Jatabase of
t « IGEA
f Software
e Building
ate Border
Ljubljana,
d Cadastre,
Overview
sodetski
> up of the
9.
>s for the
as of 23
stre Project
4.
type of the
. including
irst and the
).
Support to
SPATIAL MODELING
Dana Klimesova
Czech University of Agriculture, Prague
Faculty of Economics and Management, Dept. of Information Engineering
Kamycka 129, 165 21 Praha 6 — Suchdol, Czech Republic
klimesova@
and
Czech Academy of Sciences
Institute of Information Theory and Automation, Dept. of Image Processing
Pod vodarenskou vezi 4, 182 00 Prague 8, Czech Republic
klimes(@utia.cas.cz
Commission VI, WG VI/3
KEY WORDS: Object-oriented classification, Contextual features, Hierarchical representation, Spatial information sciences,
Class-related features, Attribute features, Group hierarchy
ABSTRACT:
The contribution deals with the contextual design of spatial data for the purposes of the regional development, local management.
The development of new information technologies, image processing techniques and knowledge-based databases together with the
geographical networks environment will provide quite new and considerably wider possibilities of using GIS. The paper describes
the role of remote sensing data and contextual modelling for the context oriented geo-information and tries to provide the framework
for the object hierarchy of classes and propose the ways of structure and behaviour modelling. GIS architecture is open to incorporate
new requirements of knowledge-based analysis and modelling.
1. INTRODUCTION
1.1 Spatial data
Only permanently comparing local resources with global
possibilities and accounting the changes of the world around
can make successful regional policy and development. The
spatial data has an extremely complex and variable structure
and the standard way to set up applicable and efficient
geographical database it takes years. On the other hand we
are surrounded by huge amount of data coming from
satellites. It means the data with excellent properties: exact,
objective, regular reachable, with very different resolution,
with wide offer of spectral channels in disposal, with
temporal component by request, ...
But to gain the useful information from remote sensed data it
needs the effective processes of data structure transformation.
Program packages for the classification, segmentation and
contextual evaluation of partial results, the detection of
changes and context sensitive analysis if possible together
with the knowledge-based databases. Many applications
incorporate the temporal component into accounting.
We can take into the consideration the context coming from
the surrounding pixels, the context between primitives, and
the context among objects and between spectral channels or
image plains and their combinations. In definite sense, the
whole image can be understood as the result of all possible
contextual relations [1-3], [5,6]. Remote sensing together
with the technologies of data processing and analysis
represent the main source of data for the creation of
geographical database. The satellite data with the resolution
Im are in disposal for many applications and the information
power continually increases.
The fine resolution of data requires new technology of data
processing and evaluation. New proper edge detection
techniques can help us to detect and interpret objects much
more exactly and simplify the process of vectorization.
Further development is based on the accessible information,
knowledge-based decision making and context sensitive
analysis application.
2. CONTEXT OF SPATIAL DATA
2.1 Object-oriented design
Methods of managing and distributing data and also the data
resolution have changed rapidly. Spatial data are collected
and processed and during the last couple of years the data
flows in and between organizations have extremely
increased. And the data management tools and techniques are
continually changing. Many users have their own specific
requirements for the access to data, with regards to the
