Nondestructive grading of black tea based on physical parameters by texture analysis

Texture is an important characteristic used in identification of objects or regions of interest in an image. This paper describes a technique to discriminate between four different grades of made black tea using textural features based on grey tone spatial dependencies. The statistical features were computed from the tea images and wavelet decomposed sub band images. The multi-layer perceptron (MLP) technique has been used for data classification and 82.33% classification accuracy was achieved. Finally, statistical analysis in the form of one way analysis of variance (ANOVA) has been employed as a validation tool to check for grading accuracy.