Movement Vector Computing of Vehicles on Mast-Mounted Camera Systems

Vehicle detection is an essential part of collecting valuable data for traffic management. Many sensors and computational architectures are used for this purpose. However, many of them have limitations in providing data in the real-time scenario. This paper surveys the architectures of similar problems. We used this survey to propose our own vehicle detection architecture using edge-computing and service-oriented architecture. The proposed architecture was implemented and tested in real-world conditions in the mast-mounted surveillance system. The deployment of services is provided via containerization.