Stefan Growe
6 CONCLUSIONS
In this contribution the use of the knowledge based image interpretation system AIDA for the analysis of multitemporal
remote sensing images was presented. General knowledge about scene objects, their structure, and their appearance in
the sensor data is stored in a semantic net. Additional temporal information about object states, their duration, their proba-
bility of occurrence, and knowledge about possible state transitions is represented by a state transition graph integrated
within the semantic net. The system exploits the temporal knowledge to predict possible successor states of scene objects
for the current image based on the object's state in the preceding image of the multitemporal sequence. Thus, the search
space is reduced which accelerates the interpretation process.
A probabilistic judgement calculus was suggested in order to compare competing scene descriptions and to select the most
promising alternative. The semantic net of instances is transformed to a Bayesian network and the measurements in the
sensor data are introduced and propagated through the network. An overall judgement is derived from the belief values
of the topmost Bayes nodes. An advantage of the approach is the capability to consider the prior probabilities given by
the temporal knowledge. In cases, where the evidence is identical for several alternatives, the most probable solution is
judged best and therefore preferred in the ongoing analysis.
The system was tested successfully for the detection of an industrial fairground in a set of four multitemporal aerial imag-
es. The fairground can be recognized, because the cycle of inactivity, construction, activity, and dismantling was observed
consecutively in the set of images. For the given example the probabilistic judgement approach was compared to an exist-
ing possibilistic method. It was shown, that the exploitation of prior probabilities increases the efficiency of the interpreta-
tion process. The number of necessary inference steps is reduced by up to 35%.
REFERENCES
Clement, V., Giraudon, G., Houzelle, S. , Sadakly, F., 1993. Interpretation of Remotely Sensed Images in a Context of
Multisensor Fusion Using a Multispecialist Architecture, IEEE Trans. on Geoscience and Remote Sensing, Vol. 31, No
4, pp. 779-791.
Dechter, R., Meiri, I., Pearl, J., 1991. Temporal constraint networks, Artificial Intelligence, Vol. 49, pp. 61-95.
Dubois, D., Prade, H., 1988. Possibility Theory: An Approach to Computerized Processing of Uncertainty, Plenum Press,
New York and London, 263 p.
Growe, S., 1999. Knowledge Based Interpretation of Multisensor and Multitemporal Remote Sensing Images, IAPRS,
Vol. 32, Part 7-4-3W6, Valladolid, Spain, pp. 130-138.
Liedtke, C.—E., Bückner, J., Grau, O., Growe, S., Tónjes, R., 1997. AIDA: A System for the Knowledge Based Interpreta-
tion of Remote Sensing Data, 3rd Int. Airborne Remote Sensing Conference and Exhibition, Copenhagen, Denmark, pp.
313-320.
Matsuyama, T., Hwang, V.S.-S., 1990. SIGMA: A Knowledge-Based Aerial Image Understanding System, Plenum
Press, New York, 277 p.
McKeown, D., Wilson, A., McDermott, J., 1985. Rule-Based Interpretation of Aerial Imagery, IEEE Trans. on Pattern
Analysis and Machine Intelligence, Vol. PAMI-7, No. 5, pp. 570—585.
Mees, W., Perneel, C., 1998. Advances in computer assisted image interpretation, Informatica — International Journal of
Computing and Informatics, Vol. 22(2), pp. 231—243.
Pearl, J., 1988. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, Morgan Kaufmann Pub-
lishers, San Francisco, 552 p.
Quint, E, 1997. MOSES: A structural approach to aerial image understanding, In: A. Gruen, E. Baltsavias, and O.P. Hen-
ricsson (ed.): Automatic Extraction of Man-Made Objects from Aerial and Space Images (II), Birkhiuser, Basel, pp.
323-332.
Sagerer, G., Niemann, H., 1997: Semantic Networks for Understanding Scences, Advances in Computer Vision and Ma-
chine Intelligence, Plenum Press, New York and London.
Stilla, U., Michaelsen, E., 1997. Semantic modelling of man-made objects by production nets, in A. Gruen, E. Baltsavias,
and O.P. Henricsson (ed.): Automatic Extraction of Man-Made Objects from Aerial and Space Images (II), Birkhiuser,
Basel, pp. 43-52.
Tonjes, R., Growe, S., Biickner, J., Liedtke, C.—E., 1999. Knowledge Based Interpretation of Remote Sensing Images
Using Semantic Nets, Photogrammetric Engineering & Remote Sensing, Vol. 65, No. 7, pp. 811-821.
Tónjes, R., 1999b. Wissensbasierte Interpretation und 3D-Rekonstruktion von Landschaftsszenen aus Luftbildern,
Fortschritt-Berichte VDI, Reihe 10, Nr. 575, VDI- Verlag, Düsseldorf
350 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000.
