A marine oil spill detection framework considering special disturbances using Sentinel-1 data in the Suez Canal

The Suez Canal is a crucial international waterway due to its strategic location. The significant traffic flow not only stimulates economic development along the coast but also leads to a high frequency of oil spill accidents, which negatively impact the ecosystem and natural resources. Synthetic aperture radar (SAR) is an important remote sensing technology for monitoring oil spills, offering all-day and all-weather capabilities. However, special disturbances (SD) caused by imaging conditions, sensor parameters, and other factors can affect image quality, reducing the accuracy and efficiency of oil spill detection. To mitigate the negative impact of SD, an original oil spill detection framework was developed, based on the analysis of these disturbances, to detect oil spills at the northern entrance of the Suez Canal from 2015 to 2019. The framework included an advantageous featureset with SD adaptability and designs a classifier, Boosting Random Support Vector Machine (BRSVM), which combines a boosting strategy with Support Vector Machine (SVM). The study found that the superiority of the featureset was pivotal in oil spill detection. The classification accuracy and F-1 score achieved by BRSVM were 94.72 % and 95.33 %, respectively, outperforming other algorithms in functionality. These results indicate that the proposed framework holds significant potential for applications requiring large-scale, automated oil spill detection.