Traffic Flow Analysis

Real-time vehicle detection and traffic parameter analysis using YOLOv8 and Flask.

Features

• Vehicle Detection: YOLOv8-based detection for cars, trucks, buses, and motorcycles
• Speed Tracking: Calculates vehicle speeds using dual detection lines
• Traffic Metrics: Real-time computation of density, flow rate, headway, and vehicle counts
• Live Dashboard: Web-based UI with streaming video feed and metrics updates
• Directional Counting: Tracks vehicles moving up and down

Tech Stack

• Detection: YOLOv8 (Ultralytics)
• Backend: Flask
• Processing: OpenCV, NumPy
• Frontend: Vanilla JavaScript, CSS

Installation

Using uv (recommended)

uv sync
``` Static assets in `static/`, templates in `templates/`
- Video files and YOLO models in root directory


### Using pip

```bash
pip install -e .

Usage

1. Start the Flask server:

python main.py

2. Open your browser to http://localhost:5000

3. The system will:

   • Load the YOLOv8 model (yolov8n.pt)
   • Process the video file (video3.mp4)
   • Stream detection results to the web interface
   • Update metrics every second

Configuration

Edit constants in main.py:

VIDEO_PATH = "video3.mp4"          # Video file path
MODEL_PATH = "yolov8n.pt"          # YOLO model (yolov8n.pt or yolov8s.pt)
FRAME_WIDTH = 1020                 # Processing width
FRAME_HEIGHT = 500                 # Processing height
ROAD_LENGTH = 100                  # Road segment length (meters)
DISTANCE_BETWEEN_LINES = 10        # Distance between detection lines (meters)

Project Structure

PythonTrafficCongestion/
├── main.py              # Flask app with YOLOv8 detection
├── static/
│   ├── script.js        # Frontend JavaScript
│   └── style.css        # UI styles
├── templates/
│   └── index.html       # Web interface
├── video3.mp4           # Input video
├── yolov8n.pt           # YOLO model weights
└── pyproject.toml       # Dependencies

Metrics Explained

• Vehicles: Current vehicle count in frame
• Speed: Average speed (km/h) from recent measurements
• Accuracy: Mean detection confidence score
• FPS: Processing frame rate
• Density: Vehicles per meter (vehicles/road_length)
• Flow Rate: Vehicle throughput
• Headway: Average spacing between vehicles (meters)
• Up/Down Count: Directional vehicle counts

Requirements

• Python 3.12+
• OpenCV
• Flask
• Ultralytics YOLOv8
• NumPy

License

MIT