san
the
J
[iC
of
ol
Sad
ive
ale
lex
in
or
ent
wo
ase
TOT
0a
uls
at
CCS
ted
ion
ap
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004
resolution, to conflicts between objects, to the quality of the
display and to legibility; the latter focuses on the content of the
data, their completeness and accuracy.
Selection of important information from a geographic database
implies a capacity for abstraction dependent on an in-depth
knowledge of the concepts connected with it. This aspect of
generalization takes the name of semantic generalization. It
refers to the meanings and functions of the cartography and
depends on the ability to identify hierarchical structures in the
geographic information.
Graphic representation. requires the use of symbols in the
identification of information; this leads us to consider both a
transformation of scale of geometric data and the choice of
graphic and textual elements in order to communicate their
meanings correctly. The latter aspect is called geometric
generalization and is governed by the relationship between
semantic generalization, the use of symbols and map scale
constraints.
To simplify the graphic representation, some operators, which
will be described below, have been developed and, quite often,
we were obliged to shift the symbols from their original
positions so as to avoid superimposition of adjacent elements.
In practice, graphic simplification must be performed only after
having acquired a good knowledge of the properties and
characteristics. of the phenomena represented by the map.
Today's research in this field is all for the purpose of reaching a
suitable degree of automation of operations; however, this goal
depends on the development of automatic model recognition
operations, a technology not yet available in this software
generation.
Decisional operations in the generalization process are still
strongly dependent on human capacity to plan, structure and
represent real world phenomena. This is due to the fact of
possessing a global knowledge of such phenomena, both of
their graphic representation and the important logical
interconnections behind their dependencies.
The Topography Section of the University of Cagliari has
performed an experiment over a two-year period for the purpose
of producing bases derived at different levels of detail starting
from the nominal scale of 1:10000. To this end, a first phase of
treatment of the cartography used was performed; following
this, an operational method for cartographic generalization in
the GIS environment was implemented. The database chosen
was the CTR of the Sardinian Regional Administration at the
nominal scale of 1:10000.
3. PROCESSING OF THE CARTOGRAPHIC BASE
The study of the processing of the numerical cartography was
performed in two stages. In the first, that of analysis, the
cartographic data were studied for classification of encoding
errors and identification of geometric incongruities. In the
second, the procedures for correction of cartographic data
(without changing the level of present knowledge) were defined
and reorganization of covers was addressed.
The final result was having available a congruent cartographic
base, the correspondence between the originals indicated in the
encoding and the graphic sign, the closed areals, the entire
transportation and fluvial networks and the organized encoding
in the covers.
4. GENERALIZATION MODELS
Subsequently, attention was focused on the generalization of the
cartographie datum through the use of the ESRI ArcGIS 8.1
package; within this environment some generalization functions
were implemented by means of the compilation of specific
routines in Visual Basic language.
73
To plan the sequence of operations for performance in the
single stages of the process, the model proposed by Beat Peter
and Robert Weibel (1999) for cartographic generalization,
Figure 1, was taken as a starting point.
pessum c sem MAP CONTROLS ANALYSIS
- Map scae - cartographie principles
+ map purpose - source Gata:
- graptis limits - data mode
- outzu: mecium - acquisition method
- symbelogy ssif n c
coabiiities
TFEMATIC GENERALIZAT.ON
- aggregstion of categories
- resamping cf raster grid
CONSTRAINTS DEFINITION
classes spatia! scope:
- graphical - object
- tecologica - category
- structural - partitionémap
- Gestalt
'
ON: CONSTRAINT? TRANSLATION
classes:
- size
- sistance and proximity
- shape
- topolcgy
. sity and cistrivution
- pattern anc aiignment
'
PROCESS MODELLING
Maeasures for Quality Zvaluatıon
m— processing sequence
functional relationshic
Figure 1. Beat Peter and Robert Weibel (1999)
In this model the generalization specifications are arrived at
through analysis of the cartographic database and its
comparison with the different constraints. This procedure
allows formalization of the objectives to be reached and address
them, controlling the degree of satisfaction of the results. In this
way, after quantifying the value of the constraints, it is possible
to choose the action which for that particular object or set of
objects best satisfies the purpose of the cartography to be
derived.
5. PROCEDURES
Presented are the functions and procedures used in
implementing the generalization process for derivation from the
scale of 1:10000 to the scales of 1:50000, 1:100000 and
1:250000.
These functions and procedures were divided into:
® functions resident in the software platform. They are
the functions aimed at the generalization which were
directly available in the program;
® Functions and procedures implemented for the first
time. They are the functions and procedures that were
implemented in the course of this work in the GIS .
environment used for the purpose of automating some
steps in the generalization procedure.
