MultiMap. This means that the geometrical accuracy
should lie in 25 - 30 meters in ICNW and 5 -10 meters in
metropolitan areas.
The data capture is carried out manually (Fig. 2.1). Analog
maps or orthoimages are scanned first. These digital raster
data should be transformed in the MultiMap reference
system and converted into the specified digitization image
format, before they can be displayed on computer screens
and then digitized. Available vector data can also be
integrated into the digitization process, so that data
capture can be speeded up.
On the basis of these data plots are generated for field
data collection (FDC). Taking these plots with them, the
field surveyors make site-the-spot investigations to collect
necessary traffic-relevant attributes and to mark possible
errors.The collected data are then put into the data base
manually.
Differential GPS is used as a supplement measure for
areas where available source materials are out of date or
no up-to-date ones are available.
An individual quality checking is catered for every step in
this vectorization process. For data base release, a
specially developed strict complex checking process is
carried out.
In order to ensure the high quality of MultiMap
permanently, updating with high quality data should take
place fast enough, before changes in reality occur again.
Instead of collecting every piece of information ourself in a
long costly process, relevant publishing houses, civil
engineering offices, driving schools and taxi associations
etc. are incorporated in order to obtain up-to-date high
quality source data. Naturally it goes without saying that all
of this information is verified, if necessary, on location by
the FDC before it is added into the database. Using this
process, we are able to release an updated database twice
a year.
2.3 Data capture through digital Photogrammetry
In many cases digital orthoimages will be the best solution
in terms of up-to-dateness, reality, accuracy and rich
information, which are essentially vital for the MultiMap. A
disadvantage compared to maps is that they don't carry
notations like street names. Another shortcoming is that at
the moment the costs of using Photogrammetry are
relatively high.
Digital and analog orthophotos have been successfully
used in our projects Austria and Switzerland respectively.
In Austria the department Land Topography of the
Bundesamt für Eich- und Vermessungswesen (BEV)
possesses different map works, which were suitable as
source materials. For the street network outside cities the
digital data from a BEV-database were used. This
database is built with photogrammetric stereo-restitution
method and currently further densified by the BEV. At that
time only an ICNW was available in addition to objects like
administrative boundaries, waters etc., so that further
source materials were needed for data densification in
metropolitan areas. For this purpose orthoimages (scale
1:10.000) in analog form were bought from the BEV, which
were then scanned in our production.
It was more difficult to obtain source materials for
Switzerland, since the Schweizerische Eidgenóssische
Vermessungsdirektion (SEVD) conducts only small scale
maps (1: 25 000 or smaller). The 1: 25 000 topographic
maps were applied for acquisition of ICNW. The large scale
maps (up to 1:10 000), needed for data densification in
metropolitan areas, are conducted respectively by the
individual federal states (Kantone). In order to keep source
materials for the planned metropolitan areas to be
homogeneous, digital black & white orthoimages (1 meter
ground resolution were ordered with the company
Swissair, which were rectified from latest photogrammetric
flights.
These digital orthoimages were then transformed from the
respective national coordinate systems into the MultiMap-
conform reference system and converted into the special
image format suitable for the on-screen digitization, before
they can be loaded into our digitization stations for the
vectorization process in our production.
The accuracy of orthoimages is usually strongly influenced
by the used basic data like the quality of DEM. During the
digitization cases were observed where the images didn't
fulfil our accuracy requirements, so that additional sources
(DGPS measurements, other map materials) were
necessary. This problem was put down to the errors
coming from the rectification processes.
In the Austria project, the digitization was made difficult
because of bad quality of several images. These images
can be divided into 2 groups: the one with bad photographic
quality (contrast, disruptive pixels) and the other negatively
influenced by unfavorable flight time (dense leaves,
unfavorable sun position). This problem was not
encountered in the Switzerland project.
With one meter resolution images, it was often very difficult
to recognize some small details such as road markings and
small streets. In future work resolutions of less than 0.5
meter should be applied.
Photogrammetric materials are currently solely applicable
large scale sources in Austria and Switzerland, owing to the
up-to-dateness and lack of other economical alternatives.
The same situation occurs partly also with the acquisition of
digital road data in the Scandinavian regions.
In general, the orthoimages proved to be effective in both
projects. In order to improve our future working process,
following points should be investigated:
- improve data exchange in digital form to avoid using
analog source materials.
- install an improved method to increase photographic
quality.
- use color orthoimages for the digitization process.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996