Chapter 6 - Oil Spill Detection Using Space-Borne Sentinel-1 SAR Imagery

Abstract

Operational monitoring of oil discharges using synthetic aperture radar (SAR) images contributes significantly toward fewer occurrences of oil spills on the sea surface. Until now commercial satellites like ERS-1/2, ENVISAT, Radarsat-1/2, TerraSAR-X, and COSMO-SkyMed were used for this purpose. However, their data availability was challenging and required specific agreement with data providers along with acquisition planning for each scene. European Space Agency's Sentinel-1 satellite is about to change the role of remotely sensed data used for the observation of marine environments. For the first time, Level-1 SAR images are freely available to the scientific community and to service providers. In addition, depending on the time requirements, Sentinel-1 images can be processed and be made available in Near Real Time (NRT) and therefore be used for operational oil spill detection. The wide availability of Sentinel-1 data encourages the development and operation of the semiautomatic NRT algorithm for the detection of oil spills since the large amount of data requires a sufficiently automatic analysis. The present chapter deals with operational aspects of oil spill detection using Sentinel-1A SAR data as case study. A well-established oil spill detection algorithm, modified in order to work with new Sentinel products, is presented. Moreover, known issues of the traditional image processing for oil spill detection are discussed and relevant solutions are suggested for further development in the new “era of Sentinels.” The present work confirms the capability of Sentinel-1A data to detect oil spills in the different modes and proves the ability to adapt known algorithms to the new dataset. Dark areas can be adequately classified using an object-based approach.