Advanced Convolution Neural Network Approach for Optimized Detection of Coffee Beans

Evaluating the color of green coffee beans is an important method for determining their consistency and market price. This measurement is generally carried out by visual inspection or by using conventional instruments with some limitations. The purpose of this thesis was therefore to create a computer vision framework that would produce CIE (Commission International d'Eclairage) L*a*b * measurements of green coffee beans and identify them according to their color. Artificial Neural Networks (ANN) were used as a transformation model was used to classify coffee beans by the Bayes Classifier, in four distinct groups: white, cane green, green and blue-green.