Evaluation of the use of hyperspectral reflectance for estimating Theanine content in tea leaves

Abstract

Theanine is the most abundant free amino acid in tea leaves (Camellia sinensis) and it is also one of important factors for assessing the quality of tea ; thus, developing an in-situ method to monitor theanine is useful for agricultural management. Some hyperspectral remote sensing techniques, especially spectral indices, have been applied for assessing vegetation properties such as pigment content and water content. In this study, searching for new indices was attempted based on hyperspectral reflectance. The newly identified index, which is expressed as a differential type of index using reflectance at 1735 nm and 1755 nm, possessed a great performance, achieving a root mean square error with leave-one-out cross validation of 0.065 mg/cm2.