Exploiting maritime Big Data, the Blue Hub · The Blue Hub – exploiting maritime Big Data . 8 ......

Joint Research Centre the European Commission's

in-house science service

The Blue Hub: https://bluehub.jrc.ec.europa.eu

Exploiting maritime Big Data,

the Blue Hub

Marlene Alvarez Alvarez,

Michele Vespe, Harm Greidanus

Marlene.alvarez@jrc.ec.europa.eu

2

Available (big) data on ship traffic:

Ship position reports

(AIS, LRIT)

Earth observation

satellites (e.g., Copernicus)

From data to knowledge

Contents

3

Irregular migration

Maritime Security

Maritime Traffic

Piracy Illegal Fishing

Oil Pollution

Maritime Situational Awareness (MSA)

Which ships are on the sea?

What are they doing?

4

Ship reporting Data

DC

INMARSAT GPS

Long Range Identification and Tracking (LRIT) Automatic Identification System (AIS)

REGULAR (space and time)

IRREGULAR (spatial/temporal gaps)

FREQUENCY Fix (to 6 hours)

FREQUENCY High (few sec to min)

ACCESS Restricted

ACCESS Free

5

Collection of AIS ship reports

Received by satellites,

when they pass over

Collected by coastal receiver

networks, continuously

Picture: FleetMon.com

Picture: exactEarth

GLOBAL

COVERAGE LIMITED

COVERAGE

CONSTELLATIONS improved frequency revisit

6

Radarsat-1 Fine © CSA/MDA

Earth observation satellites

Optical - high resolution

- Small area

- Daytime, clear skies

Radar - low resolution

- Wide area

- Through clouds and night

DigitalGlobe

Used for recognition Used for detection

7

Vessel Movement Data

Big Data (historical)

Real-Time Maritime

Surveillance

Data Fusion & Target Tracking

Routes Extraction

Anomaly Detection

(deviation from normality)

Routes prediction

Data Mining & Knowledge Discovery

Fishing grounds and fishing

efforts

Fisheries socio-economics

Mapping Off-shore and exploration

Maritime Spatial Planning

Gridding Shipping Emissions

Mapping activities at sea

Vessel profiling/ intelligence

Customs & Agencies

Ex-post policy evaluation

Geopolitical issues impact on

transport

Fisheries management

Trade Indicators

The Blue Hub – exploiting maritime Big Data

8

…

Ship Name

…

…

…

Data Fusion & Target Tracking at global scale

Track built by integrating multiple AIS data providers (ship-type color coded).

Vessel Movement Data

Real-Time Maritime

Surveillance

Data Fusion & Target Tracking

“BIG DATA”

- ~150,000 ships

carry transponders

- each may send

10,000 messages /

day

9

1,557 S-1 images (3 Oct 2014 - 11 Sep 2015)

41,928 ships detected

correlated with AIS

34,263 ships

uncorrelated with AIS

AIS Coverage

Data: Copernicus

Sentinel-1: Long term ship monitoring

10

One-month of EU LRIT CDC data, revealing the main global traffic routes and

enabling the implementation of innovative decision support tools

→ Video Link

From EU to worldwide tracking and traffic routes

Big Data (historical)

Vessel Movement Data

11

FishingDredging

Tug

Passenger

Cargo

Cargo,

Haz. Cat. X

Cargo,

Haz. Cat. Y

Cargo,

Haz. Cat. Z

Cargo,

Haz. Cat. OS

Tanker

Tanker,

Haz. Cat. X

Tanker,

Haz. Cat. YTanker,

Haz. Cat. Z

Other Type,

Haz. Cat. X

Travel Time Histogram

2-week raw AIS data

Pattern discovery

Entry area

Exit areas

Port area

1.2 1.3 1.4 1.5 1.6 1.7 1.8

50.6

50.7

50.8

50.9

51

51.1

51.2

Longitude

Latit

ude

Network representation

1.2 1.3 1.4 1.5 1.6 1.7 1.8

50.6

50.7

50.8

50.9

51

51.1

51.2

Longitude

Latit

ude

Single route analysis

Vessel pattern knowledge discovery: from raw data to support to decision

Fernandez Arguedas et al.: „Automatic Generation of Geographical Networks for Maritime Traffic Surveillance‟, Int. Conf. on Information Fusion (2014)

Big Data (historical)

Routes Extraction

Data Mining & Knowledge Discovery

Ship type distribution

Maritime Spatial Planning

12 Zampieri A., Vespe M., Westra M., Alvarez M., Greidanus H., „A future for historical LRIT data‟, International Maritime Organization (IMO) – Maritime Safety Committee (MSC) 95th session, London, 2015

Historical LRIT data can be used to predict where a vessel will be up to a few days in advance

Route prediction

Big Data (historical)

Routes Extraction

Data Mining & Knowledge Discovery

Routes prediction

13 Greidanus H, Vespe M., Alvarez M., „Detection of Anomalous behaviour in ship reporting data for improved maritime security‟, Future Security Conference, Berlin, 2016

Anomaly Detection (deviation of normality)

- Density map based on 2-y LRIT data - Overlaid real-time positions of all cargo ships - Multi-day past track from one cargo ship

14

Event-based Knowledge Discovery

Raw Data

Event map generation

Vessel based event detection

Pre-Processing

Trajectory Extraction Georeferenced grid

Vessel Objects

Event Objects

• Vessel Id • Dynamic • Static/Voyage

• Event type • Dynamic • Static/Voyage

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

15

Vessel-based Event Detection

Longitude

Latitu

de

Event Type Relevance

AOI Enter/Exit Access to protected areas or areas regulated by traffic routing schemes

Cell Enter/Exit Traffic density maps

Birth/Death Coverage maps

Start/Stop Detection of stop areas (ports, offshore platforms)

