Iranian Plate Recognizer

This repository contains a Python script for recognizing vehicle license plates using YOLO (You Only Look Once) object detection models. It consists of two main parts:

1. Plate Detection: Detects license plates in images.
2. Character Recognition: Recognizes characters on the detected license plates.

Requirements

Ensure that you have the required dependencies installed by running:

pip install -r requirements.txt

The required packages include:

• opencv-python
• torch
• ultralytics
• numpy

Files

• plate_recognizer.py: The main Python script for plate recognition.
• weights/: Folder containing the YOLO model weights files:
  • best.pt: Model for plate detection.
  • yolov8n_char_new.pt: Model for character recognition.
• parser.py: Helper script to get the paths of the model weights.

Usage

1. Place the image(s) with vehicle license plates in the same directory.
2. Run the script with the image file as input.

python plate_recognizer.py

Example Usage:

from plate_recognizer import PlateRecognizer

# Initialize the recognizer with a desired confidence threshold
recognizer = PlateRecognizer(conf_threshold=0.4)

# Recognize the plate from an image
plate_number, bbox = recognizer.recognize_plate('car.jpg')

if plate_number:
    print(f'Recognized plate number: {plate_number}')
else:
    print('No plate detected.')

For Test

input :

output:

License

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

---