Hyperspectral guidance for summer tea processing: Enhancing taste and aroma through short-term cycled heaping

Abstract

To reduce the unpleasant flavor in tea infusion made from summer fresh leaves and lessen tea resource wastage, a short-term cycled heaping method, addressing the temperature/moisture variations during prolonged heaping, was introduced to produce high-quality yellow tea with significantly enhanced taste and aroma. Furthermore, a hyperspectral approach was developed to anticipate such improvements. Specifically, short-term cycled heaping reduced summer tea's astringency and bitterness while increasing sweetness. Using near-infrared spectroscopy, the multiplicative scatter correction (MSC)–support vector machine (SVM) model predicted tea umami with 66.25% precision and richness with 77.64% accuracy via standard normal variate (SNV)–SVM, based on epigallocatechin (EGC), epicatechin (EC), and gallocatechin gallate (GCG) variations. Aroma enhancement was forecasted by a hyperspectral–electronic nose regression model, achieving over 81.50% prediction accuracy for summer tea aromas in the visible spectrum and surpassing 73.45% in the near-infrared domain, primarily attributed to pentanal, propanal, toluene, ethyl propionate, and 2,3-pentanedione variations. Overall, hyperspectral-guided summer tea processing, particularly for summer-harvested leaves, shows potential in offering a valuable screening tool, enhancing quality, and efficiency in the tea industry and promoting sustainable development.