The improvement of 3D acquisition devices (laser measurements, cameras, etc.) leads to the multiplication of avaiable geolocalised 3D data. More and more cities have a 3D mockup at their disposal. To garantee the interoperability between different data sources, standardisation work has been done on the exchange protocles (for instance, 3DPS – 3D Portrayal Service) and file formats.
Moreover, the web has evolved significantly in the last few years, it can now integrate rich content such as 3D applications. Both these factors make it possible today to share and exploit 3D urban data on the web.
My thesis, supported by the french company Oslandia, is about the 3D representation of the city on the web. More precisely, it is about fetching and visualising, using a light client, large quantity of urban data on one or more distant servers. The data is heterogenous: it can be a 3D representation of buildings (meshes) and terrain (height map), but also semantic information such as pollution rate (volumes), th elocalisation of bike stations (points) and the number of available bikes, etc.

An other problematic of my thesis is the management of the multi-scale aspect of the city. A city is not represented the same way at a small scale (street level) and at a big scale (neighbourhood or city level). Generalisation processes (adapting the representation of one or several object to the visualisation scale) allow the removal of unnecessary details and render a scene more legible. Designing such processes and handling the navigation between the resulting representations is part of my work.