Stem.
carto-
; using
rs and
n. We
| intel-
nented
"artog-
Carto-
and its
incepts
tion to
n Sys-
ction 5
set of
tion of
terpre-
mplish
be ex-
v maps
ays the
1 delet-
rtogra-
desired
he first
: on the
consid-
jects in
on uses
domain-
r chang-
tors are
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
TOURIST
Figure 1: Generalizations (Glover and Mackaness, 1999).
OP1 - Simplification: reduce the number of vertices em-
ployed to represent the element, preserving the origi-
nal appearance.
OP2 - Smoothing: displace the vertices used in the repre-
sentation, in order to eliminate small disturbances and
to capture the overall shape.
OP3 - Aggregation: join nearby elements.
OP4 - Amalgamation: join nearly contiguous and similar
areas, by eliminating borders between them.
OPS - Merging: join two or more parallel lines that are
close to each other into a single line.
OP6 - Collapse: reduce the dimension of the representa-
tion of an object.
OP7 - Refinement: discard unimportant elements, which
are close to important ones.
OP8 - Exaggeration: increase the dimensions of elements
considered important for the map.
OP9 - Enhancement: increase the dimensions of symbols
presents in the maps.
OP10 - Displacement: shift the position of a feature, in
order to make it distinct to other ones.
OP11 - Classification: group objects which share identi-
cal or similar characteristics into categories.
OP12 - Symbolization: change objects (or categories) for
symbols.
3 CARTOGRAPHIC GENERALIZATION APPLIED
TO VIRTUAL REALITY
This section shows how the concepts utilized in cartographic
generalization can be applied to virtual reality. First we
present the transposition of operators and then how domain-
specific knowledge was modelled as the rules of an expert
system. Details can be seen in (Frery et al., 2003).
201
3.1 Cartographic Generalization Operators in Virtual
Reality
OP1: these are the techniques for polygon simplification;
see (Guéziec et al., 1999, Vieira et al., 2003).
OP2: image filtering (typically low-pass) and resampling
in order to produce a new similar but smaller texture.
OP3, OP4, OPS: these operators can be built with a sin-
gle operator in virtual reality, or as the initial stage of
operators OP11 and OP12.
OP6: the result of applying OPI iteractively.
OP7: small objects (or objects defined by the user) may
not be shown in the LOD.
OPS8: the scale transformation in 3D objects.
OP9: the scale transformation in 2D objects or textures.
OP10: the translation of 3D objects.
OP11: objects are grouped into categories according to
their main features.
OP12: objects are replaced by symbols or icons defined
by the user.
3.2 The Expert System
Expert systems simulates the behavior of the expert hu-
man for the resolution, diagnostic and analysis problems
in some knowledge domain-specific (Luger, 2001).
The Cartographer builds the map generalization with his
knowledge and experience about the selection objects and
application of operators. We modelled the Cartographer's
knowledge in knowledge rules (rules if - then), and they
are applied to VRML virtual worlds for building the LODs
or 3D generalizations. Also, unimportant objects for the
theme must be discarded, so two knowledge bases were
modelled: one for the selection of important objects and
another for the application of operators.
4 THE GENERALIZATION 3D SYSTEM
Figure 2 shows the architecture of the system. The picture
represents the real world that feeds a SIG (Geographical
Information System). VRML files are generated from this
input for an specif area. The Generalization 3D System is
not concerned with these steps.
The Representation Model reads and identifies the objects
in the virtual world. The input of this step are the VRML
files, and objects are stored in a MySQL database. Com-
plex objects are are defined by the DEF node in VRML,
ant the classification of objects in simple or complex is au-
tomatically performed by the system.
The user classify the objects in one of three categories: pri-
mary, secondary and indeterminate. In the first are large
