raphic
ontains
his
ed if use
ethods
pensive
y data.
lules:
DTM;
| map,
napping
sed, point
ry of the
| which is
considered
tos (when
fusion, and
'onsidered,
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
we can say that, to generate a rigorous and complete orthophoto
of a discontinuous surface the following are required:
e the correct description of the surface:
e the availability of images that represent all the details of
the object.
If several different prospective images are available using all
the projection rays for each point (as shown in fig. 3), some
precautions have to be taken, to avoid doubling the images. For
example, point Q, has to be obtained from image Io and point
Po from Ip.
orthophoto
Figure 3: True Ortoprojection from multiple images
3. DDTM GENERATION FROM A 3D MAP
It is already known [Dequal et al., 2001], that, to produce a
complete true orthophoto of a discontinuous surface, two
starting data sets are necessary:
e aset of digital images with known interior and exterior
orientation parameters (obtained by aerial triangulation, or
by direct methods with GPS/INS sensors), to guarantee a
photographic cover of the territory that is as complete as
possible;
e the correct description of the
alternatively, by:
- a traditional DTM, completed with breaklines;
- a digital surface model (DSM), that contains superficial
geometric entities, managed by complex database logics.
- a dense DTM (DDTM).
surface, obtained,
The latter is usually the cheapest and simplest solution. It does
not require expensive stereoplotting and editing operations or
complex data base management software: it can easily be
acquired with modern instruments (laser scanner), or in a
cheaper way, with interpolation. from a 3D digital map,
whenever it exists. This kind of cartography moreover is
becoming very common in local municipality administrations,
as a base map for urban development instruments, and as a
geometric base for Geographical Information Systems. A 3D
digital map describes the territory and buildings in a
tridimensional space and therefore contains all the information
needed to generate the corresponding DDTM:
* a surface without buildings is described with height
points and contour lines;
* each area is represented with points known in the 3D
coordinates (streets, bridges, and railways are
described with arcs and nodes, whose points are
known in X, Y and Z coordinates)
* buildings are described as volumetric entities, of
uniform height, with a central height point. In this
way, the volumetric entities can be compared to
parallelepiped shapes (with the building perimeter as
the base), extruded from the ground in a vertical
direction, up to their height.
An example of 3D cartographic representation is shown in
figure 4.
539
Figure 4: An example of 3D digital cartography [Spalla, 2002]
3.1 GENEDDTM software
The GeneDDTM software, developed in Visual Fortran
language by the authors, uses information from the digital map
to generate a dense DTM of the corresponding zone. The
necessary input data are contained in a DXF format file of a 3D
digital map, with the description of the geometry (with all the
areas described as 3D closed polylines) and a code directory
concerning the height points, countour lines, the different kinds
of buildings (monumental, public, residential, industrial, and so
on), building height points, and the different kinds of areas that
have to be processed (street arcs and nodes, green areas,
playgrounds, cemeteries, and so on). Some dialogue windows
are shown in fig. 5 (the main one, data entry, and processing
windows).
rut
Discos d fee
s OUTPUT.
[ Dehecions poete
€ Ó PALA ARE ARR TE
[^ í
Gerésteire DDTN
Figure 5: Some dialogue windows of the GeneDDTM software
GeneDDTM works in four separate phases, which are displayed
in the dialogue window with a progression bar (as shown in fig.
Sy
I. Filling of the areas inherent to the buildings with their
heights, extraction of the height points from the DXF file
(group n°l), sequential access to all the polylines with a
building code, search inside group n. 1 of each
corresponding height point, giving the same height to each
node inside the single area.
2. Filling of the other areas: extraction of the height points
from the DXF file and the points that describe the contour
lines (group n. 2), sequential reading of the polylines with
a code that corresponds to the aerial entities, search for the
points in group n. 2 inside the area (group n. 3);
interpolation of the height of the DDTM inside the
considered areas based on points from group n. 3 and on
the points of the polyline.
