OCR using deep learning
Image is first cropped to the size of name plate with the help of first model trained on yolo about 2000 iteration on dataset of 233 images and then it will be predicting the character with the help of second model which is custom model trained on 20628 images dataset.
Datasets for both the models sourced from kaggle.
https://www.kaggle.com/datasets/andrewmvd/car-plate-detection https://www.kaggle.com/datasets/preatcher/standard-ocr-dataset