Vehicle Detection System

A real-time vehicle detection system using YOLOv5 and Streamlit for a user-friendly interface.

Screenshots

Detection Results

User Interface

Training Progress

Setup Instructions

Prerequisites

• Python 3.8 or higher
• CUDA-capable GPU (recommended)
• Git

Installation Steps

1. Clone the repository:

git clone [repository-url]
cd vehicle-detection-system

2. Create and activate virtual environment:

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

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

3. Install dependencies:

pip install -r requirements.txt

4. Download YOLOv5:

git clone https://github.com/ultralytics/yolov5
cd yolov5
pip install -r requirements.txt
cd ..

Dataset Preparation

1. Place your dataset in the project root directory. The expected structure is a folder named trafic_data containing train and valid subdirectories, each with images and labels folders.

2. Run the label conversion script to convert original labels to YOLO format with the correct class mapping:

python convert_labels.py

Running the Application

1. Train the YOLOv5 model:

python yolov5/train.py --img 480 --batch 24 --epochs 30 --data data/dataset.yaml --weights yolov5/yolov5n.pt --workers 4 --cache

(You can adjust parameters like --epochs and --batch as needed)

2. Start the Streamlit app:

streamlit run app.py

3. Open your browser and navigate to the provided local URL (usually http://localhost:8501).

4. Upload an image and adjust the confidence threshold for detection.

Model Choice and Approach

Why YOLOv5?

• Speed: Real-time detection capabilities
• Accuracy: State-of-the-art object detection
• Efficiency: Lightweight model (YOLOv5n) for faster processing
• Flexibility: Easy to train and fine-tune

Model Architecture

• Base Model: YOLOv5n (nano)
• Input Size: 480x480 pixels
• Classes:
  • Car
  • Truck
  • Bus
  • Ambulance
  • Motorbike
  • Bicycle
• Training Parameters:
  • Batch Size: 24
  • Epochs: 30
  • Learning Rate: Adaptive
  • Optimizer: SGD with momentum

Detection Approach

1. Image Preprocessing:

   • Resize to 480x480
   • Normalize pixel values
   • Convert to tensor format

2. Feature Extraction:

   • Convolutional layers
   • Feature pyramid network
   • Multi-scale detection

3. Post-processing:

   • Non-maximum suppression
   • Confidence thresholding
   • Bounding box refinement

Limitations and Assumptions

Technical Limitations

1. Performance:

   • CPU-only mode is slower
   • Minimum 4GB RAM required
   • GPU recommended for real-time detection

2. Detection Accuracy:

   • Works best with clear, well-lit images
   • May struggle with:
     • Extreme weather conditions
     • Poor lighting
     • Occluded vehicles
     • Very small vehicles

3. Image Requirements:

   • Minimum resolution: 480x480 pixels
   • Supported formats: JPG, JPEG, PNG
   • Maximum file size: 10MB

Assumptions

1. Vehicle Types:

   • Trained to detect: Car, Truck, Bus, Ambulance, Motorbike, Bicycle
   • Assumes vehicles are reasonably visible

2. Environment:

   • Standard road/street scenes
   • Moderate to good lighting conditions
   • Vehicles are not heavily occluded

3. Usage:

   • Single image processing
   • Static images (not video)
   • User can adjust confidence threshold

Observations and Notes

Performance Metrics

(Training metrics will be updated after training completion)

Known Issues

• Initial training attempts encountered issues with incorrect class IDs in labels, resolved by cleaning the conversion script and updating dataset paths.
• Performance may be limited on systems without a dedicated GPU.

Future Improvements

1. Model Enhancements:

   • Improve small object detection
   • Better handling of occlusions

2. Features:

   • Video processing support
   • Real-time tracking
   • Multiple model options

Project Structure

vehicle-detection-system/
├── app.py                 # Streamlit application
├── data/
│   ├── dataset.yaml      # Dataset configuration
│   ├── images/           # Training and validation images (copied from trafic_data)
│   └── labels/           # Converted training and validation labels
├── yolov5/               # YOLOv5 framework
├── requirements.txt      # Project dependencies
├── convert_labels.py      # Script to convert original labels to YOLO format
├── .gitignore            # Specifies intentionally untracked files
└── README.md            # This file

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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