Zurich – ETH researchers have developed a platform that creates 3D models of cities from images and videos. The technology can also be used to analyse when parking spaces become free. Several spin-offs have already emerged out of the research project.

The new technology platform VarCity evaluates and combines aerial photographs, 360-degree panoramic images and even standard photos and videos from YouTube or other social networks and internet platforms to create 3D models of cities. Thanks to machine learning algorithms, the technology is able to recognise image content such as buildings, streets, bodies of water, people and cars. By evaluating webcam data, it can even tell which streets are one-way only.

“The whole thing goes so far that we can perform a highly detailed analysis of traffic flows on the street and also measure pedestrian traffic, even in real time depending on the application,” VarCity project manager Hayko Riemenschneider said in a statement from the Federal Institute of Technology (ETH) in Zurich. The system also makes it very easy to count traffic as it can now be done with cameras alone.

“Our technology is dynamic. It brings life, people and vehicles to 3D city models, which is one of its major advantages over well-known applications such as Google Street View,” added Riemenschneider. 

Another advantage is that the algorithms generate anonymised information based on image data. Pedestrian flows and vehicle traffic can be shown by avatars, making the system compatible with data privacy. To showcase their work, the ETH Zurich researchers have created a 3D model of the city of Zurich, which they are now presenting in a video.

Thanks to VarCity, several spin-offs have already been founded. For example, Parquery uses this technology for its parking management system, which works exclusively with cameras installed near parking spaces and does not require parking space sensors. It shops app users in real time when parking spaces become free.

The spin-off Spectando offers virtual building inspections for the real estate market, while Casalva conducts virtual damage analyses on buildings for insurance companies. Yet another spin-off is UniqFEED, which uses methods to identify perimeter advertising and individual players during television broadcasts of sports events.