基于SLIC的哨兵1号雷达数据水体信息提取

Water body information extraction by Sentinel-1 radar data based on SLIC

针对水域面积较少时水体信息提取困难的问题,提出了以Sentinel-1A SAR为数据源,用简化的SLIC超像素分割算法对图像进行分割,根据相似度将分割后图像进行分层区域合并,将图像二值化得到最终水体信息的方法。将提取的信息与Otsu阈值法提取的水体信息进行对比分析,结果表明:水体提取算法相比Otsu阈值法有效提升了水体信息提取的效率和精度,对于防汛减灾具有重要的科学价值。将提出的方法应用于2019年7月份的株洲洪水监测,取得了较好的效果。

In view of the difficulty of extracting water information of small water body area,this paper puts forward a new method,by which Sentinel-1A SAR is used as the data source,and the simplified superpixel segmentation algorithm SLIC is used to segment the image,the segmented image is merged according to the similarity and binarized to obtain the final water body information.Compared with the water information extracted by Otsu threshold method,the proposed water extraction algorithm effectively improves the efficiency and accuracy of water extraction,and has important scientific value for flood control and disaster reduction.The method is applied in the Zhuzhou flood monitoring in July 2019 and achieves good results.