AI-Driven Self-Adaptive Smart Irrigation System Using IoT, Computer Vision and Machine Learning

Abstract

This paper presents an AI-driven self-adaptive irrigation framework integrating IoT sensing, computer vision, adaptive learning, and automated irrigation control. The proposed system continuously monitors soil moisture, temperature, humidity, water flow, and visual plant health conditions using ESP32-based sensing devices and ESP32-CAM image acquisition. Centralised FastAPI server performs feature extraction, environmental analysis and AI based irrigation decision making at the backend. The proposed system, compared with existing threshold-based conventional irrigation systems, dynamically adjusts irrigation behaviour based on feedback received from the environment and plant stress conditions.