Reconstructing legal 3D apartment models from 2D division drawings
E.C.J. de Niet (TU Delft - Architecture and the Built Environment)
JE Stoter – Mentor (TU Delft - Urbanism)
GAK Arroyo Ohori – Mentor (TU Delft - Urban Data Science)
Amir Hakim – Mentor (TU Delft - Urban Data Science)
Bala Bhavya Kausika – Mentor (Kadaster)
E. Verbree – Graduation committee member (TU Delft - Digital Technologies)
Jasper van der van der Vaart – Graduation committee member (TU Delft - Urban Data Science)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The continued growth in the complexity of apartment buildings, and the digitization of building processes have led to a rise in the use of BIM models during the architectural development phase. These models contain a lot of information, but their life cycle often ends once the building is constructed. However, this information can be used by the Dutch Cadastre, to bring legal building registration even closer to 3D reality. BIM based legal registration, researched under the name BIM Legal, is a key driver for this research. Currently in the Netherlands, the separation of ownership is registered through notarial deeds that are managed by the Dutch Cadastre. This is accompanied by division drawings that indicate how apartments are split into private apartment units and shared spaces.
To develop a full BIM Legal registration, it is necessary to also look at buildings without a BIM model. In such cases, BIM Legal models could be derived from existing legal documents, specifically division drawings. This research investigates the (semi-)automatic reconstruction of 3D legal apartment models from 2D vectorized division drawings. The proposed pipeline starts with already vectorized division drawings and applies shape-based georeferencing techniques, estimating storey heights, vertical alignment, and ends with generating 3D BIM Legal models.
The georeferencing methods tested achieved sufficient alignment across the 10 sample buildings. When outside spaces were included in the vectorized division drawing, it resulted in an average containment of 81.5%. When they were removed or non existing it resulted in 95.2% containment. Storey alignment relies on shape similarity and floor-to-floor matching, which performed well in typical cases but struggled with floors with low similarity to the floor below it. The accuracy of height estimation improves when cross sections are included in the division drawing. Otherwise averaged based on values retrieved from the 3DBAG.
The resulting models conform to the BIM Legal standard, written in CityJSON format at a LoD1+. While the schematic nature of division drawings limits the achievable level of detail, and the geometric and positional accuracy, the models offer a valuable 3D visualization of private and shared ownership spaces. Improvements to the prior vectorization would also improve the accuracy and computation time of the 3D reconstruction. Large scale testing is necessary to research the potential incorporation of BIM legal models from division drawings with a complete BIM Legal registration.
This research not only advances the automation of BIM Legal models from division drawings, but also provides methods which can be applied in other 3D reconstructions, such as georeferencing polygons with a reference to a cadastral dataset.