3-10 SAR 観測における電離層の影響 (セッション 3: 地盤沈下, 地すべり

KURENAI : Kyoto University Research Information Repository
Title
Author(s)
Citation
Issue Date
URL
3-10 SAR観測における電離層の影響 (セッション3: 地盤
沈下, 地すべり, 観測・解析技術)
島田, 政信
SAR研究の新時代に向けて (2013)
2013-02
http://hdl.handle.net/2433/173589
Right
Type
Textversion
Presentation
publisher
Kyoto University
3-10
SAR 観測における電離層の影響
島田政信
宇宙航空研究開発機構、地球観測研究センター、茨城県つくば市千現2−1−1、
tel:050-3362-4489, fax:029-868-2961, mail:shimada.masanobu@jaxa.jp
概要 2006年1月24日に打ち上げられた ALOS には、JERS-1 を継承する L-band
SAR として、更に性能向上と高機能化を果たした PALSAR が搭載され、運用を通し
て、干渉処理技術の高度化のみならず、L-band SAR の有用性の認識と将来の向上
に必要な課題が見えてきた。有用性としては、高い干渉性が確認されたこと、干渉性
を用いたデータ利用研究が増えてきたこと、(地盤沈下、火山監視、地盤沈下、干渉
性を用いた土地利用分類)、高度処理技術であるが時系列データ解析が増えてきた
ことである。一方、顕在化した課題としては電離層の認識とその取り扱いである。ある
意味では、電離層監視手法の開発と同時に補正技術の解析が重要課題になってき
た。本発表では、ALOS 運用開始から見られた電離層の現状とその補正の概要を紹
介する(予定)。
SAR観測における電離層の影響
島田政信
JAXA/EORC
平成24年9月12、13日@京大
Contents
• SARと電離層について
• 幾つかの実例
– Faraday Rotation
– TID
– Streaking (Scintillation)
• 補正方法(の紹介。しかし、、、、)
• 将来の展望
Shisen
RSP124
ScanSAR : descending
DinSAR: Ascending
3
Amplitude
image (hh
polarization)
2006/11/05
Phase (orbit and
terrain corrected
phase)
One example of lower latitude case in Brazil
Direction of the line
20060920-20061105:RSP072:Brazil
35km
35km
Unwrapped phase
S/C moving direction
S/C moving direction
Siberia Area no.1
Four corner lat/lon
71.15N,67.14E 71.32N,69.28E
68.61N,69.00E 68.77N,70.90E
Master image(sar.p_m)
coherence(sar.corr)
DInSAR(sar.ddtma)
RSP:
Date:
516
2009/02/14(Master)
2008/12/30(Slave)
Bperp:1550.9m
FBS343H‐FBS343H
Ascending
W1118562001-01
W1072516001-07
Siberia 2
4 corner lat/lon
71.21N 67.62E 71.38N 69.77E
68.67N 69.49E 68.83N 71.39E
Master(sar.p_m)
coherence(sar.corr)
DinSAR(sar.ddtma)
RSP:
Date:
515
2009/01/28(Master)
2008/10/28(Slave)
Bperp: 1937.2m
FBS343H‐FBS343H
Ascending
W1101515002-09
W1009515001-03
1. Disturbance at higher latitudes
Area
segments
Ratio(
Totals
segments %)
Siberia 546 S (273) 1274 S
(637)
Alaska
42.9%
338 S (169) 2025
16.7%
S(1012.5)
Streaks in
Coherence
phase
Coherence and
phase
2. Streaks at lower latitudes
Area
strips
Equato 1490
rial
area
Totals
strips
64500
Ratio(
%)
2.5%
Streaks in
Coherence
phase
電離層、SAR画像への現れ方、補正方法
Items
Faraday
Rotation
Scintillation
Appearance
Orientation
Rotation
appeared in SLC
Streak Noise
Location shift
SAR
Polarimetry
Polarization
Independency,
Amplitudephase
TID or low
Noise in phase or Polarization
frequency
small azimuth
independency,
component in shift
Phase only
InSAR phase
Method
BB
Fourier
Correctionamplitude –
not for location
Model-based
method
Co-registration
method
Split-widow
method
補正方法
• 振幅データ
– 2D FFT方法:但し、位相量の補正は出来ない
• 位相データ
–
–
–
–
Split Window法(Rosen et al.)
Local Max-Coregistration 法(DLR)
Model法(Meyer et al.)
