Review
Oil spill detection by satellite remote sensing

https://doi.org/10.1016/j.rse.2004.11.015Get rights and content

Abstract

This paper presents the state of the art for oil spill detection in the world oceans. We discuss different satellite sensors and oil spill detectability under varying conditions. In particular, we concentrate on the use of manual and automatic approaches to discriminate between oil slicks and look-alikes based on pattern recognition. We conclude with a discussion of suggestions for further research with respect to oil spill detection systems.

Introduction

Oil spills on the sea surface are seen relatively often. Observed oil spills correlate very well with the major shipping routes (e.g. in the Southeast Asian Waters (Lu, 2003, Lu et al., 1999), and in the Yellow and East China Sea (Ivanov et al., 2002) and commonly appear in connection with offshore installations (e.g. in the North Sea (Espedal & Johannessen, 2000). Annually, 48% of the oil pollution in the oceans are fuels and 29% are crude oil. Tanker accidents contribute with only 5% of all pollution entering into the sea (Fingas, 2001). After analysing 190 ERS-11 SAR images of the Mediterranean Sea, Pavlakis et al. (1996) found that “deliberate” oil spills appear with considerably higher frequency than oil spills corresponding to reported ship accidents. According to the European Space Agency (1998), 45% of the oil pollution comes from operative discharges from ships. When taking into account how frequent such spillages occur, controlled regular oil spills can be a much greater threat to the marine environment and the ecosystem than larger oil spill accidents like the Prestige tanker (carrying >77,000 ton of fuel oil (Oceanides Web-site, 2004) accident at Galice, northwest coast of Spain in 2002. The impact of not monitoring oil spills is presently unknown, but the main environmental impact is assumed to be seabirds mistakenly landing on them and the damage to the coastal ecology as spills hit the beach (Shepherd, 2004). Simecek-Beatty and Clemente-Colón (2004) describes how oiled birds lead to the use of SAR for locating a sunken vessel leaking oil.

Active microwave sensors like Synthetic Aperture Radar (SAR) capture two-dimensional images. The image brightness is a reflection of the microwave backscattering properties of the surface. SAR deployed on satellites is today an important tool in oil spill monitoring due to its wide area coverage and day and night all-weather capabilities.

Satellite-based oil pollution monitoring capabilities in the Norwegian waters were demonstrated in the early 1990s by using images from the ERS-1 satellite (e.g. Bern et al., 1992b, Skøelv & Wahl, 1993, Wahl et al., 1994b). A demonstrator system based on ERS for the Spanish coast was presented by Martinez and Moreno (1996). Today, RADARSAT-1 and ENVISAT are the two main providers of satellite SAR images for oil spill monitoring.

Access to an increased amount of SAR images means a growing workload on the operators at analysis centres. In addition, recent research shows that even if the operators go through extensive training to learn manual oil spill detection they can detect different slicks and give them different confidence levels (Indregard et al., 2004). Algorithms for automatic detection that can help in screening the images and prioritising the alarms will be of great benefit. Research on this field has been ongoing for more than a decade, and this paper reviews various methods for satellite-based oil spill detection in the marine environment.

As SAR is just one of many remote sensing sensors available an evaluation of the applicability of other satellite sensors for oil spill monitoring is included as well. Most studies done on airborne remote sensing techniques are excluded. For a review of airborne sensor technology for oil spill observation see Goodman (1994). The detectability of oil spills in SAR images are discussed, in terms of wind conditions, sensor characteristics and ambiguities caused by other phenomena than oil spills. Finally, our emphasis is on methodology and algorithms for oil spill detection in spaceborne SAR imagery.

Section snippets

Satellite sensors for oil spill detection

Microwaves are commonly used for ocean pollution monitoring by remote sensing. They are often preferred to optical sensors due to the all-weather and all-day capabilities, and examples of SAR-equipped satellites are presented in Table 1. Mainly spaceborne instruments are covered here, but airborne Side-Looking Airborne Radar (SLAR) is another possibility. SLAR is an older but less expensive technology than SAR, but SAR has greater range and resolution (Fingas & Brown, 1997). Airborne

Detectability of oil spills in SAR images

Oil slicks dampen the Bragg waves (wavelength of a few cm) on the ocean surface and reduce the radar backscatter coefficient.2

Methodology for oil spill detection in SAR images

We distinguish between manual approaches and automatic algorithms for oil spill detection. Detection of oil spills can be divided in (Indregard et al., 2004):

  • Detection of suspected slicks.

  • Manual verification of the slicks (oil/look-alike) and assignment of confidence levels.

This section addresses issues regarding the design of oil spill detection systems.

Automatic techniques for oil spill detection in SAR images

Several of the published papers on oil spill algorithms for SAR images (e.g. Fiscella et al., 2000, Frate et al., 2000, Solberg et al., 1999) describe a structure comparable with the one in Fig. 4.

