In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Voi. XXXVIII, Part 7B
573
Figure 5: Comparison of cluster maps extracted using object oriented analysis (left) and self-organizing maps (right) for GAC detection.
GAC regions are orange, urban areas are white, deciduous forests are light green whereas coniferous forests are dark green, and water
bodies are blue. There is high degree of similarity between these cluster maps. SOM correctly extracts urban areas (white regions
within ellipses on the right) whereas they are captured as GAC by object oriented analysis (orange within ellipse on the left). However,
inland grass, pink regions within the rectangle on the left, cannot be extracted by the SOM (orange on the right) due to the necessity of
spatial context whereas SOM clustering is pixel based.
8 CONCLUSIONS
The paper proposed a methodology for annual inventory and mon
itoring of the land which may be ‘eligible’ under SAPS in Bul
garia, using RapidEye imagery. A legal definition of “Good Agri
cultural Condition (GAC)” was introduced as a starting point for
assessment of eligible area. An object oriented classification of
multi-temporal RapidEye data was performed in order to quantify
agricultural area in GAC on annual basis. In addition to the object
oriented analysis, an alternative method based on self-organizing
maps has also been used. Preliminary results are encouraging and
they clearly indicate that multi-temporal remote sensing data can
effectively contribute to differentiate currently active and poten
tial agriculture land, and land which cannot be considered suit
able for agriculture in the context of SAPS. However further vali
dation of the methodology for the other test zones is necessary. It
is envisaged to follow-up discussions of results with the Bulgar
ian Administration.
RapidEye imagery (in terms of information content) seems to be
particularly suitable for feature detection and land cover mapping
of agriculture landscapes. As the spatial resolution doesn’t cor
respond to 1:10 000 scale, the imagery cannot be used directly
for LPIS update; however it can provide essential information on
the overall accuracy of the LPIS in relatively short time frame,
provided that the acquisition approach is adapted to the user ex
pectations. The proposed methodology may also help Bulgaria
(and Romania) to further develop their concept in respect to the
eligibility conditions currently applied under SAPS.
REFERENCES
Gamanya, R., Maeyera, P. D. and Dappera, M. D., 2007. An automated
satellite image classification design using object-oriented segmentation
algorithms: A move towards standardization. Expert Systems with Ap
plications 32(2), pp. 616-662.
Gitelson, A. A., Merzyak, M. N. and Lichtenthaler, H., 1996. Detection
of red edge position and chlorophyll content by reflectance measurements
near 700 nm. Journal of Plant Physiology.
Haboudane, D., Miller, J. R., Tremblay, N., Zarco-Tejada, P. J. and Dex-
traze, L., 2002. Integrated narrow-band vegetation indices for prediction
of crop chlorophyll content for application to precision agriculture. Re
mote Sensing of Environment 81(2), pp. 416-426.
Kohonen, T., 1997. Self-Organizing Maps. 2nd edn, Springer-Verlag
Berlin Heidelberg.
Mathieul, R. and Aryal, J., 2005. Object-oriented classification and
ikonos multispectral imagery for mapping vegetation communities in ur
ban areas. In: The 17th Annual Colloquium of the Spatial Informa
tion Research Centre (SIRC-2005), University of Otago, Dunedin, New
Zealand, November.
Milenova, L., Nedkov, R., Vassilev, V., Milenov, R, Radkov, R. and
Pironkova, Z., 2001. Preparation of land cover database of Bulgaria
through remote sensing and GIS, Carlo Travaglia FAO. Technical report,
Environment and Natural Resources Service, Sustainable Development
Department Food and Agriculture Organization of the United Nations.
Taçdcmir, K. and Merényi, E., 2009. Exploiting data topology in visu
alization and clustering of Self-Organizing Maps. IEEE Transactions on
Neural Networks 20(4), pp. 549-562.
Taçdemir, K. and Milenov, R, 2010. An automated SOM clustering based
on data topology. In: Proc. 18th European Symposium on Artificial
Neural Networks (ESANN’10), Bruges, Belgium, D-Facto, April 28-30,
pp. 375-380.
Vinal, A. and Gitelson, A. A., 2005. New developments in the remote
estimation of the fraction of absorbed hotosynthetically active radiation
in crops. Geophysical Research Letters.
Wu, C., Niu, Z., Tang, Q., Huang, W., Rivard, B. and Feng, J., 2009. Re
mote estimation of gross primary production in wheat using chlorophyll-
related vegetation indices. Agricultural and Forest Meteorology 149(6-7),
pp. 1015-1021.
