A sea-land segmentation algorithm based on Gray Smoothness Ratio

Sea-land segmentation on infrared remote sensing image has attracted a lot of research interest in the civil-military applications, such as ocean target detection, coastline extraction and so on. Unlike panchromatic satellite images, infrared remote sensing images usually have a low signal to noise ratio and suffer from weather condition like clouds and mists, so existing sea-land segmentation algorithms of infrared image are difficult to achieve satisfactory precision and efficiency. By analyzing the gray and texture feature of sea and land, this paper puts forward a sea-land segmentation algorithm based on a feature descriptor named Gray Smoothness Ratio (GSR) and a hole filling method to overcome the above problems. First, the image is segmented roughly by the SVM classifier which is based on GSR descriptor. Second, Otsu algorithm is adopted for segmentation of boundary area between sea and land. Finally, a coordinate projection method is applied to fill the isolated holes. Extensive experiments demonstrate that, in comparison with the existing relevant state-of-the-art approaches, our segmentation scheme has smaller computation complexity and better segmentation effect.