IoT-based small scale hydroponics system and lettuce disease and harvestability detection using yolov8

ABSTRACT: This study presents the development of an IoT-based hydroponics system integrated with artificial intelligence to automate the monitoring and analysis of lettuce health and growth status. Traditional methods of detecting plant diseases and determining harvest readiness rely heavily on manual inspection, which can be time-consuming, inconsistent, and prone to human error. To address these challenges, the researchers developed a smart system that combines real-time sensor monitoring with AI-powered image detection to improve accuracy and efficiency in crop management. The hardware setup includes an Arduino Uno connected to multiple sensors for measuring pH, temperature, humidity, and TDS levels, with an ESP8266 module handling wireless communication and relay control for managing pumps, lights, and essential environmental systems. The sensors achieved a margin of error ranging from 0.01 to 0.01, ensuring precise and reliable environmental monitoring. A Django-based web application was developed to remotely upload images of lettuce and monitor sensor data in real-time. The AI component utilizes YOLOv8 deep learning models to analyze images and accurately detect lettuce such as downy mildew, septoria blight, and viral infections, as well as determine harvest readiness. The models achieved a mean Average Precision (mPA@50) ranging from 0.6 to 0.9 across different classes and test conditions. The system was tested in a controlled hydroponic environment and demonstrated reliable performance in identifying plant health issues and optimizing harvesting decisions. By integrating machine learning with IoT, the project significantly improved decision-making, reduced manual labor, and highlighted the potential of intelligent automation in enhancing productivity and sustainability in modern agricultural systems.

Filipiniana