Detection and Recognition of License Plates Using Deep Learning-Based Technology and Flask

Résumé fourni par la source

Rear-view camera implementation for 2-wheelers helps people to see the vehicles that are coming from the backside without turning back. The Raspberry Pi camera is fixed to the back of the two-wheeler so that any rear-side collisions are captured and subsequently helpful to the victim. Implementing the detection of a license plate system makes it easier to take out the license plate from the captured images. However, automatic number-plate recognition systems when there are challenges, must have numerous vehicles in a single image and different lighting conditions. Additional challenges are shown in the Indian context by the different shapes and sizes of the characters on license plates. The study confronts challenges in license plate detection and recognition via YOLOv5, addressing varying plate sizes, orientations, and language diversity. This approach elevates accuracy, robustness, and resource efficiency, pushing forward automatic number plate recognition systems. This work suggests using the Inception Resnet version 2(v2) model to classify license plates from input photos to address these problems. The study suggests using the yolo-v5 model, one of the best architectures currently available to detect numerous objects simultaneously, which leads to faster and more accurate detection. The Tesseract OCR (Optical character recognition). This OCR later recognizes each character on the license plate. Implementing the Flask App makes it possible to obtain the characters from the license plate. The experimental data reported in the study supports the feasibility of the suggested work.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.