摘要：In recent years, oil spills in coastal regions have received a lot of public concern for its strong impact on the coastal ecological system. Synthetic aperture radar (SAR) is regarded as one of the most suitable sensors for oil spill monitoring for its wide-area and all-day all-weather surveillance capabilities. However, due to its special imaging mechanism, multiplicative speckle noise and dark patches caused by other physical phenomena always affect the accuracy of oil spill detection. In this work, an oil spill detection method based on dual-threshold segmentation and support vector machine was proposed. Experiments on SAR images illustrated the effectiveness of the proposed method in detecting and tracing oil spill from SAR images.
