AI-powered automated hydroponic system for smart agriculture.

Abstract

This research presents an AI-powered automated hydroponic system designed to enhance the efficiency and sustainability of modern agriculture. The system integrates real-time environmental monitoring, automated nutrient management, and AI-based disease detection to optimize plant growth and minimize manual intervention. An ESP32 microcontroller collects data from specialized sensors measuring Total Dissolved Solids (TDS), pH, temperature, and light intensity. Data is wirelessly transmitted via MQTT to an EMQX broker, subsequently processed by an ExpressJS backend, and stored in a Firebase Realtime Database. A NextJS web application provides a user-friendly dashboard for visualization, alerts, and remote control. Automation is achieved using relay-controlled peristaltic and water pumps that adjust nutrient dosing and circulation based on sensor readings. A camera module captures plant images, which are analyzed by a CNN model running on a separate AI server to detect common spinach diseases like Anthracnose and Downy Mildew, enabling early intervention. This integrated system combines IoT, cloud data management, automation, and AI-based visual inspection to offer a comprehensive solution for precision hydroponic farming. Evaluation demonstrates high accuracy in disease detection, robust system performance, and significant potential for improving crop health, yield, and reducing manual labor in diverse agricultural settings. The system, along with its full codebase, has been made publicly available to promote reproducibility.• Automated Precision Hydroponics: Combines real-time environmental monitoring, automated nutrient management, and AI-powered disease detection for optimized spinach cultivation. • Reproducible and Scalable Method: Provides a detailed, step-by-step protocol for constructing and operating the system, adaptable to various hydroponic setups and crop types. • Sustainable and Efficient Agriculture: Minimizes resource consumption, reduces manual labour, and promotes environmentally friendly practices.

Plant phenotyping relevance

水耕栽培の統合システム全体に加え、植物画像をCNNで解析してホウレンソウ病害を検出する、再利用可能な画像ベース表現型取得機能が明示されており、方法・プラットフォームの中心的構成要素である。

Code and data availability

The paper publicly releases its full codebase (ESP32 firmware, backend/frontend servers, AI model server) and uses a public Mendeley spinach disease image dataset as the phenotyping input for its CNN disease-detection analysis. All four assets are paper-specific, public, and actionable via author-provided URLs.