Digital Image Processing for Classification of Coffee Cherries

Abstract

A machine vision-based classification system to sort coffee fruits (cherries) according their ripeness stage is presented. Eight categories were defined and they include the entire coffee cherry ripeness process, from the initial stage (early green) to over-ripe and dry stages. A Bayesian classifier was implemented using a set of nine features which include color, shape and texture computed on an image of the fruit, with a 96.88% of performance using the cross-validation approach.