23rd Congress of International Council of the Aeronautical Sciences, 8-13 September, 2002, Toronto, Canada
Paper ICAS 2002-6.8.3

SEMI-AUTOMATIC 3D FLIGHT SIMULATORSCENARIOS RECONSTRUCTION FROM GEOREFERENCED AERIAL PHOTOS AND TERRESTRIAL MAPS

A. Liverani, F. Persiani, G. Piraccini, M. Tappi
University of Bologna, Italy

Keywords: flight simulation, virtual reality, scenery reconstruction

In this paper we present a completely new technique to extract 3D models from digital aerial images, by combining them with topographical maps of the site. An efficient C++ application has been developed, in order to generate 3D VRML models by exploiting both image data from on board digital camera and a database of topographical maps covering very large amounts of sites. Obviously, GPS data sets are required in order to match images with topographical maps. The described system works as follow. Aerial images are taken during traditional airplane or UAV (Unmanned Aerial Vehicle) flight. DEM data are used to generate a 3D terrain model. Then, all images are referred on the 3D terrain model and glued onto it. By matching topographical maps with aerial images an automatic recognition algorithm extracts civil and industrial buildings contours and creates 3D textured models of them

view full paper