anbul 2004 
International Archives of the Photogrammetry, Remote Sensing 
  
yrization 
  
  
  
  
  
  
and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
  
  
  
  
  
Figure 10 Examples of extraction of damaged houses based on 
the overlapping of change detection results with map. 
The registered imageries can also be searched by specification 
of type, name, time, comments, range etc, as shown in the 
dialog window in Figure 11. 
   
  
nr EM : 
pHERHM — rma 
   
  
    
  
  
  
  
- Ji ETE | | 
LET af ie 
i Es (EEE TN 
FM MESA N EHE: gE | 15 | 
77148 | SEE ise 
AE H | 
i Im UO | 
EE e oo dr E 
| Rs mo W|-RBEOW. | 
|" EOS | 
  
  
zcEeLudzL | 
Figure 11. Search setting dialog 
5. EVALUATION RESULTS 
Using a PC with a CPU of above 2.66GHz, and memory over 
312 MB, imagery of about 2GB cach can be registered in about 
30 minutes. Thus for a total of 5 aerial photos that covers about 
10 Km’, it takes less than 3 hours to register them. When 
choosing the compressed structured image format, the resulted 
hierarchical image has a size of about 200MB, which is about 
one tenth of the original one. At this compression rate, the 
deterioration of imagery's quality is not obvious, and the 
decompressing does not affect the system performance. 
The change after the disaster can be identified either by 
alternatively setting the system display time to before and after 
the disaster, or by overlapping the result of change detection 
with the imageries and map. 
Resources such as snap shot and video are also registered as 
spatial objects, which can be searched with spatial-temporal 
conditions. 
  
  
6. CONCLUSIONS 
We have proposed and implemented an image integration 
system for decision support when disaster happens. It enables 
registration and searching of all types of imageries including 
video clips and snap images, both in spatial and temporal space. 
Imagery of even giga byte size can be displayed without much 
stress, by using a hierarchical structure, with compression 
option. Simulation or change detection results of raster format 
can also be converted to polygons for efficient and precise 
analysis with standard GIS functions. 
REFERENCES: 
Kakumoto S., 1996. Development of Disaster Management 
Spatial Information System (DIMSIS). GIS for Disaster 
Management: UNCRD's 25th Anniversary Commemorative 
Programme. Proceedings of the 9th International Research and 
Training Seminar on Regional Development Planning for 
Disaster Prevention, pp.59-65 12 December 1996, Nagoya, 
Japan 
Kosugi, Y., Fukunishi, M., Sakamoto, M., Lu, W. and Doihara, 
T., 2001. Detection of sheer changes in aerial photo images 
using an adaptive nonlinear mapping. Proc. DMGIS 2001, 
pp.145-146, Bangkok. 
Lu, W., 2003. Data collection and information retrieval for 
SDS of disaster management. ISPRS Workshop on Spatial 
Analysis and Decision Making, 2003 Hong Kong. 
Sakamoto M., Yukio Kosugi, Masanori Nakamura, Munenori 
Fukunishi and Takeshi Doihara 2001. Change Detection of 
Geographical Features with An Adaptive Nonlinear Stereo 
Mapping Process, Proc. LUCC2001, Session 5, R8, CD-ROM, 
7pages, Tokyo, 2001. 
References from websites: 
BMECSST, 2002: The Research and Development Bureau of 
Ministry of Education, Culture, Sports, Science and 
Technology: About the selection of research organization of 
"Special Project for Earthquake Disaster Mitigation in Urban 
Areas" (in Japanese) 
http:/Avww.mext.go.jp/b menu/houdou/14/02/0202 13i.htm 
(accessed 19 August. 2002) 
KIWI+, 2001: KIWI+ : Simple Topology & Spatial Temporal - 
Open Database Schema (ST2-ODS) (Specification in Japanese) 
http://www.drm.jp/KiwiPLUS/index.htm 
ACKNOWLEDGEMENT: 
This work was supported by the Special Project for Earthquake 
Disaster Mitigation in Urban Area. 
A -