Cities throughout the world have invested heavily in the development of urban public spaces as a way to promote more pedestrian-friendly communities. The challenge, however, comes in understanding how successful these projects are and crafting a way of studying the usage of such spaces. Researchers at the University of Copenhagen’s Department of Geosciences and Natural Resource Management have sought to develop more precise methods for assessing these urban public spaces.
With a pilot study conducted at the Kultorvet Plaza, the research team utilized a combination of thermal cameras, computer vision technology, and GIS analytics to geo-reference the movements of pedestrians throughout the plaza. This multi-faceted process would record the movement of pedestrians 30 times per second, allowing for a spatial accuracy of approximately 25-100 centimeters, depending on where the person was in the camera’s field of vision. Once the images have been captured, they are then geo-referenced using a homography matrix, which compares the digitized coordinates to real-world coordinates.
There are several advantages to the usage of thermal cameras as opposed to traditional RGB cameras. First, they are able to work despite the availability of light, which can certainly vary in an urban space throughout the course of a day. Moreover, they keep the identities of those recorded anonymous. Yet some of the drawbacks with thermal cameras involve the difficulty in re-identifying subjects moving within space whereas a regular camera allows one to easily re-identify.
The results of the analysis found that there were discernable movement characteristics among those observed. Some of these behaviors include meeting, flocking, avoidance, and following a leader. Comparing various public spaces using this same methodology would provide insight as to the differences in public space usage. Ultimately, this technique for analyzing public spaces ideally should supplement pre-existing methods. Additionally, the data and behaviors must also be compared to qualitative data about the public space, as a means of qualifying any of the findings that may seem inherent to that area. In conjunction with other methods, this technique could help inform strategies for developing successful public space developments.
Nielsen, S. Z., Gade, R., Moeslund, T. B., & Skov-Petersen, H. (2014). Measuring Human Movement Patterns and Behaviors in Public Spaces. Poster session presented at Measuring Behavior 2014, Wageningen, Netherlands. doi: 10.13140/2.1.2247.5205
Watch the video: Capturing human movement patterns in public spaces