Proximity Discovery of illegal activities

Fishing/Steaming Mapping activities at sea

AIS on/off switching Discovery of illegal activities

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

16

Event Map Generation

Event Object

MMSI

Longitude

Latitude

SOG

COG

ROW

COL

Event timestamp

Vessel type

Event type

DWH

Presentation Layer

Analysis Cube

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

17

Event Map Generation: Results

AOI: [15ᵒ W 17ᵒ E; 43ᵒ N 46ᵒ N]

Time interval: October 1st - December 1st, 2015

Terrestrial AIS data from Italian Coast Guard

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Distribution of generated events

0.004 degree

18

Entry into cells

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

19

Entry into cells – Fishing Vessels

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

20

Entry into cells – Fishing Vessels (Fishing behaviour only)

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

21

Entry into cells – Cargo Vessels

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

22

Entry into cells – Tanker Vessels

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

23

Entry into cells – Passenger Vessels

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

24

Stop events

Alvarez M, Fernandez V., Gammieri V., Mazzarella F., Vespe M., Aulicino G., Vollero A., “AIS Event-Based Knowledge Discovery for Maritime Situational Awareness”, International Conference on Information Fusion 2016.

Event Map Generation: Results

25

Left: North- (green) and South-bound (red) traffic crossing the Indian Ocean

Right: Traffic extraction during the second semester 2009 using historical LRIT data

Analysis of geopolitical developments:

the declining impact of piracy in the Indian Ocean

Vespe M., Greidanus H., Alvarez M. : „The Declining Impact of Piracy on Maritime Transport in the Indian Ocean: Statistical Analysis of 5-year Vessel Tracking Data‟, Marine Policy, 2015.

Big Data (historical)

Data Mining & Knowledge Discovery

Mapping activities at sea

Ex-post policy evaluation

Geopolitical issues impact on

transport

26

Time series of traffic crossing the Indian Ocean in 5 years (each figure shows 6 months),

showing the effect of piracy and its progressive decline in deviating maritime traffic

Vespe M., Greidanus H., Alvarez M. : „The Declining Impact of Piracy on Maritime Transport in the Indian Ocean: Statistical Analysis of 5-year Vessel Tracking Data‟, Marine Policy, 2015.

Analysis of geopolitical developments:

the declining impact of piracy in the Indian Ocean

27

Piracy Incidents Source: EU Naval Force – Somalia

Analysis of geopolitical developments:

the declining impact of piracy in the Indian Ocean

Time series of traffic crossing the Indian Ocean in 5 years (each figure shows 6 months),

showing the effect of piracy and its progressive decline in deviating maritime traffic

Vespe M., Greidanus H., Alvarez M. : „The Declining Impact of Piracy on Maritime Transport in the Indian Ocean: Statistical Analysis of 5-year Vessel Tracking Data‟, Marine Policy, 2015.

28

Analysis of geopolitical developments:

the declining impact of piracy in the Indian Ocean

Vespe M., Greidanus H., Alvarez M. : „The Declining Impact of Piracy on Maritime Transport in the Indian Ocean: Statistical Analysis of 5-year Vessel Tracking Data‟, Marine Policy, 2015.

LRIT derived Speed Over Ground [kn] distributions change over the 5 years:

a progressive reduction of speed + the second peak (around 18 knots) disappears mid-2012

Higher speed: - reduces risk of piracy attacks - raises fuel costs

29

Track and speed profile of a trawler showing three clusters of velocities corresponding to in port, fishing (position highlighted in yellow) and steaming behaviours. Vespe M., Gibin M., Alessandrini A., Natale F. Mazzarella F., Osio G., „Mapping EU

fishing activities using ship tracking data‟, Journal of Maps, 2016.

Track and speed profile of vessels

30

1) Raw data 2) Fishing Activities

4) Real time fishing matching 3) Density of Fishing Activities

Mazzarella F., Vespe M., Damalas D., Osio G.: „Discovering Vessel Activities at Sea using AIS Data: Mapping of Fishing Footprints‟, Proc. 17th Int. Conf. on Information Fusion, 2014.

Knowledge Discovery: fishing activities

31 Vespe M., Gibin M., Alessandrini A., Natale F. Mazzarella F., Osio G., „Mapping EU fishing activities using ship tracking data‟, Journal of Maps, 2016.

EU trawlers Fishing intensity

Coverage

32 10 October 2016 32

Trajectories cleaning and splitting

Port activity and trade indicators

33 10 October 2016 33

CB-SMOT

Port activity and trade indicators

34 10 October 2016 34

DB-SCAN

Port activity and trade indicators

35 10 October 2016 35

Concave-hull and buffer

Port activity and trade indicators

36 10 October 2016 36

Semantic annotation

Port activity and trade indicators

37

Weekly Distribution

Port of Gioia Tauro Port of Genoa

Port activity and trade indicators

38

Time Distribution In/Out

Port of Gioia Tauro Port of Genoa

Port activity and trade indicators

39

Port of Genoa: link to Mediterranean Ports

Port activity and trade indicators

40

Port of Genoa: link to Mediterranean Ports

Port activity and trade indicators

41

Port of Genoa: link to Mediterranean Ports

Port activity and trade indicators

42

Conclusions

• Vessel tracking data can be used to understand,

map and quantify activities at sea, including how

they change over time

• It is possible to turn large amount of data into

elements that can be useful to policy makers

• The knowledge it is extracted directly from the

data, using data mining and knowledge discovery

techniques

43

Stay in touch

JRC Science Hub:

ec.europa.eu/jrc

Twitter and Facebook:

@EU_ScienceHub

LinkedIn:

european-commission-joint-research-centre

YouTube:

JRC Audiovisuals

Vimeo:

Science@EC

Thank you