Smart Traffic Control System

An AI-powered traffic management system using computer vision to detect vehicles and dynamically control traffic signals.

Features

• Real-time vehicle detection and counting using YOLOv8
• Dynamic traffic light timing based on vehicle density
• Emergency vehicle priority handling
• Multi-lane traffic management
• Visual overlay showing lane boundaries and vehicle statistics
• Configurable camera input (webcam or IP camera)

Requirements

• Python 3.8+
• OpenCV
• Ultralytics YOLOv8
• NumPy
• PyYAML
• A camera source (laptop webcam or smartphone camera via IP camera app)

Installation

1. Clone this repository:

git clone https://github.com/yourusername/smart-traffic-control-system.git
cd smart-traffic-control-system

2. Install the required dependencies:

pip install -r requirements.txt

3. The YOLOv8 model will be automatically downloaded when you first run the system. If you prefer to download it manually, get yolov8n.pt from the Ultralytics repository and place it in the model directory.

Configuration

All settings are stored in config.yaml:

• Camera Settings: Change the camera source, resolution, and FPS
• YOLOv8 Settings: Adjust model path, confidence threshold, and target classes
• Traffic Light Timings: Modify min/max durations and scaling factors
• Lane Definitions: Update the polygon coordinates to match your camera view

Camera Configuration

You can use either:

1. Your laptop webcam: Set camera.source: 0 in the config file
2. A smartphone IP camera app:
   • Install an IP camera app (like IP Webcam for Android or EpocCam for iOS)
   • Set camera.source: "http://your-phone-ip:port/video" in the config file

Lane Configuration

The system uses polygon coordinates to define lanes. Update the coordinates in config.yaml to match your camera view:

lanes:
  left:
    - [100, 480]  # Bottom-left point
    - [280, 100]  # Top-left point
    - [360, 100]  # Top-right point
    - [320, 480]  # Bottom-right point
  right:
    - [320, 480]  # Bottom-left point
    - [360, 100]  # Top-left point
    - [440, 100]  # Top-right point
    - [540, 480]  # Bottom-right point

Usage

Run the system with:

python main.py

Or specify a custom config file:

python main.py --config my_custom_config.yaml

Controls

• Press 'q' to exit the program

How It Works

1. Camera Input: Frames are captured from your specified camera source
2. Vehicle Detection: YOLOv8 detects vehicles in each frame
3. Lane Assignment: Each detected vehicle is assigned to a lane based on its center point
4. Traffic Light Logic:
   • Green light duration is calculated based on vehicle count
   • Emergency vehicles trigger an immediate light change (after minimum green time)
   • Yellow light provides a transition period before switching
5. Visualization: The system displays lane overlays, vehicle counts, and traffic light status

Optimization Features

• Frame skipping for better performance
• Multithreading separates detection from visualization
• Moving average smoothing for vehicle counts
• Confidence threshold filtering
• Frame resizing before inference

Troubleshooting

• Camera Access Issues: Ensure your camera is not being used by another application
• Model Loading Errors: Check that the model file exists in the specified location
• Performance Issues: Adjust the frame skip and resize values in the config to reduce processing load

License

This project is licensed under the MIT License - see the LICENSE file for details.