tanbul 2004
1 er =.
af.
Bei
v
1574
of
e analysis,
'al)
il
4. CONCLUSION
In this paper, the applicability of change detector combination
was investigated. Our assumption was that the change detection
accuracy in remotely sensed data can be increased by
combining different change detectors. Thereby, two fuzzy
change detectors based respectively on the comparative analysis
and the simultaneous analysis of multitemporal data were
combined by using the fuzzy integral. Both change detectors
used a fuzzy membership model computed by taking the
squared Mahalanobis distance from the prototypes of the
classes. Experiments using SPOT hrv data of the same area
demonstrate that the combined change detection system with the
fuzzy integral outperforms the individual change detectors. It
increases the detection rate while reducing the number of false
alarms. However, even though the usefulness of combining
change detectors was highlighted, it has been shown that in a
given land cover class, if one of the individual change detectors
gives a very poor accuracy and the second gives an important
accuracy, the precision of the combination system will be
smaller than that of the most precise system (This is the case of
the class 7). We think that this problem may be avoided when
combining more than two change detectors.
S. BIBLIOGRAPHY
Bárdossy, A., Samaniego, L., 2002. Fuzzy rule-based
classification of remotely sensed imagery, IEEE Transactions
on Geoscience and Remote Sensing, 40(2), pp.362-374.
Carlotto, M. J., 1997. Detection and analysis of change in
remotely sensed imagery with application to wide area
surveillance, /EEE Transactions on Image Processing, 6(1),
pp. 189-202.
Cho, S. B., Kim, J. H., 1995. Combining multiple neural
networks by fuzzy integrals for robust classification, /EEE
Transactions on Systems, Man Cybernetics, 25(2), pp.380-385.
Cho, S. B., 1995. Fuzzy aggregation of modular neural
networks with ordered weighted averaging operators,
International Journal of Approximate Reasoning, 13, pp.359-
375.
Deer, P., 1998. Digital change detection in remotely sensed
imagery using fuzzy set theory. PAD Thesis, Department of
Geography and Department of Computer Science, University of
Adelaide, Australia.
Lambin, E. F., Strahler, A. H., 1994. Change vector analysis in
multitemporal space : a tool to detect and categorize land cover
change processes using high temporal resolution satellite data.
Remote Sensing Environ, 48, pp. 231-244.
Liu, Q. Z., et al., 2001. Dynamic Image Sequence Analysis
Using Fuzzy Measures, /EEE transactions on systems, man, and
cybernetics-Part B: Cybernetics, 31(4), pp. 557- 572.
Rosin, P. L., 2001. Robust pixel unmixing, /EEE Transactions
on Geoscience and Remote Sensing, 39(9), pp. 1978-1983.
Verikas, A., Lipnikas, A., Malmqvist, K., Bacauskienne, M.,
Gelzinis, A., 1999. Combination of neural classifiers: A
comparative Study. Pattern Recognition Letters, Elsevier
Science, 2, pp. 429-443.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004