International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
An example of a correctly rejected grassland GIS-object is
given in Figure 8. While the GIS cropland object was covered
with vegetation on the pictures on the right hand side, the
picture in the second row on the left hand side does not show
any vegetation. As grassland is covered by vegetation all year
(Itzerott and Kaden, 2007), this GIS-object has to be grassland
and not cropland as indicated in the GIS data set. This result
can be confirmed by taken a look at the aerial image on the
upper row on the left hand side. The aerial orthoimage is used
to prove the result; it was not used within the verification
process. The decision of our system to reject the GIS-object is
correct.
Figure 8: Correct rejected grassland GIS object: aerial orthophoto,
April 2009 (top left), RapidEye CIR - 27.09.2009 (top right) -
24.04.2009 (bottom left) - 24.08.2009 (bottom right).
5. CONCLUSIONS AND OUTLOOK
The method for the verification of cropland and grassland
objects described in this paper achieved satisfactory for both
classes even that the results from the class cropland are slightly
better. In this publication we determined a suitable value for
parameter f,. t, is important to transfer the classification result
to a GIS-object and has a big influence on the verification
results. An investigation regarding the other parameters which
needs experiences from a human operator to be set still has to
be done.
Furthermore, a detailed analysis of the relevant features would
be interesting, in order to reduce the feature vector to the most
relevant features, and at the same time to reduce the necessary
numbers of training areas. Maybe the best choice of features
even could be determined during the training phase.
In addition, the approach was tested on only one multi-temporal
multi-spectral data set so far. It is interesting to see the
performance on further data sets. Especially because two out of
the three images were taken to approximately the same time
(only a few days different), the appearance of the vegetation
hardly changes. Tests showed that using only the RapidEye
images the results were comparable to using all three images;
comparable means that the time efficiency was slightly lower
and the TA a posteriori slightly higher.
In this paper we focused only on the classes cropland and
grassland. The features should be useable to achieve also good
results for further classes, e.g. the separation of different forest
types (deciduous and coniferous forest).
ACKNOWLEDGEMENTS
This work was supported by the German Federal Agency for
Cartography and Geodesy (BKG) in Frankfurt am Main.
69
REFERENCES
Albertz, J., 2001. Einfiihrung in die Fernerkundung. Wissenschaftliche
Buchgesellschaft, Darmstadt, Germany.
Baatz M and Schape A (2000) Multiresolution segmentation-an
optimization approach for high quality multiscaleimage segmentation.
Angewandte Geographische Informationsverarbeitung XII. (Eds:
Strobl, J. and Blaschke, T.), Beitrage zum AGIT- Symposium, Salzburg
2000, Karlsruhe, Herbert Wichmann Verlag, pp.12—23
Balaguer, A., Ruiz, L. A., Hermosilla, T., Recio, J.A., 2010. Definition
of a comprehensive set of texture semivariogram features and their
evaluation for object-oriented image classification. Computers &
Geoscience, 36, pp. 231-240.
Busch, A., Gerke, M., Griinreich, D., Heipke, C., Liedtke, C. E., Müller,
S. 2004. Automated Verification of a Topographic Reference
Dataset: System Design and Practical Results. IntArchPhRS, vol. 35,
no. B2, pp. 735-740.
Chang, C.C. and Lin, C.J., 2001. LIBSVM: a library for support vector
machines. htt{p:/www.csie.ntu.edu.tw/"cjlin/lLibsvm, accessed: 09/2010.
De Wit, A. J. W. and Clevers, J. G. P. W., 2004. Efficiency and
accuracy of per-field classification for operational crop mapping.
International Journal of Remote Sensing, 25(20), pp. 4091-4112.
DIN, EN, ISO (Hrsg.), Qualititsmanagement. Begriffe. DIN EN ISO
8402: 1995-08. Berlin, 1995.
EEA - European Environment Agency, 2011: Corine Land Cover.
http://www.eea.europa.eu/publications/CORO-landcover, (accessed:
06/2011).
Gerke, M. and Heipke, C., 2008. Image based quality assessment of
road databases. International Journal of Geoinformation Science (22),
No. 8, 2008, pp. 871-894.
Hall, A., Louis, J., Lamb, D., 2008. Low-resolution remotely sensed
images of winegrape vineyards map spatial variability in planimetric
canopy area instead of leaf area index. Australian journal of grape and
wine research, vol. 14, no. 1, pp. 9-17.
Haralick, R.M., Shanmugam, K., Dinstein, 1973. Texture features for
image classification. IEEE Trans. On Systems, man. And Cibernetics,
SMC-3, pp. 610-622.
Helmholz, P., Rottensteiner, F., Heipke, C., 2010. Automatic qualtiy
control of cropland and grassland GIS objects using IKONOS Satellite
Imagery. IntArchPhRS, vol. 38, no 7/B, pp. 275-280.
Itzerott, S. and Kaden, K., 2007. Klassifizierung landwirtschaftlicher
Fruchtarten. Photogrammetrie Fernerkundung Geoinformation (PFG),
vol. 2, 2007, pp. 109-120.
Lucas, R., Rowlands A., Brown A., Keyworth, S. and Bunting, P.,
2007. Rule-based classification of multi-temporal satellite imagery for
habitat and agricultural land cover mapping. ISPRS Journal of
Photogrammetry & Remote Sensing, 62(3), pp.165-185.
Müller, S., Heipke, C. and Pakzad, K., 2010. Classification of farmland
using multitemporal aerial images. IntArchPhRS, vol. XXX VIII, part 4-
8-2/W9, Haifa, pp. 70-74.
Rengers, N. and Prinz, T., 2009. JAVA-basierte Texturanalyse mittels
Neighborhood Gray-Tone Differenz Matrix (NGTDM) zur Optimierung
von Landnutzungsklassifikation in hoch auflósenden
Fernerkundungsdaten. PFG, vol. 5, pp. 455-467.
Ruiz, L.A., Fernàndez-Sarria, A. and Recio, J.A., 2004. Texture
feature extraction for classification of remote sensing data using
wavelet decomposition: A comparative study. IntArchPhRS, vol.
XXXV, pp. 1109-1115.
Simonneaux, V., Duchemin, B. , Helson, D. , Er-Raki, S. , Olioso, A.
and Chehbouni, A. G., 2008. The use of high-resolution image time
series for crop classification and evapotranspiration estimate over an
irrigated area in central Morocco. International Journal of Remote
Sensing, 29(1), pp. 95-116.
Vapnik, V.N., 1998 Statistical Learning Theory. Wiley, New York,
USA.
