Application of Various Spectral Techniques in Quantitative Analysis of Tea Quality

As the raw material of tea, the top of the most fashionable drink in the worldwide, tea has been widely distributed and adulterated. In addition, there are some problems such as nutrition loss or pesticide residue during the planting, processing and preservation of tea. In recent years, a variety of rapid development of spectroscopy technology has been widely used in the research orientation of tea for quantitative analysis of its quality, nutrition and safety due to its own safe, reliable and rapid characteristics. In this review, the latest applications of near in frared spectroscopy(NIR), Hyperspectral imaging(HSI), Raman and fluorescence spectroscopy combined with different machine learning(ML) and artificial intelligence(AI) algorithms in quantitative analysis of tea quality are reviewed and compared. Among them, NIR technology still occupies a leading position in the research of tea field, and can be used as a measurement method of various nutrients in tea. HIS and NIR can provide spatial information of tea by combining computer vision(CV) technology, and infrared spectroscopy has the potential to analyze spectral information more comprehensively. Raman spectrum and fluorescence spectrum are mainly used in the detection of crop residues and diseases, respectively. In the future, it will still be an important trend and extremely challenging to use multiple spectral techniques combined with machine learning and artificial intelligence to study the quantitative analysis of tea quality.