nbul 2004
TM, e.g.
vell as for
ive a laser
ing the z-
nbine the
is: being
-files our
1Ses some
| converts
ints in an
rages are
ate (with
Since the
ructuring:
iodes due
oroximity
lement of
proximity
the point
/e limited
d. If no
z-value 1s
blem and
s with the
os, which
the linear
enerating
here will
cted on a
ion of the
ON
tabase by
1ked with
s for the
he choice
| how the
nmetrical
detailed
It should
of them
they are
their list
IK
ser group
naking a
z-values
zes when
it locally
ht of the
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
floating point (cursor), e.g. for verifying external 2D data files
D10) automatically perform user defined consistency checks
D11) clean overshoots, undershoots and intersections, making a
vertically projected 2D topology without messing up the z-values
nor the db-linkages: dynamically — online, refreshing the cleaned
elements in the view without refreshing the images.
It should also be possible to easily create the following update
functions:
N12) replace (or delete) the geometry of an object completely,
without loosing its attributes and automatically adapt ‘length’ or
‘area’ and ‘perimeter’
N13) update according to the following process:
- for deleting an object: copy the record from the Active table to
the Old table and fill in the end date and increment the record
number; then delete the corresponding record from the Active
table
- for adding a new object: add a record to the Active table and fill
in attribute values and geometry etc.
- for modifying an existing object: copy the record from the
Active table to the Old table and fill in the delete date and
increment the record number; then modify the geometry or the
attribute values of the corresponding record in the Active table
and adapt the load date, the source(s) and the source date(s).
D14) change a (group of) selected graphical element(s) from one .
object class to another by one single mouse click; e.g. having
selected a group of paths, click on ‘become a dirt road’
D15) adapt the geometry of an object by move vertex" or
*'move the whole graphical element! and set the attribute
"big movement to 1 if the vertex or the whole graphical element
was moved by more than 5 meter.
We do not have the time nor the means to evaluate all possible
combinations of digital stereoplotters with databases. (In 2001 P.
Plugers already described 34 different models of digital
photogrammetrical workstations.) At present, our data are being
collected by digital stereoplotters SSK (ZI) using the CAD-
software MicroStation SE or J (Bentley). After transforming the
3D MicroStation designfiles into 2D Arclnfo coverages, the
topological cleaning and further identification are being
performed with ArcInfo 7 (ESRI). Regarding the licences and
maintenance contracts that we have, we started by examining the
solutions proposed by Bentley, Intergraph and ESRI.
1) MicroStation GeoGraphics (Bentley)
If it were possible to go on using the available Intergraph SSK
stereo software together with the used MicroStation CAD
software and realize the above points 1 till 14 by connecting
Microstation J through GeoGraphics 7 to a DBMS, we would not
have to purchase any supplementary licences (apart from the
DBMS), because the GeoGraphics module is included in the
maintenance contract. GeoGraphics 7 can be linked to Oracle
8iSpatial (It does not support Informix). For working with 3D-
data however, we need Oracle 9iSpatial or higher, which is only
supported by Microstation 8 + GeoGraphics 8, having a format
that 1s different from our design files. It is therefore impossible for
us to maintain the present data format for stereoplotting.
2) ISSD (ZI) + MicroStation J (Bentley)+ Dynamo (Intergraph )
+ Oracle 9iSpatial
3) Geomedia Stereo (Intergraph ) + Geomedia (Intergraph )
4) Socet Set (BAE) + ArcGis 8.3 (ESRI)
5) LPS (Leica) + ArcGis 9 (ESRI)
The comparison of the different configurations is still in
progress.
6. HOW MANY AND WHAT KIND OF
MODIFICATIONS MAY WE EXPECT IN
THE LANDSCAPE ?
We should distinguish both the frequency of changes and the
importance that users attach to these:
We would like to distinguish 2 groups of objects: those of which
the updating can be considered as important on the one hand and
the less important group on the other hand. The ‘important’ group
(indicated in bold in table 1) contains as well objects of (at least)
low importance that appear in very great number in the landscape
as objects of very high importance that rarely change. For this
group we envisage shorter updating cycles than for the less
important group. The latter contains much information that can
only be collected in the field; hence it is more labour-intensive.
One could ask oneself whether it makes sense to go on
distinguishing pastures from arable land, seeing that changes
between these soil- ids are so frequent in Belgium that these
identifications cannot be very useful. It might be better to abolish
this distinction in the future data catalogue.
Some numbers from our National Institute for Statistics: In
average there are 31 500 new buildings a year, which means that
during the first updating about 350 000 buildings will have to be
added, along with probably as many modifications to existing
buildings. There are also about 825 km of new public roads a
year, so we may expect more than 9000 km of new public roads
and an unknown number of private roads.
Importance | Very low Low
Changes
High Very High
Very frequent Arable vs. pasture | Gardens
Individual trees
Small watersurfaces
Hedges & Tree-Rows
Paths vs. dirt roads | Houses
Average frequency Forest type
Road width
Orchards
Roads, Woods
Factories, Supermarkets
Rare Culverts Slopes
Sources Water course width
Point symbols High tension, Water courses
Railroads, Churches, Town halls
Hospitals, Schools
Table 1
