Rapid and real-time detection of moisture in black tea during withering using micro-near-infrared spectroscopy

Rapid and accurate measurement of the moisture in black tea during withering is crucial for the digitalization of the processes in the black tea industry. Therefore, computational systems should be developed for the rapid detection of moisture in withered leaves. In this study, relying on miniaturized near-infrared spectroscopy (micro-NIRS) coupled with a smartphone, an Elman neural network (ENN)-based moisture-prediction model was developed. Specifically, the ENN-based moisture-prediction model incorporated principal component analysis (PCA) and was designed to perform rapid detection and analysis of the water content of withered leaves. The combination of an ENN and PCA can both embody spectral features and exhibit strong dynamic informationprocessing capabilities. The proposed approach improves the anti-interference ability and training efficiency of the model. Experimental results show that micro-NIRS is an effective and fast tool for evaluating the moisture content of withered leaves and that the proposed model is highly suited as a rapid-detection system, with a correlation coefficient of prediction of 0.99314 and a residual predictive deviation of 11.8108. Thus, this research provides a portable, accurate, fast, and non-destructive method for predicting the moisture content of withered leaves.