Parking Lots Management and Visualization in the Smart City - Digital Twin context

In recent years, extensive research has been conducted to develop traffic surveillance and monitoring applications using advanced technologies in a Smart-City context. One of the fundamental challenges in an urban city scenario is developing an efficient yet affordable parking lot management system. Such systems play a crucial role in providing the necessary data to local municipalities to make informed decisions to better manage the traffic flow while saving valuable resources and time. In this paper, we propose a novel end-to-end system architecture for a real-time video-based parking management system that can be easily scaled and enforced in a Smart-City context. Using state-of-the-art object detection and tracking methods combined with novel Static Object Detection techniques, we demonstrate a localized strategy for identifying potential parking spaces. A separate parking occupancy classification pipeline is discussed which provides fast and accurate availability status of these parking spaces. Furthermore, a 3D digital-twin application to monitor and visualize the parking spaces is demonstrated. Our experiments in a real-world scenario demonstrate the significant potential of our proposed architecture to easily scale to a given city with minimal manual efforts and lower implementation costs.