On Automatically Geotagging Archived Images

Cathal Hoare, Humphrey Sorensen

Abstract


This paper describes an automated approach to assigning a longitude and latitude - geo-tagging - historic images, opening the way to admitting their use in a range of applications. Several approaches already exist to achieve this, but all involve some element of human involvement. Websites such as Flickr and WhatWasThere allow users to upload photograph and represent its position as a marker on a map; WhatWasThere also allows a user to place the photograph as an overlay on Google Street View, placing it within its modern context. Other approaches include group projects (schools and history groups locating the location of a photograph and translating that location to a geo-tag), and crowd-sourced projects where ‘workers’ are recruited online and are paid small rewards for tagging an image; mTurk is an exemplar of such a service. These images can then be used in a range of applications; for example, StreetMuseum - an iPhone application produced by the Museum of London - which guides users on a tour of London and overlays the phone’s camera feed with historic images and information about historic events. These tools educate users about the history of their location and provide an engaging means for accessing the museum’s artifacts. However, the quantity of non-geotagged images overwhelms human based approaches and precludes the widespread adaption of these tools.

Full Text: Paper

Refbacks

  • There are currently no refbacks.




Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.