Coastline extraction algorithm for SAR images based on automatic threshold and regional growth

In order to achieve dynamic supervision and monitor the coastline, SAR image coastline automatic extraction method combined with automatic threshold and regional growth is proposed. First, the combination of the Canny edge detection and OTSU method is applied to realize the automatic threshold extraction of the water and land boundary line. Based on the automatic threshold results, the water area is operated opening and segmentation to obtain the large proportion of sea water body as a seed point for regional growth for getting the sea water, thus the part of the terrestrial water body and shadow was removed. Then, the sea-land segmentation acquired by the regional growth method is denoising, filling morphological optimization processing to fill the void caused by sea hull, windmill and island reef, so the final sea-land segmentation result was finally formed. At last, the Canny edge detection operator was used to obtain the coastline. The experimental results of Jiangsu costal zone show that the smooth and continuous coastlines can be accurately extracted by the proposed method, which has strong practicality in the SAR coastline extraction.