R APPORT
Satellittbasert radarinterferometri (InSAR) for naturfare, skred og infrastruktur
Naturfareprosjektet: Delprosjekt 3 Kartlegging, datasamordning og ROS-analyser
2015 122
2 Utgitt av:
Redaktør:
Forfattere:
Trykk:
Opplag:
Forsidefoto:
ISBN
Sammendrag:
Emneord:
Norges vassdrags- og energidirektorat Middelthunsgate 29
Postboks 5091 Majorstua 0301 OSLO
Telefon: 22 95 95 95 Telefaks: 22 95 90 00 Internett: www.nve.no
Rapport nr 122-2015
Satellittbasert radarinterferometri (InSAR) for naturfare, skred og infrastruktur
Norges vassdrags- og energidirektorat
Rouyet, Line; Lauknes, Tom Rune; Høgda, Kjell-Arild
NVEs hustrykkeri P.O.D
Figur 6 From sentinel.esa.int.
978-82-410-1174-0 ISSN 1501-2832
The project aims to evaluate the potential (advantages and limitations) of satellite Synthetic Aperture Radar Interferometry (InSAR) technology for the mapping of deformation potentially damaging and dangerous for infrastructure and population, with a focus on the main interests of the three partners of the NIFS project
SAR satellites, SAR Interferometry (InSAR), ground deformation, geohazards & infrastructure mapping and monitoring
5(3257
,6%1
,661
6SDFHERUQHUDGDULQWHUIHURPHWU\,Q6$5IRUQDWXUDO KD]DUGVODQGVOLGHVDQGLQIUDVWUXFWXUHOLPLWDWLRQVDQG
SRWHQWLDO
6DWHOOLWWEDVHUWUDGDULQWHUIHURPHWUL,Q6$5IRU QDWXUIDUHVNUHGRJLQIUDVWUXNWXUEHJUHQVQLQJHURJ
PXOLJKHWHU
$XWKRUV5RX\HW/LQH/DXNQHV7RP5XQH+¡JGD.MHOO$ULOG
352-(&71$0(1,)6,Q6$5 352-(&712
&2175$&7,1*19(6WDWHQVYHJYHVHQDQG-HUQEDQHYHUNHW
'RFXPHQW1R 'RFXPHQW7\SH5HSRUW 6WDWXV)LQDO
,6%1 ,661 1RRI3DJHV
3URMHFWOHDGHU/LQH5RX\HW 'DWH
$87+2565RX\HW/LQH/DXNQHV7RP5XQH+¡JGD.MHOO$ULOG
7,7/(6SDFHERUQH UDGDU LQWHUIHURPHWU\ ,Q6$5 IRU QDWXUDO KD]DUGV ODQGVOLGHV DQG LQIUDVWUXFWXUH OLPLWDWLRQV DQG SRWHQWLDO 6DWHOOLWWEDVHUW UDGDULQWHUIHURPHWUL ,Q6$5 IRU QDWXUIDUH VNUHG RJ LQIUDVWUXNWXUEHJUHQVQLQJHURJPXOLJKHWHU
6800$5<
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
.H\ZRUGV 6$5 VDWHOOLWHV 6$5 ,QWHUIHURPHWU\ ,Q6$5 JURXQG GHIRUPDWLRQ JHRKD]DUGV LQIUDVWUXFWXUHPDSSLQJDQGPRQLWRULQJ
1RWLFHV
38%/,6+(51RUXW
- Contents -
&217(176
,1752'8&7,21
7+(25< 0(7+2'2/2*<
2.1 Synthetic Aperture Radar Imaging ... 2
Active remote sensing ... 2
2.1.2 SAR principle and geometry ... 2
2.1.3 SAR orbit configuration ... 3
2.1.4 SAR geometrical effects ... 4
2.1.5 SAR complex images ... 5
2.1.6 SAR azimuth & range resolutions ... 7
2.1.7 Comparison of SAR satellites characteristics ... 7
2.2 SAR Interferometry principles ... 11
2.2.1 Introduction ... 11
2.2.2 Line-of-sight and detection capability ... 11
2.2.3 Phase changes between acquisitions ... 11
2.2.4 InSAR coherence and decorrelation ... 13
2.3 InSAR processing and results ... 15
2.3.1 Processing of phase components ... 15
2.3.2 Unwrapping procedure ... 17
2.3.3 Multi-interferogram techniques ... 18
2.3.3.1 Persistent Scatterer Interferometry (PSI) ... 20
2.3.3.2 Small BAseline Subset method (SBAS) ... 21
2.3.4 InSAR outputs ... 21
678'<$5($ 3.1 Hazards context ... 24
3.1.1 Events inventory ... 24
3.1.2 Focus of the study ... 25
3.2 Satellite data characteristics ... 26
3.2.1 Temporal coverages & resolutions ... 26
3.2.2 Spatial coverages & resolutions ... 27
3.2.3 Line-of-sight & geometrical effects ... 29
,16$55(68/76%(5*(1$5($ 4.1 Shadow & layover masks ... 32
4.2 Temporal decorrelations ... 35
4.2.1 Spatial variations ... 35
4.2.2 Seasonal variations ... 36
4.2.3 Coherence or amplitude dispersion filters ... 37
4.3 Ground deformation results ... 39
4.3.1 Bergen ... 39
- Contents -
4.3.1.1 Deformation maps ... 39
4.3.1.2 Time series ... 44
4.3.2 Indre Arna ... 48
4.3.2.1 Deformation maps ... 48
4.3.2.2 Time series ... 50
4.3.3 Samnangerfjorden - Kvamsskogen - Dale ... 52
',6&866,216 5.1 InSAR advantages and limitations ... 54
5.2 InSAR developments and potential ... 56
&21&/86,21 5()(5(1&(6 $33(1',;(6 8.1 RSAT-2 Standard mode images... 66
8.2 RSAT-2 Ultrafine images ... 68
8.3 TSX/TDX StripMap images ... 69
0
,1752'8&7,21
7KLVUHSRUWLVSDUWRID5HVHDUFK 'HYHORSPHQWSURMHFWLQLWLDWHGE\1RUJHV9DVVGUDJV RJ(QHUJLGLUHNWRUDW19(6WDWHQVYHJYHVHQDQG-HUQEDQHYHUNHW7KH1,)61DWXUIDUH ,QIUDVWUXNWXU )ORP 6NUHG SURMHFW LV GLYLGHG LQ VHYHQ VXESURMHFWV IRFXVLQJ RQ KD]DUGVPDSSLQJDQGPRQLWRULQJHDUO\ZDUQLQJULVNDVVHVVPHQWDQGFULVLVPDQDJHPHQW VHHZZZQDWXUIDUHQR7KLV LV SDUWRIWKH1,)6VXESURMHFW NDUWOHJJLQJVNUHGRJ IORP DLPLQJ WR SURYLGH DQ RYHUYLHZ RI WKH ZRUN LQ QDWXUDO KD]DUGV PDSSLQJ DQG DVVHVVPHQW WDNLQJ LQWR DFFRXQW WKH YDULRXV SURFHVVHV ODQGVOLGHV URFNVOLGHV DYDODQFKHVURFNIDOOVGHEULVIORZVIORRGV
7KH RYHUDOO REMHFWLYH RI WKH UHSRUW ³6SDFHERUQH UDGDU LQWHUIHURPHWU\ ,Q6$5 IRU QDWXUDO KD]DUGV ODQGVOLGHV DQG LQIUDVWUXFWXUH OLPLWDWLRQV DQG SRWHQWLDO´ LV WR HYDOXDWH WKH SRWHQWLDO DGYDQWDJHV DQG OLPLWDWLRQV RI VDWHOOLWH 6\QWKHWLF $SHUWXUH 5DGDU ,QWHUIHURPHWU\ ,Q6$5 WHFKQRORJ\ IRU WKH PDSSLQJ RI GHIRUPDWLRQ SRWHQWLDOO\
GDPDJLQJ DQG GDQJHURXV IRU LQIUDVWUXFWXUH DQG SRSXODWLRQ ZLWK D IRFXV RQ WKH PDLQ LQWHUHVWV RI WKH WKUHH SDUWQHUV RI WKH 1,)6 SURMHFWV 19( 6WDWHQV YHJYHVHQ DQG -HUQEDQHYHUNHWODQGVOLGHVURFNIDOOVPRYHPHQWVDIIHFWLQJURDGUDLOVLQIUDVWUXFWXUHHWF ,W FDQ EH VHHQ DV D FRQWLQXDWLRQ RI WKH UHSRUW LQLWLDWHG E\ 1RUVN 5RPVHQWHU WKH 1RUZHJLDQ 6SDFH &HQWHU DERXW WKH PXOWLXVH SRWHQWLDO RI UDGDU VDWHOOLWHV DQG ,Q6$5 WHFKQRORJ\ IRU PDSSLQJ DQG PRQLWRULQJ ODQGVOLGH KD]DUGV DQG LQIUDVWUXFWXUH
³.DUOHJJLQJRIRYHUYnNQLQJDYVNUHGIDUHRILQIUDVWUXNWXUYHGEUXNDYUDGDUVDWHOOLWWHURJ ,Q6$5PHWRGLNN´>6WU¡PHWDO@
7KLVUHSRUWLQFOXGHVDSUHVHQWDWLRQRIWKHFXUUHQWVWDWXVDQGXVHRI,Q6$5WHFKQRORJ\LQ GHIRUPDWLRQ PDSSLQJ D GLVFXVVLRQ RI WKH SRWHQWLDO RI WKH PHWKRGV DV ZHOO DV UHFRPPHQGDWLRQV IRU IXUWKHU ZRUNV ([DPSOHV RYHU %HUJHQ DUHD ZKHUH ZH XVHG WZR VDWHOOLWH GDWDVHWV IURP 5$'$56$7 DQG RQH VDWHOOLWH GDWDVHW IURP 7HUUD6$5
;7DQ'(0; >&HWLQLF HW DO LQ SUHS@ FRQWULEXWH WR LOOXVWUDWH GLIIHUHQW DGYDQWDJHV DQG OLPLWDWLRQV RI WKH WHFKQRORJ\ &KDSWHU LV GHGLFDWHG WR WKHRUHWLFDO H[SODQDWLRQVDERXW6$5V\VWHPVDQG,Q6$5PHWKRGV,Q&KDSWHUWKHFRQWH[WRIWKH PDLQ VWXG\ DUHD DV ZHOO DV WKH FKDUDFWHULVWLFV RI WKH DYDLODEOH GDWD DUH LQWURGXFHG ,Q
&KDSWHU H[DPSOHV RI UHVXOWV RYHU %HUJHQ DUHD DUH SUHVHQWHG DQG GLVFXVVHG 7KH
&KDSWHU LV ILQDOO\ GHGLFDWHG WR WKH PDLQ GLVFXVVLRQV DERXW WKH DGYDQWDJHV DQG OLPLWDWLRQV RI ,Q6$5 WHFKQRORJ\ DV ZHOO DV WKH IXWXUH GHYHORSPHQWV DQG SRWHQWLDO LQ WKLVILHOG
