3. the snow cover and the partly turbid atmosphere reduced the area
available for point definitions.
Therefore the orientation was based only on fifteen points, covering
that part of the model which should be examinated. For the calculations
the free net adjustment option of MOR was chosen resulting in residuals
of mar... = 5.0 um in the picture and mqryy = 123 m , mary = 41 m in the
model. However, the point distribution was not as optimal, so that an a
posteriori examination of systematic errors seems to be recommendable.
This was done comparing the DTM-data at single well distributed points
with the results of a height determination by correlation, with the
advantage that no more single image points had to be appointed to cor-
responding ground points. This reduced problems with point selection
and identification. The results of this examination are plotted in
fig.!. Obviously the residuals are of systematic character, what attests
the existence of some smaller errors in the orientation data. Therefore
later on in the accuracy control, the systematic characteristic of the
Z-residuals has to be taken into account.
4, DETERMINATION OF A DIGITAL TERRAIN MODEL
4.1 Description of the test sites
Four test sites were chosen with different functions in the investiga-
tions.
= Landshut This site covers a major part of the model. With
208*150 (=31200) equally spaced (500 m) points about
7710 km” were registrated. Obviosly the purpose of
this test site is to applicate the correlation onto
the greater part of the model space, to prove its
functionality and reliability in determinating a
great number of points.
= Deggendorf This site shows very different surface characte-
ristics. Therefore this area served to differentiate
between categories like contrast, surface topography
or object type and to investigate their influence
onto the quality of the correlation results. The
determined DTM consists of 900 (30*30) points with
180 m distance.
- Vilsbiburg Here we have varying object contents (fields, small
towns, a small sea, a river, some forests) and lower
heights differences (about 500 m). The area is
subdivided into two parts. The first one (Vilsbi-
burg1) is affected by haze in the atmosphere, while
the second (Vilsbiburg2) shows only minor atmosphe-
ric influences. The examination of these test sites
should show the influence of contrast reduction in
regions of less complicate topography. Additionally,
the dimensions of the two DTM (site 1 : 25*25 pts,
AX=AY=246m ; site 2 : 25%25 pts, AX=AY=490m) could
give an impression of the influence of varying point
density.
- BL =