3D Georeferencing of historical photos by volunteers

Abstract

Historical photographs are a very rich source of information that can be useful in a variety of different contexts such as environmental analyses, land planning and illustration of landscape evolution. However to reach this goal the images must be accurately georeferenced. In this paper we propose to use the crowd to perform the 3D georeferencing of collections of historical images. To this goal we implemented a web 3D georeferencer that offers volunteers the possibility to semi-automatically identify 1. the location of the point from where a picture has been taken, 2. the three angles: tilt, roll and yaw and 3. the field of view. A virtual web-based globe is the central element in this tool that allows both for the georeferencing in three dimensions by volunteers and for the visualization of georeferenced images to assess the landscape variation through time. In this paper we evaluate the method and the georeferencer and give suggestions for further developments and exploitation of the database.