International Archives of Photogrammetry and Remote Sensing, Vol. 32, Part 7-4-3 W6, Valladolid, Spain, 3-4 June, 1999 
Fig. 5. Diagram of a data and information mining tool. 
Fig. 6. Multiple models data and information mining tool. 
The Bayesian approach enables mining the stochastic models and 
finding the ones, which best explain the datasets. 
7. SUMMARY 
Data and information fusion and mining have as common tasks 
the information extraction and representation. The differentiation 
of the two fields is in the way how information is treated. 
Information fusion has as goal the aggregation of 
incommensurable pieces of information trying to enhance the 
quality of data interpretation. Data and information mining has as 
goal the exploration of the unexpected relationships among the 
elementary items of information extracted from the observations. 
We presented a Bayesian perspective of the two fields - 
information fusion and information mining - and proposed 
several new approaches. Part of the methods or similar 
techniques are integrated in a demonstrator system for querying 
large image archives by image content (Datcu et al., 1999). An 
interactive version is available on http://www.vision.ee.ethz.ch/ 
~rsia/. 
Bouman, C., B., Liu, 1991. Multiple resolution segmentation of 
textured images. IEEE Trans. PAMI, 13(2):99-113. 
Cowles, M.K., Carlin, B.P., 1996. Markov Chain Monte-Carlo 
convergence diagnostics: a comparative study. J. Amer. Statist. 
Soc., 91(434)883-904. 
Datcu, M., K., Seidel, M., Walessa, 1998. Spatial Information 
Retrieval From Remote Sensing Images - Part A: Information 
Theoretical Perspective. IEEE Trans, on Geoscience and Remote 
Sensing, 36(5):1431-1445. 
Datcu, M., K., Seidel, G., Schwarz, 1999. Elaboration of 
advanced tools for information retrieval and the design of a new 
generation of remote sensing ground segment systems. In: 
Machine Vision in Remote Sensing, I. Kanellopoulos (Ed.), 
Springer Verlag, in print. 
Geman, S., Geman, D., 1984. Stochastic relaxation, Gibbs 
distributions, and the Bayesian restoration of images. IEEE 
Trans, on PAMI, 6:721-741. 
Jumarie, G., 1991. Relative information. Springer Verlag. 
Lauritzen, S. L., Spiegelhalter, D.J., 1988. Local computations 
with probabilities on graphical structures and their applications 
to expert systems (with discussion). J. Rag. Statist. Soc. B, Vol. 
50, pp. 157-224. 
Li, S. Z., 1995. Markov Random Field Modeling in Computer 
Vision. Computer Science Workbench, Springer Verlag. 
Luettgen, M.R., Karl, W.C., Willsky, A.S. and R.R. Tenney, 
1993. Multiscale representations of Markov random fields. IEEE 
Transactions on Signal Processing, 41(12):3377-3396. 
MacKay, D. J. C., 1991. Maximum Entropy and Bayesian 
Methods in Inverse Problems. In: Bayesian Interpolation, pp. 39- 
66, Kluwer Academic Publisher. 
Minka, T.P., Picard, R.W., 1997. Interactive learning with a 
society of models. Pattern Recognition, Vol. 30, pp.565-581. 
Nicco, G., Walessa, M„ Datcu, M„ 1998. Topographic Mapping: 
A New Approach. Proc. of 1998 International Geoscience and 
Remote Sensing Symposium (IGARSS'98). 
Popat, K., Picard, R.W., 1997. Cluster based probability model 
and its applications to image and texture processing. IEEE Trans, 
on Image Processing, Vol. 6, pp. 268-284. 
Schroder, M., H., Rehrauer, K., Seidel, M., Datcu, 1998. Spatial 
Information Retrieval From Remote Sensing Images - Part B: 
Gibbs Markov Random Fields. IEEE Trans, on Geoscience and 
Remote Sensing, 36(5): 1446-1455. 
ACKNOWLEDGEMENTS 
The authors would like to acknowledge the contributions of 
Hubert Rehrauer, Michael Schröder, Gintautas Palubinskas, 
Marc Walessa, Sorel Stan and Andrea Pelizzari to the ETH/DLR 
Remote Sensing Image Archive (RSIA) demonstrator. 
REFERENCES 
Baldwin, D. G., Emery, W. J., Cheesman, P.B., 1998. Higher 
resolution earth surface features from repeat moderate resolution 
satellite imagery. IEEE Trans. Geoscience and Remote Sensing, 
36(l):244-255.