期刊名称：International Journal on Smart Sensing and Intelligent Systems
印刷版ISSN：1178-5608
出版年度：2016
卷号：9
期号：2
页码：709-732
出版社：Massey University
摘要：Orthodox black tea quality depends upon the amount of certain organic compounds presentand out of these, theaflavins (TF) and thearubigins (TR) are the most important ones While TF isresponsible for attractive golden colour, increased brightness and astringency in tea liquor, TR isreddish brown, reduces the brightness of tea liquor and contribute mostly for the ashy taste of theliquor with minor improvement in astringency. The rapid estimation of TF and TR thus may resolvethe problem of certain uncertainty or ambiguity that may arise during quality assessment of tea by thetea tasters. In this paper, a new method for rapid measurement of concentration of TF and TR isdescribed using a machine vision system taking images of tea liquor and employing artificial neuralnetworks (ANN). The results show good correlation of estimated values of TF and TR with the actualconcentrations obtained using ultraviolet-visible spectrophotometer (UV-VIS).
关键词：Theaflavins (TF); Thearubigins (TR); Machine Vision System; Orthodox black tea; UV-VIS;spectrophotometer; Generalised Regression Neural Network (GRNN); Back propagation multilayer;perceptrons (BP-MLP); Radial Basis Function Network (RBFN)
