IoT-based control and monitoring system for hydroponic plant growth using image processing and mobile applications

Abstract

The HydroFarm project presents an innovative IoT-based control and monitoring system for hydroponic plant growth, integrating advanced image processing techniques and mobile applications to enhance urban farming practices. This system addresses critical challenges faced by urban farmers, such as limited space and the need for precise environmental management. By employing a comprehensive approach that combines various sensors (DHT22, DS18B20, pH, TDS) with an ESP32 microcontroller, HydroFarm enables real-time monitoring of essential parameters like temperature, humidity, pH, and nutrient levels. A significant novelty of this project lies in its use of a convolutional neural network (CNN) for plant health assessment through image processing. This technique allows for accurate detection of plant conditions, categorizing leaves as healthy or unhealthy based on visual data captured via a mobile application. The application, developed in Kotlin, not only facilitates user interaction but also provides automated and manual control over nutrient delivery systems based on real-time sensor data. Testing results indicate that the HydroFarm system achieves a high accuracy rate of 96% in detecting plant health conditions, with the sensors providing accurate and consistent data to maintain effective control over hydroponic parameters. The system usability scale (SUS) evaluation yielded an impressive score of 81.875, categorizing the application as excellent and user-friendly. Overall, HydroFarm represents a significant advancement in hydroponic farming technology by integrating IoT capabilities with deep learning for enhanced decision-making and operational efficiency in urban agriculture. The findings underscore the potential for scaling this model to improve food security and promote sustainable agricultural practices in densely populated areas.

Plant phenotyping relevance

CNNによる葉の健康状態推定をシステムの中心的機能として開発・評価しており、植物状態の画像ベース計測が中心的である。

Code and data availability

The authors publicly deposited the HydroFarm plant health image dataset (labeled sehat/tidak sehat leaf images for CNN/VGG16 classification) on figshare with an explicit DOI. Supplemental code files exist but only via article supplement links, not an allowed public URL.