tanbul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
The white gaps in figure 5 show the previous positions of these 6. CONCLUSIONS
| areas. The lower image shows the inverted case, correction in
| x-direction based on a dataset imaged in y-direction. In both The new approach for the generation of true orthoimages
| cases the roofs are mapped in their correct ground position. requires two sets of image data that have been taken with a
linescanner system in two flight lines perpendicular to each
| But, what are the pros and cons of the particular methods? other. Data acquisition must be carried out nadir-looking, i.e. in
| Transforming a whole area is an important advantage of the vertical planes, thus providing data in parallel projection along
segmentation method. Hence, new ground coordinates of each the flight lines and in perspective views across. If these ideal
| point in a partitioned area are calculated. Thus, the entire image conditions are fulfilled, true orthoimages can be generated
| could be transformed, if the segmentation should works for the without any knowledge of terrain or object heights.
whole image. In contrast, by image matching methods remain
| sporadic gaps of unconsidered points, if corresponding points In order to derive an orthoimage from such data two different
| were not found. However, truly located identical points occur in approaches can be applied, image matching and image segmen-
their correct ground position, therefore. By segmentation me- tation. Preliminary studies have been carried out, and it was
| thods, a wrong partitioned area, which is not flat and contains found, that both techniques offer some advantages, but both
| significant height variances, appropriate displacements will have also their limitations. Thus, more detailed investigations
| remain. Another problem is the segmentation of confused struc- will be necessary, and especially combinations of both methods
| tured areas such as treetops, which makes difficulties. The seg- should be tested.
| mentation of a roof, even a more complicated one, is feasible,
| but searching flat area defining points of a treetop in twe ima- If a sophisticated and automated segmentation and matching
| ges is absolutely impossible. In this case, the image matching procedure is carried out, the new approach can offer a reliable
| methods may provide a more suitable result. and very productive alternative option to generate true
orthoimages. Particularly, the independence from elevation
| The mentioned problems concern one special step in the models and the coverage of other applications are significant
procedure of the true orthoimage generation and don't affect the advantages. However, for terrestrial applications special data
| main principle of the new approach. The basic advantages are acquisition systems have to be developed.
| highly visible. The new approach doesn’t require any infor-
mation about heights of the imaged features or geometrical
| modelling of the objects. The digital orthoimages, which are REFERENCES
| generated with recent methods, are particularly influenced by
several parameters. One of these factors is the geometrical reso-
| lution of the elevation model. The model needs to represent the
object height variations, depending on the accuracy require-
| ments of the final orthoimage, in order to rectify each feature to
| - its proper position. Thus, the Orthoimage is as accurate as the
digital elevation model. In general, the recent orthoimage gene-
| ration methods directly depend on the support and also the
| quality from other data sources.
Albertz, J., 2001. Einführung in die Fernerkundung — Grund-
lagen der Interpretation von Luft- und Satellitenbildern.
Wissenschaftliche Buchgesellschaft, Darmstadt.
Braun, J., 2003. Aspects on True-Orthophoto Production.
Photogrammetric Week '03, Ed. Dieter Fritsch, Herbert Wich-
mann Verlag, Heidelberg.
Brauner, J. & Breuer, M., 1998. Vom Möbiusnetz zum
Resampling — Entzerrungsverfahren in der Photogrammetrie.
Unpublished article in the photogrammetric courses of the
Technical University of Berlin.
However, by means of the new approach, a true ortho image
| containing all features like buildings, bridges and trees in their
| correct grounded position, can be derived from the involved
image data directly, without using information from other data Hofmann, O., Nave, P., Ebner, H., 1982. DPS-A Digital
| sources. A further advantage of the new approach depends on Photogrammetric System for Producing Digital Elevation
the special geometry of the pushbroom scanner images. Models and Orthophotos by means of Linear Array Scanner
| Removing the displacements leaves blind spots with blank Imagery. The International Archives of Photogrammetry and
content. Due to the fact that displacements in pushbroom Remote Sensing, Vol. 24, Part 3, pp. 216-227.
| scanner images occur just toward one direction, the resulting
| gaps are significantly smaller than in the traditionally acquired Mayr, W., 2002. Bemerkungen zum Thema ,, True Ortho-
| images by means of a metric aerial camera. The gaps can be image“. Photogrammetrie — Fernerkundung — Geoinformation,
| filled to a large extent with information from the other appro- 4/2002, pp. 237-244.
priate image, because of the upright directed displacements.
The mosaicking task to fill the gaps is at least easier than in Scholten, F., Wewel, F., Neukum, G. & Albertz, J., 1998.
Digitale Luftbildaufnahme mit der HRSC — Ein Schritt in die
Zukunfi der Photogrammetrie. Photogrammetrie — Fernerkun-
Furthermore, the new approach provides multifunctional dung — Geoinformation, 6/1998, pp. 329-335.
usability. Thus, the method is not only limited to aerial acqui-
sition of terrain surface data, but it can also be used for cap-
| turing architectural facades etc. It just requires orthogonal data
acquisition relative to the defining reference plane of the ortho-
image in two directions perpendicular to each other.
traditional approaches.
Wewel, F., 1996. Determination of Conjugate Points of Stereo-
scopic Three Line Scanner Data of Mars96 Mission. The Inter-
national Archives of Photogrammetry and Remote Sensing,
Vol. 31, Part B3, pp. 936-939.
y HRSC-
image in
d in x- |
direction,
