Roland Billen 
  
Administrative 
  
  
  
object 
P v ; 
Past caen Local administrative 
reference level 
| pr 
' Urban 00 SE | 
| Block ] 
| i 
| iere e mnm i 1 
| T 
| | C: 
: Building Vegetation Fo 
| Ensemble Ensemble vo 
dt 
| 1 I 
| Sons Enn Building Vegetation Traffic networks Public Utilities 
  
»] 2 070 0070000005 au Ca 
—— Hierarchical relation 
— Logical relation 
Figure 4. First step in the categorisation of 3D city information 
In the figure 4, the following new objects are included: - empty space; - urban block; - cadastral envelop; - cadastral 
object; - postal object; - local administrative reference level; - administrative object. A 3D urban block is composed by 
building ensemble, vegetation ensemble, empty space and potentially a part of the traffic network. Urban block, traffic 
network and public utilities represent the entire 3D physical urban space. The cadastral subdivision represents a set of 
object where the third dimension is only potentially acquired from their component objets. In this model, some 
subdivisions are directly included : - 3D politic and administrative; - 3D real estate (cadastral). The 3D “land” use (or 
better space use) can be linked to the physical object. The 2D land cover information can be, if needed, derived form the 
physical object. The 3D address information must be combined with the traffic network objects and the cells or 
buildings objects. The introduction of some abstract urban objects (Postal, Cadastral, Administrative, etc.) implies a big 
work of redefinition of actual 2D subdivisions of urban space. Note that in few cities, 3D cadastral object and some 3D 
concept of urban rules are already taken into account. But this is certainly a major challenge for the use of 3D urban 
model. 
6 CONCLUSIONS 
This primary conceptual approach of the problem of the 3D integration in urban GIS provides some interesting thought. 
A major limitation of the use of 3D model is certainly the lack of user 3D thinking. This is due to the inheritance of the 
old 2D information reasoning and structuring. The GIS conception method must integrate this technical and conceptual 
evolution. It seems that a more descriptive definition of urban objects and of their relationships is a prerequisite to an 
efficient and useful 3D GIS. Like it is (or must be) for 2D GIS, the data acquisition must correspond to specifications 
drawn up during the conceptual step of the conception method. The challenge of the use of 3D model needs a complete 
revolution of spatial thinking. The users and the specialist must redefine their spatial referential to the real one, i.e. a 3D 
world. The cost of 3D data is still high, and inaccessible for a large number of user. Therefore, public authorities should 
undertake the creation of global 3D urban databases, like it is already the case in 2D. These 3D databases should 
contain physical but also abstract urban objects. This implies a big work of redefinition of the actual 2D subdivisions of 
urban space. But it is certainly a major issue for the use of 3D urban model. Further work will formalize the common 
3D city model in order to implement a real 3D GIS. 
REFERENCES 
Breunig M., 1996. Integration of Spatial Information for Geo-Information Systems. Lecture Notes in Earth Sciences 61. 
Springer-Verlag, Berlin Heidelberg, 171 p. 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 83