NComputer Vision

Licence Plate Recognition Mobile App

Licence Plates Recognition (LPR) project is the ecosystem of 4 secure iOS and Android mobile apps with embedded Computer Vision models and API & back-end for apps management and data collection.

Aim of the Project

● Secure mobile native apps for licence plate scanning
● Scanning of license plates on-real-time without Internet
● Matching licence plates with black-lists

Results Achieved

● Secure native applications with different access levels
● Real-time on-device LPR for photo and video modes
● Immediate matching of LPR with black-lists and notifications

Main Technologies

● AI/CV/ML: OpenCV, TensorFlow Lite, C/C++
● API & web-services: Python, PostgreSQL, Docker
● Mobile: Swift, C++ for iOS and Kotlin/Java, C++ for Android
● UI/UX: Figma

Application UI/UX

The UI/UX design was developed by APRO:

What Apps Can Do

Apps, depending on their access levels (Lite or Pro), can do:

● LPR using embedded local models with performance up to 99.99+%!
● LPR by taking photos or directly in the video stream.
● Apps can collect pictures of cars, geolocation of the picture, car licence plate number, colour, manufacturer, and even model.
● Images & video corrections: apps have features to adjust image processing for environmental conditions via brightness, contrast, and zoom fitting.
● Apps have secure access: login and password with FaceID or TouchID checks
● Work without Internet (or with poor Internet connection) keeping 100% functions:
○ uses predefined black-lists
○ use secure storage for information until the synchronization with the server
● Secure work with data: apps use secure internal storage with automatic cleaning up (even if apps are in the background). A special algorithm ensures that all data is synchronized with the server before removal even if there are Internet issues.
● Attacks detection: apps can analyze their status and user activities to detect potential attacks, hooking, malware debugging and so on.

Security audit

● APRO has successfully passed the Security Audit by THALES Group with the project.