International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Methods for the automatic identification of corresponding
objects, adjusting the object geometry, and detection of changes
which occurred in reality, but are not yet integrated in one of
the data sets, will be developed. This is done with a focus on
the above mentioned data set. Geometric aspects and methods
will be described, namely the merging of segmented objects
and the adaptation of the geometry by using a rigid
transformation, followed by a mere intersection and evaluation
of the resulting elements.
In this project the German digital topographic data set (ATKIS)
will be chosen as reference, therefore the geometry of the
geoscientific maps will be adapted without using constraints
regarding accuracy or actuality so far. The approach, however,
will be extended in the near future, to also take the relative
accuracy and importance of the objects to be integrated into
account.
2. RELATED WORK
Data can be integrated and fused for mutual benefit: Walter &
Fritsch, (1999) present an approach that fuses two different data
sets with road information with the aim of mutually exchanging
attributes of the two data sets. The integration of vector data
and . raster data is - being investigded in. à
GEOTECHNOLOGIEN partner project with the aim of
enriching a 2D-vector data set with 3D-information (Butenuth
& Heipke, 2003). Data integration or data matching is also
needed for update purposes, e.g. when a data provider has to
deliver up-to-date information details to his customers (Badard,
1999).
A conflation component strategy to provide independent but
interoperable modules to solve special integration problems has
been developed by Yuan & Tao, (1999).
Integration can be used for data registration, when one data set
is spatially referenced and the other has to be aligned to it
(Sester et al, 1998). A conceptual framework for the
integration of geographic data sets, based on a domain ontology
and surveying rules, was developed for update propagation
between topographic data sets (Uitermark, 2001).
Finally, data integration is needed for the generation of
Multiple Resolution Data Bases (MRDB); in this case objects
of different geometric and thematic resolution have to be fused
(Mantel, 2002).
3. USED DATA SETS
For the research in the GEOTECHNOLOGIEN project three
data sets are used: the topographic data set ATKIS, the
geological map and the soil-science map, all at a scale of
1:25000. When going from analogue to digital maps, new
possibilities for data handling and analysis appear: basically,
the combination of different data sets in a geo-information
system (GIS) is enabled.
Simple superimposition of different data sets already reveals
visible differences (Fig. 1). These differences can be explained
by comparing the creation of the geological, the soil-science
map and ATKIS (Goesseln & Sester, 2003).
As for ATKIS the topography is the main thematic focus, for
the geo-scientific maps it is either geology or soil science, these
maps have been produced using the result of geological drills
and according to these punctual informations, areal objects have
been derived using interpolation methods based on geoscientific
models. However they are related to the underlying topography.
The connection between the data sets has been achieved by
copying the thematic information from topography to the geo-
scientific maps at that point of time the geological or soil-
science information is collected. This is done by using up
scaled copies (1:25.000 to 1:10.000) of topographic maps. The
selection and integration of objects from one data set to another
one has been performed manual and in most of the cases the
objects have been generalized by the geoscientist.
While the geological content of these data sets will keep its
actuality for decades, the topographic information in these
maps do not: In general, topographic updates are not integrated
unless new geological information has to be inserted in these
data sets.
The update period of the feature classes in ATKIS varies from
one year up to three months — in general, 10% of the objects
have to be updated pèr year (LGN 2003).
Fig. 1 : Simple superimposition of ATKIS (dark border,
hatched) and geological map GK 25 (solid fill).
The geoscientific maps have been digitized to use the benefits
of digital data sets, but due to the digitalization even more
discrepancies occurred.
Another problem which amplifies the deviations of the
geometry is the unequal data model between these data sets.
Geological and soil-science maps are single-layered data sets
which consist only of polygons with attribute tables for the
representation of thematic and topographic content, while
ATKIS is a multi-layered data-structure with objects of all
geometric types, namely points, lines and polygons, equally
with attribute tables.
These differences in acquisition, creation, modelling and
updating lead to discrepancies, making these data sets difficult
to integrate. The amount of financial and human resources
which is needed for the removal of these discrepancies can
hardly be afforded. Therefore, new methods are required which
offer an automatic or semi-automatic process capable of
detecting and removing the differences between these data sets
and supporting a human operator in this process.
In order to identify changes in the data sets and update the
changes, the following steps are needed: identification of
corresponding objects in the different data sets, classification of
possible changes, and finally update of the changes.
1250
Intern
4.1 C
Data |
differe
part o:
hetero,
integra
vertica
matchi
spatial
maps,
more)
modell
The re
maps i
conflat
master
geome
the en
ATKIS
geomet
content
data se
During
tasks. '
the hig
should
the geo
4.2 Se
At the |
— whic]
— of al
are repi
be used
mandat
be the f
Four di
Fritsch,
The int
categori
In the f
“water
and test
in all da
One airr
to real
close pa
science.