$FNQRZOHGJHPHQWV 5$'$56$7 GDWD KDYH EHHQ SURYLGHG E\ .RQJVEHUJ 6DWHOOLWH 6HUYLFHV.6$7WKURXJKDQDJUHHPHQWZLWK1RUZHJLDQ6SDFH&HQWUHDQG7HUUD6$5
;7DQ'(0;GDWDDUHRZQHGE\*OREHVDUWKURXJKDQDJUHHPHQWZLWK$LUEXVDQG DUH UHIHUUHGLQWKHWH[WWRWKHUHSRUW>&HWLQLFHWDOLQSUHS@
2
7+(25< 0(7+2'2/2*<
,QWKLV&KDSWHUZHSUHVHQWDQLQWURGXFWLRQWRVDWHOOLWH6\QWKHWLF$SHUWXUH5DGDU6$5 6$5 ,QWHUIHURPHWU\ ,Q6$5WHFKQLTXHIRUWKHGHWHFWLRQ RIJURXQG GHIRUPDWLRQ DQGWKHPDLQSURFHVVLQJVWHSVDQGPHWKRGVZLWKWKHLUUHVSHFWLYHDGYDQWDJHVDQG OLPLWDWLRQV
7KHFRQWHQWRIWKH&KDSWHULVPDLQO\EDVHGRQWKHFRQWULEXWLRQVRI>:RRGKRXVH /DXNQHV )HUUHWWL @0RUHGHWDLOHGWUHDWPHQWV RIWKH6$5SULQFLSOHFDQEH IRXQGLQHJ>&XUODQGHU 0F'RQRXJK)UDQFHVFKHWWL /DQDUL&XPPLQJ :RQJ @ 0RUH GHWDLOV WUHDWPHQWV RI ,Q6$5 SULQFLSOHV DQG SURFHVVLQJ FDQ EH IRXQGLQHJ>0DVVRQQHWDQG)HLJO%DPOHUDQG+DUWO5RVHQHWDO 5RFFDHWDO+DQVVHQ.DPSHV)HUUHWWL@
6<17+(7,&$3(5785(5$'$5,0$*,1*
$&7,9(5(027(6(16,1*
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
WRLPDJHWKH(DUWKGD\DQGQLJKW
6$535,1&,3/($1'*(20(75<
6\QWKHWLF $SHUWXUH 5DGDU 6$5 V\VWHP ZDV GHYHORSHG LQ WKH ILIWLHV ILUVW IRU DLUERUQH PLVVLRQV7KHWHFKQRORJ\DOORZVXVLQJDVKRUWSK\VLFDODQWHQQDEXWLQRUGHUWRLPSURYH D]LPXWKUHVROXWLRQLWV\QWKHWL]HVWKHHIIHFWRIPXFKORQJHUDQWHQQDXVLQJWKHPRYHPHQW RIWKHVHQVRUDORQJWKHVDWHOOLWHIO\LQJSDWKWUDFN5HVROXWLRQLQUDQJHLVREWDLQHGE\
XVLQJDKLJKEDQGZLGWKV\VWHP
$VSDFHERUQHRUDLUERUQH6$5KDVDVLGHORRNLQJJHRPHWU\DVVKRZQLQ)LJXUH$V WKH6$5PRYHVDORQJLWVDVVXPHGVWUDLJKWSDWKLWLOOXPLQDWHVDVZDWKRQWKHJURXQG E\ WUDQVPLWWLQJ D VHULHV RI HOHFWURPDJQHWLF PLFURZDYH SXOVHV 7KH DQJOH EHWZHHQ WKH UDGDUEHDPDQGWKHYHUWLFDOLVFDOOHGORRNDQJOHZKLOHWKHDQJOHEHWZHHQWKHUDGDUEHDP DQG WKH VXUIDFH QRUPDO VOLJKWO\ ODUJHU GXH WR WKH FXUYDWXUH RI WKH (DUWK LV FDOOHG LQFLGHQFHDQJOH
3
7KH6$5GDWDDUHDUUDQJHGDVWZRGLPHQVLRQDO'LPDJHVZLWKFRRUGLQDWHVH[SUHVVHG LQVODQW UDQJH GLVWDQFH IURP WKH 6$5 VHQVRU DQG D]LPXWK SRVLWLRQ DORQJ WKH 6$5 IOLJKW SDWK 'XH WR WKH VLGHORRNLQJ JHRPHWU\ DQG UHODWHG WR WKH FRQYHUVLRQ RI WKH PHDVXUHG VODQW UDQJH WR WKH JURXQG UDQJH JHRPHWULFDO GLVWRUWLRQV ZLOO DSSHDU LQ WKH 6$5 LPDJH 7KH SURFHVV RI JHQHUDWLQJ 6$5 LPDJHV ZLWK XQLIRUP SL[HO VL]H LQ JHRJUDSKLFFRRUGLQDWHV\VWHPLVFDOOHGJHRFRGLQJ
)LJXUH6LPSOLILHGJHRPHWU\RID6\QWKHWLF$SHUWXUH5DGDU6$5V\VWHP$VWKHV\VWHPPRYHVDORQJ LWVIOLJKWSDWKLWLPDJHVDQDUHDRQWKHJURXQGFDOOHGVZDWK)URP>/DXNQHV@
6$525%,7&21),*85$7,21
6DWHOOLWHVHTXLSSHGZLWK6$5VHQVRUVKDYHDQHDUSRODURUELWDWDQDOWLWXGHEHWZHHQ DQGNPDERYHWKH(DUWK¶VVXUIDFH(DUWK¶VURWDWLRQDQGVDWHOOLWHPRYHPHQWOHDGWR DQDVFHQGLQJDQGDGHVFHQGLQJLPDJLQJJHRPHWU\:LWKDQDVFHQGLQJRUELWWKHVDWHOOLWH FURVVHVWKHHTXDWRUJRLQJIURP6RXWKWR1RUWKDQGZLWKDGHVFHQGLQJRUELWLWFURVVHVWKH HTXDWRUJRLQJIURP1RUWKWR6RXWK)LJXUH
4
)LJXUH*HRPHWU\RIDVFHQGLQJOHIWDQGGHVFHQGLQJULJKWVDWHOOLWHRUELWV0RGLILHGIURP>)HUUHWWL /DXNQHV@
6$5*(20(75,&$/())(&76
:KHQ WRSRJUDSKLF IHDWXUHV H[LVW 6$5 LPDJHV DUH DIIHFWHG E\ GLVWRUWLRQV GXH WR WKH VLGHORRNLQJJHRPHWU\RIWKHVDWHOOLWH)RUHVKRUWHQLQJDSSHDUVRQWKHVORSHVIDFLQJWKH UDGDU UHVXOWLQJ LQ FRPSUHVVHG SL[HOV RQ WKH JURXQG 7KH RSSRVLWH HIIHFW JLYHV EHWWHU UHVROXWLRQ RQ VORSHV IDFLQJ DZD\ IURP WKH UDGDU )RU VWHHSORRNLQJ VSDFHERUQH UDGDU V\VWHPV WKH VODQW UDQJH GLIIHUHQFHV EHWZHHQ WZR SRLQWV ORFDWHG RQ IRUHVORSHV RI PRXQWDLQVDUHVPDOOHUWKDQWKH\ZRXOGEHLQIODWDUHDV)LJXUHD¶E¶,QWKHH[WUHPH FDVHOD\RYHUDSSHDUVZKHQWKHWRSRIDKLOOLVFORVHUWRWKHUDGDUWKDQWKHIRRWRIWKHKLOO )LJXUH E¶F¶ G¶H¶ ,Q WKLV FDVH WKH UHFHLYHG VLJQDO IURP DW OHDVW WZR GLIIHUHQW DOWLWXGHVLVDGGHGLQWRRQHVODQWUDQJHUHVROXWLRQFHOOOHDGLQJWRDYHU\KLJKUDGDU DQG DPELJXRXV UHWXUQ 7KH OD\RYHU HIIHFW FDQ HDVLO\ EH REVHUYHG LQ 6$5 LPDJHV DV DUHDV ZLWKH[WUHPHO\KLJKLQWHQVLW\2QWKHRWKHUVLGHRIWKHPRXQWDLQVKDGRZRFFXUVLQWKH DUHD QRW EHLQJ LOOXPLQDWHG E\ WKH UDGDU )LJXUH F¶G¶ 7R SDUWO\ RYHUFRPH WKLV OLPLWDWLRQ PRVW PRGHUQ VDWHOOLWHV KDYH WKH SRVVLELOLW\ WR RSHUDWH ZLWKLQ D UDQJH RI GLIIHUHQWLQFLGHQFHDQJOHVZKLOHWKH(56VDWHOOLWHVZKLFKRSHUDWHGEHWZHHQDQG IRUH[DPSOHKDGIL[HGUHODWLYHO\VWHHSLQFLGHQFHDQJOHV
5
)LJXUH([DPSOHRIJHRPHWULFGLVWRUWLRQVRYHUWKH9HVXYLXVYROFDQR(56LPDJH$SULO 7KHDQDO\VLVRIWKHUDQJHOLQH$$¶DOORZVKLJKOLJKWLQJGLIIHUHQWHIIHFWVIRUHVKRUWHQLQJOD\RYHUDQG
VKDGRZLQJ7KH\KDYHDQLPSDFWRQWKHLQWHQVLW\RIWKHEDFNVFDWWHUHGVLJQDOVLQFHLPDJHSL[HOV FRUUHVSRQGWRWHUUDLQSDWFKHVZLWKGLIIHUHQWH[WHQVLRQVEULJKWRQVORSHVIDFLQJWKHVHQVRUVGDUNRQVORSHV
IDFLQJWKHRSSRVLWHGLUHFWLRQ0RGLILHGIURP>)HUUHWWL@
6$5&203/(;,0$*(6
Radar images include both amplitude and phase components. The amplitude gives information on the reflectivity of the terrain, and is related to the amount of the energy backscattering to the receive antenna ()LJXUHD). The phase is a property related to the sensor-target distance, used in SAR Interferometry (InSAR) (see section 2.2).
Because of the sinusoidal shape of radar wave, the sensor-target distance can be expressed as a GHILQHG QXPEHU RI ZDYHOHQJWKV ZLWK DQ DGGLWLRQ RI D IUDFWLRQ RI D ZDYHOHQJWKȜFRUUHVSRQGLQJWRWKHSKDVHij'XHWRWKHWZRZD\WUDYHORIWKHZDYHWKH HIIHFWLYHZDYHOHQJWKLVDFWXDOO\Ȝ)LJXUH
6
)LJXUH$VLQXVRLGDOIXQFWLRQVLQijLVSHULRGLFZLWKDʌUDGLDQSHULRGijLVWKHSKDVHVKLIWFRPSDUHGWR WKHWUDQVPLWWHGVLJQDO'XHWRWKHWZRZD\WUDYHORIWKHZDYHWKHHIIHFWLYHZDYHOHQJWKLVȜ0RGLILHG
IURP>)HUUHWWLHWDO@
7KHSKDVHijLVWKXVMXVWDIUDFWLRQRIDZDYHF\FOHUDQJLQJEHWZHHQDQGʌPRGXOR ʌWKDWFDQEHUHODWHGWRWKHVODQWUDQJH5E\DVLPSOHOLQHDUHTXDWLRQ
߮ ൌʹߨ
ɉ ʹܴ ൌ ʹߨ
ɉ ʹΤ ܴ ൌͶߨ ɉ ܴ
/RRNLQJ DW )LJXUH E WKH SKDVH LQIRUPDWLRQ ORRNV OLNH UDQGRP QRLVH $FWXDOO\ WKH SKDVHYDOXHVRIDVLQJOH6$5DFTXLVLWLRQKDYHQRGLUHFWXVH7KHYDOXHVDUHXQLIRUPO\
GLVWULEXWHGEHWZHHQDQGGXHWRWKHVXSHUSRVLWLRQRIWKHHFKRHVFRPLQJIURPWKH PXOWLSOH UHIOHFWLRQV RQ WKH GLIIHUHQW SDUWV RI WKH WDUJHWV ZLWKLQ HDFK FHOO 7KH FRPELQDWLRQ RI WKH GLIIHUHQW SKDVH LQIRUPDWLRQ FDXVHV LQWHUIHUHQFHV ZKLFK FUHDWH DQ LPDJHDIIHFWHGE\VSHFNOHQRLVH:HZLOOVHHLQVHFWLRQKRZ6$5LQWHUIHURPHWU\LV KRZHYHUDEOHWRH[SORLWWKLVFRPSRQHQW
)LJXUH$PSOLWXGHDDQGSKDVHEYDOXHVRID(566$5LPDJHRYHU0LODQLQ-DQXDU\
FRPSDUHGZLWKDQRSWLFDOLPDJHRIWKHVDPHDUHDG0XOWLUHIOHFWLYLW\PDSFUHGXFHVWKHQRLVHE\
DYHUDJLQJWKHDPSOLWXGHYDOXHVRI6$5LPDJHV)URP>)HUUHWWL@
7
6$5$=,087+ 5$1*(5(62/87,216
The spatial resolution in SAR varies according to the sensor and is different in azimuth and range direction. The slant range resolution depends on the bandwidth of the sensor.
