基于自动阈值与区域增长的SAR影像海岸线提取算法

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

为实现对海岸线的动态监管和监测,提出了一种自动阈值与区域增长相结合的合成孔径雷达(Synthetic Aperture Radar,SAR)影像海岸线自动提取方法。首先结合Canny边缘检测和最大类间方差法(OTSU方法)实现自动阈值分割提取水陆分界线;在自动阈值提取的基础上,对水域进行开运算和图像分割,将占有较大比例的海上水体作为种子点进行区域增长,获取海上水体,从而去除陆地水体、阴影等误分为海域的部分;之后对区域增长法获取的海陆分割结果进行去噪、填充等形态学优化处理,填补海上船体、风车、岛礁等造成的空洞影像,形成最终的海陆分割结果;最后,使用Canny边缘检测算子获取海岸线。基于江苏海岸带的实验结果表明该方法可以高精度地提取出光滑、连续的海岸线,在SAR影像海岸线提取中具有较强的实用性。

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.