Abstract

The integration of Building Information Models (BIM) in environments such as Geographic Information System (GIS) is currently one of the main investigated research problem in the urban spatial information community. Among the several issues that are involved in the adequate realisation of such type of integration, one particularly relevant to the GIS world is the geo-referencing of data coming from the BIM domain. The latter, which is inherited from the Computer Aided Design (CAD) world, often relies on local coordinate systems and thereby provides models with no or wrong information with respect to their actual geographic position. Based on well-known processes in the field of computer vision, this paper proposes an automatic framework that allows to transform the local coordinates of a BIM model into its real world geographic coordinates. Starting from a geo-referenced 2D polygon of the building (footprint or vertical projection), our method directly applies a rigid motion transformation to a corresponding 2D projection generated from the BIM model in order to obtain the transformation matrix that will align the geometry of the BIM to the map's one. In comparison with the most common approaches of shape alignment and rigid body transformation, our approach requires neither the input polygons to have the same number of points nor their sequence of points to be corresponding, allowing thereby to maximize the automation of the process while still providing accurate results. Several cases of BIM geo-referencing are presented and discussed. Furthermore, our experiments show that centimetre accurary can be reached in the alignment process.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call