一般に(非常に)難しい。成功例はないといっても過
言でない。
• 偏波
– BB法で補正出来る。
SARと電離層
æ 2
ö
S = F ç - n ×R÷ Å Fg
è l0
ø
Pt G 0
Ct Gl 2 1
Pt G 2 l 2 0
Ct
Pr =
s bR
+ Pn =
s b
+ Pn
2
2
2
3
4p R
2 cosq 4p 4p R
2 cosq
( 4p ) R
Ct
Pt G 2 l 2 0
PC = A
s b
+ BPn R
2
2 cosq
( 4p ) R
f=-
4p
( rm nm - rs ns )
f
l0
4p
=( rm nm - rs nm + rs nm - rs ns )
l0
4p
4p
=nm ( rm - rs ) rs ( nm - ns )
l0
l0
æ Bperp z
ö 4p
=nm ç
+ Bpara + dr ÷ rs ( nm - ns )
l0 è rm sin q
ø l0
4p
n:屈折率(Refractive Index):
2. ポラリメトリによるTEC推定
13
1.
ポラリメトリによる解析:偏波面回転量とTECの関係
æ Z hh
ç
è Z vh
-4 pr
Z hv ö
1 l æ 1 d 3 öæ cos W sinW ö
÷= A e
ç
֍
÷
Z vv ø
r
èd 4 f 2 øè-sinW cos W ø
æ Shh Shv öæ cos W sinW öæ 1 d1 ö
×ç
÷
֍
֍
è Svh Svv øè-sinW cos W øèd 2 f1 ø
where Zij is the measurement matrix, i is the transmission polarization, j is the
reception polarization, A is the amplitude, r is the slant range, Sij is the true
scattering matrix of the target, f1 is the channel imbalance of the transmission
distortion matrix, f2 is that for the reception matrix, 1 (2) are the cross talks of
transmission, and 3 (4) are the those for the reception. Here, noise is ignored.
校正は森林とコーナー反射鏡を使用
Solutions are obtained by using Quegan’s
method and assuming  =0
14
Faraday Rotationの計算
æZ hh
ç
èZ vh
1)
Z hv ö æ cos W sinW öæS hh
÷= ç
֍
Z vv ø è-sinW cosW øèS vh
1
*
W = Arg Z LR ×Z RL
4
3)
Shv ×S*hv
S hv öæ cos W sinW ö
֍
÷
S vv øè-sinW cosW ø
æ
ö
1
-1 Z hv - Z vh
W = tan ç
÷ Freeman et al. ??
2
è Z hh + Z vv ø
Shv = Svh
2)
=
Svh ×S*vh
æ ZLL
ç
è ZRL
ZLR ö æ1 jöæ Zhh
÷= ç
֍
ZRR ø è j 1øè Zvh
) (
*
)
a ×tanW 1+ tan 2W - b 1- tan 4 W = 0
15
Zhv öæ1 jö
֍
÷
Zvv øè j 1ø
a = (Z hv + Z vh ) ×(Z hh + Z vv ) + (Z hv + Z vh ) ×(Z hh + Z vv )
b = Z hv ×Z*hv - Z vh ×Z*vh
(
手法は各種存在
*
ìb
ü
W1 = tan -1 í 1- tan 4 W 0 - tan 3W 0 ý
îa
þ
æb ö
W 0 = tan -1ç ÷
èa ø
(
)
Faraday Rotation Angel (model)
W=
K
B ×cos y ×secq 0 ×TEC
2
f
where K=2.365x104 in SI units, f is the transmission frequency (Hz),
TEC is the total electron contents (m3/m2), B is the geomagnetic
flux density (Tesla), y is the angle between the geomagnetic field
vector and the radar line-of-sight (radian), q0 the incidence angle,
and the over-bars indicate averaging.
16
Bern大学のサイトより
F.A review of ionospheric effects in low-frequency SAR — Signals, correction methods, and performance requirements
Geoscience and Remote Sensing Symposium (IGARSS), 2010 IEEE International
Causes for the stripes
Scintillation in range
Ion Density variation :Azimuth shift
Doppler Frequency:Observation target (ionosphere)
changes the Doppler frequency.
2R
f (- )
C
SAR received signal at intermediate
frequency:
æ
ö
ç 2dR / dT
÷
df
2R
2
= jw f ç +
n dn / dT ÷
2
C
dT
0
ç
÷
æ C0 ö
çè ÷
çè
÷
n
ø
nø
( )
Time variation:
Doppler
Doppler
by media
Shift in
azimuth
n(T -VpT)
n:electron density
fd
T
Shift in Az and rg
df / dT = df / d(-2R / C) ×d(-2R / C) / dT
æ 2R ' 2R dC ö
= f ¢ç + 2
÷
è C
C dT ø
Ne2
n = 1e0w 2 m
æ 2R ' 2R dn ö
= jw f ×ç + 2
è nC0 n C dT ÷
ø
æ 2R ' 2R dn ö
= jw f ×ç + 2
è nC0 n C dT ÷
ø
æ 2R ' 2R -e2 dN ö
= jw f ×ç + 2
è nC0 n C 2e 0w 2 m dT ÷
ø
æ 2R -e2 dN ö
fde = f0 ×ç 2
è n C 2e 0w 2 m dT ÷
ø
Doppler frequency due to the media
variation in azimuth
Electron density distribution
Temperature
Representative parameters for the ionosphere:
Electron mass (m):
9.109e-31kg
Electric charge (e):
1.602e-19 Coulomb
Emissivity at space (e0): 8.854e-12Fm-1
Light speed (c):299792458m-1s-1
Angular speed:2*PAI*1.27e9s-1
If we assume that dN/dT~1.0e9/m^3s-1, fde~0.2Hz at the positive slope and 0.2Hz at the negative slope. It vibrates in azimuth.