,OOXPLQDWLQJEHDPFDQEHVHHQDVWKHVXSHUSRVLWLRQRIDVHWRIVLQXVRLGDOVLJQDOVFHQWUHG DW DQRSHUDWLQJFHQWUDO IUHTXHQF\RIWKHUDGDUVHQVRU7KHEDQGZLGWK LV WKHGLIIHUHQFH EHWZHHQ WKH KLJKHVW DQG WKH ORZHVW IUHTXHQF\ RI WKH WUDQVPLWWHG VLJQDO /DUJHU LV WKH EDQGZLGWK EHWWHU LV WKH UDQJH UHVROXWLRQ 7R REWDLQ WKHJURXQG UDQJH UHVROXWLRQ WKH VODQW UDQJH PXVW EH SURMHFWHG RQWR WKH JURXQG XVLQJ D VLPSOH WULJRQRPHWULF UHODWLRQ GHSHQGLQJRQWKHLQFLGHQFHDQJOH7KHLQFUHDVHRIWKHLQFLGHQFHDQJOHIURPQHDUWRIDU UDQJHOHDGVWRDQLPSURYHPHQWLQWKHJURXQGUDQJHUHVROXWLRQ
8VLQJ WKH 'RSSOHU VKLIW GXH WR WKH PRYHPHQW RI WKH VDWHOOLWH SODWIRUP D]LPXWK UHVROXWLRQRI6$5V\VWHPVLVLQGHSHQGHQWRQWKHGLVWDQFHIURPWKHUDGDUWRWKHWDUJHW,W GHSHQGV RQ WKH OHQJWK RI WKH DQWHQQD LQ DORQJWUDFN GLUHFWLRQ 6KRUWHU DQWHQQDV \LHOG ILQHUUHVROXWLRQVVLQFHDVLQJOHSRLQWRQWKHJURXQGZLOOEHLOOXPLQDWHGIRUDORQJHUWLPH 'XHWRYDULRXVUHODWLRQVKLSVUHODWHGWRWKHSXOVHIUHTXHQF\WKHYHORFLW\RIWKHVDWHOOLWH DQG WKH DQWHQQD VL]H D ODUJHU FRYHUDJH ZLOO KRZHYHU LPSO\ D UHGXFHG D]LPXWK UHVROXWLRQ
&203$5,6212)6$56$7(//,7(6&+$5$&7(5,67,&6
(DFK6$5VHQVRUKDVLWVRZQ FKDUDFWHULVWLFVWKDWKDYHDQLPSDFW RQWKHPHDVXUHPHQW FDSDELOLW\ FRQVLGHULQJ PRYHPHQW GHWHFWLRQ GHSHQGLQJ RQ WKH GLUHFWLRQ WKH YHORFLW\
DQG WKH VL]H RI WKH PRYLQJ DUHD DWPRVSKHULF SHQHWUDWLRQ DQG VLJQDO VWDELOLW\ RQ GLIIHUHQWVXUIDFHV$VLPSOLILHGRYHUYLHZRIWKHVHHOHPHQWVLVDYDLODEOHLQ7DEOH7KH LPSRUWDQWSDUDPHWHUVRIVRPHUHOHYDQW6$5VHQVRUVDUHSUHVHQWHGLQ7DEOHDQG7DEOH
7DEOH0DLQ6$5VDWHOOLWHVHQVRUVFKDUDFWHULVWLFVDQGWKHLULQIOXHQFHVRQPHDVXUHPHQWV
SAR characteristics Related to Impact on
Sensor wavelength Used frequency band
Detection capability (dep. on the velocity of the moving area), atmosphere penetration & stability on different
surfaces Spatial resolution
(ground range &
azimuth resolutions)
Bandwidth, antenna length in along-track direction and
incidence angle
Detection capability (dep. on the size of the moving area)
Spatial coverage Antenna size, incidence angle Size of the potential analysed area
Line-of-sight Orbit and incidence angle Detection capability (dep. on the movement direction)
Temporal resolution Revisit cycle of the satellite Detection capability (dep. on the velocity on the moving area) & decorrelation
8
7DEOH5HOHYDQWVDWHOOLWH6$5PLVVLRQVZLWKPDLQVHQVRUSDUDPHWHUV0RGLILHGIURP>/DXNQHV@
0LVVLRQ (56 (QYLVDW
$6$5b 5$'$56$7b 5$'$56$7
&RXQWU\$JHQF\ (6$ (6$ &DQDGD &DQDGD
/DXQFK
(QGRIOLIH0LVVLRQGXUDWLRQ F G \UVFRQWLQH
%DQG:DYHOHQJWK>FP@ & & & &
,QFLGHQFH$QJOH>GHJ@ ± ± ± ±
6ZDWKZLGWK>NP@a ± ± ±
*URXQGUDQJHUHVROXWLRQ>P@a ± ± ±
$]LPXWKUHVROXWLRQ>P@a
5HYLVLW7LPH>GD\V@
7DEOH5HOHYDQWVDWHOOLWH6$5PLVVLRQVZLWKPDLQVHQVRUSDUDPHWHUV0RGLILHGIURP>/DXNQHV@
0LVVLRQ $/26
3$/6$5I
7HUUD6$5;
7DQ'(0;
&2602
6N\0HG 6HQWLQHODE
&RXQWU\$JHQF\ -DSDQ *HUPDQ\ ,WDO\ (6$
/DXQFK
(QGRIOLIH0LVVLRQGXUDWLRQ \UV \UVFRQWLQH \UVFRQWLQH \UV
%DQG:DYHOHQJWK>FP@ / ; ; &
,QFLGHQFH$QJOH>GHJ@ ± ± ± ±
6ZDWKZLGWK>NP@a ± ± ± ±
*URXQGUDQJHUHVROXWLRQ>P@a ± ± ± ±
$]LPXWKUHVROXWLRQ>P@a ± ±
5HYLVLW7LPH>GD\V@ J K
D6LJQLILFDQWYDULDWLRQVDFFRUGLQJWRWKHDFTXLVLWLRQPRGH E2QO\6WULS0DSPRGHLQFOXGHG
F(56RSHUDWLQJZLWKRXWJ\URVLQFH-DQXDU\DIIHFWLQJLQWHUIHURPHWULFDSSOLFDWLRQV
G7KHRUELWZDVFKDQJHGLQ2FWREHUHQGLQJWKHLQWHUIHURPHWULFWLPHVHULHV HDespite the official life duration, the satellite is still operating.
I2QO\6SRWOLJKWDQG6WULS0DSPRGHVLQFOXGHG
J&OXVWHURIVDWHOOLWHVZLWKDWKHRUHWLFDOGD\VUHYLVLWWLPH K:LWKWKHDGGLWLRQRI6HQWLQHOEWKHUHYLVLWWLPHZLOOEHGD\V
9
7DEOHDQG7DEOHKLJKOLJKW WKHVLJQLILFDQW YDULDELOLW\RIWKHLQFLGHQFH DQJOHVZDWK ZLGWK DQG VSDWLDO UHVROXWLRQV HYHQ IRU D VSHFLILF VHQVRUFigure 6 LOOXVWUDWHV GLIIHUHQW RSHUDWLQJPRGHVRIWKHQHZ(XURSHDQVDWHOOLWH6HQWLQHO
)LJXUH)RXU6HQWLQHOPRGHVKLJKOLJKWLQJWKHKLJKYDULDELOLW\RIVSDWLDOFRYHUDJH)RUHDFKPRGHWKH UDQJH D]LPXWKUHVROXWLRQDVZHOODVWKHLQFLGHQFHDQJOHYDU\DOVR)RU1RUZD\WKH,:6PRGHZLOOEH
PDLQO\XVHG)URPVHQWLQHOHVDLQW
)LJXUH VKRZV H[DPSOHV RI 6$5 LPDJHV IURP (QYLVDW $6$5 5$'$56$7 8OWUDILQHDQG7HUUD6$5;6WULS0DSPRGHZLWKGLIIHUHQWUHVROXWLRQV7KHGLIIHUHQFHLQ GLVWLQJXLVKDEOHIHDWXUHVLVREYLRXV,QDGGLWLRQWKHLPSDFWRIWKHGLIIHUHQWOLQHRIVLJKW /26FDQEHKLJKOLJKWHGHVSHFLDOO\WKHLQIOXHQFHRIWKHLQFLGHQFHDQJOHRQWKHOHYHORI UDGDUVKDGRZERWKLQFUHDVLQJIURPDWRF
10
)LJXUH6$5LPDJHVRYHU1RUGQHV7URPV1RUZD\,PDJHVIURPGLIIHUHQWVHQVRUVLOOXVWUDWLQJ GLIIHUHQWVSDWLDOUHVROXWLRQV7KHVDWHOOLWHVORRNWRZDUGWKH:HVWGHVFHQGLQJJHRPHWU\$OOLPDJHVKDYH
EHHQJHRFRGHGWRWKHVDPHSL[HOVL]HD(QYLVDW$6$5E5$'$56$78OWUDILQHF7HUUD6$5;
6WULS0DSGDHULDOSKRWR)URP>/DXNQHV@
11 6$5,17(5)(520(75<35,1&,3/(6
,1752'8&7,21
As discussed in section 2.1.5, the phase information of a single image is not directly usable. However, phase differences between two (or more) different acquisitions are meaningful. This is the main topic of SAR interferometry (InSAR). By computing phase differences between two acquisitions from two different locations, it is possible to build a Digital Elevation Model (DEM), while by computing phase changes between two acquisitions at two different times, it is possible to detect ground deformation along the line-of-sight (LOS). This is the focus of this report. 7KHPHWKRGRORJ\KDVEHHQPDLQO\
GHYHORSHG LQ WKH ¶V DIWHU ODXQFK RI WKH (6$ (XURSHDQ 6SDFH $JHQF\ (56 VDWHOOLWHV ZKLFK WULJJHUHG D ORW RI ,Q6$5 UHVHDUFK DFWLYLW\ ,Q WKLV SHULRG ,Q6$5 ZDV VXFFHVVIXOO\XVHGWRVWXG\VXUIDFHGLVSODFHPHQWVGXHWRHJJODFLHUG\QDPLFV>*ROGVWHLQ HWDO@YROFDQRGHIRUPDWLRQ>0DVVRQQHWHWDO$PHOXQJHWDO@DQG HDUWKTXDNHV >0DVVRQQHW HW DO 0DVVRQQHW DQG )HLJO @ 0RUH UHFHQWO\ WKH SRWHQWLDORIGLIIHUHQWLDO ,Q6$5KDVEHHQLQYHVWLJDWHGWRVWXG\ODQGVOLGHV>%HUDUGLQRHW DO+LOOH\HWDO6WUR]]LHWDO5RWWDQG1DJOHU&ROHVDQWLDQG :DVRZVNL/DXNQHV@
/,1(2)6,*+7$1''(7(&7,21&$3$%,/,7<
The analysis of phase changes between two acquisitions at two different times can provide information about ground deformation along the line-of-sight (LOS) of the SAR sensor (Figure 8). InSAR LV RQO\VHQVLWLYHWR GLVSODFHPHQWVWKDWKDYHD FRPSRQHQWLQ WKH/26GLUHFWLRQZKLFKdepends on the flying orientation of the satellite (track) and the incidence angle of the radar beam (see section 2.1.2). SWHHSHUWKHLQFLGHQFHDQJOHV OHDG WR EHWWHU VHQVLWLYLW\ WR YHUWLFDO GLVSODFHPHQWV /RRNLQJ WRZDUG WKH :HVW D GHVFHQGLQJ RUELW JLYHV PDLQO\ QRQGLVWRUWHG FRYHUDJH LQ ZHVW IDFLQJ VORSH DQG DQ DVFHQGLQJFRYHUVPDLQO\HDVWIDFLQJVORSHV6HQVLWLYLW\LVYHU\ORZLQFDVHVZKHUHWKH DFWXDOVXUIDFHGLVSODFHPHQWYHFWRULVQHDUSHUSHQGLFXODUWRWKH/26'XHWRWKH1RUWK 6RXWKRUELWGLUHFWLRQWKHVHQVLWLYLW\WRVXUIDFHGLVSODFHPHQWLQWKLVSODQHLVQHDU]HUR
3+$6(&+$1*(6%(7:((1$&48,6,7,216
Due to the modulo-2π and the effective wavelength λ/2 explained in section 2.1.5, there is an ambiguity if the phase difference between the two acquisitions is higher than λ/2.
Thus, in practice, for a sensor with a wavelength of 5.55 cm with a revisit time of 24 days (as RADARSAT-2), a moving area with velocity higher than 2.8 cm in 24 days will induce a phase ambiguity.
