Automatic Number Plate Recognition

This is our solution repo hosting the code and approach used for Mosaic'21 PS2.

With the ever-increasing need for transportation globally, the amount of vehicles on the road is piling everyday. The need for automatic traffic monitoring has never been so high. Your task is to implement an automatic number plate recognizer in an unconstrained condition that considers occlusion, poor quality of images, and other spatial variations in image data.

Approach & Optimizations

The whole pipeline is a 3 stage process:

• License Plate Detection from unconstrained environment images using YOLOv5
• Segmentation of Characters from cropped license plate image using openCV
• character Recognition of Individual contours using a CNN network

A YOLOv5 is used to predict the bounding box around License plate. Then the cropped license plate image is processed using:

• noise removal, extra lines filtering, Adaptive histogram equalization (CLAHE)
• a four point perspective transform to handle different viewing angles

A rule based character segmentation is performed using contours. We add several checks over expected dimensions of charcter blobs to select only those contours that are characters. Then the segmented characters are passed through a CNN classifier for recognition

Run Locally

Clone the project

  git clone https://github.com/arch-raven/Automatic-Number-Plate-Recognition.git

Go to the project directory

  cd Automatic-Number-Plate-Recognition

Install dependencies

  pip install -r requirements.txt

Make predictions

  python src/main.py --image_path path/to/licensePlate

Procedure

Training Metrics & Snapshots

Acknowledgements

• Ultralytics YOLOv5

Our Team:

Lavish Bansal

Aditya Kumar

Ayush Singh