COMPUTER-ASSISTED GENERALIZATION - AN IMPORTANT SOFTWARE-TOOL IN GIS
Bernd M. Powitz
Institute of Cartography, Hannover University
Germany
ISPRS Commission IV
The present expansion of GIS has also a grand effect on computer-assisted cartography.
Nowadays there is a great demand for special software-tools and applications. GlS-users do
need algorithms and computer-solutions for the main items of generalization and visualization.
At the Institute of Cartography (IfK) at Hannover University it is a long tradition to work on
research projects in the field of computer-assisted generalization. The main emphasis is
put on the development of practical computer solutions. The actual repertoire of the IfK's
generalization software called CHANGE consists of a lot of program modulis and algorithms
for the automatic map design of large scale maps.
At the IfK individual products for cartographic preparation of the traffic network and the
buildings have been developed. A lot of difficult problems of cartographic generalization and
creation of digital cartographic models (DCM) have yet been solved (eg. smothing, elimination,
simplification, symbolization, selection, classification, exaggeration, combination etc).
KEY WORDS: Generalization, Visualization, GIS-software
1. INTRODUCTION
The current boom in the field of geographical
information systems (GIS) and the rapid
development in graphic data processing have
many consequences concerning computer
cartography. Mainly for visualization and
graphic output of spatial GlS-data suitable
software solutions are necessary. These
software modules are indispensable
components in GIS as alongside of data
aquisition, data storage and data management
etc. varied and flexible possibilities for
graphic presentation of spatial information
have to be provided. This brings new chances
and challenges for computer cartography and
familiar science disciplines.
Cartographic generalization is an imperative
means in the visualization of spatial GlS-data.
The generalization or visualization process
can be divided in single elementary steps
as selection, simplification, symbolization,
displacement etc. apart from the necessary
data preprocessing and data structuring.
At present the automation of generalization
for large scale applications is a subject
of intensive research. Computer-assisted
procedural solutions of particular genera-
lization tasks refering to selected feature
classes are nowadays available (e.g. generali-
zation-software of Institute of Cartography
(IfK) at Hannover University called CHANGE).
The aim is to construct a modular program
system in the frame of GIS through the link
of indepentent software modules that also
takes into account the interactions between
single processes and thus is able to support
effectively the visualization of spatial
information and of course the graphic
production of modern maps.
664
2. HISTORY
For many years several research projects at
the Institute of Cartography (IfK) at Hannover
University have contained problems and tasks
from the field of computer-assisted generali-
zation (Powitz, 1990). In the seventies the
prior endeavours of IfK research were
definitions, theoretical basis and methodical
analysis in cartographic generalization.
The eighties were the decade of program
developments and realisation of different
computer-assisted generalization procedures.
The current tasks and research projects
of the nineties are the provision of efficient
robust software-modules, the arranging
of complex system-components as powerful
software tools in GIS, the practical tests
of generalization programs and above all
the use of modern computer technologies
as for example object-oriented or rule-based
methods and procedures.
3. SOFTWARE-CONCEPT AT THE IFK FOR
COMPUTER-ASSISTED GENERALIZATION
For a constructive graphic presentation
of spatial GlS-data several software modules
for modelling, generalization and visualization
are necessary. These cartographically
demanding components have to be integrated
in GIS as powerful tools. Flexible graphical
output in various scales and an efficient
use of GIS-information are above all
solutions in computer-assisted cartographic
generalization. Cartographic generalization
is absolutely necessary where for faultless
legibility and good interpretation of graphic
representation in smaller scales on one side
given minimum dimensions have to be kept
—— (m (o 0 FN edn ped m=