Machine vision-based analysis for black tea quality evaluation

Abstract

Currently, visual quality evaluation of black tea production depends only on expert labor judgement. This allows biases and inconsistency in the end product caused by unstandardized methods. Therefore, it is proposed that machine vision could be applied in order to standardize the quality evaluation method. The goal of this research is to build a machine vision-based system to classify quality classes of black tea. In this research black tea classes used are class I, class II, and class III without considering the grade for each class. A neural network classifier was used to classify black tea images with seven nodes, one hidden layer. Ten features were used as an input for the classifier, RGB, index RGB, R/G, R/B, magenta, and yellow. From the results, the system achieved an accuracy of 100%. Application of machine vision technology in black tea production could not only support sustainability in agriculture, but also can also be used for precision agriculture in practical fields.