Designing the quality of coffee bean detection application using Hue Saturation Intensity

Abstract

This study discusses the design of applications that can detect the quality of raw coffee beans both good and bad based on the value of HSI (Hue Saturation Intensity) in the coffee beans using digital image processing and Backpropagation artificial neural networks. The process of identifying coffee beans is based on the intensity of the HSI value that is owned by coffee beans. The HSI value is converted from RGB values, then the training process is performed on backpropagation artificial neural networks to recognize good quality seeds and poor-quality seeds. The testing phase is carried out using an interface designed in the MATLAB R2013a software. Based on the results of testing, it was found that this application was able to detect samples of coffee beans that were not properly trained.