SAR Image Oil Spill Detection Based on Maximum Entropy  Threshold Segmentation

In recent years, marine oil spills have occurred frequently, and the use of SAR remote sensing images for oil spill detection is of great significance. In this paper, a SAR image oil spill detection algorithm based on maximum entropy threshold segmentation was proposed. The algorithm is simple and suitable for on-board platform, which can achieve efficient and accurate detection. Since SAR imaging has inherent speckle noise, filtering is first required to suppress noise. The land area existing in the image will affect the oil spill detection. Depend on latitude and longitude information, we mask the land area before performing a sliding window operation. Inside the window, we determine the optimal segmentation threshold for background and oil spill based on maximum entropy. Finally, The small block area generated by the segmentation is filtered, and the adjacent regions are merged according to the distance information. The GF-3 satellite images were used for verification test, and compared with other algorithms, the algorithm can meet the requirements of real-time and accuracy of remote sensing image detection.