Exploring the 3D BAG: How to define it and to what extend can it automatically be created using open data
P5 presentation: 10.45 7th of November BK-CZ-E
The System of Key Registers is the main source of information for all governmental organizations in the Netherlands. It includes information that is used by the government on a regular basis, such as company names, personal data and spatial information. The Basisregistraties Adressen en Gebouwen (BAG) is part of this system, and stores all buildings and addresses within the Netherlands. It comprises non-spatial information such as the year of construction of a building, but also the 2D representation of buildings and units as polygons and points. Because of the increasing densification of the urban environment, it is more difficult to model the reality on a flat map without losing important information. As such, the 2D representation of the BAG has drawbacks and an improvement of this model is necessary.
The aim of this thesis was to explore the needs and possibilities to improve the BAG representation from a 2D into a 3D-model. I have especially focused on the main features within the BAG that are most influenced by the shift from a 2D to a 3D representation: the Panden (buildings) and the Verblijfsobjecten (units). In this thesis, I have firstly proposed a definition for the 3D BAG, and subsequently investigated to which extent the current available open data can be used to create 3D geometries of the buildings and units.
By the use of literature studies and interviews with experts in the field, I recognized a shift within the government regarding the use of 3D data. However, current efforts are scattered and on different scales. Here I proposed a 3D model that connects spatial information of different key registers harmonizing these efforts, and also solving current problems with the 2D representation of the BAG. Hence, it provides new possibilities for the data Moreover, this thesis I proposed methodology for creating 3D geometries of the BAG buildings and units. It is based on the use of different, nationwide spatial datasets that are made available as open data by the government. This methodology resulted in a workflow that can theoretically create a nationwide 3D model of the BAG. The workflow consists of different steps, starting with the classification of underground buildings and followed by the calculation of the number of storeys, placement of the units over the storeys and dividing the storeys themselves. The workflow was implemented and tested for three areas to investigate its applicability to different environments; the highly densified area of The Hague, the urban area of Hoofddorp and the rural of Vlist-Schoonhoven. The resulting dataset of The Hague was validated using a reference dataset provided by the municipality.
Results of the implementation of a 3D model of the BAG validated the potential of the designed methodology. For most buildings in the Netherlands, a 3D model can be created with reasonable accuracy. However, this thesis also reveals uncertainties within the BAG that makes it impossible to create accurate 3D models for all buildings and units with the developed methodology. When aiming at improving the results and overcoming these uncertainties a wide approach is needed. This approach should not solely focus on improving the developed methodology, but also on the acquisition of data that provides more information about the location and geometry of buildings and units.
Main mentor: Hugo Ledoux
Second mentor: Jantien Stoter
Third mentor: Martijn Meijers