SmartEpiStock: An AI-Driven Solution for Warehouse Monitoring

Abstract

This dissertation introduces SmartEpiStock, an innovative solution for grain silo monitor ing based on the integration of three key technologies: the Internet of Things (IoT), Deep Learning, and ontologies. The system is designed to optimize grain preservation by moni toring storage conditions in real time such as temperature, humidity, and CO, automatically detecting anomalies using a MobileNetV3 model trained on grain images, and assessing risks through logical rules embedded in a semantic ontology. An intuitive mobile application en ables users to view alerts and silo status, thereby improving farmers’ responsiveness. The proposed approach overcomes the limitations of traditional storage methods by combining perception, analysis, and reasoning into a smart, autonomous, and scalable system.