A dew computing-based smart tomato storage monitoring framework

Abstract

Post-harvest tomato storage faces significant challenges due to the lack of intelligent infrastructure, resulting in high spoilage rates for low-income farmers. This paper presents an Internet of Things (IoT)-based dew computing architecture for real-time monitoring, control, and optimization of stored tomato shelf life in storage warehouses. Each warehouse is divided into zones where tomatoes are stored in side-by-side crates. Temperature, humidity, pressure, pH, gas, color, and motion sensor data are collected through zone-level data sensing modules and transmitted locally to a warehouse-level Dew edge node via the long-range wide area network (LoRaWAN) module. The Dew edge node processes data for localized decisions and synchronizes with the cloud using a Dew computing-based Message Queuing Telemetry Transport (DewMQTT) back-off mechanism, enabling reliable global access and zone-wise monitoring. We introduce the Shelf-Life Tomato Cluster Net Optimizer (SL-TCNO) algorithm, integrated with a multilayer perceptron model, to manage cooling ventilation, provide transportation recommendations, detect fungus disease, predict tomato health, and monitor daily freshness levels. The system also detects rodent activity in each zone and triggers real-time alerts. It achieves an error rate of 0.069% for tomato health prediction and 0.045% for fungus detection, with a DewMQTT throughput of 3.76 messages/second. Results demonstrate high accuracy, adaptability, and scalability across multiple zones and warehouses, outperforming existing post-harvest storage solutions in enhancing the shelf life of stored tomatoes.