License Plate Detection

Overview

This project enables accurate detection and recognition of license plates from video and image data. By leveraging YOLO for object detection and PaddleOCR for optical character recognition (OCR), the system can locate and interpret license plates efficiently. To enhance video processing reliability, FFmpeg is used to handle video inputs, addressing certain limitations encountered when using OpenCV alone.

Workflow

The License Plate Recognition Project is a robust and efficient system designed for detecting and reading license plates in images or from recorded video. This project combines:

• YOLOV8n (You Only Look Once) for fast and accurate license plate detection.
• PaddleOCR for reliable text recognition, supporting multiple languages.
• FFmpeg re-encode the video cause OpenCV and skvideo libraries, which you likely used to generate or manipulate the video, may encode MP4 files using codecs that are not fully supported by the HTML5 video player TML5 video players, which Streamlit uses to display videos, only support certain codecs reliably H.264 for video

Componenet

• best.pt: model weights YOLOv8n
• Deployment.py: model deployment
• pp.pt: functions

Installation

To run this project, install the required dependencies:

pip install -r requirements.txt

To run Deployment:

streamlit run Deployment.py