Automated number plate scanning system for road crime detection

Abstract

This report presents the design, development, and implementation of an Automatic Number Plate Recognition (ANPR) system aimed at enhancing road offense detection and improving traffic law enforcement in Uganda. The system leverages computer vision, Optical Character Recognition (OCR), and real-time data processing to automatically capture, identify, and verify vehicle number plates at strategic checkpoints or traffic stops. The primary objective of the project is to reduce manual inefficiencies in traffic policing, minimize human error, and provide a reliable digital platform for tracking road offenses. The ANPR system is built using Python, OpenCV, and Easy OCR for image processing and plate recognition, with a SQLite/Postgres database backend for secure data storage and retrieval. A web-based interface developed using Flask allows administrators and law enforcement officers to manage user accounts, monitor offenses, and generate reports. Data collected from vehicle owners, police officers, and regulatory bodies was analyzed to understand user expectations and system requirements. The system also incorporates security protocols to safeguard sensitive data and supports real-time notifications and reporting features to aid timely decision-making by authorities. While the system demonstrates significant potential in streamlining traffic enforcement, challenges such as environmental factors, privacy concerns, and technical limitations are acknowledged. Recommendations for future work include integration with national vehicle registration databases, enhanced AI-driven recognition models, and mobile-based deployment. This project contributes to ongoing efforts in digital governance, public safety, and smart transportation by offering a scalable and efficient solution to address road discipline and automate fine management in Uganda.