Sea-land segmentation via hierarchical region merging and edge directed graph cut

Abstract

Separating an optical remote sensing image into sea  and land areas is very challenging yet of great importance to the coastline extraction and subsequent object detection. In this paper, we propose a hierarchical region merging approach to automatically extract the sea area  and employ edge directed graph cut (GC) to accomplish the final segmentation. Firstly, an image is segmented into superpixels and a graph-based merging method is employed to extract the maximum area of sea region (MASR). Then the non-connected sea regions are identified by measuring the distance between their superpixels and the MASR. When modelling the pairwise term in GC, we incorporate edge information between neighboring superpixels to reduce under--segmentation.  Experimental results on a set of challenging images demonstrate the effectiveness of our method by comparing it with the state-of-the-art approaches.