m. manunta berardino p., bonano m., de luca c., de novellis v.,...
TRANSCRIPT
M. Manunta, Berardino P., Bonano M., De Luca C., De Novellis V., Elefante S., Fusco A., Lanari R., Manzo M., Ojha C., Pepe A., Sansosti E., Zinno I., Casu F.
CNR IREA, Via Diocleziano 328, 80124 Napoli (Italy)
Achieved accuracies:
• ≈ 1 - 2 mm/year on the mean deformation velocity
• ≈ 5 - 10 mm on the single displacement
exploiting interferograms characterized by a “small baseline” in order to limit the noise (decorrelation) phenomena, thus maximizing the number of investigated pixels;
PhU operation is usually performed by applying MCF or EMCF techniques.
using no a priori or model information on the investigated deformation signal;
The SBAS approach allows to produce “long term” deformation times-series by:
SBAS approach: key points
Interferometric Pair Selection
Geometrical Registration
Spectral Diversity
Compensation
Interferogram Generation
Phase Unwrapping
Time-series Generation
Deformation Time-series
S1 SBAS-DInSAR Block Diagram
Data Ingestion
Acquisition Time
S-1A SBAS approach: pair selection
The number of interferograms is about 3*(Number of acquisitions)
Orbital tube of Sentinel-1 should be very short, therefore no perpendicular baseline constraint needs to be applied. To get redundancy of interferograms, each acquisition is coupled with the 3 following scenes in time.
Orbit Registration: Rigid offset. It is retrieved in one point exploiting orbit and target location (DEM) information.
Coherence Maximization: Rigid offset. It is retrieved in one patch by maximizing the number of coherent points.
Geometric Registration: it is performed point by point, using orbit and location (DEM) information.
Spectral Diversity Compensation: the residual phase, estimated through Spectral Diversity method, is compensated directly from interferograms, without performing again the interpolation of slave images.
Co-Registration Scheme
Interferometric Pair Selection
Geometrical Registration
Spectral Diversity
Compensation
Interferogram Generation
Phase Unwrapping
Time-series Generation
Deformation Time-series
S1 SBAS-DInSAR Block Diagram
Data Ingestion
Spectral Diversity Scheme
Inter-Swath overlapping: the 2 bursts have a shift of about 500 samples, corresponding to 800 Hz Hz
Samples
500 Samples
800 Hz
Inter-Swath Overlap
∆dop≅800 Hz
∆dop≅2400 Hz
Shift max: ± 0.3 samples
Shift max: ± 0.1 samples
Spectral Diversity Scheme
The spectral diversity method is applied: • first in the inter-swath areas, to
estimate larger shifts; • then in the inter-burst areas, to
accurately estimate the residual shift.
This double step allows us to limit phase ambiguity problems in the shift estimation.
Spectral Diversity Scheme
Acquisition Date
Shift of Ann. St. Vec. [samples]
Shift of Rest. St. Vec. [samples](S1A_OPER_AUX
_RESORB) Orbit Coherence Orbit Coherence
20-10-2014 0.0 0.0 0.0 0.0 01-11-2014 6.806 6.981 6.963 6.996 13-11-2014 1.546 1.296 1.276 1.293 25-11-2014 7.110 7.335 NA NA Mean Error 0.21 0.02
Orbit Information: Annotated vs. Restituted Restituted: 10-sec State Vectors Annotated: 1-sec State Vectors
Burst Misalignments
B1 B2
Start Time of B2 usually does not correspond to a B1 sample:
TstB
1
TstB
2
This misalignment has to be properly compensated in the registration phase.
St. Dev. Residual Shifts [samples] Interferogram With Registration Without Registration
01122014_18012015 0.0021 0.0068 01122014_25122014 0.0013 0.0100 01122014_30012015 0.0013 0.0080 14102014_01122014 0.0018 0.0141 14102014_25122014 0.0011 0.0159 14102014_26102014 0.0011 0.0182 18012015_30012015 0.0015 0.0090 25122014_18012015 0.0012 0.0093 25122014_30012015 0.0009 0.0093 26102014_01122014 0.0020 0.0153 26102014_18012015 0.0019 0.0127 26102014_25122014 0.0017 0.0091
Mean Standard Deviation 0.0015 0.0115
Burst Misalignments
Fogo Volcano (Cape Verde) 03112014_27112014 b ⊥≅ 16 m
S-1A First Results: Fogo Volcano (Cape Verde)
S-1A First Results: Hawaii (USA)
S-1A First Results: Hawaii (USA)
Azim
uth
07-Dec-2014 12-Jan-2015
Sentinel Asc
S-1A First Results: Mt. Etna
Azim
uth
22-Dec-2014 02-Jan-2015
TSX Asc 131
S-1A First Results: Mt. Etna
Baseline Acq. Time Perp. [m] Parallel [m] Along track [m]
04042013 -27 -30 14
28042013 -120 -98 -13
22052013 -65 -53 -22
15062013 70 30 -12
09072013 129 81 -13
02082013 0 0 0
26082013 -117 -99 8
19092013 50 -36 10
13102013 -35 -103 29
06112013 124 21 20
30112013 -23 -82 9
TOPS SBAS results: Mexico City RS2 campaign
Azimuth
Rang
e
Several RS2 TOPS scenes are affected by significant orbit errors
RS2 TOPS interferograms show good coherence (24 days revisit time)
TOPS SBAS results: Mexico City RS2 campaign
RS2 TOPS interferograms: orbital parameter correction
Azimuth
Rang
e
Pepe et al., SBAS-Based Satellite Orbit Correction for the Generation of DInSAR Time-Series: Application to RADARSAT-1 Data, TGRS, 2011
Mea
n ve
loci
ty [c
m/y
r]
>20
<-20
TOPS SBAS results: Mexico City RS2 campaign
S-1A First Results: Istanbul (Turkey)
S-1A First Results: Sakurajima Volcano (Japan)
S-1A First Results: Hawaii (USA)
Mea
n ve
loci
ty [c
m/y
r]
>25
<-25
Upcoming steps: S1 SBAS chain within GEP
The S1 SBAS processing chain is currently under development to be integrated within ESA Geohazards Exploitation Platform (GEP).
https://geohazards-tep.eo.esa.int
S1 Processing on demand (G-POD)
Archives
G-POD User
Web Interface Computing Power
P-SBAS service (or other tools)
EPOS User
G-POD Web Portal of P-SBAS service
Paper 179 - Session title: PSI and DInSAR (3) 11:30 A Cloud Computing Approach for Big DInSAR Data Processing through the P-SBAS Algorithm
Paper 161 - Session title: PSI and DInSAR (3) 11:50 P-SBAS Service within ESA G-POD Platform for Unsupervised on Demand DInSAR Processing
http://gpod.eo.esa.int - [email protected]
Welcome Sentinel-1