1 2004
of the
1at the
is. not
quired
issues
ng the
Global
types
ntains
r than
| to be
ymatic
ot be
6. For
bases.
>, it is
ämter.
m the
le the
: data
d' are
m the
erman
in city
s and
ick. of
Other
ke are
k
orded.
vedish
otype.
ded in
a' and
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
5.2 Different geometry types
In Finland the forest is not available as an area feature. The
large scale topographic map depicts forest by a point symbol, so
does the topographic database (see Figure 3). Area colours (in
the map) or area features (in the database) are used for other
land cover features such as agricultural land, mire, meadow and
water body. The built-up area and forest share the background
(white colour in the map) The human map reader can
distinguish these two land cover classes from each other by
interpreting the symbols for individual houses and trees, but this
method is not easy to implement in a geographical information
system (GIS).
Figure 3: Finnish Forest as point feature type
5.3 Differences in conceptual modelling
Some feature types and attributes of the Global Schema have no
or only ambiguous relationship with the Local Schema because
of the significant differences in conceptual model, related e.g.
to watercourses. The GiMoDig Global Schema proposes a
classification into feature types 'Water except Inland' (i.e. Sea,
Ocean), 'Watercourse' (which aggregates 'River', 'Canal' and
Ditch) and 'Lake/Pond'. Sweden does not distinguish 'Ocean'
from 'Lake' and from area-type 'River', instead they put all
instances into one feature type 'Vatten' and only separate area-
type features from line-type features. Finland has a similar
approach but distinguishes at least the "Water with Current'
(Watercourse) from 'Water without Current' (Ocean, Lake). For
visualisation purpose this generalised approach does not matter
since all water area is displayed in same blue colour, but for
GIS applications the approach of Sweden and Finland may lead
to difficulties. :
Denmark does not make any distinction between ‘Airports’,
Railway stations’ or 'Power Plants. They are all together
recorded as 2700 "Technical Area'.
Another problem in Denmark is the 3119 'Recreational Area'.
This feature type comprises all kind of recreational areas, e.g.
"Amusement Park', 'Campground' and also 'Park'. There is no
attribute that further specifies the instances according to the
type of the recreational area.
The criteria for classification of roads into primary, secondary
or local road differs between countries. Germany applies the
official administrative categories. The other countries classify
by the width of the road. There is no uniform European
classification scheme except for the few trunk routes with 'E'-
numbers.
5.4 Differences in data modelling
Concerning data modelling there also exist differences between
the four participating countries.
The German dataset differs from the data model of the
Scandinavian partners especially for roads with separated lanes.
The NMAs of Sweden, Finland and Denmark portray the lanes
in each direction with individual centre lines. In contrast, the
German dataset models the motorways by a complex feature
that is composed of road body (one centre axis) and road lanes
(two centre lines). For the Global Schema we decided to include
only the individual lanes and skip the centre axis of the German
motorway, with the lanes inheriting the attributes from the
complex feature and the road body.
Concerning railway features each track is modelled separately
in Finland, Sweden and Denmark. It appears that Finland shows
the complete layout whereas Sweden and Denmark apply some
generalisation. Double track railways appear as two parallel
lines in Finland and Denmark. In Sweden and Germany the
railway is recorded as a single centre line if tracks are parallel,
the number of tracks is stored as an attribute. In case of tracks
not being parallel the German data model applies a complex
feature like the motorway. In case of complex railway yards or
stations the German data model applies an area-type feature
railway station’ where onky the major track is continued
through the area and connected with the railway lines leading
to/from the station. In the Global Schema we keep both
approaches as it is neither feasible to generalise the
Scandinavian track layout nor to add the missing tracks to the
German data.
5.5 Level of Detail
From the sample data in the cross-border test areas it becomes
obvious that the level of detail differs from country to country.
Germany and Denmark have great differences in their
hydrographic network. The network of drainage canals and
ditches in Denmark is much denser than in Germany. The
difference is less significant in hill areas where the selection
criteria for natural watercourses appear quite homogenous at
both sides of the border.
The hydrographic network at the border between Sweden and
Finland can be compared with the German-Danish example.
The network of artificial watercourses appears much denser in
Finland than in Sweden. Another problem illustrated in the
example is the lack of connectivity in the Finnish and Swedish
network. We can recognize many gaps and isolated objects (see
Figure 4).