12
)LJXUH'HWHFWLQJVXUIDFHGLVSODFHPHQWZLWK,Q6$5$QDUHDRQWKHJURXQGLVLPDJHGDWWZRGLIIHUHQW WLPHWDQGW$SRWHQWLDOVXUIDFHGHIRUPDWLRQRFFXUULQJEHWZHHQWKHDFTXLVLWLRQWLPHVZLOOOHDGWRDQ LQWHUIHURPHWULFSKDVHWHUPǻ5G7KHFORXGDWWKHDFTXLVLWLRQWLOOXVWUDWHVWKDWDWPRVSKHULFHIIHFWVFDQ
DIIHFWWKHUDGDUSURSDJDWLRQSURGXFLQJDQDGGLWLRQDOSDWKGHOD\)URP>/DXNQHV@
0RUHRYHUWHUUDLQGHIRUPDWLRQLVRQO\RQHRIWKHFRPSRQHQWVRIWKHSKDVHGLIIHUHQFH,Q DGGLWLRQ WKH WRSRJUDSKLF FRPSRQHQW GXH WKH GLVWDQFH EHWZHHQ WZR RUELWV FDOOHG EDVHOLQH DQG H[SUHVVHG DV % LQFigure 8), the atmospheric effects and other noise contributions has to be taken into account. The interferometric phase difference can thus be expressed as >%HUDUGLQRHWDO@:
ο߮ ൌ Ͷߨ ߣ
ܤୄ
ܴݏ݅݊ߠݖ Ͷߨ
ߣ οܴௗ ο߮ௌ ο߮ௗ
ZKHUHߣLVWKHUDGDUZDYHOHQJWK]LVWKHHOHYDWLRQRIWKHWDUJHWSRLQWDERYHDUHIHUHQFH SODQHܤୄLVWKHGLVWDQFHEHWZHHQWKHWZRRUELWVSHUSHQGLFXODUEDVHOLQHο5GLVVXUIDFH GLVSODFHPHQW DORQJ WKH /265 LV WKH VODQW UDQJH GLVWDQFHߠ LV WKH LQFLGHQFH DQJOH ο߮$36LVWKHGLIIHUHQFHLQDWPRVSKHULFSDWKGHOD\DWPRVSKHULFSKDVHVFUHHQ$36DQG ο߮GHFRUULVDWHUPLQFOXGLQJRWKHUQRLVHFRQWULEXWLRQVGHFRUUHODWLRQVRXUFHV
,QRUGHUWRVWXG\VXUIDFHGLVSODFHPHQWVWKHWRSRJUDSKLFFRQWULEXWLRQKDVWREHUHPRYHG 7KLV LV PRVW RIWHQ GRQH E\ XVLQJ DQ H[WHUQDO KLJKSUHFLVLRQ 'LJLWDO (OHYDWLRQ 0RGHO '(07KHUHPRYLQJRIDWPRVSKHULFHIIHFWVLVRQHRIPDMRULVVXHRI,Q6$5SURFHVVLQJ WKDWZLOOEHGLVFXVVHGLQVHFWLRQ
Figure 9 shows an example of an interferogram between two ENVISAT ASAR acquisitions in 2003-2004. The results highlight the co-seismic displacement field originating from an earthquake in Bam, Iran. The values are expressed in modulo-2π (-π to + π) corresponding to λ/2 (28 mm or -14 to 14 mm). This interferogram is thus
13
wrapped, highlighting a succession of patterns called fringes. In section 2.3, we will show how to convert these cyclic phase differences to absolute phase differences and thus continuous distance differences (unwrapping procedure).
)LJXUH([DPSOHRI6$5LQWHUIHURJUDPXVLQJLPDJHVIURP(QYLVDW$6$5VHQVRUૃȀ PPZLWK DVPDOOSHUSHQGLFXODURUQRUPDOEDVHOLQH%QRIPDGHVFHQGLQJRUELWDQGDORRNDQJOHRIGHJ7KH DUHDRILQWHUHVWLVLQVRXWKHDVWHUQ,UDQZKHUHWKHWRZQRI%DPZDVGHYDVWDWHGRQ'HFHPEHUE\
DQHDUWKTXDNH7KHILUVWLPDJHZDVDFTXLUHGRQ'HFHPEHUDQGWKHVHFRQGRQ)HEUXDU\
GD\VRIWLPHLQWHUYDORUWHPSRUDOEDVHOLQH%W7KHUHVXOWVKLJKOLJKWWKHFRVHLVPLFGLVSODFHPHQWVILHOG RULJLQDWLQJIURPWKHHDUWKTXDNH)URP>)HUUHWWL@
,16$5&2+(5(1&($1''(&255(/$7,21
7KH SKDVH DFFXUDF\ LQ 6$5 LQWHUIHURPHWU\ LV PDLQO\ DIIHFWHG E\ SKDVH QRLVH DQG GHFRUUHODWLRQ 3KDVH GHFRUUHODWLRQ LV GXH WR FKDQJHV LQ SRVLWLRQ RI LQGLYLGXDO VFDWWHUV ZLWKLQWKHUHVROXWLRQFHOODQGLVRQHRIWKHPDLQOLPLWDWLRQVIRUVXFFHVVIXOXVHRI,Q6$5 'HFRUUHODWLRQ LV PDLQO\ GXH WR HLWKHU 6$5 LPDJLQJ JHRPHWULF HIIHFWV VSDWLDO GHFRUUHODWLRQRUWHPSRUDOEDFNVFDWWHULQJFKDQJHVWHPSRUDOGHFRUUHODWLRQ
6SDWLDOGHFRUUHODWLRQLVUHODWHGWRWKHVSDWLDOEDVHOLQHEHWZHHQWKHVHQVRUDWWKHGLIIHUHQW DFTXLVLWLRQV 7HPSRUDO GHFRUUHODWLRQ LV GXH WR FKDQJHV LQ JHRPHWULFDO RU HOHFWULFDO SURSHUWLHVRIWKHVXUIDFHDVIXQFWLRQRIWLPHEHWZHHQWKHDFTXLVLWLRQV7KHVHFKDQJHV PD\ EH FDXVHG HJ E\ PRYLQJ SDUWV RI YHJHWDWLRQ HURVLRQ RQ WKH ODQG VXUIDFH RU DJULFXOWXUDODFWLYLW\)RUHVWLVVKRZQWRKDYHDJHQHUDOO\ORZSKDVHFRUUHODWLRQHYHQIRU DVKRUWWHPSRUDOEDVHOLQHZKLOHXUEDQDQGDULGDUHDVVKRZXVXDOO\KLJKVLJQDOVWDELOLW\
14
HYHQ IRU DFTXLVLWLRQ WLPH LQWHUYDOV ORQJHU WKDQ RQH \HDU 7HUUDLQ FRQWDLQLQJ YDULDEOH OLTXLG ZDWHU VXFK DV HJ DUHDV FRYHUHG ZLWK ZHW VQRZ ZLOO DOVR KDYH GLIIHUHQW VFDWWHULQJSURSHUWLHVIURPRQHREVHUYDWLRQWRWKHQH[W7KXVWKHVFHQHVDFTXLUHGGXULQJ ZLQWHUVHDVRQFDQEHXQXVDEOHLIVQRZIDOORFFXUV7RRYHUFRPHWKLVOLPLWDWLRQDUWLILFLDO FRUQHU UHIOHFWRU FDQ EH LQVWDOOHG WR PRQLWRU DUHDV RI LQWHUHVW GXULQJ WKH ZKROH \HDU )LJXUH
7KHWHPSRUDOGHFRUUHODWLRQSKHQRPHQRQLVGHSHQGHQWRQWKHUDGDUZDYHOHQJWKORQJHU ZDYHOHQJWKVDUHOHVVVHQVLWLYHWRVPDOOVFDOHVXUIDFHVFDWWHULQJFKDQJHVKRZHYHUZLWK UHGXFHGVHQVLWLYLW\WRGLVSODFHPHQW
$TXDOLW\PHDVXUHRIWKHLQWHUIHURPHWULFSKDVHLVWKHFRPSOH[FRUUHODWLRQFRHIILFLHQWRU FRPSOH[FRKHUHQFH_Ȗ_7KHYDOXHVRI_Ȗ_DUHEHWZHHQDQGZKHUHDFRKHUHQFHYDOXH RIFRUUHVSRQGVWRSHUIHFWSKDVHFRUUHODWLRQEHWZHHQWKHWZRPHDVXUHPHQWV&RKHUHQFH YDOXHVOHVVWKDQXQLW\FRUUHVSRQGWRUHGXFHGSKDVHFRUUHODWLRQ)LJXUH
)LJXUH2YHUYLHZRILQVWDOODWLRQRIDFRUQHUUHIOHFWRULQ7URPV&RXQW\7KHUHIOHFWRUVSURYLGHVWDEOH VLJQDODQGDUHFRYHUHGWRSUHYHQWVQRZDFFXPXODWLRQ
Figure 11: Example of coherence map over Bergen. It clearly appears that the urban area has high coherence values (yellow-red), while the areas with vegetation or water are affected by low coherence
(blue).
15 ,16$5352&(66,1*$1'5(68/76
,Q6$5 SURFHVVLQJ LQYROYHV PDQ\ VWHSV ,UUHOHYDQW SKDVH FRPSRQHQWV KDYH WR EH UHPRYHG VHFWLRQ DQG ZUDSSHG LQWHUIHURJUDPV KDYH WR EH FRQYHUWHG LQWR XQZUDSSHG LQWHUIHURJUDPV UHSUHVHQWLQJ FRQWLQXRXV GLVWDQFH YDULDWLRQV VHFWLRQ 7R RYHUFRPH WKH PDLQ ,Q6$5 OLPLWDWLRQV GLIIHUHQW PXOWLLQWHUIHURJUDPV WHFKQLTXHV ZHUHGHYHORSHGLQWKHODVWGHFDGHVDQGWKHPDLQZLOOEHSUHVHQWHGLQVHFWLRQ7KH IRFXV LV RQ WKH PDLQ VWHSV DQG PHWKRGV LQFOXGHG LQ WKH *6$5 SURFHVVLQJ FKDLQ GHYHORSHGDW1RUXW>/DUVHQHWDO@
352&(66,1*2)3+$6(&20321(176
7KH FKDOOHQJH LQ ,Q6$5 LV WR VHSDUDWH WKH ZDQWHG VLJQDO HJ GHIRUPDWLRQ RU WRSRJUDSK\IURPWKHVXPRIDOOSKDVHFRQWULEXWLRQV)RUWHUUDLQGLVSODFHPHQWVWXGLHV WHPSRUDOGHFRUUHODWLRQFDQEHFRQVLGHUHGDVDUDQGRPQRLVHVRXUFHZKLOHHUURUVLQWKH '(0 XVHG WR UHPRYH WKH WRSRJUDSKLF SKDVH RUELWDO HUURUV DQG DWPRVSKHULF FKDQJHV ZLOOLQWURGXFHDVSDWLDOO\FRUUHODWHGSKDVHHUURUFRQWULEXWLRQ
7RLOOXVWUDWHWKH HIIHFWV RIVRPHSKDVHFRPSRQHQWV LWLV XVHIXOWR FRQVLGHUILUVW VRPH LGHDOFDVHVZKHUHQRLVHDQGDWPRVSKHULFHIIHFWVDUHQHJOLJLEOH,QFigure 12WKHFDVHD VKRZV DQ LQWHUIHURJUDP EXLOW ZLWK WZR DFTXLVLWLRQV DW GLIIHUHQW WLPHV EXW PDGH DW WKH H[DFWVDPHORFDWLRQSHUSHQGLFXODUEDVHOLQH DQGZLWKRXWWRSRJUDSKLFFRQWULEXWLRQ ,IQRGHIRUPDWLRQ RFFXUUHGEHWZHHQWKHWZRVFHQHVWKHSKDVHGLIIHUHQFH ZLOOEH]HUR RYHUWKHZKROHLQWHUIHURJUDP,QWKHFDVHEWKHSHUSHQGLFXODUEDVHOLQHLVVWLOO]HURDQG WKH WHUUDLQ IODW EXW D GHIRUPDWLRQ RFFXUUHG EHWZHHQ WKH WZR VFHQHV KLJKOLJKWLQJ D SDWWHUQ LQ WKH PLGGOH RI WKH LQWHUIHURJUDP ,Q WKH FDVH F WKH DUHD LV QRW DIIHFWHG E\
JURXQG GHIRUPDWLRQ EXW D GLIIHUHQFH EHWZHHQ WKH WZR RUELWV LV LQWURGXFHG QRQ]HUR SHUSHQGLFXODU EDVHOLQH 7KLV FUHDWHV D VXFFHVVLRQ RI IULQJHV LQ WKH UDQJH GLUHFWLRQ )LQDOO\LQWKHFDVHGWRSRJUDSKLFIHDWXUHVDUHDGGHGWRWKHVFHQDULRF7KLVDGGVDQ DGGLWLRQDO SKDVH FRPSRQHQW ZKLFK KDV WR EH UHPRYHG LQ FDVH RI WHUUDLQ GLVSODFHPHQW VWXGLHV
,QDGGLWLRQZHKDYHWRWDNHLQWRDFFRXQWWKHDWPRVSKHULFHIIHFWV$UDGDULQWHUIHURPHWHU PHDVXUHV WKH SKDVH GLIIHUHQFH ZLWK DFFXUDF\ RQ WKH RUGHU RI D IUDFWLRQ RI WKH ZDYHOHQJWK PRUH WKDQ DFFXUDWH HQRXJK WR EH LQIOXHQFHG E\ DWPRVSKHULF SDWK GHOD\
3KDVHSURSDJDWLRQGHOD\GXHWRDWPRVSKHULFYDULDELOLW\LVRQHRIWKHPDLQHUURUVRXUFHV LQ UHSHDWSDVV ,Q6$5 ,W LV FRPPRQ WR GLYLGH WKH DWPRVSKHULF SDWK GHOD\ LQWR RQH FRPSRQHQW FRPLQJ IURP WXUEXOHQW PL[LQJ SURFHVVHV DQG D VWUDWLILHG FRPSRQHQW FRUUHODWLQJ ZLWK HOHYDWLRQ 7XUEXOHQW PL[LQJ FRPHV IURP PL[LQJ SURFHVVHV LQ WKH LQKRPRJHQHRXV DWPRVSKHUH ZKLOH VWUDWLILFDWLRQ UHVXOWV IURP YDULDWLRQV LQ WKH YHUWLFDO UHIUDFWLYHLQGH[SURILOH7KHVHFRQGLVFRUUHODWHGZLWKWKHORFDOWRSRJUDSK\)LJXUH VKRZVKRZDWPRVSKHULFFRPSRQHQWVDVZHOODVDSKDVHQRLVHFRQWULEXWLRQFDQDIIHFWDQ LQWHUIHURJUDP 7KH DWPRVSKHULF HIIHFWV DUH WHPSRUDOO\ XQFRUUHODWHG EXW VSDWLDOO\
FRUUHODWHGFRQWUDU\WRQRLVHFRPSRQHQWV,Q)LJXUHGDQLQWHUIHURJUDPFRPELQLQJ
