License Plate Recognition and Vehicle Tracking System

This project is designed to detect vehicles, recognize license plates, and track vehicles using a combination of YOLO object detection and DeepSORT tracking. The system is capable of recognizing Swedish license plates and fetching relevant vehicle data from an external source.

demo.mp4

Features

• Vehicle Detection: Detects cars, motorcycles, buses, and trucks using the YOLOv8 model.
• License Plate Recognition: Recognizes license plates using OCR and verifies Swedish license plates.
• Vehicle Tracking: Tracks vehicles across frames using the DeepSORT tracking algorithm.
• Data Fetching: Fetches vehicle-related data from a website based on the recognized license plate.

How it Works

1. Vehicle Detection: The YOLO model detects vehicles in the video feed.
2. Tracking: The detected vehicles are tracked with unique IDs.
3. License Plate Detection: The system detects license plates within the vehicle bounding boxes.
4. License Plate Recognition: OCR is used to recognize and validate the text on license plates.
5. Data Fetching: If a valid license plate is found, vehicle information and owner is fetched from Biluppgifter.se and Ratsit.

Dependencies

• ultralytics: For YOLO-based object detection.
• cv2: OpenCV for video processing.
• deep_sort_realtime: For vehicle tracking.
• paddleocr: For license plate text recognition.
• requests & BeautifulSoup: For scraping vehicle data from a website.

Running the Project

1. Follow the steps bellow or use INSTALL.bat to install automatically.
2. Install python (3.10.0 to 3.10.15)
3. Download and install Cuda 11.7
4. Install Torch

   Option 1 (GPU)

   pip3 install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/test/cu118

   Option 2 (CPU)

   pip3 install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/test/cpu

5. (If you selected Option 1) Install paddlepaddle-gpu:

   python -m pip install paddlepaddle-gpu==2.6.1.post117 -f https://www.paddlepaddle.org.cn/whl/windows/mkl/avx/stable.html

6. Install dependencies:

   pip install -r requirements.txt

7. Run the main script to process a video file:

   python main.py

Files

• main.py: Core logic for detecting, tracking, and recognizing vehicles and license plates.
• utils.py: Helper functions for OCR, data fetching, and license plate validation.

License

This project is licensed under MPL 2.0.