»090v 
nage object 
Itural plot 
AMEWORK 
use classification 
ing techniques to 
ter, segmentation 
1e region growing 
. These basic land 
d, grass, parts of 
open water, etc. 
n constituents of 
| patches of bare 
ie, or a part of a 
vork, labelling of 
ts produced at an 
basic land cover 
objects would be 
bjects. Relatively 
t of backyards of 
B3. Istanbul 2004 
      
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
Residential Area | 
  
  
Non-residential Area 
Large-scale 
  
  
Forest 
  
  
  
Farmland 
  
  
  
  
  
Office Building 
  
  
  
  
House 
A 
  
  
  
| À 
  
  
  
  
  
Backyard 
  
  
  
  
  
Rice field, 
fallow ete 
Plant community 
  
  
  
  
  
Manmade 
object 
  
  
  
  
  
  
| | 
  
Soil Grass 
  
  
  
  
  
Tree 
  
Le 
a 
  
  
Pixels with different spectral properties 
Small-scale 
  
  
Figure 5. Schematic of the object relationships among the image and geographic features 
houses. On the other hand, relatively large patches of bare soil 
may be part of fallow plots. 
Spectral and textural information would be also utilized for 
detailed land use mapping in the framework of the object-based 
classification. Labelling plant community types such as 
deciduous forests, bamboo forests, etc. needs spectral and 
texture analyses of image objects. 
A land use mapping of the study area was demonstrated using 
the above classification strategy, and the result was compared to 
an existing land use map which was compiled from the results 
of aerial photo-interpretation and ground truth data. The land 
use map created through the object-oriented approach showed a 
good correspondence with the existing map. 
6. CONCLUDING REMARKS 
This study discussed the correspondence between image objects 
produced with a multi-scale segmentation technique and land 
use classes commonly observable in rural land uses in Japan, 
through an actual image analysis using IKONOS data, and 
suggested a land use classification framework employing the 
object-oriented approach. Regarding the future work, the 
proposed classification framework would be tested in the land 
use mapping projects in the different areas for the validation 
and generalization of the classification strategy. 
References 
Baatz, M. and Schápe, A., 2000. Multiresolution segmentation: 
an optimization Approach for high quality multi-scale image 
segmentation. In: Strobl, J. and Blaschke, T. (Eds.): 
Angewandte Geographische Informations verarbeitung XII, 
Wichmann-Verlag, Heidelberg, pp.12-23. 
Blaschke, T., Strobl, J. 2002. What’s wrong with pixels? Some 
recent developments interfacing remote sensing and GIS. 
GeoBIT/GIS: J. Spatial inform. Decision Making, No. 6/2001, 
pp. 12-17. 
Usuda, Y., et al. 2003. A study on the optimization of image 
segmentation in object-Oriented Classification. Proceedings 
of the annual conference of the Japan Society for 
Photogrammetry and Remote Sensing in 2003 pp. 125-128. 
(in Japanese) 
van der Sande, C. J, et al. 2003. A segmentation and 
classification approach of IKONOS-2 imagery for land cover 
mapping to assist flood risk and flood damage assessment. Int. 
J. Applied Earth Observation and Geoinformation, Vol. 4, pp. 
217-229.