16
WKHGLIIHUHQWSKDVHFRPSRQHQWVLVSUHVHQWHG,WLQFOXGHVJURXQGGHIRUPDWLRQFRPSRQHQW IURP Figure 12 E EDVHOLQHUHODWHG FRPSRQHQW IURPFigure 12 F WRSRJUDSKLF FRPSRQHQWIURPFigure 12GDWPRVSKHULFHIIHFWVIURP)LJXUHDDQGEDQGQRLVH IURP)LJXUHF
)LJXUH([DPSOHVRI,Q6$5FRQILJXUDWLRQVDQGUHVXOWLQJLQWHUIHURJUDPVD6DPHDFTXLVLWLRQ JHRPHWU\IODWDUHDQRDWPRVSKHULFHIIHFWVDQGQRJURXQGGHIRUPDWLRQE6DPHDFTXLVLWLRQJHRPHWU\
IODWDUHDQRDWPRVSKHULFHIIHFWVEXWJURXQGGHIRUPDWLRQF)ODWDUHDQRDWPRVSKHULFHIIHFWVDQGQR JURXQGGHIRUPDWLRQEXWGLIIHUHQWDFTXLVLWLRQJHRPHWULHVSHUSHQGLFXODUEDVHOLQHG1RDWPRVSKHULF
HIIHFWVDQGQRJURXQGGHIRUPDWLRQEXWWRSRJUDSKLFDQGEDVHOLQHUHODWHGFRPSRQHQWV)URP>)HUUHWWL
@
17
)LJXUHD([DPSOHRIDWPRVSKHULFWXUEXOHQFHHIIHFWVRQD6$5LQWHUIHURJUDPE([DPSOHRI VWUDWLILHGFRPSRQHQWUHODWHGWRWKHVDPHWRSRJUDSK\DV)LJXUHGF([DPSOHRIQRLVHFRQWULEXWLRQ
DIIHFWLQJDQLQWHUIHURJUDPG)LQDOH[DPSOHFRPELQLQJWKHGLIIHUHQWSKDVHFRQWULEXWLRQVSUHVHQWHGLQ )LJXUHEG )LJXUHDF)URP>)HUUHWWL@
81:5$33,1*352&('85(
$ ZUDSSHG LQWHUIHURJUDP LV FRPSRVHG E\ D VXFFHVVLRQ RI IULQJHV ZKHQ WKH SKDVH H[FHHGVKDOIWKHZDYHOHQJWK7KHSURFHVVRIUHVWRULQJWKHFRUUHFWPXOWLSOHRIʌWRHDFK SRLQWRIWKHLQWHUIHURPHWULFSKDVHLPDJHRUZLWKHDVLHUZRUGVWo convert cyclic phase difference into continuous phase differenceLVFDOOHGSKDVHXQZUDSSLQJ)LJXUHWRS 7KH SURFHGXUH XVHV WKH DVVXPSWLRQ WKDW WKH WUXH GLVSODFHPHQWV ILHOG KDV D VSDWLDO FRQWLQXLW\DQGWKXVWKHQHLJKERXULQJSL[HOVDUHVXSSRVHGWREHZLWKLQDSKDVHF\FOHRI WKHRWKHUV8VLQJDUHIHUHQFHSRLQWZLWKKLJKFRKHUHQFHDQGLISRVVLEOHNQRZQDVVWDEOH RU ZLWK D NQRZQ GHIRUPDWLRQ UDWH WKH SKDVH GLIIHUHQFHV DUH WKXV VSDWLDOO\ LQWHJUDWHG XVLQJGLIIHUHQWPHWKRGVDFFRUGLQJWRWKHFKRVHQDOJRULWKP
7KLVKLJKOLJKWVWKUHHSRWHQWLDOLVVXHV
x ,Q6$5GDWDDUHQRWDEVROXWHEXWUHODWLYHWRDUHIHUHQFHSRLQWWKDWKDVWREHFKRVHQ FDUHIXOO\ ,I WKH UHIHUHQFH SRLQW VXSSRVHG WR EH VWDEOH LV LQ UHDOLW\ PRYLQJ WKH ZKROHGDWDVHWZLOOEHDIIHFWHGE\DVKLIWRIYDOXHV
x ,IZHXVHLQWHUIHURJUDPVIRUPHGXVLQJODUJHWHPSRUDORUVSDWLDOEDVHOLQHVDUHDVFDQ EH GHFRUUHODWHG GXH WR FKDQJHV LQ VFDWWHULQJ SURSHUWLHV ZLWKLQ WKH UHVROXWLRQ FHOO EHWZHHQ WKH WZR DFTXLVLWLRQV 6XFK GHFRUUHODWLRQ HIIHFWV FDQ FRQWDPLQDWH ODUJH DUHDVLQWKHLQWHUIHURJUDPVDQGFUHDWHGLVFRQWLQXRXVFRKHUHQWSDWFKHV7KLVPDNHV WKHUHWULHYDORIDEVROXWHSKDVHDYHU\FKDOOHQJLQJWDVN
x ,IWKHPRYHPHQWVDUHVSDWLDOO\GLVFRQWLQXRXVIRUH[DPSOHLQWKHFDVHRIORFDOL]HG TXLFNHYHQWWKHSURFHGXUHFDQIDLOWRUHWULHYHFRUUHFWVROXWLRQV
18
7KXVWKLVVWHSLVRQHRIWKHPRVWFKDOOHQJLQJDVSHFWVIRUVXFFHVVIXODSSOLFDWLRQVRI6$5 LQWHUIHURPHWU\
,Q )LJXUH ERWWRP DQ H[DPSOH RI ZUDSSHG DQG FRUUHVSRQGLQJ XQZUDSSHG LQWHUIHURJUDP LV SUHVHQWHG7KHSDUWVDIIHFWHGE\GHFRUUHODWLRQHIIHFWVDUHDVXQGHUDQ DSSOLHGFRKHUHQFHWKUHVKROGLQZKLWHRQGDUHLQWKLVFDVHVPDOO
)LJXUH7RS3ULQFLSOHRIWKHXQZUDSSLQJSURFHGXUHD5HSUHVHQWDWLRQRIDVOLFHWKURXJKVRPH IULQJHVRQDQLQWHUIHURJUDP7KHUHVXOWVDUHZUDSSHGDURXQGʌE7KHDEVROXWHSKDVHLVIRXQGDIWHUDQ
XQZUDSSLQJSURFHGXUH)URP>:RRGKRXVH@%RWWRP([DPSOHRIZUDSSHGLQWHUIHURJUDPRYHUDQ RSHQSLWPLQHLQ86$FDQGWKHFRUUHVSRQGLQJXQZUDSSHGLQWHUIHURJUDPG)URP>)HUUHWWL@
08/7,,17(5)(52*5$07(&+1,48(6
0XOWLLQWHUIHURJUDPV WHFKQLTXHV ZHUH GHYHORSHG RYHU WKH ODVW GHFDGHV LQ RUGHU WR RYHUFRPH PRVW RI WKH OLPLWDWLRQV RI WKH VWDQGDUG VLQJOH LQWHUIHURJUDP DSSURDFK DV WHPSRUDO DQG JHRPHWULF GHFRUUHODWLRQV DWPRVSKHULF HIIHFWV HWF 7KH\ DUH JHQHUDOO\
GLYLGHGLQWRWZRPDLQJURXSV
x 0HWKRGV EDVHG RQ ORFDWLQJ3HUVLVWHQW 6FDWWHUHUV 36V UHIHUUHG WR DV3HUVLVWHQW 6FDWWHUHU ,QWHUIHURPHWU\36,PHWKRGV >)HUUHWWL HW DO )HUUHWWL HW DO Colesanti et al,. 2003; :HUQHUHWDO+RRSHUHWDO.DPSHV3UDWL HWDO@
x 0HWKRGVEDVHGRQVSDWLDOFRUUHODWLRQDQGGLVWULEXWHGVFDWWHULQJ'6UHIHUUHGWRDV 6PDOO %$VHOLQH 6XEVHW 6%$6 PHWKRGV>/XQGJUHQ HW DO %HUDUGLQR HW DO
19
0RUDHWDO6FKPLGWDQG%UJPDQQ/DQDULHWDO3UDWLHW DO6DQVRVWLHWDO/DXNQHVHWDO@
7KHWZRPHWKRGVUHIHUWRWKHGLIIHUHQWVFDWWHULQJPHFKDQLVPVWKDWFDQEHH[SHFWHGIURP DSL[HOXQGHUVWXG\)LJXUH$FRKHUHQWSRLQWOLNHWDUJHWLVDVFDWWHUHUWKDWGRPLQDWHV WKHVFDWWHULQJIURPWKHUHVROXWLRQFHOO'XHWRWKHEURDGVFDWWHULQJOREHSRLQWWDUJHWVDUH YLVLEOHIURPDODUJHVSDQRIYLHZLQJDQJOHV$3HUVLVWHQW6FDWWHUHU36LVDWDUJHWZKRVH DPSOLWXGH DQG SKDVH LV TXLWH FRQVWDQW DV D IXQFWLRQ RI WLPH ,I D SL[HO LQFOXGHV RQH VFDWWHUHU WKDW LV EULJKWHU WKDQ WKH RWKHU RQHV WKH SL[HO SKDVH YDULDQFH LV VLJQLILFDQWO\
UHGXFHGZLWKUHVSHFWWRWKHFDVHRIGLVWULEXWHGVFDWWHULQJIURPDKLJKQXPEHURISRLQW WDUJHWV)LJXUHWRS$36FDQEHWKHFRUQHURIDEXLOGLQJRUDURFNIRUPDWLRQLQUXUDO DUHDV )RU D GLVWULEXWHG VFDWWHULQJ PHFKDQLVP WKH PHDVXUHG VFDWWHULQJ LV WKH FRKHUHQW VXPPDWLRQRIDOOLQGLYLGXDOVPDOOVFDWWHUHUVZLWKLQWKHUHVROXWLRQFHOO
)LJXUH'LIIHUHQWVXUIDFHVFDWWHULQJPHFKDQLVPV7RSDSL[HOZLWKGLVWULEXWHGVFDWWHULQJ'6 PHFKDQLVPWRSOHIWDGRPLQDWLQJSRLQWVFDWWHUHUWRSPLGGOHDQGDSL[HOZLWKRQHVFDWWHUHUWKDWLV
EULJKWHUWKDQDOORWKHUWRSULJKW7KHSKDVHGLVSHUVLRQIRUWKH36SL[HOVLVPXFKVPDOOHUWKDQIRU GLVWULEXWHGVFDWWHULQJSL[HOV)URP>/DXNQHV)HUUHWWL@
20
3(56,67(176&$77(5(5,17(5)(520(75<36,
PSI methods uses a stack of interferograms generated with the same master scene (Figure 16, left) to analyze the temporal phase evolution of individual coherent point scatterers (Figure 16, ULJKW >)HUUHWWL HW DO )HUUHWWL HW DO @ $OO LQWHUIHURJUDPVDUHFRPSXWHGDWIXOOUHVROXWLRQDQGDOOLQWHUIHURPHWULFFRPELQDWLRQVDUH HPSOR\HGHYHQWKRVHH[FHHGLQJWKHFULWLFDOEDVHOLQH7KHJHQHUDWHGLQWHUIHURJUDPVDUH RIWHQ VHYHUHO\ GHFRUUHODWHG +RZHYHU E\ VHDUFKLQJ IRU SL[HOV WKDW LQFOXGH D VWDEOH VFDWWHUHU RYHU ORQJ WLPH LQWHUYDOV D QHWZRUN RI VR FDOOHG 3HUVLVWHQW 6FDWWHUHUV FDQ EH HVWDEOLVKHGZKHUHGLVSODFHPHQWVLJQDODQG'VFDWWHUHUSRVLWLRQFDQEHUHVROYHGThe PS can be selected looking at amplitude dispersion below a defined threshold.6XFKD VWDWLVWLFDOPHDVXUHUHTXLUHVDODUJHVWDFNRILPDJHV
)LJXUH/HIW36EDVHOLQHSORWSHUSHQGLFXODUEDVHOLQHYVDFTXLVLWLRQWLPH7KHPDVWHUVFHQHLV
(DFKOLQHFRUUHVSRQGVWRDQLQWHUIHURJUDPEHWZHHQWZRDFTXLVLWLRQGDWHV)URP>/DXNQHV@
5LJKW36,DQDO\VLVRIDVWDFNRIPXOWLWHPSRUDO6$5DFTXLVLWLRQVXVLQJDVSDUVH36JULGUDWKHUWKDQWKH DQDO\VLVRIDOOLPDJHSL[HOV)URP>)HUUHWWL@
PSI procedure includes two networks of Persistent Scatterers. The first PS network is built using a sparse grid of very highly stable targets (low amplitude dispersion) in order to remove irrelevant components such as atmospheric effects. Using a large stack of interferograms (Figure 16, b) and based on the assumption that aWPRVSKHULFFRPSRQHQW LVVSDWLDOO\FRUUHODWHGEXWXQFRUUHODWHGLQWLPHWKH$36$WPRVSKHULF3KDVH6FUHHQLV HVWLPDWHGLQWHUSRODWHGDQGVXEWUDFWHGIURPHDFKLQWHUIHURJUDP7KHVHFRQG36QHWZRUN LV EXLOW DSSO\LQJ D QHZ DPSOLWXGH GLVSHUVLRQ WKUHVKROG KLJKHU )RU HDFK VHOHFWHG 36 PHDQYHORFLW\DQGWLPHVHULHVDUHFDOFXODWHGFigure 16, c
21
60$//%$6(/,1(68%6(70(7+2'6%$6
Similar to PSI, SBAS method uses a stack of interferograms to remove or reduce the effects of irrelevant phase components. However, it incorporates a larger number of interferograms, not from a unique master scene as for PSI (Figure 17, left) [Berardino, et al. 2002]. Spatial and temporal baselines below chosen thresholds are applied in order to reduce geometric and temporal decorrelation, as well as the effects due to errors in the used DEM. Moreover, a complex multilooking procedure is applied at the beginning of the processing chain. This reduces the spatial resolution, but improves the signal quality of pixels characterized by GLVWULEXWHGVFDWWHULQJ (Figure 17, right).
