Real-time Occupancy Detection Of On-street Parking Spaces Based On An Edge Device

We developed a system that can detect on-street parking spaces in real-time. The system uses a positioning module, a camera, and NVIDIA Jetson Xavier NX to build the device components on which pruned You Only Look Once v5, Simple Online and Realtime Tracking and counting algorithm are deployed to complete the collection of occupancy, geographic location, and other information of urban on-street parking spaces. Our device can be integrated into the intelligent dashboard cameras onboard taxis or online cars to form a monitoring network. Such a network enables a dynamic update of urban on-street parking occupancy information. More importantly, a long-term collection of this information allows for an analysis of the occupancy rate which is invaluable in terms of on-street parking space planning.