Change in Doppler ->
Azimuth shift mainly
very slightly in range.
DfD
Dy =
vg
- fDD
f
1Hz
0.2Hz
:
:
:
fDD=-500Hz/s
Vg=6.7km/s
y
13m
2.6m
Geometric evaluation using the corner reflector.
CRs in Amazon are used for the location shift and the resolution.
Geolocation
2006/w
2007/w
Resolution
2008/w
Azimuth and Range shifts
S/C
Rare
Smaller TEC, slower C, projected
nearer
k
dr = 2 DTEC
f
k=40.28 m3/s2
10^9*500000*40.28/1.27e9^2
=12.4m
-fD
Dense
+fD
Rare
1x10^9/m^3/s
N
SAR imaging (azimuth compression) is
not affected by the scintillation
SAR and Scintillation
R1
q=23.4568 degrees
N
1
2
3
R0
D
D
191.3
270.7
331.5
Plasma Bubble
SAR
q
Scintillation line
2R0 = 2R1 + nl
nl ( H - z ) z
D=
H cos q
Sub satellite track
ì ( H - z ) z nl ü
q = cos í
2 ý
D þ
î H
-1
Synthetic aperture line
Simulation: Ranging
SAR
N 0 e2
N = 1me0w 2
Distance
N1
R0
N1(1012/m3)>N2(109/m3)
N2
Bubble
R1
R0 n0 R1n1 R2 n0
+
+
c
c
c
Intensity increase
T=
Normal
R3
R2
Distance change
DR
Abnormal
Normal
R0 n0 R1n0 R3n0
T=
+
+
c
c
c
Assumption on the electron density
distribution in the bubble
Case1 : Density jump in the ellipsoid
N1
N2
Case2 : radius dependent density
distribution in the ellipsoid
N1
2
ì
r1 (f,q ) ü
1
N=
N
N
N
í 2 ( 2
ýdf
1)
ò
f2 - f1 f1 î
r0 (f ) þ
f
N2
Geomagnetic lines
Inclination
14.79 degrees
Declination=-14.185 deg.
Comparison of the density variations on N1 and N2
N2 1011
N1
1012
○
1011
1010
109
○
×
×
1010
a = 7km+1.5kmx3
b = 10km+1.5kmx3
Difference of the electron density should be at least 10^12~10^11.
×
Simulated image for the Amazon case
Electron density model: case2
Inclination: 14.79 degrees
Declination:-14.185 degrees
Number of bubbles:
1 large + 4 smalls
Measured cross section of the electron density
Prof. Watkins at IGARSS2009
Modelによる解法
• 相当大変:IGARSS2013 でF. Meyerに期待が
かかったが、成功例の話はでなかった(残念)
Correction of ionospheric distortions in low frequency interferometric SAR data
Jun Su Kim; Danklmayer, A.; Papathanassiou, K.; IGARSS), 2011 IEEE International
代表的な論文
Meyer, F.A review of ionospheric effects in low-frequency SAR — Signals,
correction methods, and performance requirements
Geoscience and Remote Sensing Symposium (IGARSS), 2010 IEEE International
Digital Object Identifier: 10.1109/IGARSS.2010.5654258
Publication Year: 2010 , Page(s): 29 – 32
Measurement and mitigation of the ionosphere in L-band Interferometric SAR
data
Rosen, P.A.; Hensley, S.; Chen, C.
Radar Conference, 2010 IEEE
Digital Object Identifier: 10.1109/RADAR.2010.5494385
Publication Year: 2010 , Page(s): 1459 – 1463
Correction of ionospheric distortions in low frequency interferometric SAR
data
Jun Su Kim; Danklmayer, A.; Papathanassiou, K.; IGARSS), 2011 IEEE
International
Page(s): 1505 - 1508
Digital Object Identifier: 10.1109/IGARSS.2011.6049353
まとめ及び将来の展望
• 厄介な問題である。
• 唯一FRだけが解決されている。
• 解決方法はまだ定まっていない。
• >PALSAR-2 の85MHzは希望かも。