UAV-based People Location Tracking and Analysis for the Data-Driven Assessment of Social Activities in Public Spaces

Abstract

In sustainable high-density cities, public spaces play an important role in supporting social and community health and wellbeing. Amidst ongoing urbanisation, it is of increasing importance to study public space interaction patterns and placemaking processes that contribute to the quality of life of urban residents. This paper reports on the development of a new methodology for the computational tracking and analysis of social activities in urban spaces, using Computer Vision Object Detection (CVOD) techniques to create digitalised pedestrian trajectory data. Referring to concepts from humanistic geography and time geography, our method offers a new platform for data-driven urban place studies, detecting co-presence and social interaction in relation to urban morphology. This paper focuses on the development of Machine Learning protocols, algorithms for tracing and mapping pedestrian trajectories in a georeferenced photogrammetry model, and computational analysis of co-presence. The resulting workflow forms a foundation for future research around detecting, analysing and quantifying behavioural parameters, to evaluate the ability of public spaces to support social interaction and placemaking.