A lightweight, OpenCV based tool for Number Plate Recognition (ANPR) and Reading. This project uses YOLOv3 for vehicle/plate detection and Tesseract OCR for character recognition.
- Vehicle Detection: Uses YOLOv3 (Darknet) to identify vehicles and plates.
- Image Processing: Adjustable pipeline for grayscale conversion, noise reduction, thresholding and other preprocessings.
- OCR engine: Tesseract configuration for license plate alphanumeric strings.
- Web Interface: Easy-to-use Flask UI for uploading and processing images.
- Python 3.12
- Flask (Web Framework)
- OpenCV (Image Processing)
- PyTesseract (OCR Engine)
- PDM (Package & Dependency Manager)
- Install Python 3.12
- Install PDM
- Install Tesseract OCR
- Install Project dependancies:
pdm install
This project utilizes the YOLOv3 object detection architecture.
- Model Architecture: YOLOv3 (You Only Look Once) by Joseph Redmon and Ali Farhadi.
- Framework: Darknet, an open-source neural network framework written in C and CUDA.
- Weights & Config: These parameters are based on the original pre-trained weights released into the Public Domain by the authors.
Special thanks to the open-source community for maintaining various implementations of the Darknet weights and configurations.