Cooperation Unit - Timorese Academic Journal of Science & Technology

Abstract: The Coffee Bean is the most important agricultural product in East Timor and it serves as the primary source of income for many Timorese. However, almost all procedures of cultivation and processing have not automated yet. At present, the classification of defective coffee bean is achieved by human inspector based on the size, shape, colors and so on. The purpose of this research is to classify coffee bean images automatically using image processing and machine learning techniques. As the machine learning technique, we use a conventional method of Support Vector Machine (SVM).  The SVM is one of the powerful machines learning method for classification. The result shows that defective coffee bean is selected with 97 % accuracy using the SVM.

Keywords: Support Vector Machine, Coffee Bean, Image Classification.