Artificial Intelligence based Building Attributes Enrichment in OpenStreetMap using Street-view Images

Abstract

This work aims to improve OSM building attributes using street-view images. As OSM data are open and street-level photos can be taken with standard cell phones, our approach is neither geospatially restricted nor economically discriminated. In addition, crowdsourced platforms, such as Flickr, Unsplash, and Mapillary, provide huge amounts of street-view images that contain valuable building attribute information. We seek to facilitate open data and citizen science and encourage people to map for their communities.