le tr.
eth.
tion
»pre-
eter-
rtain
t ob-
nt in
rates
e re-
pe).
orom
as to
next.
jgre-
ters,
e ex-
lying
> ag-
mple
jects
two
rep-
n be
bject
| ob-
task
rela-
1 the
riable
Differences: In the case of model generalization, the iden-
tification of the objects is already given (e.g. buildings,
parcels of land, streets). Thus the learning starts al-
ready with objects instead of mere polygons. However
in the course of the learning it might become necessary
to create new or intermediate objects. In this way the
interpretation capability can be exploited as well.
5 SUMMARY AND CONCLUSION
After an introduction in the importance of multiple represen-
tations in GIS, some of the problems when generating such
descriptions were presented. Special emphasis was put on
the transition between different levels of detail. This has been
identified as a learning problem in multiple representation.
Starting from a similar approach in image interpretation, a
concept for the learning of generalization rules was pre-
sented. This transfer is possible since both domains base on
a description of objects in terms of an object-class hierarchy
with complex object relations. Depending on the type of ob-
ject and its methods, different actions can take place. These
actions in the first case help to identify and interpret the ob-
jects, in the second case they are applied for the derivation
of other levels of detail.
References
Barrett, R., Ramsay, A. & Sloman, A. [1985], POP-11, A
Practical Language for Artificial Intelligence, Ellis Hor-
wood Ltd., Chichester, West Sussex, England.
Beard, K. & Mackaness, W. [1991], Generalization opera-
tions and supporting struktures, in 'Auto-Carto 10, Tech-
nical Papers ACSM/ASPRS Annual Convention', Vol. 6,
pp. 29—45.
Bobick, A. & Bolles, R. [1989], Representation space: An
approach to the integration of visual information, in ‘Pro-
ceedings of the CVPR 1989’, pp. 492—499.
Bundy, G., Jones, C. & Furse, E. [1995], Holistic generaliza-
tion of large-scale cartographic data, in Müller, Lagrange
& Weibel [1995], pp. 106—119.
Hahn, M. [1989], Automatic measurement of digital terrain
models by means of image matching techniques, in ‘Pro-
ceedings of the 42th Photogrammetric Week’, Vol. 13 of
Series of the Institute of Photogrammetry, Stuttgart Uni-
versity, Stuttgart, pp. 141-151.
Harbeck, R. [1995], Überblick über Konzeption, Aufbau und
Datenangebot des Geoinformationssystem ATKIS, in
E. Kophstahl & H. Sellge, eds, 'Das Geoinformationssys-
tem ATKIS und seine Nutzung in Wirtschaft und Verwal-
tung - Vortráge des 2. ADV-Symposiums ATKIS am 27.
und 28. Juni 1995 in Hannover’, Nieders.
773
Jager, E. [1990], Untersuchungen zur kartographischen
Symbolisierung und Verdrängung im Rasterddatenfor-
mat, PhD thesis, Fachrichtung Vermessungswesen, Uni-
versitat Hannover.
Müller, J.-C., Lagrange, J.-P. & Weibel, R., eds [1995], GIS
and Generalization - Methodology and Practice, Taylor
& Francis.
Müller, J., Weibel, R., Lagrange, J. & Salge, F. [1995], Gener-
alization: state of the art and issues, in Müller, Lagrange
& Weibel [1995], pp. 3-17.
Quinlan, J. [1986], ‘Induction of decision trees’, Machine
Learning1(1), 81-106.
Ruas, A. & Lagrange, J. [1995], Data and knowledge mod-
elling for generalization, in Müller, Lagrange & Weibel
[1995], pp. 73-90.
Sester, M. [1990], Multiscale representation for knowledge
based recognition and tracking of objects in image se-
quences, in ‘International Archives of Photogramme-
try and Remote Sensing’, Vol. 28, ISPRS, Comm. lll,
Wuhan.
Sester, M. [1995], Lernen struktureller Modelle fiir die Bild-
analyse, Vol. C441, Deutsche Geodätische Kommission,
München.
van Oosterom, P. [1995], The GAP-tree, an approach to 'on-
the-fly" map generalization of an area partitioning, in
Müller, Lagrange & Weibel [1995], pp. 120—132.
Walter, V. & Fritsch, D. [1995], Matching techniques for road
network data in different data models, in J. Soliman &
D. Roller, eds, '28th International Symposium on Au-
tomotive Technology and Automation', Automotive Au-
tomation Limited, Croydon, England, pp. 633—640.
Weibel, R. [1995], Three essential Building blocks for au-
tomated generalization, in Müller, Lagrange & Weibel
[1995], pp. 56-70.
Weibel, R., Keller, S. & Reichenbacher, T. [1995], Overcom-
ing the knowledge acquisition bottleneck in map gener-
alization: The role of interactive systems and computa-
tional intelligence, in A. Frank & W. Kuhn, eds, ‘Spatial
Information Theory’, Springer, Berlin, pp. 139-156.
Winkelhausen, H. [1995], ATKIS-Daten als Grundlage eines
Planungsinformationssystems fiir die Region Bodensee-
Oberschwaben, in ‘ATKIS-Daten und ihre Nutzung,
DVW-Mitteilungen, Heft 2’, pp. 85-96.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
