International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
   
MRDB 
  
    
   
  
  
WFS 
Request N 
WES p 
"| Request 
  
LoD low low LoD data 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
low LoD data 
  
  
  
  
  
  
  
Figure 3. Receive data for multiscale map from MRDB via 
repeated WFS-request 
5. CASE STUDIES 
This section describes several applications of utilising an 
MRDB for creating individual maps in real-time. 
5.1 Variable-scale maps 
A major problem with mobile maps is their small displays, 
which puts high demands on the selection of cartographic data 
to be shown. This becomes problematic when the user requires 
a considerable amount of cartographic information. In personal 
navigation, for example, users often need both a detailed map of 
the area surrounding the current position as well as an overview 
map. This means, in cartographic terms, that the user requires 
both large-scale and small-scale cartographic data. To 
overcome this problem a map can be created where the scale in 
the middle (close to the user’s position) is larger than the scale 
in the border areas of the display (Figure 4). Details of the type 
of variable-scale map used in this case study are given in Harrie 
et al. (2002). 
  
Small-scale data 
   
    
User’s 
position 
€ 
     
  
  
   
Large-scale 
data 
   
  
  
  
Figure 4. The figure illustrates the circular cap (with 
radius ry) where large-scale data is shown and 
the area outside this cap where small-scale data 
is shown. (Harrie et al., 2003) 
The main problem of creating a good variable-scale map is that 
the level of detail in the cartographic information should be 
different in the centre of the map than in the border area of the 
display. For this an MRDB is very useful. 
Figure 5 shows a variable scale map created in real-time using 
an MRDB. The map was briefly created as follows. First a WFS 
request was performed to the small-scale layer in the MRDB. In 
the processing layer then the built-up areas in the centre of the 
map were identified. By using the MRDB links a new request 
was performed that requested the buildings (from a large-scale 
layer in the MRDB) within these built-up areas. These buildings 
then replaced the built-up areas. Finally, the variable scale map 
was created by a coordinate transformation. 
  
  
  
Figure 5. A variable-scale map created using an MRDB. 
To create a variable-scale map, such the one in Figure 5, is 
fairly fast. The process that takes most time in our experiments 
is the transferring of data from the server to the client. Since the 
built-up areas objects contain much less data compared to 
corresponding building objects, the MRDB is an efficient 
approach for this application. It would take longer time to 
request all building objects and generalise the outer part of the 
map (by aggregation) than utilising an MRDB. And most likely, 
the result is visually more appealing if an MRDB is used. 
5.2 Emphasise landmarks and points of interest 
For general purposes maps, which serve for a lot more than 
only one person, the basic principle exists to present similar 
objects in a similar manner (Hake et al. 2002). The advantage 
of the non-printed mobile maps is to take the special needs of 
the user into account while creating the map. The user is 
interested in certain objects more than in other objects. For 
instance while navigating the user is interested in landmarks, 
eye-catching objects or points of interest located within a 
certain distance from the users position. 
To emphasise these objects in the map several possibilities are 
cogitable. In order to visualise one individual building in the 
generalised data set a possibility is to use graphic variables like 
colour, size or the shape of the object (Figure 6 left). At this 
point the problem occurs that in a certain scale the objects are 
generalised which e.g. means the buildings are for instance 
amalgamated which in turn means that we can't localise a single 
object to point out a landmark or a point of interest (Pol). 
Alternatively, the position of the Pol's can be marked with a dot 
or special symbol (Figure 6 right). 
  
Figure 6. Emphasise Pol using color (left) or symbol (right). 
This however hides underlying structures and prevents exact 
identification of buildings. Thus, a combination of this 
individual building and the generalised background buildings is 
desirable, given in Figure 7 (right) with detailed buildings and 
aggregated builtup areas. 
138 
Inte 
  
This 
Firs 
requ 
of 
gen 
orig 
lead 
deta 
On 
obje 
han 
rem 
the « 
In tl 
prov 
The: 
of t 
obje 
of a 
53 
Ano 
the 
user 
map 
locat 
is pa 
resct 
navi, 
infor 
certa 
Fron 
objec 
This 
with 
reque 
linke 
deep! 
plan 
5.4 
In so 
not c 
exam 
whicl 
map « 
in the 
"nam 
can a