FCM-Based Image Segmentation of Sea Surface Oil Spills

Oil spills cause severe pollution to the marine environment. Unmanned aerial vehicles (UAVs) equipped with high-resolution photographic equipment can be utilized for monitoring sea surface oil spill conditions. To accurately identify oil spill areas in aerial images under complex circumstances, the potential oil spill regions in the images first need to be segmented. A two-step segmentation algorithm is proposed in this paper: First, corresponding color models are constructed based on the distinct color characteristics of different oils in the YCbCr color space. Then, the Fuzzy C-Means (FCM) algorithm with optimized parameters is employed for segmentation to extract suspicious regions containing oil spills from aerial scenes.