|
|
|
Disasters/Pollution
|
Characteristics of Satellite SAR Images in the Damaged Areas Due to the Hyogoken-Nanbu Earthquake
Extraction of Damaged Area by Satellite SAR
The relationship between the building damage area by the field survey and the satellite SAR images revealed the trend regarding the severe damage ratio. The trend in the area of a high severe damage ratio was obtained when the difference of the backscattering coefficients becomes high and the correlation coefficient becomes low. We tried to extract damaged area by calculating the correlation coefficient and the difference of the backscattering coefficients. The areas corresponding the severe damage ratio larger than 60% and the areas of nondamaged were selected because these areas exhibit the biggest difference in the relationship between the correlation coefficient and the difference of the backscattering coefficients. A linear discriminate analysis is used to classify the areas, and the result of the classification is shown in fig. 6. The building damage distribution by the field survey is damaged buildings. The comparison of the two figures indicated that the classification result from the satellite SAR image almost corresponds to the severely damaged areas by the field survey.
Fig. 6 Classification result that allocated Areas of the severe damage ratio larger 60% and the areas of nondamaged by satellite SAR images
Fig7 Distribution of hard -hit areas based on building damage data by BR1
Conclusions
This paper reports on quantitative evaluation about the characteristic of satellite SAR images of the damaged area using images observed before and after the 1995 Hyogoken-Nanbu Earthquake. As a result, the relationship between the building damage ratio of a district block and the satellite SAR images was found to have a trend regarding the severe damage ratio. This trend in the area of a high severe damage ratio was obtained when the difference of the backscattering coefficients becomes high and the correlation coefficient becomes low. These results suggest that the changes in the earth's surface resulted from severe building damage within a particular area. One possible explanation is that hard-hit areas became open space from the result of building removal at the time of satellite SAR observation, four months after the earthquake. The result from area classification using this trend almost corresponded to the actual severely damaged area by a field survey. In spite of this good agreement regarding the mean characteristic of satellite SAR image, a large amount of randomness still exists in each classification category. Hence, Further study it necessary until a practical use of satellite SAR images for earthquake damage surveys in urban areas.
Acknowledgment
ERS-1/SAR images used in this research are owned by the ESA and used in cooperation with the National Space Development Agency of Japan.
Reference
-
Building Research Institure (1996). Final Report of Damage Survey the 1995 Hyogoken-nanbu Earthquake (in Japanese).
- Japan Society of Photogrammetry and emote Sensing (1998). Synthetic Aperture Radar Image, Asakura Publishing (in Japanese).
- National Space Development Agency of Japan (1996). JERS SAR/ERS AMI IMAGE Data Format Manual.
-
Ozawa, S., Murakami, M., Fujiwara, S., and Tobita, M. (1997), Synthetic aperture radar interferogram of the 1995 Kobe earthquake and its geodetic inversion, Geophysical Research Letters, vol. 24 No 18, pp2327-2330.
-
Remote Sensing Society of Japan (1992). Illustration Remote Sensing, Japan Society of Photogrammetry and Remote Sensing (in Japanese).
-
Yonezawa, c. and Takeuchi, S. (1998). Detection of Damaged Build-up Areas by the 1995 Hyogoken-Nanbu Earthquake using Ers-1/SAR Intensity Images, Photogrammetric Engineering and Remote Sensing, Vol. 37 No. 4, pp57-61 (in Japanese)
|
|
|
|
|
|
|
