assification
Maximum
) for stand
for LVQ in
S with the
ented, the
le training
tability of
rthermore,
sults with
their very
iven in ta-
Yosaicking
jtting line
us
IS
yrtant to
to verify
cess was
e practi-
ossibility
he prob-
but also
periment
because
collecting the necessary new experience is very time consum-
ing. Furthermore, we doubt that the results will be signifi-
cantly better simply because some hidden problems with the
new classifier will be overlooked easily.
As to the mosaicking, in general we would prefer to have
a separate training set for each separate satellite image. In
our case, the applied approach was the only possibility of
composing a classification mosaic due to unsufficient ground
truth in the Eastern and Western scene. The results of the
overlaping area proved the approach to be successful.
REFERENCES
[Bähr, 1985] Bahr, H. P., 1985. Digitale Bildverar-
beitung - Anwendung in Photogrammetrie
und Fernerkundung. Karlsruhe.
[Coenradi, 1992] Coenradi, B., 1992. Waldklassifizierung
und Totholzkartierung im | Nationalpark
Bayerischer Wald unter Verwendung von
Landsat-TM- und digitalen Zusatzdaten.
Diplomarbeit Universität Bonn, p. 199.
[Colby, 1991] Colby, J. D., 1991. Topographic Normaliza-
tion in Rugged Terrain. Photogrammetric
Engineering & Remote Sensing, vol. 57, no.
5, pp. 531-537.
[Hertz, 1991] Hertz, J., Krogh, A., Palmer, R. G., 1991.
Introduction to the Theory of Neural Com-
putation. Addison-Wesley.
[Horler & Ahern, 1986] Horler, D.N.H., Ahern, F. J., 1986.
Forestry Information Content of Thematic
Mapper Data. Int. J. Remote Sensing,
vol. 57, no. 3, pp. 505-428.
[Keil, 1990] Keil, M., Schardt, M., Schurek, A., Win-
ter, R., 1990. Forest Mapping Using Satel-
lite Imagery. The Regensburg Map Sheet
1:200.000 as Example. ISPRS Journal
of Photogrammetry and Remote Sensing,
no. 45 (1990) pp. 33-46, Amsterdam.
[Kohonen, 1989] Kohonen, T., 1989. Self-Organization
and Associative Memory. Springer-Verlag,
Berlin-Heidelberg-New York-Tokio, 3rd edi-
tion.
[Kohonen, 1995] Kohonen, T., 1995. Self-Organizing Maps.
Springer-Verlag, Berlin-Heidelberg-New
York- Tokio.
[Niblack, 1985] Niblack, W., 1985. An Introduction to Dig-
ital Image Processing. Strandberg-Verlag,
Birkergd, Danemark.
[Schardt, 1989] Schardt, M., 1989. Forest Classification
with TM Data in the Area of Freiburg, Fed-
eral Republic of Germany. Proceedings of
a Workshop on "Earthnet Pilot Project on
Landsat Thematic Mapper Applications",
S. 251-259, Frascati.
[Schardt, 1990] Schardt, M., 1990. Verwendbarkeit
von Thematic Mapper Daten zur Klassi-
fizierung von Baumarten und natürlichen
Altersklassen. DLR-Forschungsprojekt,
(DLR.FB 90-44), Oberpfaffenhofen.
[Schardt, 1993a] Schardt, M., Kenneweg, H., Sagischewski,
H., 1993. Forest Stand Characteristics to be
Applied to Monitoring Lingering Disasters
by Satellite remote sensing and GIS. Pro-
ceedings of the Symposium "Inventory and
Management Techniques in the Context of
Catastrophic Events", State College, Penn-
sylvania, June 21-24.
[Schardt, 1993b] Schardt, M., Martin, K., Keil, M., 1993.
Mapping of Biotic Damages Using Satellite
Data. Proceedings of the Symposium “In-
ventory and Management Techniques in the
Context of Catastrophic Events”, State Col-
lege, Pennsylvania, June 21-24.
[Song, 1996] Song, Hee-Heon, Lee, Seong-Whan, 1996.
LVQ Combined with Simulated Annealing
for Optimal Design of Large-set Reference
Models Neural Networks, Vol. 9, No. 2, pp
329-336, Elsevier Science Ltd.
[Weiss, 1991] ^ Weiss, Sholom, M., Kulikowski, Casimir, A,
1991. Computer Systems That Learn Mor-
gan Kaufmann Publisher Inc., San Fran-
Cisco.
605
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
