Classification of Chinese Famous Tea Base on Visible and Near Infrared Hyperspectra Imaging

A total of 180 samples from six varieties of typical tea (30 for each variety) were collected for hyperspectral image classification. The first 2 principal components (PCs) explained over 97% of variances of all spectral information. Gray-level co-occurrence matrix (GLCM) analysis was implemented on the 2 principal component (PC) images to extract 24 textural feature variables in total. Least squares-support vector machine (LS-SVM) classification models were developed to classify Chinese famous tea based on (i) spectral variables, (ii) textural variables, (iii) spectral and textural combined variables, respectively. Satisfactory average correct classification rate (CCR) of 100% for the prediction samples based on (iii) was achieved, which was superior to the results based on (i) or (ii). The experimental results indicate that the proposed method is effective to recognize different types of Chinese famous tea by hyperspectral imaging technique. The overall results indicate that VIS/NIR hyperspectral imaging technique is promising for the reliable classification of Chinese famous tea.