ul 2004
ildings
le map
RDB.
re 5, is
riments
nce the
ared to
fficient
ime to
t of the
t likely,
re than
similar
vantage
eeds of
user is
ts. For
imarks,
ithin a
ties are
' in the
les like
At this
ects are
nstance
a single
t (Pol).
th a dot
(right).
ts exact
of this
dings is
ngs and
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Ungeneralised
Figure 7. Emphasise Pols using MRDB:
buildings (left). Pol in full detailed shape, re-
generalised remaining buildings (right).
This can be achieved using the following process:
Firstly we identify the building block (Figure 6 left) and then
request the individual detailed buildings the block is composed
of from the MRDB (Figure 7 left). Afterwards, a new
generalisation is triggered where the landmark object stays in
original shape and the surrounding objects are aggregated,
leading to the final representation (Figure 7 right). A more
detailed description can be found in Elias et al. (2004).
On the one hand the user gets detailed information of the
objects of interest with a high level of detail and on the other
hand he or she gets generalised spatial information of the
remaining objects which are adequate for the actual scale and
the display size (Figure 7 right).
In this case we preserve a clear map view and at the same time
provide all the necessary information covering a larger area.
These kinds of applications can be realised using the structure
of the underlying MRDB. Because only a low number of
objects has to be transferred and generalised the whole process
of adapting the map can be done in real-time.
5.3 Information drilling
Another use case which benefits from the MRDB-structure is
the "information drilling" scenario. The mobile or non-mobile
user may be interested in a certain area or a certain object in the
map. For instance the user can request the buildings which are
located inside a certain built-up area or a single building which
is part of an amalgamated object in the actual map. For example
rescue parties might be interested in a large scale map to
navigate to the person who needs help as well as in detailed
information of some parts of the map, i.e. the environment of a
certain building.
From the users point of view it is only necessary to click on the
object or position he or she wants to drill for more information.
This click activates a request sending the ID (identification)
with the number of the built-up area. A servlet passes this
request on the WFS by requesting only those objects which are
linked to the requested object. The user can determine the
deepness of drilling which can go down to a detailed building
plan linked with a certain building.
5.4 Getting attributes from linked objects
In some situations it may be useful to get information which are
not connected directly to the objects in the actual map. One
example may be the name of a city, district or postcode area
which is not typically an attribute of the buildings within the
map of a certain scale. However if we go down to smaller scales
in the MRDB the object built-up area or city owns the attribute
"name of the city" and thus the buildings linked with this object
can access this information (Figure 8).
139
Building: Restaurant
Address: Leinaustr. 7
A RIA E
Figure 8. Information from small scale (upper left and
right) accessed for large scale presentation using
the links in MRDB.
That means the information the user can get is not limited to the
attributes stored with a certain feature. Because of the links
between the objects in the database the user has indirect access
to all the information stored in the database.
6. DISCUSSION & CONCLUSION
The motivation for this paper was to circumvent the
disadvantage of small display devices by generating adaptive
maps containing only the necessary data visualised in an
adequate way. This leads to generalising the data processed in
real-time like we aim at in the GiMoDig-project. As the
generalisation process can't be handled completely in real-time
because of its complexity we need pre-computed data, stored in
an MRDB.
As demonstrated by several examples, the benefits of an MRDB
can be used for different applications. If we concentrate on
mobile applications it serves to customise the map. The objects
of interest are visualised in a more detailed way. On the other
hand the user gets access to all information related to several
scales. The whole architecture combining a WFS and an MRDB
allows for fast and mobile access to all spatial information
stored in the database. In this study we have shown that an
MRDB can be used to enhance the information visualisation on
a small display. It is also shown that an approach based on
repeated WFS-queries can be used for accessing the MRDB in a
real-time environment.
We implemented, tested and presented only some possible use
cases based on such a combination. In general the benefit is that
there is no need to store all the information on the client device
and at the same time the client has access to all the information
available in the database.
The future work will concentrate on extending the MRDB to
obtain a multiscale database starting at very a high level of
detail (1:5k) up to a small scale (1:1Mill.). Additionally the
middle layer will be upgraded to contain an intelligent rule
