Can we “see” the neighborhood-built environments from a UAV?

Abstract

Health-related built environments in a neighborhood are significant to the health of the residents. In order to quantitatively examine built environments at the neighborhood scale, the operational unit neighborhood should be clearly defined in advance. This paper assessed a newly developed neighborhood concept, sidewalk-homogenous neighborhood, which corresponds to the economic profile and the health behaviors of residents.  This study examined the applicability of the sidewalk-homogenous neighborhood in the process of detecting and evaluating built environments at the street level in four study sites. Sidewalks and greenery as the representations for built environments were classified on UAV images using the combined applications of geographic information systems (GIS) and deep learning. The study addressed that the sidewalk-homogenous neighborhood is a practical, operational unit to identify the spatial disparities of built environments at the neighborhood scale. In addition, the study revealed the inequality in Unmanned Aerial Vehicle (UAV) research opportunities between rich and poor neighborhoods.