Potency of Image Color Textural Feature for Granular Size Estimation of Ground Coffee

摘要

Determining particle size of ground coffee is generally conducted with laboratory equipment, such as the use of sieves, which of course, the method is costly and time consuming. On the other hand, a continuous secondary coffee processing requires a rapid measurement method for accessing its quality product for all stages, include on the grinding process. Therefore, it is necessary to develop a robust and rapid method for the particle size estimation on grinding stage in the secondary coffee processing. This study aims to evaluate an image processing technique for determining particle size of the ground coffee. In this study, we focused on evaluating some image features which have possibility to estimate the coarseness of the ground coffee product. The features included color and textural features extracted from an image of ground coffee powder. A simple image capture box, equipped with controlled light source and a camera was prepared. This study used varied samples of ground coffee with different particle size placed on a static small container inside the captured box. The sample was then captured at a constant lighting level. The image was then analyzed. Twenty-six color and textural features were extracted and then evaluated in term of the granular size. The study showed that a number of texture features had a strong relationship to the granular size. Thus, these features are potentially used for estimating the product particle size.