Abstract: With the development of digital city and smart city construction, the integration of building
information modeling (BIM) and geographic information system (GIS) has received much attention
from a wide academic circle. The current integration mainly focuses on the conversion of both
geographic and semantic information between the two data standards, IFC and CityGML, but there
are problems, such as data error and loss, lacking geometric-semantic coherence and poor application
extensibility. This paper proposed a multi-scale 3D city data model which takes both entity and
geographical objects into account, and studied the extraction, processing and transformation method
of spatial and semantic data of BIM and GIS. Accordingly, the integrated application framework was
designed, verifying and preliminarily applied to the 3D visualization platform. It is advantageous in
the realization of a total fusion of BIM and GIS information in terms of geometry, semantics and
precision as well as the avoidance of the information loss caused by the traditional data
transformation. It also has an advantage over the multilevel storage, loading and displaying of
multi-scale spatial and semantic data and helps to achieve efficient integration of a large scale of
building and city data with high accuracy.

Key words: building information modeling, geographic information system, IFC, CityGML, data
integration