Diseño e implementación de un prototipo clasificador de granos de cacao usando Visión Artificial y Machine Learning

Abstract

The objective of this degree project is to design and implement a cocoa classifier prototype using Artificial Vision and Machine Learning using open sources resources and technologies. The initial step was to train a neuron using Teachable Machine to identify its fermentation level according to the characteristics of the cocoa bean and also ignore other elements inherent to the target to be analyzed, once the neuron was trained, the analysis was carried out in real time to then get to the classification The basic idea of implementing these types of concepts in these types of activities is to demonstrate how malleable they can be when used in areas that apparently would not be directly correlated with this level of technology. As for the system, it will be an amalgam of both logical and physical tools, since in a structural way it will be made up of servo motors, a conveyor belt, a camera, a LED system, an infrared sensor and others; in terms of logic, Python will be used as the base language and libraries such as Tkinter, Numpy, Tensorflow, among others, will be used.