DEVELOPMENT OF AI BASED COLD STORAGE SYSTEM

Postharvest losses caused by storage related issues, remains one of the greatest challenges facing agricultural productivity in Nigeria, particularly for perishable crops such as tomatoes and bananas. This study presents the design, fabrication, and evaluation of an Artificial Intelligence (AI)-based cold storage system aimed at reducing postharvest losses through intelligent monitoring and environmental control. The system integrates the Internet of Things (IoT), machine learning, and computer vision technologies to automatically regulate temperature, humidity, and air quality during storage. It employs a Raspberry Pi 4B microcontroller as the central processing unit, interfaced with DHT11 sensors for real-time environmental monitoring and a camera module for fruit quality assessment. Thermoelectric cooling using Peltier modules was adopted to achieve efficient temperature regulation within a range of 8–12 °C and humidity levels of 85–90% relative humidity (RH). Locally sourced materials such as ABS plastic, polyurethane foam, and aluminium were used for fabrication to ensure cost-effectiveness and adaptability to rural conditions. Experimental tests revealed that the system maintained stable cooling performance with minimal energy consumption (average 40 W) while achieving an AI-based classification accuracy of 91% in identifying fruit ripeness and spoilage. The total fabrication cost of ₦273,810 confirms its affordability relative to conventional compressor-based systems. This study demonstrates the feasibility of deploying affordable, AI-driven storage technologies to enhance food preservation, improve market value, and support sustainable agriculture in Nigeria. It further establishes a practical framework for integrating smart technologies into postharvest management systems for smallholder farmers.