CMRT09: Object Extraction for 3D City Models, Road Databases and Traffic Monitoring - Concepts, Algorithms, and Evaluation
120
4.6 Model analysis
Many more activities might be implemented on the integrated
model, since all kinds of queries may be posed on it. Thus, a
further processing activity is the analysis of the integrated model.
This activity is not mentioned in the feature diagram in figure 2,
since it is carried out as part of nearly every other integrated
model processing activity. All tests concerning existing model
elements and/or their properties or relationships are model anal
ysis steps. There can also be transversal analyses, regardind co
herences of a whole model, a part of a model (e.g. one building)
or a special view to a model (e.g. geometry). Using the TGraph
structure traversal and analysis of the integrated model is done
repeatedly during runtime using the graph API and/or GReQL
queries. Transversal analyses are particularly supported by the
graph API. For example it offers iterators for all nodes or edges
of a special type (and it subtypes) in the whole model.
Listing 1 : Building analyser results.
Building 1
Id : 2
Name: Example building
Description: Example building model for testing.
Year of construction : not known
Year of demolition : not known
Number of appearances : 0
Number of building installations: 0
Number of building parts: 0
Number of boundary surfaces : 9
Number of wall surfaces : 4
Number of roof surfaces : 4
Number of ground surfaces : 1
Number of openings: 6
Number of doors : 1
Number of windows: 5
Number of 3d faces: 15
Number of 3d points : 34
Lowest 3d point: Point 1: (0.0, 0.0, 0.0)
Highest 3d point: Point 9: (2.0, 1.0, 4.0)
Height : 4
Width: 4
Depth : 5
Volume: 68
To demonstrate the usage of querying with GReQL an additional
component BuildingAnalyser was developed, that writes in
formation about all buildings of the integrated model into a .txt-
file. The file contains different kinds of information. At first there
is semantic attribute information like name, description and year
of construction/demolition of the building. Moreover there is se
mantic entity information like the number of wall, ground and
roof surfaces, the number of doors and windows, and so on. Fur
thermore there are geometric information like the count of points
and faces of the building geometry or the lowest and highest point
of a building. And there is inferred semantic information com
puted using semantic background knowledge in combination with
geometry information, like the building height, the building vol
ume, and so forth (listing 1).
5 CONCLUSIONS AND FUTURE WORK
This paper showed how geometric, topological, semantic and ap
pearance information can be integrated in one integrated graph
model. The class of models was defined by an integrated model
schema. Graph representation gives rise to all kinds of algorith
mic processing, some examples of which were given, including
model creation, improvement, transformation, analysis and ex
port. Using a lightweight Java component model some example
components were implemented and illustrated based on a simple
example.
Though the example has toy character, it should suffice to demon
strate the wide range of manipulation possibilities given by an
internal integrated graph representation for the enhancement of
urban object models. Since TGraph technology is easily applica
ble to graphs containing millions of elements, the approach scales
to a wide range of applications.
The integrated model was developed in the context of a project
for object-recognition (Falkowski et al., 2009). It forms the basis
for the application of efficient graph-matching algorithms in this
context.
The integrated model schema is still under construction. But it
is easily modifiable and each of the three parts can be replaced
by different variants. Further goals are the enhancement of the
schema for the full CityGML base profile (CityGML[full]) and
the support for other urban object description languages, like KM-
UCOLLADA (section 1.1). Here the tasks are the change and en
largement of the integrated model schema and the adaption of all
existing processing components. Some of the described activi
ties could be splitted to more processing steps. A lot of them can
be composed to interesting combined processing activities. And
there could even be interactive processing components.
Further research topics could be the supplement of more com
plex model parts to an existing integrated model or the integra
tion of two different integrated models. Another interesting field
is the inference of semantics from geometric, topological and/or
appearance information.
ACKNOWLEDGEMENTS
This work has been carried out in close cooperation with Peter
Decker, Dietrich Paulus and Stefan Wirtz from the Work Group
Active Vision as well as Lutz Priese and Frank Schmitt and
from the Laboratory Image Recognition, both at the University
of Koblenz-Landau.
REFERENCES
Cox, S.. Daisey, P., Lake, R., Portele, C. and Whiteside, A., 2001.
OpenGIS Geography Markup Language (GML) Implementation Speci
fication. Technical Report 3.1.1, Open Geospatial Consortium, Inc.
Czarnecki, K. and Eisenecker, U. W., 2000. Generative Programming:
Methods, Tools and Applications. ACM Press/Addison-Wesley Publish
ing Co., New York. NY, USA.
Ebert, J.. Riediger, V. and Winter, A., 2008. Graph Technology in Re
verse Engineering, The TGraph Approach. In: R. Gimnich, U. Kaiser,
J. Quante and A. Winter (eds), 10th Workshop Software Reengineer
ing (WSR 2008), GI Lecture Notes in Informatics, Vol. 126, Gl, Bonn,
pp. 67-81.
Falkowski, K.. Ebert, J., Decker. P, Wirtz, S. and Paulus, D., 2009. Semi
automatic generation of full CityGML models from images. In: Geoin-
formatik 2009, ifgiPrints, Vol. 35, Institut für Geoinformatik Westfalische
Wilhelms-Universität Münster, Osnabrück, Germany, pp. 101-110.
Foley, J. D., van Dam, A., Feiner, S. K. and Hughes, J. F., 1990. Computer
Graphics. Principles and Practice. Addison Wesley.
Gamma, E., Helm, R., Johnson, R. and Vlissides, J., 1995. Design Pat
terns: Elements of Reusable Object-Oriented Software. Addison-Wesley
Longman Publishing Co., Inc., Boston, MA, USA.
Groeger, G., Kolbe, T. H., Czerwinski, A. and Nagel, C., 2008. OpenGIS
City Geography Markup Language (CityGML) Encoding Standard. Tech
nical Report 1.0.0, Open Geospatial Consortium Inc.
Herring, J., 2001. Topic 1: Feature Geometry (ISO 19107 Spatial
Schema). Technical Report 5.0, Open Geospatial Consortium Inc.
Muller, D. E. and Preparata, F. P, 1978. Finding the Intersection of two
Convex Polyhedra. Theor. Comput. Sei. 7, pp. 217-236.
Pohl, K., Böckle, G. and van der Linden, F., 2005. Software Product Line
Engineering. Springer, Berlin.
