Characters Segmentation and Recognition for Vehicle License Plate

Characters segmentation and recognition using OpenCV and deep learning

Prerequisites

• Python 3.5.2
• Keras 2.2.4
• Tensorflow 1.13.1
• OpenCV 4.4.0
• Numpy 1.16.2

Usage

The pre-trained deep learning model used in this project is taken from my other project https://github.com/minhthangdang/CharactersRecognition. The model is too big for github, so you can download it from https://drive.google.com/file/d/1ojsMc-VYSFKromwSm4Qx3kRpPmi2klOV/view?usp=sharing. Alternatively you can train it yourself from scratch using my repository above.

Copy the pre-trained weights into the same folder of this project.

To test on an image, just run:

python plate_read.py --model characters_model.weights --image plate1.jpg

Where characters_model.weights is the pre-trained model and plate1.jpg is the test image.

The result will look like this:

Demo video (click to open youtube page):

Full tutorial is available at http://dangminhthang.com/computer-vision/characters-segmentation-and-recognition-for-vehicle-license-plate/