Combining VGI with viewsheds for photo tag suggestion

Abstract

In this research we attempt to develop an improved method for tagging of digital photos, making the tagging process simpler and more accurate for users, and thus ultimately increasing the quality of the VGI data as a whole. Our method combines Volunteered Geographic Information (VGI) in the form of tags on photo sharing websites (such as Flickr and Panoramio) with a visibility analysis of the new photo to be tagged. This analysis is performed by combining the photo metadata with a Digital Surface Model and building footprints. This approach is used to derive a ranked set of suggestions for a given photograph. To deploy the system, a webservice was implemented.