🤖

AgroDew

Flutter
CNN
TFLite
IOT

🌾 AgroDew: Smart Irrigation System for Farmers

AgroDew is an intelligent irrigation system designed for small and mid-scale farmers. It enables remote control of water sprinklers via a Flutter-based mobile app, provides real-time weather updates, schedules irrigation, detects plant diseases using a supervised machine learning model, and supports offline functionality through SMS commands using GSM modules ensuring usability even in low-connectivity rural areas. The system integrates IoT (ESP32, sensors, relays), TFLite for on-device ML, and MQTT/REST protocols for seamless communication. AgroDew helps save water, reduce crop loss, and improve farming efficiency.

🚀 Features

  • Weather Monitoring:
    • Displays real-time weather updates using a WeatherAPI
    • Integrates location services (geolocator, geocoding packages) to auto-fetch weather for current farm location.
  • Irrigation Control System:
    • Realtime Communication: Firebase Realtime Database
  • Plant Disease Detection (ML Integration):
    • Integrates a Supervised Machine Learning model trained on image datasets (e.g., PlantVillage).
    • Model built using TensorFlow Lite (TFLite) for on-device inference.
  • Offline SMS Control Mode:
    • Implements SMS-based command structure using:
      • sms Flutter Plugin
      • Backend microcontroller with GSM module (SIM900A)
    • Commands like IRRIGATE_ON, IRRIGATE_OFF, STATUS, parsed and executed locally on the device.

🔌 Hardware & Embedded System (IoT)

  • Microcontroller: Arduino UNO with WiFi/GSM module support
  • Sensors:
    • Soil moisture sensor (capacitive or resistive)
    • DHT11/DHT22 for temperature & humidity
  • Communication Modules:
    • WiFi (ESP32) for online control
    • GSM (SIM900A) for offline/SMS control
  • Relay module: To control high-voltage irrigation motor/pump

🧠 Machine Learning (Disease Detection)

  • Dataset: PlantVillage or custom labeled dataset of leaf images
  • Model: CNN (Convolutional Neural Network) using:
    • TensorFlow / Keras for training
    • Exported and quantized to TensorFlow Lite for mobile deployment

🛰 Key Advantages

  • Works in remote areas with no internet through GSM-based SMS fallback.
  • Supports automated and scheduled irrigation, reducing water wastage.
  • Provides plant disease alerts, reducing crop loss via early detection.
  • Built using cross-platform Flutter to ensure reach across low-end devices.

📜 License

This project is part of a Freelancing.