International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B6. Istanbul 2004
Located within this larger area is the Garden Kingdom of Des-
sau-Worlitz, a UNESCO Cultural Heritage site. The Garden
Kingdom is a product of the Age of Enlightenment in the 18th
century and portrays the connection among nature, regional
planning, and architecture. The socially-minded Prince Leopold
Friedrich Franz of Anhalt-Dessau developed an artistically-
enhanced landscape between 1760 and 1817, with the help of
his advisor and friend Friedrich Wilhelm von Erdmannsdorff.
In addition to its aesthetic appeal, the entire Garden Kingdom
should be understood as an educational program. The parks and
individual buildings were connected with each other through
levees, avenues, oak trees, meadows and fruit orchards. Visual
corridors and plantings along the avenues increasingly
developed a network, which connected Wórlitz and the other
parks. These include, in chronological order, the Luisium,
Georgium, the Sieglitz Park and the Kühnau Park. Pre-existing
parks, such as the Tiergarten, the Mosigkau Palace, and the
Oranienbaum Park were worked into the landscape plan. Efforts
are now underway to raise dams; however, dams located in
historical visual corridors will either not at all or only to a lesser
extent be raised. The data to be gathered with the project could
be used to simulate e.g. the effect of that decision.
If critical gauges at both rivers are to be included, the test area
would have a size of 42x20 km?; it is bordered by the following
four geographical border lines:
west: 12°03'10, east: 12°40'02" eastern longitude,
north: 51?54'06", south: 51°42'18" northern latitude.
Smaller meaningful areas would be that of the Garden Kingdom
rounded to a rectangular area (25x15 km?) or the extended
confluence area (17x10 km?).
Direct sensor orientation based on GPS/INS (inertial navigation
systems) provides a high flexibility in mission planning and
appears to obviate the need to fly regular block structures.
However, the determination of exterior orientation parameters
can reach accuracies similar to those of standard photogram-
metric procedures only if an overall system calibration (includ-
ing GPS/INS components and camera self-calibration) and a
correct GPS/INS data processing (including efficient GPS/INS
error control and datum problems) is guaranteed for the specific
mission site. It is for this reason that a small calibration site
consisting of targeted control points will be arranged within the
project. The flight planning is based on a regular block
structure, however, the use of the calibration site would also
enable investigations in regard to accuracy of exterior ori-
entation parameters using GPS/INS.
S. PROJECT PARAMETERS
A DTM being part of the national survey activities is available;
it consists of a grid of points spaced at 10 m intervals and is
quoted to have an elevation accuracy of oz — t 0.5 m. This
DTM does not meet the requirements of hydrographical engi-
neers in regard to either point density or elevation accuracy.
One important objective is therefore the generation of a three
dimensional digital terrain model of the project area with as
high an accuracy as meaningfully achievable. If laser scanning
is assumed as method of providing a standard, an elevation
accuracy of oz 7 € 0.15 m is an acceptable assumption.
It is assumed that photogrammetric evaluation of aerial photo-
graphs on an analytical plotter using the (static) raster meas-
212
urement method yields elevation accuracies of o; = + 0.15%0oxc
xmp = + 0.15 m; this translates for a wide-angle film camera
with a focal length c = 153 mm to mp = 6536. The correspond-
ing swath width is 1503.28 m. Assuming accuracies for the
horizontal coordinates of well defined natural points of oy
[um] = + 10xmg, one obtains oxy = + 6.5 cm. First published
results indicate that improved accuracies can be obtained when
using digital aerial cameras.
Digital and film cameras are proposed to be used to obtain
comparable image data. After consultation with a private firm
interested in purchasing a digital mapping camera it was de-
cided to plan the project for identical swath width for all sen-
sors. This has the advantage that pass and control points can be
located into the areas of overlap between adjacent swaths.One
could also have chosen identical ground pixel size, however,
this would have increased sensor-specific concerns, especially
in regard to the line scanners. In addition, an investigation with
aerial test-area images scanned at different pixel sizes showed
that photogrammetric accuracies are only marginally related to
the scanning pixel size in the range from 14x14 to 28x28 um?.
Of the considered sensors, the ADS40 is most limited in regard
to obtaining large-scale images, it is proposed to plan the sets of
image data assuming for all the same swath width of 2400 m
obtainable with the ADS40 at a scale of 1 in 30770. This swath
width can be obtained with the other systems as follows:
HRSC-AX at 1 in 30770, DMC at | in 14285, UltraCamD at |
in 23188, DSS at 1 in 65168 and aerial film camera at 1 in
10435. The corresponding flying heights would be 1923 m,
4616 m, 1714 m, 2319 m, 3584 (or 2281) m and 1597 m. With
an aerial camera with ¢ = 153 mm an elevation accuracy of o7
— t 0,1596oxcxmg = + 24 em can be expected. In order to also
gain insight into a possible change in elevation accuracy as a
result of automatic measurement, it is proposed that at least the
aerial photographs should also be measured by an operator
using a static or dynamic grid sampling procedure. First results
for the DMC indicate that elevation accuracies in the order of
oz = + 0,08%oxexmp can be obtained; this translates for the
indicated project parameters to 07 = + 13,7 cm, an accuracy
similar to that of LIDAR.
The pixel sizes on the ground would be as follows for the dif-
ferent systems: ADS40 10x10 cm? with staggered line sensors
and 20x20 cn? without, HRSC-AX 20x20 cn, DMC 17x17
cm?, UltraCamD 21x21 cm? DSS 59x59 cm? and aerial pho-
tographs scanned at 14x14 pm? 15x15 cm?. All these values
are, with the exception of that for the DSS, smaller than the
pixel size of 40x40 cm? used in the creation of orthophotos for
the state mapping authority of Saxony-Anhalt which are the
base for the maintenance of the official topographic GIS called
ATKIS and the derivation of the digital topographic base map
at the scale 1 in 10000.
Planning for a swath width of 2400 m with 30% sidelap and,
where applicable, 60% forwardlap in case of the traditional
aerial photography would result for the large project area in 12
flight lines of 42 km length or of the following number of
images for the different farm cameras: DMC 80, UltraCamD
67, DSS 44 and film camera 44. In addition, there would be
several cross fights.
