A Hierarchical Intelligent Highway Surveillance System Based on YOLOv10 and StrongSORT
Ziming Huang, Mingzhi Mao · 2025
With the acceleration of urbanisation and the continuous increase in traffic flow, traditional traffic monitoring methods, which rely on manual inspection, have become inadequate for modern traffic management needs. To address this dilemma, a hierarchical intelligent highway surveillance system is designed and implemented, comprising the data, model, service, and application layers. The data layer collects multi-source data and performs preprocessing such as cropping, labelling, and storing to support downstream tasks. The model layer processes video data using YOLOv10 and Strongsort to achieve multi-object tracking, vehicle classification, accident detection, and speed estimation. The service layer offers lightweight task management and data processing interfaces, while the application layer features an intuitive user interface that invokes these interfaces. Experimental results show that the system exhibits high detection accuracy and good real-time performance on various tasks. In the future, model structures will be optimised, and more advanced technologies will be integrated to enhance system intelligence.