On the Development of a Customizable Crowd Sensing System for Public Spaces Using IoT Cloud Services

Before deciding to visit a public facility such as a library with limited seating, it is often convenient to know how crowded the location is before arriving. Various sensor systems on the market can detect the presence of people in particular areas of a building, but if we want to install such a system on the premises of a public space, there are limitations to consider such as cost, ease of installment, readjustments, removal, etc. In this paper, we describe our work in progress on the development of a compact crowd sensor system based on the Raspberry Pi microcontroller which can be outfitted with a camera for computing the approximate number of people appearing in a snapshot of a selected area and judging the general level of crowdedness at periodic intervals. To protect the privacy of individuals in public spaces, once the camera snapshots have been processed to identify human figures, they are discarded and only the timestamped crowd level results are posted on a cloud database for viewing by remote users. Unlike commercial products which are sometimes mounted firmly in a building structure, the camera and microcontroller unit of the prototype can be held in one hand and positioned in a convenient location on a wall, desk, shelf, etc. in library study areas and cafeteria facilities on a college campus.