"Time is the only resource we cannot recover." > WildfireGuard AI leverages Deep Learning to detect early signs of forest fires from video feeds, enabling rapid response and environmental protection.

WildfireGuard AI is a real-time computer vision application designed to analyze video streams for smoke signatures and thermal anomalies indicative of forest fires. Built with Streamlit and TensorFlow, it features a responsive "Command Center" UI with simulated telemetry and live inference logging.

• Real-Time Inference: Frame-by-frame analysis using a custom-trained CNN (Convolutional Neural Network).
• Dynamic Confidence Control: Adjustable sensitivity threshold via the sidebar to reduce false positives.
• Immersive UI: Custom CSS styling with a futuristic, dark-mode aesthetic (Neon Red/Orange).
• Live Telemetry: Simulated dashboard metrics for wind, ping, and area coverage.
• System Logs: Scrolling terminal-style logs tracking detection events and timestamps.
• Safety Mechanisms: Automatic temporary file cleanup and user-controlled stop functionality.

• Core Logic: Python 3.10
• Frontend: Streamlit (Custom CSS injected)
• Computer Vision: OpenCV (cv2)
• Deep Learning: TensorFlow / Keras
• Data Manipulation: NumPy

git clone [https://github.com/Safae26/WildfireGuard-AI.git](https://github.com/Safae26/WildfireGuard-AI.git)
cd WildfireGuard-AI

# Windows
python -m venv venv
venv\Scripts\activate

# Mac/Linux
python3 -m venv venv
source venv/bin/activate

pip install -r requirements.txt

Note: Ensure you have the forest_fire.keras model file in the root directory.

streamlit run app.py

WildfireGuard-AI/
├── app.py                # Main application entry point (Streamlit)
├── =wildfire_detection_model.keras     # Trained Deep Learning Model (Required)
├── requirements.txt      # Python dependencies
├── README.md             # Project documentation=

The system utilizes a Convolutional Neural Network (CNN) trained on a dataset of satellite forest fire imagery. The model performs binary classification:

• Preprocessing: Frames are resized to (350, 350) and normalized.
• Inference: The model outputs a probability score (0.0 - 1.0).
• Thresholding: If Probability > Threshold (default 0.5), the system triggers a CRITICAL ALERT.

• Integration with live satellite API feeds.
• GPS coordinate mapping for detected fires.
• SMS/Email alert notifications via Twilio or SMTP.
• YOLOv8 implementation for bounding-box localization.

Contributions are welcome! Please open an issue or submit a pull request for any improvements. Fork the Project Create your Feature Branch (git checkout -b feature/NewFeature) Commit your Changes (git commit -m 'Add some NewFeature') Push to the Branch (git push origin feature/NewFeature) Open a Pull Request

Safae Data Science & AI Student | Full-Stack Developer

Built with ❤️ and ☕ by Safae.