Assessing the Built Environment Using Omnidirectional Imagery

Several methods are used to examine the characteristics of the built environment as potential determinants of physical activity—surveys of community residents, observation by trained observers, and interpretation of geographical information system (GIS) imagery. The last category includes multiple images of an area that allow auditing of environment characteristics such as sidewalks, ramps, crosswalks, bike lanes, traffic-calming devices, shade trees, graffiti and litter.

These researchers compared built environment audits derived from in-person field assessments with three types of omnidirectional imagery: 1) commercially provided new images collected with the field audit; 2) commercially provided archived imagery from previous (6 to 12 months old) audits; and 3) publicly available imagery from Google Street View. They evaluated neighborhoods in Indianapolis and St. Louis that were socioeconomically stratified.

They found that the archived omnidirectional imagery was comparable to field audits for most (84%) of the characteristic assessed, with slightly higher agreement for Google Street View (86%) and newer images (87%). Items that had lower level of agreement included aesthetic items such as trash or graffiti, as well as sidewalk condition.

Omidirectional imagery offers a viable and efficient alternative to field audits for characteristics of the built environment important to public health researchers.