Indonesian Black Tea Classification Using Fourier-Transform Near-Infrared Spectroscopy and a Principal Component Analysis

摘要： Indonesian black tea (Camellia sinensis) is processed based on several standard methods to achieve a certain grade. There are more than 25 grades produced including Dust one (D1), Fanning's (FANN) and Pekoe Fanning's (PF). In general, the physical appearances are similar to each other, which makes it difficult to be differentiated from visual observations by non-experts. During the packaging process, a tea grade can easily be packed using a bag with a different label of the grade. In the global market, such a mistake should be avoided before shipping. Since shipping usually involves a large number of bags, then rapid, accurate, and nondestructive methods are critical. This paper discusses a method for classifying tea grades rapidly, accurately and nondestructively using Fourier-transform near-infrared spectroscopy (FT-NIRS) and a principal component analysis (PCA). Sixty-nine tea samples from grade Dust one (D1), Fanning's (FANN) and Pekoe Fanning's (PF) were separated into two groups; 46 samples were used for the training sets and 30 samples for the validation sets. The analysis successfully differentiated that of 3 different grades, indicating that the method is a potential alternative for tea classification in quality control processes.