ccurrence
ential for
vith high
listic and
nal Baye-
1e overall
ikelihood
g training
| decision
propriate
H——+—+
0 l 2 km
Legend
Industrial Area
Docks
Residential Area
Other Areas /
Unclassified
rd
à
D d
Figure 9: Result of maximum-likelihood classification
Land-Use Class Reference Classified Number Producers Users
Totals Totals Correct Accuracy Accuracy
Residential Area 120 72 66 533.0.% 91.7%
Docks 24 38 16 66.7% 42.1%
Industrial Area 12 39 12 100.0% 30.7%
Other Areas 100 107 97 97.0% 90.7%
Overall Classification Accuracy = 74.6%
Table 10: Accuracy report for maximum-likelihood classification
Land-Use Class Reference Classified Number Producers Users
Totals Totals Correct Accuracy Accuracy
Residential Area 90 72 70 77.8 % 97.2%
Docks 34 38 30 88.2% 79.0%
Industrial Area 26 39 24 92.3% 61.5%
Other Areas 106 107 103 97.2% 96.3%
Overall Classification Accuracy = 88.7%
Table 11: Accuracy report for the neural network classification using region-based co-occurrence matrices
5. REFERENCES
Benediktsson, J.A., Swain, P.H., and Ersoy, O.K., 1990. Neural
Network Approaches Versus Statistical Methods in Classifica-
tion of Multisource Remote Sensing Data. IEEE Transactions
on Geoscience and Remote Sensing, 28(4), pp. 540-552.
Bischof, H., Schneider, W., and Pinz, A.J., 1992. Multispectral
Classification of Landsat-Images Using Neural Networks. IEEE
Transactions on Geoscience and Remote Sensing, 30(3), pp.
482-490.
75
Bock, S., 1995. Ein Ansatz zur polygonbasierten Klassifikation
von Luft- und Satellitenbildern mittels künstlicher neuronaler
Netze. Kieler Geographische Schriften 91, Kiel.
Chen, K.S. et al., 1993. Classification of Multispectral Imagery
Using Dynamic Learning Neural Network. In: Proceedings of
the International Geoscience and Remote Sensing Symposium
(IGARSS'93), Tokyo, Japan, Vol. 2, pp. 896-898.
Chester, M., 1993. Neural Networks: A Tutorial. Prentice Hall,
Englewood Cliffs, NJ-USA.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
