> is to be viewed with
s. Glasses with the
e requisite for stereo
1oney.
'IELD CAMPAIGN
fter the end of the field
ne single fields were
saicking is the process
zer image. The result is
scale of 1 : 5. This map
:ause of the larger scale
is not essential as the
ovided with the single
he mosaicking. Firstly,
trast and colors of the
ipproximately selecting
g step by step with the
> will be geometric
| caused by the fact that
otos. This cannot easily
n cuts or unsharpnesses
the procedure used for
ethods to perform this
ine the single rectified
ing software. In the
rated which define the
ed for the final mosaic.
tools for this purpose
—
400.058
400.074
nda
that are often used in photogrammetry or remote sensing
packages. These packages offer different methods for adjusting
radiometry of the images, e.g. hard cutlines or soft transition
within the overlapping areas. For this project, ERDAS
Imagine® was used to generate the mosaic. Soft transitions
were used which led to a certain unsharpness in these areas. The
position information which is required to define the geometric
position of the image in the common reference system is known
from the rectifying process. If unknown yet, the images have to
be referenced now.
Photogrammetry. As described above, the locations were
completely imaged with stereo models using a photogrammetric
camera. Before plotting, exact coordinates of the reference
points have to be determined. Image coordinates of all marked
points and several tie points were measured using an analytical
plotter. Together with the level measurements and the tape
measurements, a bundle adjustment was calculated to compute
the coordinates. Those could be defined with an accuracy of 3
mm which is a result better than needed in this project. Smaller
image scales would probably have been sufficient for the
accuracy of the bundle adjustment, but the interpretation of the
images was easier and the local conditions during image capture
in combination with the existing lenses for the camera led to
take the images in the way described.
After these preparations, the stereo plotting could be done. A
Zeiss P3 analytical plotter was used in combination with
MicroStation® as connected CAD system; the later processing
of the maps was performed using AutoCAD®. The outlines of
every single bone were plotted with 3D polylines. For every
400.066
100.085
400041
400.074
400.054
/
Figure 4. Part of the vector
400.011 |
400.062
Se
1 : Se ae
map. Scale approx. 1 : 5
piece, the height of selected points was measured by plotting a
small circle to this position. In AutoCAD®, a Lisp program was
used to draw small dots at these positions and to place texts
designating the height of the plotted circles there, allowing to
specify the text size in the plot. This procedure resulted in
speeding up the time for placing these texts significantly. In the
larger location more than 1200 height spots were assigned (cf.
fig. 4).
The final maps of the two locations were plotted at a scale of
1 : 5 with height indices accurate to few millimeters, which will
allow the repositioning of the single findings with a sufficient
accuracy in all three dimensions.
RESULTS
The results of the documentation process are different types of
products. The locations are completely covered with analogue
metric stereo models which were the basis for the plotting of the
metric plans. These images can be archived for rather long
periods of time, whereas in the archiving of digital image data
there are still uncertainties concerning the durability of media
and the future availability of suitable hard- and software. In
combination with the sketches of the situation, the
measurements and the coordinates of the marked points, in the
worst case the whole geometry of the locations can be
reconstructed using these analogue sources.
The derived products are CAD maps, rectified image maps and
anaglyph images, both in digital and analogue form. The
durability of the analogue prints is limited in comparison to that
of the analogue images mainly because of color fading of the
—383—
