Oussama MOSLAH has successfully obtained his engineering diploma in networking and telecommunications on June 2005 from INSAT in Tunisia. He also obtained a master degree in image, vision and perceptual systems on September 2006 from Cergy Pontoise University and ENSEA in France. He is currently a third year PhD student in THALES Security Solutions and Services company. His main research topics are image-based modeling and rendering techniques. He is working on the Cap Digital Business Cluster Terra Numerica project which aim is to reconstruct the Paris city from various data types such aerial images, laser scan, photographs and building footprints.

AG 4, MMCI   Info

AG Audience

English

30 Minutes

Saarbrücken

The context of this talk is the growing interest in urban models reconstruction from various sources including aerial images, laser scanned data, and façade images. Many applications raise based on the 3D city model such as GPS-based location services, virtual tourism, simulation, security, transport control and video games. Large scale urban environnements reconstruction is typically done in different steps. First, an elevation map is computed from aerial images. Then, a rought 3d model of buildings is reconstructed based on the elevation map, an approximate knowledge of the building footprints and an a priori of the architectural styles. At this step, we can already get an approximate 3d model of the city textured with aerial images. Finally, in order to get a more precise and realistic 3d city model we use laser scanned data and façade images to reconstructed high detailed buidling models. Our work is mainly related to this part and consists on a model-based approach that uses shape grammars and procedural modeling techniques to provide a framework able to describe and encode the variety of the architectural styles in Paris. The model (a set of rules) is then optimized towards a rectified façade image using both bottom-up detection and top-down proposals.

Thorsten Thormählen

+49 681 9325-417

--email hidden

passcode not visible

logged in users only