The production of reliable 3D urban models is essential in developing autonomous cars and drones as well as for digital urban management. If vehicles are equipped with suitable sensors, they are capable of updating the 3D models in real time, thus integrating the fourth dimension - i.e. time.

Lay summary

With the help of complex algorithms, an on-board computation system installed on the ScanVan transforms the constant flow of images obtained from an omnidirectional camera into point clouds representing the surroundings of the moving vehicle. Systematising all these components results in a complete workflow from the acquisition of images through to the calculation of 3D point clouds.

With current photogrammetric techniques, scanning buildings is a routine task. However, none of these techniques can cope with something the size of a city. The major challenges of such an undertaking lie in the quantity of data to be acquired and the computing power needed to transform it into 3D models.

The aim of this project is to improve the production of 4D urban models by developing the ScanVan – a vehicle capable of continuously scanning towns in three dimensions. Tested on a real, life-size case, this prototype will make it possible to confirm the theoretical choices made with respect to processing the enormous quantity of data acquired. Ultimately, the project will produce a 4D urban model of the city of Sion (Valais, Switzerland).

The open access data set produced by the ScanVan can be used to explore potential new services that could benefit from it, such as the autonomous vehicle projects currently being trialled in the city of Sion. At a more general level, this project constitutes the first step in establishing 4D models of towns, since it puts forward solutions capable of processing the large quantity of data involved.