|
|
|
Poster Session 2
|
Identification of Damaged areas Due to the 1995 Hyogoken-Nanbu Earthquake Using Satellite Optical Images
Extraction of Damage Distribution Using LANDSAT Images
We tried to extract the earthquake damage distribution from the DN changes using the LANDSAT images, because the LANDSAT images were characteristics of spectral patterns before and after the earthquake.
Damage distribution was extracted by discriminant analysis using the DN after the earthquake and the difference in DN before and after
the earthquake as the
representative variables. The
minimum distance classification
method was used as the
discriminant analysis. The
discriminant table is shown in
table 6. the liquefied or burned
area was had a comparatively
high coincidence ratio. The
distribution of the heavy damage
area (black part in Fig. 4) was in
relatively good agreement with
that of the actual damage survey
result (Fig. 5), but the coincidence ratio of the building damage area was not very high.
Figure.4 : Damage distribution estimated from TM images
Figure.5 : Distribution of hard-hit areas based on building damage data by BRI
| Classified damaged area |
Number of pixels |
Digital number : Average (sd.) |
| 1 |
2 |
3 |
| Liquefied |
500 |
57.5(5.8) |
46.9(7.6) |
40.3(7.7) |
| Burned |
500 |
49.6(2.8) |
36.2(2.6) |
30.4(2.8) |
| Heavy Damage |
500 |
51.0(3.5) |
37.5(3.1) |
31.9(3.3) |
| Slight Damage |
500 |
51.1(3.7) |
36.9(3.2) |
31.2(3.3) |
| No Damage |
500 |
49.5(3.2) |
36.2(2.8) |
31.8(2.9) |
Table 4: Average digital number corresponding to HRV bands for the Classified damage area
| Classified damaged area |
Number of pixels |
Digital number : Average (sd.) |
| 1 |
2 |
3 |
| Liquefied |
500 |
3.0(9.1) |
5.7(10.3) |
6.1(10.2) |
| Burned |
500 |
-0.9(3.6) |
-0.8(3.0) |
-0.3(3.0) |
| Heavy Damage |
500 |
0.3(4.5) |
0.5(3.7) |
0.9(3.6) |
| Slight Damage |
500 |
0.9(4.7) |
0.1(3.7) |
042(3.5) |
| No Damage |
500 |
0.1(3.8) |
0.0(3.2) |
0.1(3.2) |
Table 5: Average digital number difference corresponding to HRV bands for the classified damaged areas
| |
Classified damaged area |
Estimation from TM images |
No. of pixels |
Coinci-
dence ratio (%) |
Lique-
fied |
Burned |
Heavy Dmg. |
Slight Dmg. |
No Dmg. |
| Survey Result |
Liquefied |
224 |
11 |
29 |
15 |
21 |
300 |
75 |
| Burned |
6 |
239 |
28 |
8 |
19 |
300 |
85 |
| Heavy Damage |
14 |
42 |
144 |
51 |
49 |
300 |
48 |
| Slight Damage |
20 |
27 |
78 |
108 |
67 |
300 |
36 |
| No Damage |
17 |
36 |
53 |
46 |
148 |
300 |
49 |
| No. of Pixels |
281 |
355 |
332 |
228 |
304 |
300 |
58 |
Table 6: Discrimination of damage using ground survey data and TM
Images
Conclusions
We examined the spectral characteristics in area damaged by the 1995 Hyogoken Nanbu Earthquake, using LANDSAT/TM and SPOT/HRV images. The digital numbers of the liquefied area were high in the range of visible to infrared bands, whereas those of the burned area became low in the visible light range. The heavy damaged area shown a trend similar to the liquefied area. The extracted damaged distribution using the minimum distance classification agreed with damage survey results.
Acknowledgement.
We thank Prof. Saburoh Midorikawa of Tokyo Institute of Technology and Ms. Asako Inanaga of Remote Sensing Technology Center of Japan for their helpful advice. The LANDSAT and SPOT images used in this study were provided by Space Imaging EOSAT, SPOT IMAGE, and NASDA.
Reference
-
Building Research Institute (1996). Final Report of Damage Survey of the 1995 Hyogoken-Nanbu Earthquake (in Japanese).
-
Hamada, M., isoyama, R. and Wakamatsu, K.(1995). The 1995 Hyogoken-Nanbu (Kobe) Earthquake, Liquefied, Ground Displacement, and Soil Condition in Hanshin Area, Association for Development of earthquake Prediction.
-
Hosokawa, M. and Zama, S. (1998). Comparison of Landcover Mapping Method Using Satellite
Data for Extracting Area Damaged by 1995 Hyogoken-Nanbu Earthquake, Technical Report of National Research Institute of Fire and Disaster, No.85, pp.10-21 (in Japanese).
-
Inanaga, A., Tanaka, S., takeuchi, S., Takasaki, K., and Suga, Y. (1995). Remote Sensing Data for Investigation of Earthquake Disaster, Proc. of the 21st Annual Conf. of the Remote Sensing Society, pp.1089-1096.
-
Mitomi, Y. and Takeuchi, S. (1995). Analysis of spectral Feature of the Damaged Areas by liquefaction and Fire Using Airborne MSS Data, Proc. of the 18th Japanese Conf. on Remote Sensing, pp.117-118 (in Japanese).
-
Sudo, N., Tada, t., Nakano, R., Cho, K., Shimoda, h., and Sakata, T. (1995). Multi Stage Remote Sensing on the Great Hanshin Earthquake Disaster Survey, proc. of the 18th Japanese Conf. on Remote Sensing, pp.115-116 (in Japanese).
-
Yoshie, T. and Tsu, H. (195). Satellite Data Processing to Delineate the Densely Built-up Areas Damaged Disastrously By 1995 Hyogoken-Nanbu Earthquake, Proc. of the 18th Japanese Conf. on Remote Sensing, pp.119-122 (in Japanese).
|
|
|
|
|
|
|
