|OUYANG, QIN - Jiangsu University|
|WANG, LI - Jiangsu University|
|KANG, RUI - Nanjing Agricultural University|
|CHEN, QUANSHENG - Jiangsu University|
Submitted to: Food Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/17/2021
Publication Date: 2/4/2021
Citation: Ouyang, Q., Wang, L., Park, B., Kang, R., Chen, Q. 2021. Simultaneous quantification of chemical constituents in matcha with visible near infrared hyperspectral imaging technology. Food Chemistry. https://doi.org/10.1016/j.foodchem.2021.129141.
Interpretive Summary: Match is a type of green tea obtained in powder form, which is cultivated and processed under special conditions with requirement of covering and shading tea leaves for at least twenty days before harvest, resulting in unique flavor and brilliant green color due to increasing chlorophyll and amino acid content in leaves. Matcha is healthy diet and becomes increasingly popular in recent years. The main functional constituents in matcha provide special flavor and health benefits such as antifatigue, anticancer and antioxidant. Therefore, detection of chemical constituents in matcha is necessary for quality evaluation. Because current traditional wet chemistry methods are not cost-effective and time-consuming, more accurate, rapid, and high-efficient detection methods are demanding for simultaneous quantification of several chemical constituents in matcha. Visible near-infrared hyperspectral imaging (VNIR-HSI) technology is an effective analytical technique for fast and non-invasive food analysis. In this study, characteristic spectra measured from matcha samples with VNIR-HSI were used to establish models for predicting multiple chemical constituents in matcha. The models showed promising performance to predict major chemical constituents in matcha.
Technical Abstract: Matcha is specially grown tea powder and contains several chemical constituents including caffeine, tea polyphenols (TP), free amino acids (FAA), and chlorophyll that provides vibrant green color, intense umami flavor, astringency and bitterness, which are quality factors in matcha. The present work aimed to assess the feasibility of predicting multiple chemical constituents in matcha using visible near infrared hyperspectral imaging (VNIR-HSI) technology. Regions of interest (ROI) were first extracted to calculate the representative mean spectrum of each sample. Then, standard normal variate (SNV) was applied to correct the characteristic spectra for data preprocessing. Competitive adaptive reweighted sampling (CARS) and bootstrapping soft shrinkage (BOSS) were comparably used to optimize models, which were built based on partial least squares (PLS), called as CARS-PLS and BOSS-PLS, respectively. Results showed that BOSS-PLS models outperformed CARS-PLS and simply PLS models, with correlation coefficient in prediction dataset (Rp) of 0.8987 for caffeine, 0.8425 for TP, 0.8912 for FAA, 0.9118 for ratio of TP to FAA, and 0.9205 for chlorophyll, respectively. The proposed approach suggested that VNIR-HSI technology has the potential as a rapid and non-destructive alternative for simultaneous quantification of chemical constituents in matcha.