ISPRS, Vol.34, Part 2W2, “Dynamic and Multi-Dimensional GIS”, Bangkok, May 23-25, 2001 
INCORPORATING 3D GEO-OBJECTS INTO AN EXISTING 2D GEO-DATABASE: 
AN EFFICIENT USE OF GEO-DATA 
Jantien STOTER 
Tel:+31 -15-2788136, Fax:+31 -15-2782745, E-mail:j.e.stoter@geo.tudelft.nl 
Peter VAN OOSTEROM 
Tel:+31 -15-2788136, Fax:+31 -15-2782745, E-mail:oosterom@geo.tudelft.nl 
Section GIS Technology, Department of Geodesy 
Faculty of Civil Engineering and Geosciences 
Thijsseweg 11, P.O. Box 5030, 2600 GA Delft 
The Netherlands 
Keywords: 3D data models, 3D GIS, 3D spatial features, geo-DBMS, SQL for spatial data types 
abstract 
GISs are changing from static modelling in 2D to dynamic modelling in multi-dimensions. Yet, it will take an infinite time until the world 
as been defined within a full three- or four-dimensional approach. Besides that, a 2D approach still suffices in many cases. A relevant 
question is what kind of detail and resolution are required to satisfy the actual needs. 
2D geo-information is available at large amounts, at different scales and covering many application domains. An approach to combine 
2D and 3D geo-data in one DBMS, including storage and query can support an efficient use of data. Therefore at the Department of 
Geodesy, Delft University of Technology a research is started on this topic in order to look for possibilities and accompanying 
complications. 
The background of this research is the need for more than just a 2D approach by the Dutch Cadastre (the Netherlands’ Kadaster). In 
countries with an intensive use of land, like in the Netherlands, there is a growing interest to use space under and above the surface. 
With this interest, there is also a need to develop and maintain insight in the situation under and above the surface. Since this detailed 
information is not needed for rural land and furthermore there is a lot of information available in the second dimension, a system is 
required which is able to add 3D geo-objects in the current 2D geo-DBMS. 
In this research, concepts are developed to integrate 3D objects in the current geo-database defined in 2D. The integration consists of 
data definition, data manipulation and data querying of 2D and 3D data in one environment. The concepts are being translated into 
prototype implementations. These implementations cover the following aspects: 
modelling and storage of 3D geo-objects into a geo-DBMS based on 2D parcels (data definitions) 
collecting, preparing and inserting (generalising/converting) 3D data 
querying and visualising the data (administrative, 2D, 3D and in combination with each other) 
The research will finally lead to the implementation of all these aspects in one system: a geo-DBMS supporting 2D geo-objects and 3D 
geo-objects as well as in combination with each other. 
This article describes both the background of the research as well as the progresses made so far. The basic concepts of a geo-DBMS 
are discussed, including the recent standardisation efforts by ISO and the OpenGIS Consortium. The current 2D model of the 
Netherlands’ Kadaster is explained, which also includes a discussion of some possible 3D model extensions. Some of the issues 
involved in 3D geo-data collection and preparation are described, either based on CAD models created by designers or on surveying. 
Important in this is the insertion of the 3D data into the 2D geo-DBMS and the definition of the 'horizontal zero level’. Some examples of 
integrated querying and viewing of 2D and 3D data are given. Finally, conclusions and future work are described. 
1. INTRODUCTION 
In science as well as in applications, there is a growing interest 
in modelling the world in multi-dimensions. In this scope, GISs 
are changing from static modelling in two dimensions to dynamic 
modelling in three dimensions. However, a full and complete 
three (or four) dimensional model of the world is a too heavy 
solution in many applications. Thereby, not only the 
development of such models and representations should be 
considered, but also the availability and the effort that it takes to 
collect the needed data to make the developed models 
operational. Finally, incorporating the developed concepts into 
present systems and working processes is an important factor 
for reconsideration. 
A relevant question is what kind of detail, or map-scale is 
required to satisfy the actual needs. Geo-information in two 
dimensions is available at large amounts, at different scales and 
covering many themes. In addition, an approach to combine two- 
dimensional and three-dimensional temporal data in one 
database, including storage and query, can support an efficient 
use of data. Therefore at the Department of Geodesy, Delft 
University of Technology a research is started on this topic in 
order to look for possibilities and accompanying complications. 
The background of this research is the need for more than just a 
two-dimensional approach by the Dutch Cadastre (the 
Netherlands’ Kadaster). In the Netherlands, there is a growing 
interest to use space under and above the surface because 
there is a high pressure on the available land. With this interest, 
there is also a need to develop a spatial Database Management 
System (DBMS) to maintain insight in the situation under and 
above the surface. 
This detailed information is often not needed for rural land and, 
what is more important, there is a lot of valuable information 
available in the second dimension. Therefore, a system is 
needed which is able to integrate 3D information with 2D 
information. 
This article describes the model that has been developed so far 
to integrate 3D objects in the current geo-database defined in 
2D. This conceptual model consists of data definition (data 
models), data manipulation (inserting, deleting, and updating) 
and data querying of 2D and 3D data in one environment. This 
research is started as a practical request of the Netherlands’ 
Kadaster. However, the research question covers the 
fundamental problem of combining newly required 3D data with 
conventional 2D data, which is available in a wide range and 
does still suffice in most cases. To understand the developed 
concepts, first the practical case used in the research is 
explained in Section 2 of this paper. Section 3 discusses the 
basic concepts of a geo-DBMS, including the recent 
standardisation efforts by ISO and the OpenGIS Consortium. 
The current 2D model of the Netherlands’ Kadaster is explained