Forest Fire Detection System

Overview

This repository contains an IoT-enabled Forest Fire Detection System designed to monitor environmental conditions and predict fire risk in real time. It includes a Flutter-based mobile application, sensor data fetching using NodeMCU, and a basic machine learning model - xgboost for fire prediction.

The system monitors temperature, humidity, and smoke levels, and visualizes the data in a user-friendly mobile interface.

---

Key Features

• Real-Time Monitoring: Tracks temperature, humidity, and smoke levels from sensors.
• Mobile Dashboard: Flutter app displays live data and predictions in a clean UI.
• ML-Based Prediction: XGBoost model trained to classify whether sensor readings indicate fire.
• IoT Integration: NodeMCU fetches and transmits data using Wi-Fi.
• Database Storage: Sensor values and predictions are stored in a MySQL database.
• Resident Alert System: Button in the app to notify nearby residents with safety instructions.
• Expandable: Future integration with AR/VR fire spread simulations and remote server dashboards.

---

Technologies Used

• Flutter: For the mobile UI.
• NodeMCU: For data collection from DHT11, MQ-2 sensors, etc.
• Python & jupyter: For training and running the ML model.
• MySQL: For storing and syncing sensor data.
• C and Arduino IDE: For programming microcontrollers.

---

Tech Stack

Technology	Purpose
Flutter	Mobile app for real-time monitoring
NodeMCU	IoT microcontroller for sensor interfacing
MQ-2 Sensor	Smoke detection
DHT11 Sensor	Temperature and humidity measurement
Python	Machine learning model development
XGBoost	ML model for fire risk classification
MySQL	Backend database
C (Arduino IDE)	Microcontroller programming

---

Pre-requisites

To run this project, ensure you have the following installed:

• Flutter SDK (latest version)
• Python 3.7+
• Arduino IDE (for NodeMCU)
• Hardware: DHT11, MQ-2, NodeMCU, Breadboard, Jumper wires
• Database: MySQL account for backend data storage

---

How It Works

1. Sensor Layer
   The DHT11 and MQ-2 sensors measure environmental conditions and transmit data via NodeMCU.

2. Data Transmission
   NodeMCU sends sensor readings over Wi-Fi to a backend server (or directly to the mobile app via Firebase/MySQL).

3. Prediction Engine
   The ML model (XGBoost) predicts fire risk using the incoming data.

4. User Interface
   A Flutter-based mobile app displays:

   • Live data (temperature, humidity, smoke)
   • Fire prediction result
   • Alert functionality for residents

---

Machine Learning Details

• Algorithm: XGBoost Classifier
• Input Features: Temperature, Humidity, Smoke Levels
• Training Dataset: Custom collected sensor readings under fire and non-fire conditions
• Accuracy: ~96.4%
• Evaluation Metrics: Precision, Recall, F1-score, Confusion Matrix

The model is lightweight and optimized for fast predictions on limited hardware setups.

---

Installation and Setup

---

Demo of the Project
https://github.com/Bevinaa/Forest-Fire-Detection-System/demo.mp4

---

Screenshots

Mobile Application	Fire Prediction

---

Analytics Dashboard

---

Contact

Author: Bevina R
Email: bevina2110@gmail.com
GitHub: Bevinaa

---