indeso workshop romy

(1)

Institut Mines-Télécom

SAR-VMS-AIS Correlation to

Identify Illegal Fishing Vessels in

Indonesia

Romy ARDIANTO

Master of Science in Computer Science and

Decision Systems (CSDS)

(2)

Internship Supervisors

06-Oct-17 Soutenance de stage –Romy ARDIANTO 2

Guillaume HAJDUCH

Yannis HARALAMBOUS

▪ Head of Expertise and

Innovation Department

at CLS Brest

▪ Service Manager of the

Sentinel-1 Mission

Performance Center

▪ Senior Professor at Computer

Science Department,

TELECOM Bretagne

▪ Head of Master of Science in

Computer Science and

Decision Systems (CSDS)

Program

(3)

Master’s Degree in France

(4)

Outline



Context

• INDESO project

• CLS company

• Objective of the internship



Access to satellite data



Estimation of VMS and AIS positions based on SAR

echoes



Correlation of satellite data (SAR-VMS-AIS) and SAR

detection performance characterization



Characterization of the non-VMS fishing fleet



Conclusion

(5)

Context

(6)

Monitoring of marine and fisheries

resources against illegal fishing activities



INDESO project (Infrastructure Development of Space

Oceanography)

• Collaborative project between France and Indonesia (2013

–

2017)

• Mission:

─

To protect our marine ressources against illegal fishing activites

and establish sustainable fisheries in Indonesia



Moratorium of ex-foreign fishing vessels

• Temporary suspension of issuance of fishing licenses for all

Indonesian waters for vessels constructed abroad (2014

-2015

(7)

INDESO



Build, develop and operate

an operational center for monitoring & forecast



Implement

a large capacity building plan to transfer technology & knowledge



Expand & enhance

result-oriented downstream applications

06/10/2017 Soutenance de stage –Romy ARDIANTO 7

Fight IUU Fishing

Monitor Fish Stocks

Monitor Coral reefs

Shrimp Farming

Seaweed Farming

Integrated Coastal

Zone Management

(8)

CLS (Collecte Localisation Satellites)



Created in 1986, a subsidiary of CNES, Ifremer &

ARDIAN



Pioneer provider of monitoring and surveillance

solutions :

• Sustainable management of fisheries

• Environmental monitoring

• Maritime safety and security



506 collaborators :

• 336 in France (Toulouse, Brest and Bidart)

• 170 headquarters and offices around the world

(9)

Organizational Chart

(10)

Objective of the internship

To perform a joint analysis of VMS (Vessel Monitoring System) and

AIS (Automatic Identification System) data with ship targets in SAR

imagery (Synthetic Aperture Radar)

• Evaluate correlation coefficient

• SAR detection performance characterization

• Identify illegal fishing vessels present in SAR imagery

Software applications written in Python were developed in order to

perform this joint analysis automatically

(11)

Methodology



Venn Diagram Correlation of SAR, VMS and AIS

SAR correlated with VMS

VMS

AIS

SAR

AIS correlated with

VMS

SAR correlated with

AIS

SAR correlated with

VMS and AIS

Illegals

Illegals detected by SAR,

not reported by AIS

Illegals not detected by

SAR, reported by AIS

Illegals detected by SAR,

reported by AIS

(12)

Access to satellite data

(13)

Satellite Data

(14)

Estimation of VMS and AIS

positions based on SAR echoes

(15)

Interpolation/extrapolation of VMS and AIS

data based on SAR echoes

Methodology



Linear interpolation/extrapolation

• Based on 2 successive points having the date closest to

the date of SAR echoes



Geodetic calculation used: Vincenty's formulae

• Widely used in geodesy to calculate the distance between

two points on the surface of a spheroid



accurate to

within 0.5 mm

• Consist of 2 methods:

Direct

and

Inverse

─

Direct method

: computes the location of a point that is a

given distance and azimuth (direction) from another point

─

Inverse method

: computes the geographical distance and

azimuth between two given points

(16)

Interpolation



Algorithm

1. Selection of VMS/AIS data with ± 3h on the date of SAR echo

2. Selection of 2 positions closest to the date of SAR echo

3. Apply inverse method to calculate the geographical distance between point A

and B

4. Calculate the interpolation distance (based on the geographical distance

between point A and B, time interval between A and B, and time interval

between interpolated point et point B)

5. Apply direct method to determine geographical coordinate of the interpolated

point

2014/12/07 09:00:00

2014/12/07 08:50:00

2014/12/07 09:10:00

Point A

Interpolated point

Point B

on the date of

SAR echo

(17)

Extrapolation



Algorithm

1. Selection of VMS/AIS data with ± 3h on the date of SAR echo

2. Selection of 2 positions closest to the date of SAR echo

3. Apply inverse method to calculate the geographical distance and azimuth

between point A and B

4. Calculate the extrapolation distance (based on time difference between the

date of point B and SAR echo, and the speed of vessel point B)

