As part of the new Delft AI initiative, the new 3D Urban Understanding (3DUU) lab has started as a collaboration between Dr. Liangliang Nan of the 3D Geoinformation Group at the BKE Faculty, and Dr. Julian Kooij at the Intelligent Vehicles Group of the 3ME Faculty.
The 3DUU Lab will develop new methods and techniques to automatically recognise and model objects in the built environment in 3D. We combine data from different sources, such as aerial photos and laser scanners on vehicles. Using AI, we will allow computers to learn and work with them so that we can automatically process and interpret large amounts of urban data. Using the complementary expertise and data sources in both groups, the lab will combine research “in AI” to enable novel applications “with AI”. The 3D models of the built environment created in this way will not only allow us to map the current state of the environment very accurately, they will also be suitable for making simulations of future events. Examples include potential floods, plans for redevelopment and how heat spreads across the city. Architects, urban planners and others involved in the development of cities can use the data to better tackle urban problems such as air pollution and heat islands. The techniques can also be used for intelligent vehicles that use advanced sensors to map urban environments for navigation, and to easily build 3D models.
The lab will host four new PhD students: two at the BKE faculty and two at 3mE in the Intelligent Vehicles Group.
For more information, see the 3DUU lab website (under construction).