Planning and development of urban areas require careful consideration of many aspects such as noise, wind, light ingress, and sustainability. Nuisance created by factors like these is often unexpected and becomes apparent only after construction has been completed.
Now that 3D data is made available of all 10 million buildings in the Netherlands by researchers at TU Delft, which involves among others our Principal Investigator Jantien Stoter, effects like these can be simulated before the construction phase.
“An accurate, up-to-date virtual 3D copy of the city is needed to understand urban processes and study the impact of different future scenarios. This asks for 3D data beyond visualization.”
Professor & AMS PI
Urban digital twin to simulate ‘what-if’ scenarios
The 3D models of a city, also called an urban digital twin, can be used to virtually simulate ‘what-if’ scenarios. These simulations allow us to determine the impact of different design options on the living environment during the design process. The design can be adjusted based on the findings – this way better-informed and evidence-based decisions can be made that lead to an optimal design.
Within domains such as (solar) energy, wind and ventilation, noise and air quality, virtual 3D models have long been used. Now with this virtual model of the living environment, there is a dataset available that all domains can use directly and freely. The data and underlying technology have been made available as open data and open-source software to stimulate potential innovations as much as possible.
A new method for fully automatic building reconstruction
Using a new building reconstruction method, 3D models of all 10 million buildings in the Netherlands were generated in a fully automated manner. Until recently, reconstruction required semi-automatic processing, which made it difficult to reconstruct large areas in 3D.
The method uses 2D building footprints from the national Basic Register of Addresses and Buildings (BAG) as input and the developed service is therefore called 3D BAG.
To derive information on height, a point cloud (in Dutch: Actueel Hoogtebestand Nederland) was used. This point cloud data is collected using ‘laser altimetry’: from an aircraft or helicopter a laser beam is sent to the earth and depending on the response time the height at that point is determined. This information is available for the whole of the Netherlands as open data.
Flexible algorithms and artificial intelligence
The main challenge was to develop algorithms that are flexible enough to automatically generate a simple and accurate 3D model for any type of building, regardless of the shape of the roof. In addition, the algorithms must be able to robustly process all kinds of source data with varying qualities.
This requires algorithms that can take the technical limitations into account, such as height data with varying quality obtained from an airplane or helicopter via laser scanning. To illustrate, heights of buildings in dense urban areas cannot always be observed when (partially) covered by another building. Therefore, height points are sometimes missing in the point cloud data.
The researchers want to address these kinds of technical limitations, for example by filling in gaps in the point cloud data by using artificial intelligence.
Netherlands’ Cadastre, Land Registry and Mapping Agency
The stable version of these 3D BAG developments at TU Delft is forwarded to the 3D base-data facility of the Netherlands’ Cadastre, Land Registry and Mapping Agency (in Dutch: Kadaster). The models can be used as a basis in the digital information infrastructure of the new Environment and Planning Act.
Furthermore, in projects with various partners, the researchers work on an optimal connection of the 3D data to real-life urban applications. An important application, for example, is noise simulation. For this purpose, a national dataset was recently generated in collaboration with the Kadaster and National Institute for Public Health and the Environment (RIVM).
The technology for the 3D BAG service was developed by the 3D Geoinformation research group, Department of Urbanism, in various research projects, with funding from the European Commission (ERC), the Netherlands Organization for Scientific Research (NWO), and Amsterdam Institute of Advanced Metropolitan Solutions (AMS Institute). The algorithms are improved further in collaboration with partners such as RIVM, Rijkswaterstaat and Kadaster.
The 3D models can be viewed and downloaded via the 3D viewer that has also been developed as part of this project. For more information about 3D BAG, please reach out to: Balázs Dukai or Jantien Stoter.
Urban Modeling in Higher Dimensions
Responsible Urban Digitization
Solving Amsterdam’s urban issues would benefit from one 3-D model with continuously updated information. However, scientists so far have matched unique data to their own simulating models. Research on one integrated solution is underway.