Sistema de clasificación de granos de cacao frescos basados en visión computacional

Abstract

In Ecuador, farmers from small cocoa farms perform the classification of fresh cocoa beans with pulp in a traditional way, i.e., through their senses. In addition, they do not use any kind of technology that makes their production faster. The objective of this project is to propose a fresh cocoa beans classification approach based on computer vision for farmers, which will be very helpful in the process of removing cocoa beans pulp allowing them to estimate efficiently the quality of these beans. For this purpose, we describe a methodology to classify cocoa beans by using the k-means clustering method and morphological operations to remove the background of the images, then we use a Bag of Visual Words (BoVW) approach as feature extractor and finally we classify these features using a Support Vector Machine (SVM) classifier with k-fold cross-validation. As a result, we achieved a great performance of the SVM classifier with linear kernel function by keeping a low false positive rate, which demonstrated the viability of using computer vision for this task.