Acs

CCTV Accident Detection with YOLO and Streamlit

This project is a CCTV accident detection application using the YOLO model, Streamlit, and Google Generative AI. The app analyzes video footage to detect accidents, generates detailed reports, and provides a user-friendly interface for viewing and managing accident reports.

Features

<video controls autoplay src="https://cdn-uploads.huggingface.co/production/uploads/671903e9916f48b2c41863b1/3gAwk7SAp9-9vNLMphOaa.mp4"></video>

• Real-time Accident Detection: Uses the YOLO model to detect accidents in video footage.
• Video Annotation: Annotates the video with detected objects and displays it in a Streamlit app.
• Automated Report Generation: Generates and saves accident reports with details like class, confidence score, and timestamp.
• Accident Description: Uses Google Generative AI to generate a professional accident description based on the detected incident.
• Report Management: Displays saved reports and associated images, with an option to generate or view detailed descriptions.

Project Structure

Prerequisites

• Python 3.7+
• best_model_without_freeze.pt (YOLO model weights)
• Google API key for Generative AI (stored in a .env file)

Installation

1. Clone the repository:

   git clone https://github.com/your_username/cctv-accident-detection.git
   cd cctv-accident-detection
   

2. Install the required libraries:

   pip install -r requirements.txt
   

3. Set up the .env file: Create a .env file in the project root and add your Google API key:

   GOOGLE_API_KEY=your_api_key_here
   

4. Download the YOLO model weights: Place the best_model_without_freeze.pt file in the project root.

Running the Application

1. Start the Streamlit app:

   streamlit run main.py
   

2. Navigate to the app: Open your browser and go to http://localhost:8501.

Usage

1. Video & Prediction Page:

   • Upload a video file to analyze.
   • The app will display the annotated video and detect accidents.
   • If an accident is detected, a report is generated and saved.

2. Reports Page:

   • View saved accident reports with details like class, confidence, and timestamp.
   • See associated images for each report.
   • Generate or view accident descriptions using Google Generative AI.

Example

• Detected Accident Report:

  Class: Accident
  Confidence: 0.85
  Time: 2024-10-13_14-30-00
  Description: Two vehicles involved, with visible damage on the front of one vehicle. Road conditions are wet, and traffic is partially blocked.
  

Dependencies

• opencv-python
• streamlit
• pandas
• ultralytics
• numpy
• google-generativeai
• python-dotenv

Future Enhancements

• Add support for different video formats.
• Implement email notifications for generated accident reports.
• Integrate with a cloud storage service for report storage.

Contributing

1. Fork the repository.
2. Create a new branch (git checkout -b feature/YourFeature).
3. Commit your changes (git commit -m 'Add some feature').
4. Push to the branch (git push origin feature/YourFeature).
5. Open a Pull Request.

License

This project is licensed under the MIT License.

Acknowledgements

• Thank you, Tuwaiq Academy for your guidance and for fostering an environment that allowed us to successfully complete this project.
• YOLO (You Only Look Once)
• Streamlit
• Google Generative AI