Enabling post-recording deep georferencing of walkthrough videos : an interactive approach

Abstract

The usage of large scale databases of georeferenced video stream has an infinite number of applications in industry and research since efficient storing in geodatabases to allowing the performing of spatial queries. Due to the fact that the video capturing devices have become ubiquitous, a good source for the acquisition of a lot of video contents is the crowdsourced approach of Social Media. However, these social apps usually do not support geo metadata or it is very limited to a single location on Earth. In other cases, the regular user usually does not have the required hardware and software to capture video footage with a deep georeference (position and orientation in time). There is a clear lack of methods for the extraction of that spatial component in video footage. This study proposes and evaluates a new method for the manual capture and extraction of the spatial geo-reference in the post production phase of video content. The proposed framework is based on a map-based user interface synchronized with the video stream. The efficiency and usability of the resulting framework were evaluated performing a user study, in addition, the resulting geo-metadata of the manual extracted georeference has been compared with the one previously captured by hardware in order to evaluate the goodness of the method.