Characterizing Street-level Accessibility at Scale

Poorly maintained sidewalks pose considerable accessibility challenges for people with mobility impairments. Despite comprehensive civil rights legislation for Americans with disabilities, many city streets and sidewalks in the U.S. remain inaccessible. The problem is not just that sidewalk accessibility fundamentally affects where and how people travel in cities, but also that there are few, if any, mechanisms to determine accessible areas of a city a priori. In this talk, I will introduce scalable data collection methods for acquiring street-level accessibility information using a combination of crowdsourcing, computer vision, machine learning, and Google Street View. Our overarching goal is to transform the ways in which accessibility information is collected and visualized for every sidewalk, street, and building façade in the world.

Date:
Speakers:
Kotaro Hara
Affiliation:
University of Maryland