Determinación del Grado de Fermentación del Cacao mediante diferentes técnicas de visión artificial

Abstract

This article presents the results of the application of several computer vision methods on the fermentation process of cocoa beans. The analyzed methods are: Vector quantification, k-means, fuzzy grouping, average displacement. With these methods, the training and classification of the images of the cocoa beans in the fermentation process has been developed, these methods have been codified in m files of MATLAB and have allowed to estimate the level of fermentation of the cocoa from the information of the color in RGB format with 75% certainty. This work is based on the technique of sensory profiles, used empirically by farmers in charge of cocoa cultivation.