Research Explore the world of Geo-information
  • Back to the timeline
  • Jul 2013

    Geodan en Open Data

    3D Roof detection and analysis from AHN2 and BAG

    Properties and roof shapes obtained automatically

    Geodan en Open Data

    By cleverly combining data sources, a host of options emerges. Take, for example, the Basic Register of Addresses and Buildings (BAG) and version 2 of the National Height Model of the Netherlands (AHN2). Various types of 3D objects can be detected from these files. This enables more information to be added to 2D data sources and more realistic visualisations can be made.

    Using the AHN2, amongst other things, roof shapes and the properties of roofs, such as roof areas and roof pitches can be detected. By combining these with the BAG data, additional information is available for each building. This information can be used for various purposes. Such as to determine the suitability for solar panels or for calculating maintenance costs. Realistic visualisations of buildings can also be created.

    How does it work?

    To detect roofs, a modified version of the RANSAC algorithm is used. This algorithm was developed to extract objects from a 3D point cloud.

    The AHN2 is a 3D point cloud that (almost) covers the whole of the Netherlands. This file is obtained using LIDAR technology, which measures heights from an aeroplane. The accuracy of the AHN2 enables properties such as roof shapes, roof pitches and areas to be determined. Dormer windows and chimneys can also be detected by combining the AHN2 with the BAG, this information can be linked to the correct buildings. It is also possible to detect other objects using this technology, as has been demonstrated by the 3D tree database.

    Unique attributes

    • Automatic detection and analysis of roofs based on the AHN2 and BAG.
    • Flexible algorithm that also works with other data sources (both 2D and 3D) which can be modified according to local variations.
    • Information can be used for various purposes, such as cost calculations and potential solar panel yield.
    • In combination with the 3D tree database, a unique 3D data set is available that can be used to, for example, make shadow calculations.

    Read more about 3D-Roof detection and other 3D-research at Geodan (in Dutch)

  • Back to the timeline