Cacao Pod Infection Level Classification Using Transfer Learning

Abstract

This research aims to address the constant threat to cacao farming, particularly from diseases like black pod rot, by utilizing Artificial Intelligence (AI) and Deep Learning algorithms. An experimental research design method was used to train a convolutional neural network that can classify three levels of cacao pod infection: low, moderate, and severe. Under controlled conditions, the model achieved an accuracy of 91% in correctly classifying the cacao pod condition. The study recommends further research to integrate the model with hardware monitoring and surveillance devices to enable real-time classification of cacao pod conditions in the actual field. With this capability, farmers can respond quickly and effectively to mitigate production loss caused by cacao diseases. This research presents a promising approach to leveraging AI in cacao farming, with potential benefits for farmers, researchers, and the broader community.