SAE-MCVT: A Real-Time and Scalable Multi-Camera Vehicle Tracking Framework Powered by Edge Computing
Yuqiang Lin, Sam Lockyer, Florian Stanek, Markus Zarbock, Adrian Evans, Wenbin Li, Nic Zhang
TL;DR
SAE-MCVT tackles city-scale Multi-Camera Vehicle Tracking by combining edge-based intra-camera processing with a central, self-supervised inter-camera association mechanism. The edge pipeline uses a light-weight YOLO detector, a non-embedding single-camera tracker, geo-mapping, and dynamic feature extraction to generate compact metadata; the central server performs tracklet re-merging and cross-camera association guided by a learned spatial–temporal camera link model and periodic optimization. Key contributions include an end-to-end edge–cloud architecture, a dynamic workload adjustment scheme, a self-supervised camera link model based on KDE transition-time densities, and a public RoundaboutHD dataset benchmark, achieving real-time operation on 2K 15 FPS streams with IDF1 around $61.97\%$. The framework demonstrates scalable real-time MCVT suitable for city-scale deployments while preserving privacy and reducing bandwidth by transmitting metadata only. The practical impact lies in enabling safe, efficient urban traffic monitoring, anomaly detection, and suspect-vehicle retrieval at scale.
Abstract
In modern Intelligent Transportation Systems (ITS), cameras are a key component due to their ability to provide valuable information for multiple stakeholders. A central task is Multi-Camera Vehicle Tracking (MCVT), which generates vehicle trajectories and enables applications such as anomaly detection, traffic density estimation, and suspect vehicle tracking. However, most existing studies on MCVT emphasize accuracy while overlooking real-time performance and scalability. These two aspects are essential for real-world deployment and become increasingly challenging in city-scale applications as the number of cameras grows. To address this issue, we propose SAE-MCVT, the first scalable real-time MCVT framework. The system includes several edge devices that interact with one central workstation separately. On the edge side, live RTSP video streams are serialized and processed through modules including object detection, object tracking, geo-mapping, and feature extraction. Only lightweight metadata -- vehicle locations and deep appearance features -- are transmitted to the central workstation. On the central side, cross-camera association is calculated under the constraint of spatial-temporal relations between adjacent cameras, which are learned through a self-supervised camera link model. Experiments on the RoundaboutHD dataset show that SAE-MCVT maintains real-time operation on 2K 15 FPS video streams and achieves an IDF1 score of 61.2. To the best of our knowledge, this is the first scalable real-time MCVT framework suitable for city-scale deployment.