Real-Time Automatic License Plate Recognition Using YOLOv8, SORT Tracking, and Temporal Data Interpolation

Researchers present an automated license plate recognition system combining YOLOv8 object detection, SORT multi-object tracking, and temporal data interpolation to improve real-time video processing in traffic monitoring. The five-stage pipeline addresses challenges like variable lighting, high vehicle speeds, and occlusion that traditionally degrade recognition accuracy and tracking consistency.

This research addresses a critical gap in computer vision infrastructure for traffic management systems. The proposed pipeline solves practical limitations that have historically prevented widespread deployment of ALPR technology in real-world conditions. By integrating YOLOv8 nano models for both vehicle and license plate detection, the system achieves computational efficiency suitable for edge deployment while maintaining detection accuracy across challenging environmental conditions.

The incorporation of SORT tracking represents a shift toward kinematic-aware object tracking that maintains spatial-temporal coherence across video frames. This is particularly significant because disjointed tracking paths have plagued previous ALPR implementations, causing missed recognitions and false matches. The offline temporal interpolation mechanism for bounding boxes reconstructs fragmented detection sequences, effectively converting spotty detection patterns into continuous tracking paths.

For the traffic management and smart city sectors, this architecture enables practical ALPR deployment at scale. The system's ability to handle acute camera angles, rapid illumination changes, and occluded plates directly addresses failure modes documented in production systems. This has immediate implications for law enforcement, toll collection, parking management, and autonomous vehicle monitoring applications.

The technical maturity demonstrated here suggests the next phase involves field validation across diverse geographic and climatic conditions. Practitioners should monitor adoption rates in municipal traffic systems and integration into existing surveillance infrastructure. The efficiency gains from YOLOv8 nano versus larger models make this approach particularly suitable for resource-constrained deployments in developing regions.

• →End-to-end pipeline combining YOLOv8, SORT tracking, and temporal interpolation significantly improves real-time ALPR performance in challenging conditions.
• →Offline temporal bounding box interpolation reconstructs fragmented detection paths, addressing a major failure point in traditional ALPR systems.
• →YOLOv8 nano model enables efficient edge deployment while maintaining accuracy across variable lighting, camera angles, and vehicle speeds.
• →The system's ability to handle occlusion and environmental constraints removes practical barriers to ALPR adoption in smart city infrastructure.
• →Multi-stage architecture separates vehicle detection, plate localization, and character recognition for modular optimization and improved accuracy.