School of Information Science and Engineering, Department of Computer Science and Engineering, Yanshan University, Qinhuangdao 066004, China
2. School of Mechanical and Electrical Engineering, Shijiazhuang University, Shijiazhuang 050035, China
3. Department of Computer Science and Technology, Tangshan Normal University, Tangshan 063000, China
Abstract：The adulteration of edible oil seriously threatens consumers' physical health and disrupts the social market order. Therefore, developing effective methods for identifying adulterated edible oil is crucial to establishing a safe and reliable food supply chain and enhancing consumer welfare. This article studies a method to identify adulterated edible oil using sesame oil as a case study. The study first formulated three types of contaminated sesame oil by adding sesame flavor, corn oil, soybean oil, and rapeseed oil. The FLS920 steady-state fluorescence spectrometer was then employed to collect 3D fluorescence spectrum data from 45 experimental samples, including these three types of contaminated sesame oil and different brands of pure sesame oil. Subsequently, two-dimensional features were extracted from the experimental samples using the 2D-LDA method. The principle of nearest-neighbor classification was applied to identify adulterated edible oils accurately. Moreover, the proposed method was compared with the PARAFAC-QDA and NPLS-DA methods. The results demonstrated that the 2D-LDA method effectively extracted two-dimensional features characterizing adulterated sesame oil. These features facilitated maximum separation of different classes of experimental samples in the projection subspace. Simultaneously, they allowed experimental samples of the same class to cluster closely in the subspace. The distinct characteristics of these features enhanced sample separability in the low-dimensional subspace, resulting in 100% identification accuracy. In contrast, the PARAFAC-QDA and NPLS-DA methods achieved 85% and 95% discrimination accuracies, respectively. Hence, the 2D-LDA method outperformed these two methods in identifying edible oil adulteration, offering a simpler and more accurate identification process and results. This study provides an efficient and feasible new solution for on-site food safety supervision.
Key words：Edible oil; 3D fluorescence spectroscopy; 2D-LDA;Adulteration identification
姜海洋，崔耀耀，贾彦国，谌志鹏. 基于三维荧光光谱结合2D-LDA的食用油掺假鉴别研究[J]. 光谱学与光谱分析, 2024, 44(11): 3179-3185.
JIANG Hai-yang, CUI Yao-yao, JIA Yan-guo, CHEN Zhi-peng. Identification of Adulterated Edible Oils Based on 3D Fluorescence
Spectroscopy Combined With 2D-LDA. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(11): 3179-3185.
[1] Nam Y S, Noh K C, Roh E J, et al. Anal. Lett., 2014, 47(7): 1190.
[2] Zhang L X, Shuai Q, Li P W, et al. Food Chem., 2016, 192: 60.
[3] Ozulku G, Yildirim R M, Toker O S, et al. Food Control, 2017, 82: 212.
[4] Wang T, Wu H L, Long W J, et al. Food Chem., 2019, 293: 348.
[5] Xu J, Liu X F, Wang Y T. Food Chem., 2016, 21: 72.
[6] PAN Zhao, CUI Yao-yao, WU Xi-jun, et al(潘 钊, 崔耀耀, 吴希军, 等). Chin. J Lumin.(发光学报), 2018, 39(4): 568.
[7] Pan Zhao, Li Ruihang, Cui Yaoyao, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2021, 245: 118948.
[8] KONG De-ming, LI Yu-meng, CUI Yao-yao, et al(孔德明, 李雨蒙, 崔耀耀, 等). Acta Optica Sinica(光学学报), 2019, 39(3): 033004.
[9] Rennhard H H. J. Agric. Food Chem., 1971, 19(1): 152.
[10] Hu Y, Wu H L, Yin X L, et al. Anal. Chim. Acta, 2018, 1021: 28.
[11] da Silva A C, Soares S F C, Insausti M, et al. Anal. Chim. Acta, 2016, 938: 53.
[12] Li M, Yuan B. Pattern Recogn. Lett., 2005, 26: 527.
[13] Kong Deming, Cui Yaoyao, Kong Lingfu, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2020, 228: 11779.
