o.55, pp.53-6, 22. SAR DETECTIO OF BUILDIG DAMAGE AREAS DUE TO EARTHQUAKES USIG SATELLITE SAR ITESITY IMAGES Masashi MATSUOKA and Fumio YAMAZAKI An imaging radar called synthetic aperture radar (SAR) has rapidly become one of the major tools for observing the earth s surface and its change. One of the distinctive characteristics of SAR systems is their capability of recording both amplitude and phase of the backscattering echoes from objects on the earth s surface in all weather conditions, day and night, and independence of sun illumination. SAR installed on the satellites J and observed Kobe area before and after the 995 Hyogo-ken anbu, Japan earthquake. Backscattering properties such as the difference and the correlation of the backscattering coefficient called Sigma-nought from the damaged areas by the earthquake were investigated. By analyzing the time-series SAR data, we clarified that the decreasing of the values of the difference and the correlation in the hard-hit areas was due to the surface changes by the event. Then, we developed an automated method to detect hard-hit areas based on the discriminant analysis using pre- and post-event SAR images. Furthermore, the accuracy of the proposed method was verified by comparing the results from detected areas and the one of the actual damage survey. Keywords : Synthetic Aperture Radar (SAR) Intensity Images, Building Damage Area, Backscattering Characteristics, the 995 Hyogo-ken anbu Earthquake, (European Remote Sensing Satellite), J(Japanese Earth Resources Satellite) SAR 995 J ) SAR * ** SAR 2 2),3) 4),5) 2 2 B 2 2 6) SAR 4) 995 European Remote Sensing Satellite Deputy Team Leader, Earthquake Disaster Mitigation Research Center, IED, Dr. Eng. Team Leader, Earthquake Disaster Mitigation Research Center, IED, Dr. Eng. / 8
Satellite (Pre-event) B Satellite (Post-event) Microwave Transmitting & Receiving Strong Reflection Weak Reflection Intact Building Collapsed Building Open Space SAR, J J Frequency 5.3GHz (C-band).275GHz (L-band) Wavelength 5.7 cm 23.5 cm Polarization VV HH Incidence Angle 23 38.7 Orbit Inclination 98.5 97.7 Swath Width km 75 km Spatial Resolution 3 m (3 look) 8 m (3 look) Peak Electric Power 4.8 kw 325 W Recurrence Period 35 days 44 days J 992// 992//22 993/8/8 Á 994/6/3 Á 994//2 994/5/7 Á 995/2/5 995/5/23 995/5/4 995/8/ Á 996/2/27 Á 996/4/2 Á 997/9/ 997/8/7 7) 4 J Japanese Earth Resources Satellite SAR SAR SAR SAR J S/ 2 J 2 7) 994 2 995 5 23 9 J 2 3 J 995 2 5 8) J SAR ) SAR 9), ) 7) ), J 3m 2) 2) J 994 2 r = 2 Iai i= i= Ia Ib Ia Ibi i i i i= i= i= 2 Ia i 2 Ibi i= i= 2 Ib i () 2 / 8
Osaka Bay Á ÂÁ ÃÁÄÅ 4) % % 995 5 23 J d 2 db 3) a b i 7 x 7 Îa i, Îb i d = log Iˆ ai log Iˆ b i (2) J 995 5 4 SAR 2 SAR GIS SAR 4) GIS 2% r Ia i, Ib i a b i 7 x 7 49 2 3 J J J D J 7) 3 / 8
