ban scenes
posed end-
2osed prob-
multisensor
he chances
; the avail-
lly increase
rease, mul-
emonstrate
acted from
/ leads into
the object
n the sen-
comparison
able.
in the pre-
iat deserve
to combine
In other
it data and
the combi-
is when to
tional con-
take laser
ita sources
n fact, we
range data
ction is an
reo, shape
A combi-
processing
. As a con-
to include
A obtained
'd but still
1-5. The
id the dark
their envi-
, cars (1))
ysical phe-
nomena in object space. Depending on the level of grouping,
extracted features convey information that can be related to
physical phenomena in the object space. Obviously, features
extracted from different sensors should be fused when they
have been caused by the same physical property. Generally,
the further the spectral bands are apart the lesser the fea-
tures extracted from them are caused by the same events.
On the other hand as the level of abstraction increases, more
and more different phenomena are described and need to be
explained.
As pleasing as the object recognition paradigm on the con-
ceptual level is, its implementation on the algorithmic level
is flawed. As pointed out by several researchers (e.g. Mayer
1998), the models currently used for describing objects to
be mapped, are weak. Often, there is a representational in-
compatibility between data and object model which, in turn,
causes the matching to fail. With multispectral and multi-
sensor data available, objects can be modeled more distinctly
and, equally important, closer to what one can extract from
sensory input data.
7 Acknowledgements
The authors would like to thank William Krabill, NASA WFF
and Jim Lucas, NGS, for their help and great support in ob-
taining the multisensor data set at the Ocean City study site.
We also greatly appreciate the help of those who acquired
and processed the data.
References
Abidi, M. A., and R. C. Gonzalez, 1992. Data Fusion in
Robotics and Machine Intelligence. Academic Press, Inc.,
San Diego, CA, 546 pages.
Adams, J. B., M. O. Smith, and P. E. Johnson, 1986. Spec-
tral mixture modeling: a new analysis of rock and soil types
at the Viking Lander site. Journal of Geophysical Research,
Vol. 91. No. B8, pp. 8098-8112.
Benediktsson, J. A., J. R. Sveinsson, and P. H. Swain, 1997.
Hybrid consensus theoretic classification. /EEE Transactions
on Geoscience and Remote Sensing, Vol. 35, No. 4, pp.
833-843.
Chavez, P. S., 1989. Radiometic calibration of Landsat The-
matic Mapper multispectral images. Photogrammetric Engi-
neering and Remote Sensing, Vol. 55, No. 9, pp. 1285-1294.
Cloutis, E. A., 1996. Hyperspectral geological remote sens-
ing: evaluation of analytical techniques (review article). /n-
ternational Journal of Remote Sensing, Vol. 17, No. 12. pp.
2215-2242.
Csathó, B.M., W.B. Krabill, J. Lucas, T. Schenk. A multi-
sensor data set of an urban and coastal scene. Paper in this
proceédings.
Harsanyi, J. C., and C. I. Chang, 1994. Hyperspectral im-
age classification and dimensionality reduction: an orthogo-
nal subspace approach. /EEE Transaction on Geoscience and
Remote Sensing, Vol. 32, pp. 779-785.
Le Hégarat-Mascle, S., |. Bloch, and D. Vidal-Madjar, 1997.
Application of Dempster-Shafer evidence theory to unsu-
pervised classification in multisource remote sensing. /EEE
Transactions on Geoscience and Remote Sensing, Vol. 35,
No. 4, pp. 1018-1031.
Hepner, G. F., T. Logan, N. Ritter, N. Bryant, 1990. Ar-
tificial neural network classification using a minimal training
set: comparison to a conventional supervised classification.
Photogrammetric Engineering and Remote Sensing, Vol. 56,
No. 4, pp. 469-473.
Jensen, J. R., 1983. Urban/suburban land use analysis. ln
Manual of Remote Sensing, Colwell, R. N. (editor-in-chief),
2nd edition, American Society of Photogrammetry, Vol II, pp.
1571-1666.
Krabill, W. B., R. Thomas, K. Jezek, K. Kuivinen, and S.
Manizade, 1995. Greenland ice sheet thickness changes mea-
sured by laser altimetry. Geophysical Research Letters, Vol.
22, No. 17, pp. 2341-2344.
Lee, T., J. A. Richards, and P. H. Swain, 1987. Probabilistic
and evidential approaches for multisource data analysis. IEEE
Transactions on Geoscience and Remote Sensing, Vol. GRS-
25, pp. 283-293.
Mayer, H., 1998. Automatische Objektextraktion aus digi-
talen Luftbildern. Deutsche Geodàtische Kommission, Reihe
C, Nr. 494, 133 pages.
Merényi, E., B. M. Csathó, 1998. Experiments on classifying
multispectral data sets with neural networks for large scale,
urban scenes.
Nadler, M., and E. Smith, 1993. Pattern recognition engi-
neering. John Wiley & Sons, Inc., New York, Ny, pp. 299-
302.
Schistad Solberg, A. H., T. Taxt, and A. K. Jain, 1996. A
Markov random field model for classification of multisource
satellite imagery. /EEE Transactions on Geoscience and Re-
mote Sensing, Vol. 34, No. 1, pp. 100-113.
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 341
