CI P A 2003 x/x"' International Symposium, 30 September - 04 October, 2003, Antalya, Turkey
249
Abbaye d'Aulne
Orthofotoplan XI
Frank Benoit & Dennis Devriendt, 2001
Figure 4. Example of an Orthophotplan
3. PROBLEMS IN THE METHODOLOGY AND
POSSIBLE SOLUTIONS
The methodology described in the previous paragraphs suffers
some problems. These problems can be divided in two main
categories, according to the solution.
3.1 Problems that can be solved only with other
photographs
When parts of the wall on a photo are covered with vegetation,
this results in a false DEM and according orthophoto. Although
the surface beneath the vegetation can be interpolated, a more
accurate DEM will be possible is the vegetation is removed
before the photographs are taken.
Problems due to the central projection of the photograph and
the relief of the object, e.g. occlusions, can only be solved by
making photographs from other points of view.
As mentioned before, a number of stereopairs were taken with a
certain inclination of the optical axis. This leads to a different
resolution over the image. No quantitative analysis was done on
the effect of tilt on the quality of the orthoimages. Although we
can say that the absolute orientation is harder in the upper parts
of tilted images, and the matching algorithm gives poor results
in those areas, it is certainly possible to use tilted images up to a
certain limit of the angles. In figures 5 and 6 the effect of tilt on
the image geometry and the resolution becomes clear.
Figure 5. Left image of a tilted stereopair
Mr*
Figure 6. The orthophoto derived from a tilted stereopair