3-3 (95/5/23-94//2) J (95/2/5-94/5/7) J (95/5/4-94/5/7) A B C D E F G.5 (95/5/23-94//2) J (95/2/5-94/5/7) J (95/5/4-94/5/7) A B C D E F G Damage Level Damage Level 2 A 6.25% 2.5% 25% 5% % % B G % 4 3m % % % 995 5 4 4 994 5 7 J 995 2 5 5 7 J J 5 b 3 4 5 a 995 2 5 3 994 5 7 995 5 23 4 994 2 994 6 3 2 A 995 5 23 994 2 994 6 3 2 J SAR SAR SAR SAR 994 2 J 994 5 7 2 b a 6 % G A 4 / 8
3-3 (G) Damage Ratio = J 3-3 (A) o Damage J 993 994 995 996 997 993 994 995 996 997 Aquisition Date Aquisition Date %.5 (G) Damage Ratio = J.5 (A) o Damage J 993 994 995 996 997 993 994 995 996 997 Aquisition Date Aquisition Date % % 995 db.5db.4.6db J 997 % 7 %.5 3 %.35.25.5 6 % J 2 SAR 5 % 2 J 995 2 5 994 5 7 % 8 5 / 8
.5 J (95/2/5-94/5/7) (G) D =.5 J (95/2/5-94/5/7) (A) o Damage.5 (95/5/23-94//2) (G) D =.5 (95/5/23-94//2) (A) o Damage -5 5-5 5-5 5-5 5 J % 995/2/5 994/5/7 % 995/5/23 994//2 % 2 8 umber of Pixels Classified into D= Satellite J Post-event Image 95/5/23 95/2/5 95/5/4 Pre-event Image 92// 93/8/8 94/6/3 94//2 92//22 94/5/7 92//22 94/5/7 (A) o Damage (Total: 2 pixels) 2 26 2 25 24 43 27 47 27 52 27 2 2 47 2 77 (G) D = (Total: 2 pixels) 86 97 963 3 9 974 874 654 268 647 462 957 69 48 747 2 % % 4 8 3 2 6% % 32% 3 4 % 9 7) 69% 5) % 995 5 23 992 993 8 8 994 6 3 994 2 2 4 995 5 23 994 6 3.5 J 95/5-94/ 95/5-94/6 95/5-93/8 95/5-92/ 95/2-94/5 95/2-92/ 95/5-94/5 95/5-92/ -5 5 % J 995 2 5 992 22 994 5 7 995 5 4 995 5 4 995 5 23 994 6 3 995 2 5 3 6 / 8
% 3 % 5% % 7% 2 7% 4 J 995 5 4 4 995 2 4 3 % J 995 2 5 992 22 995/5/23 994/6/3 SAR 995 5 23 a 994 2 4dB 2 995 5 4 J b 994 5 7 c 3 995 2 5 J J 995/5/4 994/5/7 J 995/2/5 994/5/7 SAR 6) 3 7 / 8
J b c SAR S/ SAR J SAR J 2 CD-ROM 8, 26 ) 2 3 GIS WORLD o., pp.74-77, 999. 2) Massonnet, D., Rossi, M., Carmona, C., Adragna, F., Peltzer, G., Fiegl, K., and Rabaute, T.: The Displacement Fields of the Landers Earthquake Mapped by Radar Interferometry, ature, Vol.364, pp.38-42, 993. 3) pp.9-96, 995. 4) Yonezawa, C. and Takeuchi, S.: Detection of Urban Damage Using Interferometric SAR Decorrelation, Proc. International Geoscience and Remote Sensing Symposium, IEEE, CD-ROM, 999. 5) Matsuoka, M. and Yamazaki, F.: Use of Interferometric Satellite SAR for Earthquake Damage Detection, Proc. 6th International Conference on Seismic Zonation, EERI, CD-ROM, 2. 6) Matsuoka, M. and Yamazaki, F.: Interferometric Characterization of Areas Damaged by the 995 Kobe Earthquake Using Satellite SAR Images, Proc. 2th World Conference on Earthquake Engineering, CD-ROM, 2. 7) 995 SAR o.546, pp.54-6, 2.8. 8) SAR pp.75-9 995. 9) 996. ) 989. ) Lee, J.S.: Digital Image Enhancement and oise Filtering by Use of Local Statistics, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol.PAMI-2, o.2, pp.65-68, 98. 2) 99. 3) J SAR / AMI IMAGE 996. 4) 7, 996. 5) Radarsat International Inc.: http://www.rsi.ca/. 6) SAR pp.37-4, 999. ) 8 / 8