Title :
A System for Automatic Identification of Oil Spill in ENVISAT ASAR Images
Author :
Tian, Wei ; Shao, Yun ; Wang, Shiang
Author_Institution :
State Key Lab. of Remote Sensing Sci., Jointly Sponsored by the Inst. of Remote Sensing Applic. of Chinese Acad. of Sci. & Beijing Normal Univ., Beijing
Abstract :
An automatic oil spill identification system was introduced in this paper. It was designed to work independently to detect oil spills in ENVISAT ASAR level 1b images (of WS, IMP and APP mode). The absolute calibration, geometric correction without any GCPs, dark patches segmentation based on wind force levels, dark patches feature extraction, drifty trend prediction and identification confidence evaluation are the essential components of the system. The system works well for its accuracy of 85% in discriminating between oil spills and look-alikes in the case of 60 scenes of examined images.
Keywords :
feature extraction; geophysical techniques; geophysics computing; image segmentation; marine pollution; oil pollution; remote sensing by radar; wind; APP mode; ENVISAT ASAR level 1b images; IMP mode; WS mode; absolute calibration; automatic oil spill identification system; dark patches feature extraction; dark patches segmentation; drifty trend prediction; geometric correction; identification confidence evaluation; wind force levels; Calibration; Equations; Feature extraction; Image segmentation; Petroleum; Radar scattering; Remote sensing; Sea surface; Surface discharges; Surface morphology; Envisat ASAR; Oil spill; drift; identification system; wind;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 2008. IGARSS 2008. IEEE International
Conference_Location :
Boston, MA
Print_ISBN :
978-1-4244-2807-6
Electronic_ISBN :
978-1-4244-2808-3
DOI :
10.1109/IGARSS.2008.4779621