Advances of Vis/NIRS and imaging techniques assisted by AI for tea processing

Abstract

Tea is one of the most popular drinks due to its distinct flavor and numerous health benefits. The quality of tea is closely related to production processing. Human sensory evaluation is the conventional method for quality monitoring in tea processing. However, this method is subjective and susceptible to environmental influences. Therefore, visible/near-infrared spectroscopy (Vis/NIRS) and hyperspectral imaging (HSI) techniques offer great potential due to their rapid detection speed, nondestructive, low cost, and simple operations. Artificial intelligence (AI) is one of the most promising methodological approaches for spectral analysis and decision-making of automated production. Vis/NIRS and HSI techniques assisted by AI further promote the progress of quality monitoring in tea processing. This paper reviewed the updated applications of Vis/NIRS and HSI techniques assisted by AI for quality monitoring in tea processing from 2019 to 2025. In particular, the tea production process, theories of Vis/NIRS and HSI techniques, and AI algorithms in spectral analysis are briefly introduced. Furthermore, the recent applications of Vis/NIRS and HSI techniques assisted by AI in tea processing quality monitoring are summarized and discussed. Finally, the challenges and future trends of Vis/NIRS and HSI techniques associated with their practical application in the tea industry are presented.

Keywords: Quality monitoring; chemical components; deep learning; machine learning; nondestructive detection; sensory attributes.