This paper proposes a low-level image feature extraction and data analysis technique for rapid assessment of quality of tea to overcome the drawback of human perception-based organoleptic method of sensory panel during ‘tea tasting’. An image capturing system has been developed to capture the image of tea liquor samples under controlled illumination. The experiment has been performed using 36 CTC (cut-tear-curl) tea sample collected from the Tocklai Tea Research Institute, Jorhat, Assam. Firstly, 27 low-level image features have been extracted using three different colour models and are then analysed using principal component analysis (PCA) and linear discriminant analysis (LDA) for visualization of underlying information. Finally, different classification models based on statistical regression techniques, e.g. multiple linear regression (MLR), principal component regression (PCR) and partial least square regression (PLSR), are investigated to find out a correlation between extracted image features with the tea tasters’ score.
CITATION STYLE
Akuli, A., Pal, A., Dey, T., Bej, G., Santra, A., Majumdar, S., & Bhattacharyya, N. (2020). Assessment of black tea using low-level image feature extraction technique. In Advances in Intelligent Systems and Computing (Vol. 1112, pp. 453–467). Springer. https://doi.org/10.1007/978-981-15-2188-1_36
Mendeley helps you to discover research relevant for your work.