Primary Data Analysis Node (P-DAN) Report

3rd Joint Project Team Meeting for Sentinel Asia STEP3 (JPTM2016)

Dr. Masahiko NAGAI

Associate Professor, Center for Spatial Information Science, The University of Tokyo, Japan

Space Agencies Disaster ManagementAgencies

Disasters

SatelliteObservation

DataUtilization

Information Provision

Data Request

DPN DAN End User

Space Agencies Disaster ManagementAgencies

Disasters

SatelliteObservation

DataUtilization

Information Provision

Data Request

DPN DAN End User

Activities for DAN Network for University and

Research Institution Advanced analysis InSAR, etc. Training for instructor

Activity for End User Support as a helpdesk Basic RS and GIS process Map making

4

Strategy of Capacity Building

Disaster Management AgenciesNetwork for Universities and Research Institutions

University of Tokyo

Keio University

Tokyo University of Marine Science and Technology

AIT

University of Philippines

Indonesia University

Chulalongkorn University

NationalUniversity of Laos

etc.

Cloud-based System for Archive, Integration and Analysis

Satellite data

5

Strategy of Capacity Building

University of Tokyo

Keio University

Tokyo University of Marine Science and Technology

AIT

University of Philippines

Indonesia University

Chulalongkorn University

NationalUniversity of Laos

etc.

Disaster Management AgenciesNetwork for Universities and Research Institutions

contents

e-Learning(course）

e-Learning(material）

e-Learning System of G-SPASE

System configuration

6

Summary of methodology of InSAR processing

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

SAR Interferometry

Range Doppler processing

Range Compression

Azimuth Compression

Single look complex (SLC)

Multi look image

SAR processing

RAW data Sensor Parameter Processing Parameter

7

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

Co-registration of SLC Data: co-registration is the process of fitting one SARimage accurately upon another SAR image. Co-registration of complex imagesinvolves two steps.

1) the location of each pixel in the slave image is changed with respect to themaster image.

2) the amplitudes and phases of the sensors are recalculated by interpolation.

GMTSAR is used for this research using cross correlation algorithm (xcorr) forregistration. “xcorr” uses window size of 64 pixels and has never failed toprovide accurate co-registration even in cases where the interferometriccoherence is close to zero.

Summary of methodology of InSAR processingSummary of methodology of InSAR processingSAR Interferometry

8

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

Summary of methodology of InSAR processingSAR Interferometry

Concept of repeat orbit interferometry

The phase difference to point on the ground is related to the range difference

;Therefore

- range from reference track to reflectorB - total baseline distance between reference and repeat track - look angle - angle between the baseline vector and the tangent plane

= Bcos(-)

= Bsin(-)

-

9

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

Summary of methodology of InSAR processingSAR InterferometryCorrection of earth curvature and topographic phase

Interpolate the topography

For other information, it can be computed from SAR datab(a) - radius of the spacecraft orbit for thereference imageB(a) - the length of the baseline(a) - the orientation of the baseline

topography (range, azimuth)

10

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

Summary of methodology of InSAR processingSAR InterferometryPhase due to earth curvature and topography

Law of cosine

It can be rewritten as

Phase to range =

Reference: GMTSAR (2011)

= -B sin (-)

11

It can be rewritten as

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

SAR Interferometry

Phase unwrapping: The interferometric phase shows many discontinuities,which originate the classical fringe pattern. The interferometric phase isalmost linearly proportional to the topographic height. The difficulty is thatthe interferometric phase is only known to module 360 degrees. If the phasevariation exceeds 360 degrees, it wraps around to 0 degrees. It is necessaryto unwrap the phase in the interferogram by using several algorithms forconverting the interferometric phase to topographic phase.

12

Preprocess

Registration

InterferogramGeneration

Phase Unwrapping

Geocoding

SAR Interferometry

The final step is to geocode all the products by transforming from the range/azimuth coordinate system of the master image to longitude and latitude.

Geocode

Amplitude image

Interferogram image13

• Range Doppler Processing

• Single Look complex (SLC)

• Multi look image

• Color enhancement• Classification• Change detection• Deep learning

DInSARDInSAR and ALOSand ALOS2 2 imagesimages

Image 1 on 2014/10/04

Image 3 on2015/05/02

Image 2 on2015/02/21

2014 2015

Oct Nov Dec Jan Feb Mar Apr May Jun

25 April 2015EQ 7.8 Mw

InSAR Pre-Seismic InSAR Co-Seismic

DEM DEM

Pre-seismic Co-seismic

12 May 2015EQ 7.2 Mw

More Images

InSAR Co-Seismic

DEM

Co-seismic

1970 - 03 October 2014 04 October 2014 – 20 February 2015 21 February 2015 – 02 May 2015

HH DInSAR HV DInSAR HH DInSAR HV DInSAR HH DInSAR HV DInSAR15

16

Azimuth Direction

LOS

West East

Line of sight displacement

N

High : 1530 mm.

Low : near 0

Line of sight displacement

17

Summary of Summary of DInSARDInSAR results results Pr

e-se

ism

icCo

-sei

smic

Correlation Interferogram Unwrapped phase (LOS)

LOS = Line of sight

Earthquake in Japan 2011 (5500 Scenes), Flood in Thailand 2011 (1500 Scenes),

Earthquake in Nepal 2015 (more than 8000 Scenes)

Oh My God!

〜

A Solution; Cloud-based System for Archive, Integration and Analysis

！Co-work on the same platform!

Share data, tools, ideas and experiences!

More Participations and Contributions

！！

http://gestiss.org

Emergency Data Analysis team

Faculty in RSGIS: 7Faculty in DPMM: 2Student in RSGIS: 15Nepalese: many

Satellite Data: Dec. 4, 14:00Analyzed Product: Dec 4, 17:00

Satellite Data: Dec. 16, 14:40Analyzed Product: Dec 16, 18:00