CIP A 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey
The distance between the centres of gravity of 14 lakes in the
Pfyffer's relief and in the current reference data is very diverse
and reaches the values from 129 m to 1460 m. However, when
the wrong position of the lakes is not considered (after the
translation to the reference centre of gravity), the common area
in respect to the reference data is only in two cases less than
60%. The total relative shape dissimilarity also shows rather a
good result: only in three cases this ratio is more than 60%.
Useful information about the representation of the lake form in
the historical data is given by the not common areas divided by
the reference perimeter (average shape difference). This
measure reaches the value of less than 200 m by the 80% of the
analysed data, what is significantly better than the relief
parameters listed in Table 2. This fact underlines the good local
accuracy of the relief features after their absolute position had
been corrected.
4.5 Comparison of the height models
The DTM of Pfyffer's relief transformed to the national
coordinate system is compared with the current DTM. The
procedure is performed as a comparison of the reference
regular raster and the irregular point cloud of the historical
data (nearest neighbour interpolation). According to
expectations the biggest height differences are located in the
southern alpine part of the relief (Figure 10). The unusually big
values in the histogram should not be understood as errors in
Pfyffer's height measurement: they can rather be explained by
the wrong position of several mountains. After removing the
worst 2% of the data set as outliers (white spots in the lower
part of the Figure 10), the average height difference in absolute
values reaches 177 m (corresponding to 1.6 cm in the relief).
This is a very respectable number for area-wide height
measurements at the end of 18 th century, because the maps at
that time usually do not contain any height information.
Figure 10. The difference DTM and a histogram
(current DTM minus Pfyffer's relief)
5. CONCLUSIONS
On the example of the relief model of Switzerland, constructed
by Franz Ludwig Pfyffer in 1762-1786, methods for 3D
reconstruction and accuracy analysis of historical reliefs were
shown. An interesting feature of the photogrammetric part of
the project is the unusual combination of aerial and close-range
applications: a big landscape mapping process is performed in
an artificial environment of the cellar room of a museum. Since
a very precise DTM is required, manual measurements of the
surface were applied. The complete digital data set was
archived for the documentation of the cultural heritage.
For the first time the relief of Franz Ludwig Pfyffer has been
quantitatively evaluated. As there is an absolute lack of written
documents concerning the relief construction, the applied
procedures represent the only possible way of the exploration
of this chapter of the Swiss history. The a-posteriori standard
deviation of spatial vector on identical points, reaching 3.2 cm
within the 26 m 2 big relief shows a surprising accuracy of the
relief constructed long time before the first Swiss national
triangulation network was established. In particular, the precise
representation of the third dimension is remarkable for that
time. The presented approaches extend the procedures used in
the history of cartography to the vertical coordinate direction.
In addition, some innovative methods for the analysis of
polygonal features of the old maps and relief models as well as
for the visualization are suggested.
REFERENCES
Baletti, C., 2000. Analytical and quantitative methods for the
analysis of the geometrical content of historical cartography.
International Archives of Photogrammetry and Remote
Sensing, Vol. XXXIII, Part B5, p. 30-37.
Beineke, D., 2001. Verfahren zur Genauigkeitsanalyse für
Altkarten. Dissertation, Heft 71. Universität der Bundeswehr
München, Fakultät für Bauingenieur- und Vermessungswesen,
Studiengang Geodäsie und Geoinformation, Neubiberg.
Fuse, T., Shimizu, E., Morichi, S., 1998. A study on geometric
correction of historical maps. International Archives of
Photogrammetry and Remote Sensing, Vol. XXXII, Part 5, p.
543-548.
Imhof, E., 1981. Bildhauer der Berge. Ein Bericht über alpine
Gebirgsmodelle in der Schweiz. Published by SAC, p. 107-110.
Leu, H. J., 1788: Supplement zu dem allgemeinen helvetisch
eidgenössischen oder schweizerischen Lexikon. In: Zeiger, F.,
1933. Luzern im Spiegel alter Reiseschilderungen, 1757-1835.
Eugen Haag Luzern, p.23.
Niederöst, J., 2002a: Landscape as a Historical Object: 3D-
Reconstruction and Evaluation of a Relief Model from the 18th
Century. International Archives of Photogrammetry, Remote
Sensing and Spatial Information Sciences, Vol. XXXIV, Part
5/W3.
Niederöst, J., 2002b: Das Relief der Urschweiz von Franz
Ludwig Pfyffer: Digitale Dokumentation und vermessungs
technische Aspekte. Cartographica Helvetica 26.
Niederöst, M., 2003: Detection and Reconstruction of
Buildings for Automated Map Updating. Dissertation. Institute
of Geodesy and Photogrammetry, ETH Zurich. IGP Mitteilung
Nr. 78.
Shimizu, E., Fuse, T., Shirai, K., 1999. Development of GIS
integrated historical map analysis system. International
Archives of Photogrammetry and Remote Sensing, Vol.
XXXII, Part 5-3W12, p. 79-84.
594
