Quantitative Prediction and Visualization of Key Physical and Chemical Components in Black Tea Fermentation Using Hyperspectral Imaging

Abstract Fermentation is a key process that affects the quality of black tea. In this study, we discussed the changes and influencing factors of key endoplasmic components at different positions of stacked fermented leaves, and the effects of different preprocessing, variable selection and intelligent algorithm on the model performance are compared, the quantitative prediction model of main endoplasmic components of Congou black tea under different fermentation time series was established, finally, the content distribution is depicted in different colors. The results show that the RPD values of the random forest (RF) prediction model constructed using the optimal variables of theafuscin, thearubigin, catechin, caffeine, and soluble sugar were 3.40, 2.21, 5.71, 1.46, and 2.89, respectively. The RPD values of the support vector machine (SVR) prediction model constructed using the optimal variables of theaflavin and the phenol ammonia ratio were 3.78 and 2.91, respectively. Furthermore, the visualization process successfully displayed the distribution of various quality indicators of the samples at different time periods. These research results lay a theoretical foundation for advancing the judicious processing of black tea.