)LJXUH/HIW6%$6EDVHOLQHSORWSHUSHQGLFXODUEDVHOLQHYVDFTXLVLWLRQWLPH(DFKOLQHFRUUHVSRQGV WRDQLQWHUIHURJUDPEHWZHHQWZRDFTXLVLWLRQGDWHV>/DXNQHV@5LJKW(IIHFWRIFRPSOH[PXOWL ORRNLQJ1/ VTXDUH[ZLQGRZRQZUDSSHGSKDVHYDOXHVZLWKGLIIHUHQWFRKHUHQFHOHYHOV
0RGLILHGIURP>)HUUHWWL@
,QRUGHUWRSURYLGHDPHDQYHORFLW\PDSZLWKUHGXFHGDWPRVSKHULFQRLVHVKRUWEDVHOLQH LQWHUIHURJUDPV FDQ EH VLPSO\ DYHUDJHG DQG ZHLJKWHG E\ WKH WHPSRUDO GLIIHUHQFHV EHWZHHQWKHVFHQHV7KLVUHIHUVWRWKHVWDFNLQJPHWKRG>*DEULHOHWDO6DQGZHOO DQG3ULFH3HOW]HUHWDO@6%$6PRGHOVZHUHODWHUGHYHORSHGLQRUGHUWRWDNH LQWRDFFRXQWWKHWHPSRUDOHYROXWLRQRIWKHGHIRUPDWLRQDQGSURYLGHFRPSOHWHWLPHVHULHV
>%HUDUGLQRHWDO3HSHHWDO/DXNQHVHWDO@
,16$52873876
Different InSAR results can be produced. The most common output is a deformation map over the area in interest. The map can display the annual mean velocity as in Figure 18 or specific displacements between two defined times when mean velocity value does not make sense (by example displacements between June and September 2012 if the stack uses only snow-free scenes in a mountainous region). As discussed above, according to the processing method, each element of the map will represent a selected
22
Persistent Scatterer below a defined amplitude dispersion threshold (for PSI, as in Figure 18, top) or a multilooked pixel above a defined coherence threshold (for SBAS, as in Figure 18, bottom) for the area visible from the satellite (without shadow) and unaffected from geometrical distortions (without layover). For each selected PS or SBAS pixel, time series usually displayed on a displacement/time graph can be extracted (Figure 18, bottom). These two outputs allow providing complementary information about the spatial distribution of the movement and its temporal evolution.
In addition, derived products such as coherence map, intensity map, shadow mask, layover mask can be generated.
)LJXUH7RS([DPSOHRI,Q6$5PHDQYHORFLW\PDSRYHU%M¡UYLND2VORIURP36,SURFHVVLQJXVLQJ 7HUUD6$5;GDWDVXSSOLHGE\$LUEXV'HIHQVH 6SDFH,QIRWHUUD*PE+%RWWRP([DPSOHRI,Q6$5
PHDQYHORFLW\PDSRYHU1RUGQHVURFNVOLGH7URPV&RXQW\1RUZD\IURP6%$6SURFHVVLQJXVLQJ 5$'$56$7LPDJHVDQG,Q6$5WLPHVHULHZKLWHFLUFOHVLQWKHJUDSKFRPSDUHGWR'*36FRQWLQXRXV
PHDVXUHPHQWVEOXHGRWVLQWKHJUDSK
23
It should be reminded that the results are relative to a reference point and the satellite is only able to measure 1D displacement along its LOS. However, having available data from different acquisition geometries over the same area, it is possible to retrieve multi- dimensional information. In practice, using satellites data only, due to the near-polar passes of all the SAR sensors, two geometries (ascending/descending) are available, allowing only a 2D decomposition (vertical/horizontal components in the East-West LOS-plane). An example over Nordnes rockslide (Troms region) is SUHVHQWHGLQ)LJXUH
)LJXUH7RS3ULQFLSOHRI'GHFRPSRVLWLRQXVLQJDVFHQGLQJDQGGHVFHQGLQJJHRPHWULHVRYHU1RUGQHV URFNVOLGH%RWWRP([DPSOHVRIPHDQYHORFLW\PDSVDORQJWKH/26XVLQJGDWDDSSUR[LPDWHVDWHOOLWH SDWKVGLVSOD\ZLWKGDVKHGZKLWHDUURZVDQGDIWHUKRUL]RQWDOYHUWLFDOGHFRPSRVLWLRQ)URP>(ULNVHQ
@76;GDWDXVHG'/5
24
678'<$5($
After discussion between the partners of the projects, it appeared that Bergen area including the town and various sectors located within 100 km toward the East is an interesting case-example to illustrate the potential of InSAR technology for ground deformation detection. In section 3.1, the natural hazards context of the region is introduced, as well as the main sectors of interest highlighted by 1RUJHVYDVVGUDJVRJ HQHUJLGLUHNWRUDW19(6WDWHQVYHJYHVHQDQG-HUQEDQHYHUNHW,QVHFWLRQWKHPDLQ FKDUDFWHULVWLFVRIWKHDYDLODEOHVDWHOOLWHGDWDDUHSUHVHQWHG
+$=$5'6&217(;7
(9(176,19(1725<
Bergen area is interesting for its natural hazards context. Looking at the inventory of the historical events, it clearly appears that the natural hazards are a significant issue for the management of infrastructure and the safety of population in this region ()LJXUH).
)LJXUH([WUDFWRIWKHLQYHQWRU\RIQDWXUDOKD]DUGVHYHQWVLQ1RUZD\URFNIDOOVJUD\ODQGVOLGHV EURZQGHEULVIORZV\HOORZURDGVGHWHULRUDWLRQEODFNDYDODQFKHVZKLWHLFHIDOOVGDUNEOXH
XQVSHFLILHGWXUTXRLVH)URPZZZVNUHGQHWWQRGRZQORDG
25
Unsurprisingly the main clusters of registered events are located along the main infrastructure in areas characterized by significant relief. The density is especially high along the E16 road from Arna to Evanger via Dale (top-right part of the )LJXUH), as well as around the NE part of Samnagerfjorden (Årland, Haga) and over Kvamskogen (middle of the )LJXUH). It appears that most of the events are classified as rockfalls, landslides or avalanches.
)2&862)7+(678'<
The project description highlighted three main areas covered by the available datasets and interesting because of the density of events and the importance of infrastructure (roads and rail).
They are located within an approximate 50 km distance around Bergen:
x E16 and rail Arna-Dale, eventually to Evanger (and also on the other side of the fjord, along the south part of Osterøy);
x Fv7 over Kvamsskogen, especially around Norheimsund (Tokagjelet);
x The area at the head of Samnangerfjorden.
These are covered at least by one of the available satellite dataset. However, due to several limitations regarding to the characteristics of the datasets (see section 3.2) and the region (see sections 4.1 & 4.2), the main focus of the ground deformation analysis will be placed on the western part of the region including Bergen and Indre Arna (see sections 4.3.1 & 4.3.2).
26
6$7(//,7('$7$&+$5$&7(5,67,&6
7KHUHVXOWVSUHVHQWHGLQ&KDSWHUDUHEDVHGRQWKHDQDO\VLVRIWKUHHGLIIHUHQWGDWDVHWV IURP WZR VHQVRUV 5$'$56$7 56$7 DQG 7HUUD6$5;7DQ'(0;
76;7'; 7KH UHVXOWV XVLQJ WKH VHFRQG VHQVRU DUH IURP >&HWLQLF HW DO LQ SUHS@ 7KH PDLQ GLIIHUHQFHV EHWZHHQ WKH WZR VDWHOOLWHV FDQ EH IRXQG LQ 7DEOH DQG 7DEOH VHFWLRQ 7KH PRVW LPSRUWDQW HOHPHQWV WR WDNH LQWR DFFRXQW DUH WKH IUHTXHQF\EDQGDQGFRUUHVSRQGLQJZDYHOHQJWK&Ȝ FPIRU56$7YV;Ȝ FPIRU76;7';DQGWKHUHYLVLWWLPHGD\VIRU56$7YVGD\VIRU76;7';
,Q DGGLWLRQ WKH DYDLODEOH VHWV LQFOXGH GLIIHUHQW DPRXQW RI VFHQHV DFTXLUHG XVLQJ GLIIHUHQWPRGHV
7KHHOHPHQWVUHODWLYHWRWKHGLIIHUHQWWHPSRUDO VSDWLDOFRYHUDJHVDQGUHVROXWLRQVDUH SUHVHQWHG WKHUHDIWHU LQ VHFWLRQV DQG 7KH HOHPHQWV UHODWLYH WR WKH GLIIHUHQW OLQHRIVLJKWDQGWKHLULPSDFWRQJHRPHWULFDOGLVWRUWLRQVDUHGLVFXVVHGLQVHFWLRQ 7DEOHV KLJKOLJKWLQJ WKH DFTXLVLWLRQ WLPH VSDWLDO DQG WHPSRUDOEDVHOLQHV RI HDFK VFHQH DUHDYDLODEOHLQ$SSHQGL[HV7DEOH7DEOH 7DEOH5HPHPEHUWKDWWKHWKHRU\
DERXW WKH UHODWLRQVKLSV EHWZHHQ WKHVH GLIIHUHQW SDUDPHWHUV DQG WKH DGYDQWDJHV DQG OLPLWDWLRQVRI,Q6$5UHVXOWVZHUHSUHVHQWHGLQ&KDSWHUDQGVXPPDUL]HGLQ7DEOH
7(0325$/&29(5$*(6 5(62/87,216
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
7KH UHSHDWSDVV LQWHUYDO DQG WKH VHQVRU ZDYHOHQJWK KDYH DQ LPSDFW RQ WKH GHWHFWLRQ FDSDELOLW\ GHSHQGLQJ RQ WKH VSHHG RI WKH PRYLQJ DUHD VHH VHFWLRQV 0RUHRYHU WKH UHVXOWV KDYH QRW WKH VDPH WLPH UHIHUHQFHV GXH WR WKH GLIIHUHQW WHPSRUDO FRYHUDJHV7RFRPSDUHWKHUHVXOWVWKHPDSVDUHH[SUHVVHGIRUWKHWKUHHGDWDVHWVLQWHUP RI PHDQ DQQXDO YHORFLW\ EXW LW KDV WR EH NHSW LQ PLQG WKDW WKH PHDQ YDOXHV DUH QRW FDOFXODWHGRQWKHVDPHWHPSRUDOEDVLV
27
7DEOH7(0325$/&29(5$*(6$1'5(62/87,216 6DWHOOLWH %DQG
ZDYHOHQJWK 5HSHDWSDVV
LQWHUYDO 'DWDVHWPRGH 1XPEHURI
LPDJHV 7LPHSHULRG
56$7 &Ȝ FP GD\V 6WDQGDUG
8OWUDILQH
76;7'; ;Ȝ FP GD\V 6WULS0DS
)URP>&HWLQLFHWDOLQSUHS@
)LJXUH%DVHOLQHSORWSHUSHQGLFXODUEDVHOLQHYVDFTXLVLWLRQWLPHIRUWKHGDWDVHWVKLJKOLJKWLQJWKH WHPSRUDOYDULDELOLW\RIWKHGDWDVHWVEOXHFLUFOHV56$76WDQGDUGPRGHUHGVTXDUHV56$7 8OWUDILQHPRGHJUHHQGLDPRQGV76;7';6WULS0DSPRGH>&HWLQLFHWDOLQSUHS@%DVHGRQ
LQIRUPDWLRQDYDLODEOHLQ7DEOH7DEOH 7DEOHLQ$SSHQGL[HV
63$7,$/&29(5$*(6 5(62/87,216
(DFK GDWDVHW KDV D GLIIHUHQW FRYHUDJH RI WKH UHJLRQ RI LQWHUHVW DQG GLIIHUHQW VSDWLDO UHVROXWLRQV 7DEOH )LJXUH 56$7 6WDQGDUG PRGH KDV D PHGLXP UHVROXWLRQ UDQJH[D]LPXWKP[PEXWWKHVFHQHVFRYHUDODUJHDUHDRINP[NP WR 1RUKHLPVXQG %ROVWDG¡\UL DQG 1DPGDO DW WKHLU HDVW HGJH )RU 56$7 8OWUDILQH PRGHWKHUHVROXWLRQLVFOHDUO\EHWWHUUDQJH[D]LPXWKP[PEXWWKHVFHQHV FRYHU RQO\ D VPDOO DUHD LQFOXGLQJ %HUJHQ NP [ NP )RU 76;7'; 6WULS0DS PRGHWKHUHVROXWLRQLVHYHQEHWWHUWKDQ56$78OWUDILQHPRGHUDQJH[D]LPXWK P[PDQGWKHVL]HRIWKHVFHQHVNP[NPDOORZVDFRYHUDJHWR9DNVGDORJ +DJD DW WKHLU HDVW HGJH ,W VKRXOG EH WKXV QRWHG WKDWin the eastern part of the region, only RSAT-2 Standard mode is available. Because of its low resolution, it makes impossible the detection of small moving areas.