5. Apply direct method to determine geographical coordinate of the extrapolated

point

2014/12/07 09:10:00

2014/12/07 08:50:00

2014/12/07 09:00:00

Point A

Point B

Extrapolated point

on the date of

(18)

Estimation of VMS position (1/2)

Vessel 1

Vessel 2

(19)

Estimation of VMS position (2/2)

(20)

Correlation of satellite data

(SAR-VMS-AIS) and SAR performance

characterization

(21)

SAR Doppler effect (1/2)



Doppler effect (Doppler shift)

• Definition: the change in frequency or wavelength of a

wave for an observer who is moving relative to the

wave source

• Radar applications:

─

Speed measuring

─

Moving target indication

• Illustration of Doppler effect:

(22)

SAR Doppler effect (2/2)



Doppler effect application to estimate the real position of

SAR echo

(23)

Distance parameters (Dneigh and Dmax)

Dmax

Dneigh

SAR echo

Interpolated/extrapolated VMS/AIS <= Dmax

Interpolated/extrapolated VMS/AIS <= Dneigh

Dneigh

Maximum distance when selecting

interpolated/extrapolated VMS/AIS positions

near from SAR echos (without Doppler effect)

Dmax

(24)

Automatic coupling of SAR echo and VMS data (Test the

algorithm)

(25)

Automatic coupling of SAR echo and VMS data (1/2)



Histogram of distance

Dneigh = 10 km

Dmax = 5 km

Dmax

(26)

Automatic coupling of SAR echo and VMS data (2/2)

(27)

Automatic coupling of SAR echo and AIS data (1/2)



Histogram of distance

Dvoisin = 5 km

Dmax = 2 km

(28)

Automatic coupling of SAR echo and AIS data (2/2)

(29)

SAR detection performance

characterization

(30)

SAR-VMS correlation

Correlation

percentage

between

SAR

and

VMS

=

242

514 ´

100%

=

47.08%

Nb of SAR

echoes

Nb of VMS

SAR correlated

with VMS

SAR not correlated

with VMS

VMS not correlated

with SAR

4042

514

242 3800

272

(31)

SAR-AIS Correlation

Correlation

percentage

between

SAR

and

AIS

=

238

307 ´

100%

=

77.52%

Nb of SAR

echoes

Nb of AIS

SAR correlated

with AIS

SAR not correlated

with AIS

AIS not correlated with

SAR

9044

307

238 8806

69

(32)

Characterization of the

non-VMS fishing fleet

(33)

Methodology



Venn Diagram Correlation of SAR, VMS and AIS

SAR correlated with VMS

VMS

AIS

SAR

AIS correlated with

VMS

SAR correlated with

AIS

SAR correlated with

VMS and AIS

Illegals

Illegals detected by SAR,

not reported by AIS

Illegals not detected by

SAR, reported by AIS

Illegals detected by SAR,

reported by AIS

(34)

Results



AIS fishing vessels present in Indonesia waters

No.

MMSI

Vessel name

Country

Length [m]

Width [m]

1 412440276

FU YUAN YU 386

Chine

48

7

2 412440275

FU YUAN YU 385

Chine

48

7

3 412419967

TAIYUANYU9009

Chine

45

8

(35)

FU YUAN YU 386 (MMSI 412440276)

(36)

FU YUAN YU 386 (MMSI 412440276)

Hongkong, China

(27/04/2014)

Arafura Sea, Indonesia

(08/05/2014

–

26/11/2014)

Conakry, Guinea

(31/01/2015)

1

(37)

Conclusion



The software applications developed made it possible to perform this join

analysis automatically



Correlation percentage of SAR-AIS had better results than the correlation of

SAR-VMS, due to:

• Time interval between sending two messages

• Average size of vessels



Approach used allowed identifying illegal fishing vessels present in SAR

imagery



Perspective:

• Taking into account wind condition to the analysis



Recommendation:

• Apply the analysis results to support illegal fishing monitoring in Indonesia

(38)

Contribution to INDESO illegal fishing

application

06/10/2017 Modèle de présentation Télécom Bretagne 38

Correlating radar imagery with VMS and AIS data to identify illegal fishing vessels

Suspected illegal fishing vessel (no VMS)

Suspected illegal vessel (no AIS)

(39)

Behind the scenes

06/10/2017 Modèle de présentation Télécom Bretagne 39

Soutenance :

(40)

(41)

Thank you

(42)

Annexe

(43)

Interpolation linéaire et non linéaire

Position réelle de la balise

Position balise interpolée

Trajectoires

(44)

Navires de pêche illégale balisés par l’AIS présents dans

les eaux territoriales indonésiennes

(45)

(46)