Color and Texture Classification of Green Tea Using Least Squares Support Vector Machine (LSSVM)

Abstract

This work presented an approach for color and texture classification of green tea using Least Squares Support Vector Machine (LSSVM). Color features extracted from histogram of every channel in RGB and HSI color space, texture features computed from Grey Level Co-occurrence Matrix (GLCM) of every channel in RGB and HSI color space, and different combinations of the color and texture features, were used respectively as input data set for the LSSVM classifiers. The classification performances of these different methods were compared. The results show that the combined color and texture features from HSI color space give the best performance with accuracy of 96.33% for prediction unknown samples in testing set. Based on the results, it can be concluded that combined color and texture features coupled with a LSSVM classifier can be a fast and non-destructive technique efficiently utilized to classify green tea.