28
$VGLVFXVVHGLQVHFWLRQGHSHQGLQJRQWKHSURFHVVLQJPHWKRGVWKHILQDOUHVROXWLRQ FDQ FKDQJH 8VLQJ 3HUVLVWHQW 6FDWWHUHUV ,QWHUIHURPHWU\ 36, WKH UHVXOWV DUH LQ IXOO UHVROXWLRQ UG FROXPQ LQ 7DEOH 8VLQJ 6PDOO %$VHOLQH VXEVHW 6%$6 WKH UHVXOWV KDYHWREHPXOWLORRNHGWKFROXPQLQ7DEOHDQGWKHUHVROXWLRQLVWKXVUHGXFHGWK FROXPQLQ7DEOH
7DEOH63$7,$/&29(5$*(6$1'5(62/87,216 6DWHOOLWH 'DWDVHW
PRGH )XOOUHVRO
UDD] 6%$6PXOWLORRN
UDD] 6%$6UHVRO
UDD] 6FHQHVL]H UDD]
56$7 6WDQGDUG PP PP NPNP
8OWUDILQH PP PP NPNP
76;7'; 6WULS0DS PP PP NPNP
)URP>&HWLQLFHWDOLQSUHS@
)LJXUH6SDWLDOFRYHUDJHDQGVXPPDU\RIPDLQFKDUDFWHULVWLFVRIWKHWKUHHVDWHOOLWHGDWDVHWVLOOXVWUDWHG KHUHXVLQJWKHPHDQLQWHQVLW\LPDJHVRIWKHDUHDVFRYHUHGIRUHDFKGDWDVHW7KHDVFHQGLQJVDWHOOLWHVDUH IO\LQJSDUDOOHOWRWKHOHIWHGJHRIWKHLPDJHVWKHGHVFHQGLQJVDWHOOLWHLVIO\LQJSDUDOOHOWRWKHULJKWHGJHRI
WKHLPDJH7KHDUURZVUHSUHVHQWWKH/26RULHQWDWLRQ76;7';>&HWLQLFHWDOLQSUHS@
7KHGLIIHUHQWVSDWLDOFRYHUDJHVDQGUHVROXWLRQVKDYHREYLRXVO\DQLPSDFWRQWKHVL]HRI WKHSRWHQWLDODQDO\]HGDUHDDVZHOODVRQWKHGHWHFWLRQFDSDELOLW\GHSHQGLQJRQWKHVL]H RIWKHPRYLQJDUHD,QRUGHUWRLOOXVWUDWHWKHHIIHFWVRIWKHGLIIHUHQWVSDWLDOUHVROXWLRQVRI
29
WKHWKUHHGDWDVHWVRQWKHJURXQGIHDWXUHVGHWHFWLRQDFRPSDULVRQRIWKHPHDQLQWHQVLW\
LPDJHVRYHU%HUJHQLVDYDLODEOHLQ)LJXUH
)LJXUH(IIHFWRIWKH6$5UHVROXWLRQRQWKHTXDOLW\RIJURXQGIHDWXUHVGLVSOD\)URPWRSOHIWWRERWWRP ULJKWRUWKRSKRWR PHDQLQWHQVLW\LPDJHVIRU56$76WDQGDUGPRGH56$78OWUDILQHPRGH
76;7';6WULS0DSPRGH>&HWLQLFHWDOLQSUHS@
/,1(2)6,*+7 *(20(75,&$/())(&76
(DFKGDWDVHWKDVDGLIIHUHQWPHDVXUHPHQWGLUHFWLRQOLQHRIVLJKW/26DFFRUGLQJWRWKH WUDFNRULHQWDWLRQDQGWKHLQFLGHQFHDQJOHRIWKH6$5JHRPHWU\7DEOH )LJXUH )RU56$76WDQGDUGPRGHWKHVFHQHVRIWKHGDWDVHWZHUHDFTXLUHGXVLQJDQDVFHQGLQJ JHRPHWU\VDWHOOLWHIO\LQJIURPWKH66(WRWKH11:ZLWKD1RULHQWDWLRQ7KH /26 LV WKXV WRZDUG (1( ZLWK D TXLWH VWHHS YLHZ LQFLGHQFH DQJOH ș )RU 56$78OWUDILQHPRGHDGHVFHQGLQJJHRPHWU\ZDVXVHGVDWHOOLWHIO\LQJIURPWKH1 1(WRWKH66:ZLWKD1RULHQWDWLRQ7KH/26LVWKXVWRZDUG:1:ZLWKD YLHZ PRUH REOLTXH WKDQ IRU 56$7 VWDQGDUG PRGH LQFLGHQFH DQJOH ș )RU 76;7';6WULS0DSPRGHWKHGDWDZHUHDFTXLUHGXVLQJDQDVFHQGLQJJHRPHWU\ZLWK DQ RULHQWDWLRQ VOLJKWO\ PRUH WRZDUG WKH 1 WKDQ 56$7 6WDQGDUG PRGH 1
30
RULHQWDWLRQ7KH/26LVDOVRWRZDUG:1:EXWZLWKDYLHZFOHDUO\PRUHREOLTXHWKDQ 56$7LQFLGHQFHDQJOHș
7KH OLQHRIVLJKW /26 KDV DQ LPSDFW RQ WKH Getection capability. Indeed, due to the ability of the satellite to measure only 1D displacement along the LOS, any moving area with a direction perpendicular to the LOS will be missed, and more the directions of the displacements diverge than the LOS, more they are underestimated.
7DEOH6$5*(20(75,(6$1'/,1(2)6,*+7/2625,(17$7,216 6DWHOOLWH 'DWDVHWPRGH 2UELW ,QFLGHQFHDQJOH 7UDFNDQJOH /26RULHQWDWLRQ
56$7 6WDQGDUG $VFHQGLQJ
8OWUDILQH 'HVFHQGLQJ
76;7';
6WULS0DS $VFHQGLQJ
)URP>&HWLQLFHWDOLQSUHS@
)LJXUH6FKHPDWLFUHSUHVHQWDWLRQRIWKHVDWHOOLWHVWUDFNVDQGOLQHRIVLJKWRYHU%HUJHQ
The LOS has also an impact on the geometrical distortions according to the local topography (see section 2.1.4). In )LJXUH, the Digital Elevation Model (DEM) of the area is presented as well as shaded reliefs generated using the information about the SAR track and its line-of-sight to define the orientation of the illumination source. It appears that with an ascending geometry, the west-facing relief with a slope perpendicular to the LOS will be affected by layover effects, while the east-facing relief
31
can be affected by shadow (and the opposite for descending geometry). The areas affected by layover and shadow thus vary according to the datasets and require the generation of masks. They will be presented and compared in section 4.1.
)LJXUH'LJLWDO(OHYDWLRQ0RGHODQGVKDGHGUHOLHIJHQHUDWHGFRQVLGHULQJWKHUHVSHFWLYH/26RI 56$76WDQGDUGPRGH56$78OWUDILQHPRGHDQG76;7';6WULS0DSPRGH>&HWLQLFHWDOLQ
SUHS@DVLOOXPLQDWLRQVRXUFHFRQVLGHUHGWREHDWLQILQLW\%\GHIDXOWKLOOVKDGHVDUHH[SUHVVHGDVLQWHJHUV EHWZHHQVKDGRZDQGEHDPSHUSHQGLFXODUWRWKHORFDOVORSH7KHFDOFXODWLRQPHWKRGLVDYDLODEOH
LQ>%XUURXJK 0F'RQHOO@
32
,16$55(68/76%(5*(1$5($
This Chapter presents results from InSAR analysis using the case study of Bergen area.
As briefly described in section 2.3.4, the most interesting results concern usually the ground deformation and can be expressed as deformation maps or time series (section 4.3). But other derived produces can also be generated. They are mainly related to geometrical effects and temporal decorrelation. Because they are valuable to better understand advantages and limitations of InSAR technology, they will be presented in sections 4.1 and 4.2.
6+$'2: /$<29(50$6.6
Shadow and layover masks can be generated depending on the line-of-sight and the local topography. The areas affected by these geometrical effects will be removed of the final analysis.
Over Bergen area, the datasets are overall not highly affected by shadow due to the relatively vertical views of the satellite, but the layover effect has a high impact.
Depending on the different lines-of-sight, layover varies significantly according to the datasets. Due to its small incidence angle, the RADARSAT-2 (RSAT-2) Standard mode dataset is highly affected by layover on east facing slopes, but little by shadow. For RSAT-2 Ultrafine mode (descending geometry), the layover effect occurs RQ ZHVW IDFLQJ VORSH 7KH 7HUUD6$5;7DQ'(0; 76;7'; 6WULS0DS PRGH GDWDVHW DVFHQGLQJ JHRPHWU\ >&HWLQLF HW DO LQ SUHS@ LV OHVV DIIHFWHG E\ OD\RYHU DQG PRUHE\VKDGRZWKDQ56$76WDQGDUGPRGHGXHWRLWVKLJKHULQFLGHQFHDQJOH
The shadow and layover masks for the three datasets are presented in )LJXUH and )LJXUH,Q)LJXUHDQG)LJXUHVPDOOHUDUHDVDUHKLJKOLJKWHG]RRPVRIWKHEOXH VTXDUHV RQ )LJXUH )LJXUH LQ RUGHU WR JLYH D EHWWHU YLHZ RI WKH GLIIHUHQFHV EHWZHHQWKHGDWDVHWV)LJXUHOHIW compares RSAT-2 Standard mode and TSX/TDX StripMap mode (RSAT-2 Ultrafine not available at this location). Both datasets have an ascending geometry with quite similar LOS orientation toward NE, but WKH GLIIHUHQW LQFLGHQFHDQJOHVLQGXFHVLJQLILFDQWGLIIHUHQFHVLQWKHH[WHQWRIWKHJHRPHWULFDOHIIHFWV ODUJHULQFLGHQFHDQJOHIRU76;7';StripMap modePRUHDUHDVDIIHFWHGE\VKDGRZ DQGOHVVE\OD\RYHU)LJXUHULJKWFRPSDUHVWKHWKUHHGDWDVHWV,WVKRZVWKDWOD\RYHU HIIHFWLVSUHVHQWRQZHVWIDFLQJVORSHVIRUDVFHQGLQJJHRPHWU\DQGHDVWIDFLQJVORSHVIRU GHVFHQGLQJ
Overall it clearly appears the use of the RSAT-2 Standard mode is strongly limited by the layover. The mask affects a large proportion of the scenes (16 % on land). It reduces the areas available for the analysis, but also makes them discontinuous and thus more difficult to process.