The importance of the wind vector was emphasised in Section 3.2, and Salvatori et al. (2003) include two additional steps of wind direction estimation and wind speed calculation. Manual wind estimation was included by Solberg et al. (1999). SAR image calibration, land masking, speckle reduction and

Conclusion and suggestions for further work

Synthetic aperture radar is the most applicable spaceborne sensor for operational oil spill detection, mostly because of its all-weather/all-day detection capabilities and wide coverage. It can operate from light wind to wind speeds up to 12–14 m/s, but the maximum wind speed for oil slick detection depends on oil type and age. Sensors operating in wide swath mode with a spatial resolution of 50–150 m are found to be sufficient and allow covering large ocean areas efficiently.

The largest

Acknowledgements

This work is performed as a part of a PhD study funded by the Norwegian Research Council and the Norwegian Defence Research Establishment. The authors would like to thank the Oceanides project, in particular KSAT, for the SAR scenes used for illustration purposes.

References (83)

  • T.-I. Bern et al.

    Oil spill detection using satellite based SAR: Experience from a field experiment

  • C.M. Bishop

    Neural networks for pattern recognition

    (1995)
  • K.W. Bjerde et al.

    Oil spill detection in SAR imagery

  • J. Canny

    A computational approach to edge detection

    IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI

    (1986)
  • L.Y. Change et al.

    A multiplayer-multiresolution approach to detection of oil slicks using ERS SAR image

  • C.F. Chen et al.

    The use of satellite imagery for monitoring coastal environment in Taiwan

  • J.C. Curlander et al.

    Synthetic aperture radar, systems and signal processing. Wiley series in remote sensing

    (1991)
  • Dokken, S. T. (1995). Optimal bruk av avanserte radarsatellitter. Master's thesis, Institutt for Tekniske Fag, Norges...
  • C. Elachi

    Introduction to the physics and techniques of remote sensing

    (1987)
  • ERS-2 Web-site, (2004). http://earth.esa.int/services/pg/pgerssarpri.xml, accessed 20 October...
  • H. Espedal

    Detection of oil spill and natural film in the marine environment by spaceborne SAR

  • Espedal, H. A. (1998). Detection of oil spill and natural film in the marine environment by spaceborne synthetic...
  • H.A. Espedal et al.

    Detection of oil spills near offshore installations using synthetic aperture radar (SAR)

    International Journal of Remote Sensing

    (2000)
  • H.A. Espedal et al.

    Satellite SAR oil spill detection using wind history information

    International Journal of Remote Sensing

    (1999)
  • European Space Agency

    Oil pollution monitoring

    ESA brochure: ERS and its applications—Marine

    (1998)
  • European Space Agency, (2002). ASAR product handbook. Tech. rep., European Space Agancy-ENVISAT Product Handbook, Issue...
  • M. Fingas

    The basics of oil spill cleanup

    (2001)
  • B. Fiscella et al.

    Oil spill detection using marine SAR images

    International Journal of Remote Sensing

    (2000)
  • J. Fortuny-Guasch

    Improved oil spill detection and classification with polarimetric SAR

  • F.D. Frate et al.

    Neural networks for oil spill detection using ERS-SAR data

    IEEE Transactions on Geoscience and Remote Sensing

    (2000)
  • K.S. Friedman et al.

    GoMEx—an experimental GIS system for the Gulf of Mexico Region using SAR and additional satellite and ancillary data

  • M. Gade et al.

    Measurements of the radar backscattering over different oceanic surface films during the SIR-C/X-SAR campaigns

  • M. Gade et al.

    Marine pollution in European coastal waters monitored by the ERS-2 SAR: a comprehensive statistical analysis

  • M. Gade et al.

    On the detectability of marine oil pollution in European marginal waters by means of ERS SAR imagery

  • A. Gasull et al.

    Oil spills detection in SAR images using mathematical morphology

  • F. Girard-Ardhuin et al.

    Oil slick detection by SAR imagery: potential and limitation

  • R.M. Haralick

    Statistical and structural approaches to texture

    Proceedings of the IEEE

    (1979)
  • Hovland, H. A., Johannessen, J. A. (1994). Norwegian surface slick report, Final report to Norwegian Defence Research...
  • H.A. Hovland et al.

    Slick detection in SAR images

  • C. Hu et al.

    MODIS detects oil spills in Lake Maracaibo, Venezuela

    EOS, Transactions, American Geophysical Union

    (2003)
  • M.-K. Hu

    Visual pattern recognition by moment invariants

    IEEE Transactions on Information Theory

    (1962)
  • Cited by (939)

    • Natural oil seep systems in the Aegean Sea

      2024, Marine and Petroleum Geology
    View all citing articles on Scopus
    View full text