Acoustic-based Stress Monitoring and Automated Water Delivery System

This paper presents an intelligent and automated system designed to monitor plant stress using acoustic and environmental parameters, supported by an enhanced Convolutional Neural Network (CNN) combined with a Bidirectional Long Short-Term Memory (Bi-LSTM) predictive model. The system employs a microphone sensor to detect acoustic stress patterns and a DHT11 sensor to measure ambient temperature and humidity. These data streams are processed by an Arduino microcontroller and transmitted to a Python-based server. The deep learning model analyzes the incoming time-series data, capturing both spatial and temporal dependencies, to predict the plant’s health status as either Healthy or Unhealthy with improved accuracy. Upon detecting stress, the system automatically activates a water pump for irrigation. A visually appealing web dashboard provides real-time visualization of sensor readings and model predictions. This approach demonstrates the potential of combining IoT, acoustic sensing, and advanced machine learning techniques for efficient and intelligent agricultural automation.