What is automatic number-plate recognition?
Automatic number-plate recognition (ANPR) uses computer vision technology to read license number plates on vehicles in video streams from standard CCTV cameras. State-of-the-art deep learning algorithms provide rapid and fully automated number plate detection and recognition in real-time.
A number plate recognition system is a key component of security and transportation systems for vehicle identification. Such systems have many purposes and can be used in large-scale computer vision systems by tollway authorities, parking authorities, and law enforcement.
Features of automatic number-plate recognition (ANPR)
Deep learning based number plate recognition uses deep neural networks to read multiple number plates in real-time video streams. The visual vehicle identification method is robust and applicable for large-scale use across many locations.
- Automated vehicle detection in real-time videos from common CCTV surveillance cameras.
- Detection of multiple number plates and reading of the number plate to obtain a text string that is queried against a database.
- Deep learning models are robust to different angles, weather, and lighting conditions.
- Edge AI enables distributed on-device machine learning is fully privacy-preserving.
- Decentralized, real-time processing does not need to send visuals to the cloud and is robust to connection loss (offline).
The value of ANPR with Deep Learning
Video-based number-plate recognition with deep learning is important for a wide range of applications that require vehicle identification to increase safety and security, automation, and productivity.
- Automation: Automatic vehicle identification with common security cameras. No additional sensors are required, given security cameras provide a clear image of the number plates.
- Efficiency: High-performing deep learning algorithms allow reading multiple number-plates in parallel.
- Save costs of manual personnel to watch multiple video streams of surveillance cameras.
- Performance: Increase identification quality through fully automated, fast, and contactless vehicle identification.
- Scalable: Computer vision is a highly scalable method to identify vehicles across multiple locations using numerous cameras.