fahad-h36 / alpr-pakistan

Automated License Plate Recognition, specific to the Pakistani Number Plates

Real-Time ALPR (Automatic License Plate Recognition)

Real-Time ALPR (Automatic License Plate Recognition) is a project aimed at detecting and recognizing license plates in real-time video streams. It utilizes YOLOv8 for license plate detection and manga-ocr for optical character recognition (OCR) to extract the license plate number.

Features

• Real-time detection and recognition of license plates in video streams.
• Utilizes YOLOv8 for license plate detection.
• Uses manga-ocr for optical character recognition (OCR) to extract license plate numbers.
• Supports detection of single license plates.
• Specifically designed for 6-digit license plate numbers.

Requirements

• Python 3.x
• OpenCV
• ultralytics
• manga-ocr

Installation

1. Clone this repository:

   git clone https://github.com/Fahad-H36/ALPR-Pakistan.git

2. Install dependencies:

   pip install -r requirements.txt

Usage

Using a Video File

1. Prepare your video for processing.

2. Run the main script:

   python main.py --video path_to_your_video.mp4

   Replace path_to_your_video.mp4 with the path to your video file.

   The script will process the video, detect license plates, and recognize the plate numbers in real-time.

Using a Camera

1. Run the main script with the --mode c option:

   python main.py --mode c --cam <camera_number> --video "./"

   Use --cam only if you have multiple cameras connected.
   
   This will start capturing video from your camera, detect license plates, and recognize the plate numbers in real-time.

Output

The output will be an annotated video with detected license plates and recognized plate numbers. You can find the output video in the output directory.

Example

Acknowledgments

• Roboflow - ALPR Dataset for providing the dataset used to train YOLOv8.
• manga-ocr by kha-white

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Contact

Fahad Soomro - [email protected]