Detection and Classification of Man-Made Offshore Objects
ProjectsCompany Profile
Contact
Detection and Classification of Man-Made Offshore Objects
The DeMarine sub-project DEKO focuses on the development of services for the automatic detection and classification of ships and offshore artificial objects on the basis of Earth Observation data acquired by the German remote sensing satellites TerraSAR-X and RapidEye. The objectives are 1st the development of reliable detection algorithms and 2nd the definition of effective and customised service concepts.
With the globalisation, ship traffic is continuously rising, demanding for enhanced traffic monitoring to enable safe shipping. In addition, illegal maritime activities and threats have been increasing significantly in recent years calling for enlarged maritime surveillance. As today’s monitoring systems like coastal RADAR and terrestrial AIS have a limited range of 30 to 40 km, only a small part of the entire exclusive economic zone can be continuously supervised. Therefore, the benefit of the complementary use of Earth observation satellites for maritime surveillance has been recognised.
Definition of Service Concepts
Based on user consultations different service concepts have been analysed with regard to their potential to satisfy existing user needs. Amongst these concepts, a centralised service, being directly operated in the satellite ground receiving station, showed the most advantages: it enables information delivery in near-realtime due to the possibility of optimised image processing and information transfer. Moreover, a multi-mission concept integrating multiple high spatial resolution radar satellites, is recommended being the prerequisite for a frequent, reliable, weather and daytime independent monitoring service.
Development of algorithms for TerraSAR-X and RapidEye
Large satellite imagery is evaluated in 2 subsequent steps: (1) Screening (detection of object candidates with the sea area as background) and (2) Signature analysis (object-background separation, object parameter estimation and classification). While the first step has to be optimized towards high object detection reliability and fast data throughput, the 2nd step is responsible for reducing false alarms and extraction of correct object characteristics, e.g. object position, size, and heading. By combining SAR and optical measurements, a more robust object description is achieved as well as a higher monitoring frequency.
Ship detection and profile of Radar backscatter
(Click to on the picture to enlarge it)
Analysis of wind farms for anomaly detection
(Click on the picture to enlarge it)
Ship monitoring in the North Sea
As test sites for algorithm development the Strait of Dover and the harbour of Bremerhaven have been used due to dense traffic of varying ship types (ferries, tankers, freighters, etc.) and the availability of coastal radar and terrestrial AIS (ship identification system) for validation purposes. The ship detection algorithm extracts the position and size of ships. In some cases, the ship wake was extracted too, thus enabling to correct the ship position and heading plus to estimate the vessels’ speed. The algorithms produce results in common xml or GIS exchange formats, e.g. kml. Therefore, they can easily be imported into commercial GIS SW as well as be visualised by common desktop browsers or Google Earth. The geo-position allows the overlay of background data (e.g. coastal shore line, AIS data) and the object parameters are accessible interactively.
Surveillance of wind farms
For inspections and surveillance purposes, static objects like wind farms can be continuously monitored and compared to reference data (change detection).
In the frame of the DeMarine-DEKO project two wind farms in the front of Rostock and Copenhagen were subject of tests. For demonstration purposes the TerraSAR-X data were modified to simulate damages on the wind turbine. In a first step, the developed surveillance system detects and measures the objects being present in a predefined area containing a wind turbine. In a second step, the object shape is compared to reference data of the wind farm. In case of discrepancies between both data sets an error file is generated. This information could be further used to alarm maintenance units.
Project Description:
DeMarine Security is a German GMES gateway project concentrating on the development of satellite based ship detection services. Besides the technical aspects, the project aims to strengthen the position of German users and service providers within the frame of the European “Global Monitoring of Environment and Security” (GMES) program.
The sub-project DEKO concentrates on the following topics:
- Development of detection and classification algorithms for TerraSAR-X and RapidEye images
- Definition of effective, customised ship & object detection service concepts
Contact
Dr. Stephane Estable
Tel.: +49(0)421 5395574
Fax: +49 (0)421 539285574
stephane [dot] estable [at] astrium [dot] eads [dot] net
www.astrium.eads.net
Astrium GmbH
Airbus Alle 1
28199 Bremen, Germany
Dr. Stefan Knabe
Tel.: +49 (0)7545 83177
Fax: +49 (0)7545 81337
stefan [dot] knabe [at] infoterra-global [dot] com
www.infoterra.de
Infoterra GmbH
Claude-Dornier-Strasse
88090 Immenstaad, Germany 
Gunter Saur
Tel.: +49 (0)721 6091230
Fax: +49 (0)721 6091413
guenter [dot] saur [at] iosb [dot] fraunhofer [dot] de
www.iosb.fraunhofer.de
Fraunhofer-Institut fur Optronik, Systemtechnik und Bildauswertung
Fraunhoferstrase 1
76131 Karlsruhe, Germany
The DeMarine-Security DEKO project is sponsored by the German Space Agency, Grant No. 50 EE 0810.
