ation "
charest
being
for the
00,000
ies of
4;
15,000,
'or the
) using
'anning
for the
name;
| film):
e four
ftware,
on 400
riewing
ith the
ersion
00 dpi
ding to
he off-
down-
nt the
ed the
eferred
of the
'encing
tained
Logic
ed the
tiscope
files on
rencing
. sheet
asovski
ar one,
because they are differences between the formulas
used by DTM and those of the routine implemented in
the library GEOLIB;
- manual digitizing of the vector elements (on the
screen) resulting vector files (spaghetti vectors);
- creation of the topology for the obtained vectors by
manual digitizing, resulting vector files topologically
structured.
- it was performed the Vectorisation algorithm testing.
D. Scanning the aerial photograms using the
DESKSCAN software, with the following scanning
parameters: resolution 800 dpi and TIFF format of the
output files (with 256 grey levels). The phase included the
following successive operations:
- establishing the number of photograms of the interest
area (photograms having 33% length coverage and
66% transversal coverage):
- photograms identification:
- aerotriangulation points marking by transposition:
- measuring the aerotriangulation points having as
result files with photogrammetric coordinates for the
triangulation points;
- aerotriangulation adjustment;
- scanning the photograms with the DESKSCAN
software, resulting raster files in TIFF format by grey
levels;
- correction the resulting files after the scanning. using
the CAL TIFF software.
E. Georeference of scanned aerial photograms and
satellite image, using Mu/tiscope software's tools: within
the Multiscope environment they were opened two
windows: one containing vectors and the other - the
image: georeferencing was made in the same way as the
files in manual digitizing. The successive operations were:
- TIFF files by grey levels import in the Muftiscope
environment, resulting raster files by grey levels in
the internal format of the software:
georeferencing the scanned images and the satellite
one, resulting metric files;
- satellite image classifying (unsupervised and
supervised), being obtained raster files on which can
be distinguished different elements categories.
F. Manual digitizing of the Bucharest city plan and the
topology of the obtained vectors building-up. using
TopoLogic software's tools. The operations in this phase
were as follows:
- digitizing the Bucharest city plan on the scale
1:15,000 using the a software designed and produced
in Romania (ArcSys), resulting vector files in the DXF
format (spaghetti vectors);
- import of vector date from the exchange format DXF
in the TopoLogic environment (resulting files in the
internal format of the software).
Also, geometric elements attributes input in the database
was performed from the TopoLogic environment.
G. Graphically editing the vector elements
corresponding to the map on the scale 1:100,000 with
those corresponding to the plan on the scale 1:15,000
(node additions, deletion, thickening. generalizing etc),
resulting updated vector files. The operation is performed
when the vectoring result is not correct.
H. Topology creation for the obtained data set, using the
Topol ogic tools and database creation.
I. Overlapping the vector data set obtained after the
topology creation on the georeferenced Images;
emphasizing the content elements geometry modifications
and their operation. It was checked out if the Multiscope
software accepts as input data the vector files obtained in
the Topol ogic environment.
J. Results checking In the field. It was performed a
randomly checking.
24 Conclusions regarding tests performed
1. It were obtained test areas and thematic plans updated
on the scale 1:15,000 for the down-town area and on the
scale 1:100,000 for the extra urban areas of the
Bucharest municipium.
2. The integrated system works on SUN Sparc machines.
3. It was remarked a good integration of the 3
components.
4. Even in the Multiscope and Topol ogic specifications a
connection with a digitizer is permitted, the two integrated
modules in the prototype of VIRGOS system do not permit
the connection with this device.
5. Time estimation. For a test area of one map sheet at
scale 1:100,000 it were obtained the following time
consuming:
vectorisation : 2 days
editing of vector elements: 10 days
raster data processing, using
Multtiscope module in order to
extract areal and linear features: 5 days
- photointerpretation and vectorial
elements updating using TopoLogic: 10 days
data input into the attributive data
base: 16 days
TOTAL 43 days
After 6 months of training on this system, a skilled
operator was able to update a map sheet at scale
1:100.000 using a Spot P image in 30 working days
(6 weeks x 5 days/week x 5 h/day).
If itis used a SPOT XS image instead a SPOT P image, it
might be performed and the multispectral analyze, such
as Vegetation Index (VI):
VI=XS3/XS2 2T (1)
where T represents a Limit Value, set-up by experiments.
3 VIRGOS VERSUS OTHER MAPPING AND
REMOTE SENSING SYSTEMS
The comparation between systems can be very easy to
performed due to very large amount of variables which
must be taken into account. It is performed a comparation
between VIRGOS and Arc/Info-Imagine.
From technical point of view:
a. Multiscope has the advantage that all the data are in
ASCII form and any user can handle this data without
any restriction from outside Multiscope environment
but, there are a lot of files and it is very hard to trace
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 733