33
)LJXUH$UHDVDIIHFWHGE\OD\RYHU VKDGRZIRU56$76WDQGDUGPRGHDVFHQGLQJJHRPHWU\7KH GDWDVHWLVKLJKO\DIIHFWHGE\OD\RYHUHIIHFWVEXWOLWWOHE\VKDGRZGXHWRLWVVWHHSLQFLGHQFHDQJOH
)LJXUH/HIWDUHDVDIIHFWHGE\OD\RYHU VKDGRZIRU56$78OWUDILQHPRGHGHVFHQGLQJJHRPHWU\
/D\RYHUHIIHFWRFFXUVRQZHVWIDFLQJVORSHVFRQWUDU\WR56$76WDQGDUGPRGHGXHWRWKHGLIIHUHQW JHRPHWU\5LJKWDUHDVDIIHFWHGE\OD\RYHU VKDGRZIRU7HUUD6$5;7DQ'(0;6WULS0DSPRGH DVFHQGLQJJHRPHWU\>&HWLQLFHWDOLQSUHS@7KHGDWDVHWLVOHVVDIIHFWHGE\OD\RYHUDQGPRUHE\
VKDGRZWKDQ56$76WDQGDUGPRGHGXHWRWKHODUJHULQFLGHQFHDQJOH
34
)LJXUH=RRPLQWKHZHVWRI9DNVGDOXSSHUULJKWEOXHVTXDUHLQ)LJXUH )LJXUH$UHD FRYHUHGE\56$76WDQGDUGPRGHDQG76;7';6WULS0DSPRGH>&HWLQLFHWDOLQSUHS@
)LJXUH=RRPRQ%HUJHQORZHUOHIWEOXHVTXDUHLQ)LJXUH )LJXUH$UHDFRYHUHGE\56$7 6WDQGDUGPRGH56$78OWUDILQHPRGH 76;7';6WULS0DSPRGH>&HWLQLFHWDOLQSUHS@
35
7(0325$/'(&255(/$7,216
63$7,$/9$5,$7,216
As described in section 2.2.4, InSAR results can be affected by temporal decorrelations due to changes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n the eastern part, the patches of medium coherence corresponding to the mountainous parts without dense vegetation are discontinuous which makes difficult the production of relevant InSAR results.
)LJXUH/HIWVLPSOLILHGYHJHWDWLRQPDSRYHU%HUJHQDUHDFODVVHVEDVHGRQYHJHWDWLRQPDSIURP
>-RKDQVHQ -RKDQVHQHWDO@5LJKW([DPSOHRIPHDQFRKHUHQFHPDSIURP56$7 6WDQGDUGPRGHGDWDVHWXVLQJRQO\LQWHUIHURJUDPVZLWKDWHPSRUDOEDVHOLQHRIGD\V
36
6($621$/9$5,$7,216
The coherence varies also in time. For example, tHUUDLQFRQWDLQLQJYDULDEOHOLTXLGZDWHU VXFKDVHJDUHDVFRYHUHGZLWKZHWVQRZZLOODOVRKDYHGLIIHUHQWVFDWWHULQJSURSHUWLHV IURPRQHREVHUYDWLRQWRWKHQH[W7KXVLQDUHDVDIIHFWHGE\VQRZGXULQJWKHZLQWHUWKH FRKHUHQFHFDQYDU\DORWDFFRUGLQJWRWKHVHDVRQ7KHZLQWHUVFHQHVFDQEHXQXVDEOHLI EDUHURFNVRUDUWLILFLDOLQIUDVWUXFWXUHDOORZLQJKLJKVLJQDOVWDELOLW\DUHQRWSUHVHQWDERYH WKHVQRZOHYHO,Q)LJXUHWZRH[DPSOHVRIVLPSOHLQWHUIHURJUDPVLQVXPPHUDQGLQ ZLQWHU EHWZHHQ 56$7 6WDQGDUG PRGH DFTXLVLWLRQV ZLWK D WHPSRUDO EDVHOLQH RI GD\VDUHSUHVHQWHGDQGFRPSDUHGZLWKWKHVQRZFRYHUDWWKLVSHULRG
)LJXUH([DPSOHVRIWKHHIIHFWVRIVQRZFRYHURQWKHFRKHUHQFH56$76WDQGDUGPRGHGDWDVHW 7RSOHIWFRKHUHQFHPDSEDVHGRQLQWHUIHURJUDPLQVXPPHU7RSULJKWFRKHUHQFHPDSEDVHGRQ
LQWHUIHURJUDPLQZLQWHU%RWWRP6QRZGHSWKDWWKHVDPHSHULRGDVWKHLQWHUIHURJUDPV)URP ZZZVHQRUJHQRGRZQORDG
37
2QWKHOHIWRI)LJXUHWKHLQWHUIHURJUDPZDVJHQHUDWHGXVLQJVFHQHVLQVXPPHU DQGRQULJKWXVLQJVFHQHVLQZLQWHU 7KHVXPPHULQWHUIHURJUDPVKRZVJRRGFRKHUHQFHYDOXHVRQWKHXUEDQ DUHDVDVZHOODVLQWKHDUHDVZLWKEDUHURFNVRUORZYHJHWDWLRQ7KHZLQWHULQWHUIHURJUDP VKRZV VLJQLILFDQW WHPSRUDO GHFRUUHODWLRQ RQ WKH ZKROH HDVWHUQ SDUW GXH WR WKH VQRZ H[FHSWRQXUEDQDUHDV7he seasonal variations can require removing winter scenes. In areas where snow affects large areas over a long time period, this is a significant limitation that can severely reduce the amount of usable interferograms
&2+(5(1&(25$03/,78'(',63(56,21),/7(56
In practice the coherence map is also used to select pixels that will be used for the ground deformation analysis. During the SBAS processing, in addition to the layover/shadow masks, a coherence mask is thus generated using a coherence threshold defined by the operator )LJXUH
)LJXUH([DPSOHRIVHOHFWHGSL[HOVIRUWKH6%$6DQDO\VLVRYHU%HUJHQIRUHDFKGDWDVHW76;7';
>&HWLQLFHWDOLQSUHS@GHSHQGLQJRQWKHDSSOLHGFRKHUHQFHWKUHVKROGDQGWKHOD\RYHU VKDGRZ PDVNV
38
The selected pixels correspond finally to those outside the layover/shadow masks with a coherence value above a defined coherence value (for ex. 0.3) for a minimum percent (for ex. 25 %) of the XVHG LQWHUIHURJUDPV The choice of the thresholds is a balance between amount of selected pixels and good coherence. A too high coherence threshold can indeed lead to too small areas of analysis organized in discontinuous patches.
For PSI processing, instead of a coherence threshold, the amplitude dispersion is used to select the points used for the analysis. As for the coherence threshold in SBAS procedure, the choice of the thresholds is a balance between low amplitude dispersion and amount of points. )LJXUHVKRZVDQH[DPSOHRISULPDU\DQGFRPSOHWHQHWZRUNV RYHU%HUJHQXVLQJWKH56$78OWUDILQHGDWDVHW+HUHWKHDSSOLHGDPSOLWXGHGLVSHUVLRQ WKUHVKROG LV IRU WKH SULPDU\ QHWZRUN )LJXUH OHIW DQG IRU WKH FRPSOHWH QHWZRUN)LJXUHOHIW
)LJXUH([DPSOHRISULPDU\OHIWDQGFRPSOHWHULJKW36QHWZRUNVEDVHGRQWZRDPSOLWXGH GLVSHUVLRQWKUHVKROGV56$78OWUDILQHGDWDVHW
39
*5281''()250$7,215(68/76
The ground deformation results are presented separately for different subareas. In section 4.3.1, the results on Bergen are analyzed. Although this area was not specified in the main areas of interest for the partners (see section 3.1.2), it is the only sector covered by RSAT-2 Ultrafine mode and it is interesting to have one case study comparing the three datasets. Moreover the coherence is good in this area due to the urban infrastructure which give overall better results than most of the areas in the eastern part. In section 4.3.2, results in Indre Arna are analyzed. In section 4.3.3, some results from the eastern part of area are introduced (Samnangerfjorden - Kvamsskogen - Dale). However, it has to be noted that due to the limitations described in sections 3.2, 4.1 and 4.2, especially related to the spatial resolution, the geometrical effects and the coherence, the reliability of the results are highly reduced.
The main deformation maps presented thereafter are based on stacking and PSI methods (see section 2.3.3). A SBAS method is used to retrieve the time series presented in sections 4.3.1.2 and 4.3.2.2. All the maps displayed the mean velocity in millimeter per year along the line-of-sight. Negative values correspond to deformation away from the SAR sensor. Positive values correspond to deformation toward the SAR sensor. The millimetric interval that was chosen for the color scale highlights that the displacements over the area are very small. It should be noted that for display purpose the scales vary from a dataset to another. The results differences are due to the different wavelengths and revisit times of the sensor, as well as the different lines-of-sight, spatial resolution, temporal coverage of the datasets.
%(5*(1
'()250$7,210$36
The availability of three datasets on Bergen makes this area interesting to compare InSAR results. ,Q)LJXUHXSOHIWDPDSKLJKOLJKWV%HUJHQDQGWKHGLIIHUHQW]RRPV OHWWHUV$&WKDWZLOOEHXVHGIRUWKHFRPSDULVRQ,Q)LJXUHXSULJKWDQGERWWRP VWDFNLQJ UHVXOWV RIWKHWKUHHGDWDVHWV DUH SUHVHQWHG IRUWKH]RRP $2YHUDOOLWDSSHDUV WKDW WKH DUHD LV QRW DIIHFWHG E\ VLJQLILFDQW JURXQG GHIRUPDWLRQ +RZHYHU VRPH VPDOO PRYLQJDUHDVDUHGHWHFWHGE\VHYHUDOVHQVRUV7KHLUVL]HDQGYHORFLW\YDULHVHVSHFLDOO\
GXH WR WKH GLIIHUHQW VSDWLDO UHVROXWLRQV OLQHRIVLJKW DQG WHPSRUDO GLVWULEXWLRQ RI DYDLODEOH LPDJHV ,W DSSHDUV IRU H[DPSOH WKDW 56$7 6WDQGDUG PRGH GHWHFWV VORZHU DQGVSDWLDOO\VPRRWKHUGHIRUPDWLRQWKDQWKHWZRRWKHUVHQVRUVGXHWRLWVORZHUVSDWLDO UHVROXWLRQ
40
)LJXUH8SOHIWORFDWLRQPDSRYHU%HUJHQZLWK]RRPVRIWKHDQDO\]HGDUHDV8SULJKWVWDFNLQJUHVXOWV IRU56$76WDQGDUGPRGH]RRP$%RWWRPOHIWVWDFNLQJUHVXOWVIRU56$78OWUDILQHPRGH]RRP
$%RWWRPULJKWVWDFNLQJUHVXOWVIRU76;7';6WULS0DSPRGH]RRP$>&HWLQLFHWDOLQ SUHS@%DFNJURXQGV1RUJHLELOGHURUWKRSKRWR /26UHODWHGVKDGHGUHOLHIV
7KH)LJXUH FRUUHVSRQGVWRWKH]oom B on downtown. VDULRXVVPDOOPRYLQJ DUHDV DUHSUHVHQWRQWKHWKUHHVHWVRIUHVXOWVDW0¡KOHQSULV6ROKHLPVYLNHQDURXQGWKH6WRUH /XQJHJnUGVYDQQRQWKH1RUWKRIWKH%UDQQ6WDGLRQComparing the results of the three datasets and the PSI and stacking methods, it appears that the moving sectors detected thanks to stacking are less well displayed by PSI due to small amount of Persistent Scatterers in the these areas.
41
)LJXUH,Q6$5UHVXOWVRQ]RRP%VHH)LJXUH/HIWVWDFNLQJUHVXOWV5LJKW36,UHVXOWV7RS 56$76WDQGDUGPRGH0LGGOH56$78OWUDILQHPRGH%RWWRP76;7';6WULS0DSPRGH>&HWLQLF
HWDOLQSUHS@%DFNJURXQGV1RUJHLELOGHURUWKRSKRWR /26UHODWHGVKDGHGUHOLHIV
42
)LJXUH corresponds to the zoom C in the area of the airport. Here only the two RSAT-2 datasets are available. The moving area in the North of the landing strips appears to be a gravel pit, while the areas highlighted along the strips are artificial banks. The comparison of the two datasets clearly highlights the impact of the spatial resolution on the delineation and rate of movement. Comparing the results of PSI and stacking methods, it appears that the moving sectors detected thanks to stacking are less well displayed by PSI due to small amount of Persistent Scatterers aside the main artificial infrastructure. In the eastern part, some small patterns and especially on RSAT-2 Ultrafine dataset are probably due to unwrapping errors due to the patchy coherent area.
)LJXUH,Q6$5UHVXOWVRQ]RRP&VHH)LJXUH/HIWVWDFNLQJUHVXOWV5LJKW36,UHVXOWV7RS 56$76WDQGDUGPRGH%RWWRP56$78OWUDILQHPRGH%DFNJURXQGV/26UHODWHGVKDGHGUHOLHIV
