Rapid and non-destructive cinnamon authentication by NIR-hyperspectral imaging and classification chemometrics tools.

Autor: Cruz-Tirado JP; Department of Food Engineering, School of Food Engineering, University of Campinas, Campinas, SP, Brazil., Lima Brasil Y; Department of Food Engineering, School of Food Engineering, University of Campinas, Campinas, SP, Brazil., Freitas Lima A; Department of Food Science, School of Food Engineering, University of Campinas, Campinas, SP, Brazil., Alva Pretel H; Escuela de Ingeniería Agroindustrial, Facultad de Ciencias Agropecuarias, Universidad Nacional de Trujillo, Av. Juan Pablo II s/n, Trujillo, Peru., Teixeira Godoy H; Department of Food Science, School of Food Engineering, University of Campinas, Campinas, SP, Brazil., Barbin D; Department of Food Engineering, School of Food Engineering, University of Campinas, Campinas, SP, Brazil., Siche R; Escuela de Ingeniería Agroindustrial, Facultad de Ciencias Agropecuarias, Universidad Nacional de Trujillo, Av. Juan Pablo II s/n, Trujillo, Peru. Electronic address: rsiche@unitru.edu.pe.
Jazyk: angličtina
Zdroj: Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy [Spectrochim Acta A Mol Biomol Spectrosc] 2023 Mar 15; Vol. 289, pp. 122226. Date of Electronic Publication: 2022 Dec 09.
DOI: 10.1016/j.saa.2022.122226
Abstrakt: Cinnamon is a valuable aromatic spice widely used in pharmaceutical and food industry. Commonly, two-cinnamon species are available in the market, Cinnamomum verum (true cinnamon), cropped only in Sri Lanka, and Cinnamomum cassia (false cinnamon), cropped in different geographical origins. Thus, this work aimed to develop classification models based on NIR-hyperspectral imaging (NIR-HSI) coupled to chemometrics to classify C. verum and C. cassia sticks. First, principal component analysis (PCA) was applied to explore hyperspectral images. Scores surface displayed the high similarity between species supported by comparable macronutrient concentration. PC3 allowed better class differentiation compared to PC1 and PC2, with loadings exhibiting peaks related to phenolics/aromatics compounds, such as coumarin (C. cassia) or catechin (C. verum). Partial least square discriminant analysis (PLS-DA) and Support vector machine (SVM) reached similar performance to classify samples according to origin, with error = 3.3 % and accuracy = 96.7 %. A permutation test with p < 0.05 validated PLS-DA predictions have real spectral data dependency, and they are not result of chance. Pixel-wise (approach A) and sample-wise (approach B, C and D) classification maps reached a correct classification rate (CCR) of 98.3 % for C. verum and 100 % for C. cassia. NIR-HSI supported by classification chemometrics tools can be used as reliable analytical method for cinnamon authentication.
Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
(Copyright © 2022 Elsevier B.V. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: