The
r the
\. So
ts as
tures
d by
ntain
ands,
elect
5 are
Nhich
t 3-D
lle et
as la
ssent
cette
e aux
Is qui
tives
lessin
s par
points
pis de
posés
nées,
)n de
n des
INTRODUCTION
A Geographic Information System (GIS) is a system that
is used to manage and represent data for the description
of a part of the earth.
In the most existing GIS packages the third dimension is
not considered as a real dimension, but is only
considered as a supplementary dimension or rather as a
supplementary information.
At present, the well used GISs contain data that apply
only to two dimensions in the specific ground system and
that are completed with height information. These GISs
are called 2.5-D or 2-D * 1-D (Bill, Fritsch, 1991) (Kraus,
1991).
A Digital Terrain Model (DTM), in which only the third
dimension is significant, can be a part of these GISs and
is another important, well defined and often integrated
concept.
The aim of most existing 3-D systems is the viewing of 3-
D perspectives to perform urban planning, insertion of
urbanistic projects into urban areas. A real 3-D model
involves the use of a structure where for each point the x,
y and z coordinate have the same significance
(Rongxing, 1994).
1. OBJECTS IN URBAN LANDSCAPES
GISs are used to manage objects that describe a specific
part of the earth reality.
In an urban landscape there are artificial objects which
have a very complex architecture. This is the reason why
these objects contain very complex shapes.
Objects contain often very significant characteristics in
the third dimension because they are often higher than
wide. A house or an apartment building can not be
efficiently modelled with an area boundary. They have
got heights, several floors, cellars of which the fittings out
could be significant to be represented and managed.
There are also other types of objects that are very
complex, like bridges, tunnels or underground buildings,
and that in addition have got overlapping parts. This kind
of objects needs more than one level for their description.
Our system has to allow also the managing of such
complex objects. This is only possible if the third
dimension is considered as a real dimension and as a
real part of the objects.
2. GEOMETRIC RELIABILITY
The important factors in the definition of a system are the
geographic position in the 3-D space and the geometric
particularities of the data. These particularities will make
it possible to oppose the systems based on a reliable
geometry with the systems expected for the only
l'epresentation and visualisation.
The 3-D data are acquired by using surveying and
photogrammetric techniques with the principal objective
of achieving and preserving a reliable object. This can be
performed by storing as data all the measures obtained
during the acquisition.
The storage of all ground measures allows to get primary
acquired data in reply to specific geometric queries.
461
The object representation could be slightly different than
the measures done because it results from the fitting out
of the measures to geometric and topologic constraints.
3. BASIC STRUCTURE
In wide area systems that do not contain very complex
and subtle elements, a Digital Terrain Model (DTM) can
be used as a basic structure of the 3-D model. In this
case, the DTM is the structure that holds up the objects.
In a real and complete 3-D model, the DTM is the inferior
part of the superficial objects but is also the superior part
of the underground objects or, more generally, will
determine the intersection between the ground and the
objects placed over, under or in it.
This DTM is very important and is to be considered as
the basic concept of this modelling. It is used to define
and to build the ground terrain surface, but can also be
used to form and to build complex surfaced shapes
(Kager, Halmer, Heitzinger, 1996).
4. THE MODELLING OF 3-D OBJECTS
In order to achieve geometric reliability, solids from the
concept of constructive solid geometry can not be the
only primitives used for efficient 3-D modelling.
Nevertheless, this concept can give us a library of initial
geometric shapes.
To supplement this concept, new types of complex
objects can be hierarchically defined. These new complex
object types are in fact the result of the combination of
blocks by using geometric constructive operators. Blocks
are themselves a combination of basic primitives which
are, at last, split up to increase the precision grade.
4.1. Envelope of the blocks
Each block placed in the 3-D space is surrounded by an
envelope (smallest surrounding prism). The envelopes
make it possible to integrate objects which are not
definitively formed into the 3-D space. These envelopes
are also used for the selection of the objects by their
graphic representation.
There are two types of envelopes : the validated one that
contain definitively formed objects and the not validated
one that contain not definitively formed objects.
These different envelopes ask the problem of continuity
management between adjoining envelopes. Two
adjoining objects contained in two validated envelopes
will define several common intersections on the objects
themselves.
Two adjoining and not definitively formed objects that are
contained in two not validated envelopes will define their
intersections on the envelopes and not on the objects
themselves.
In the not validated envelopes all the construction
parameters of each object must be stored with the
objects so that their constructions can be fulfilled in a
next step.
These different concepts can be extended to the
primitives of objects that can that way be validated or
not.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
