GF-3与Sentinel-1洪灾淹没信息提取

Extraction of Flood Submergence Information from GF-3 and Sentinel-1

以斯里兰卡2017-05-25开始的洪灾为研究对象,利用灾中和灾后GF-3和Sentinel-1获取的多期数据,采用符合研究区域特征的最优提取方法,提取了多期洪水淹没范围。并利用同时期的光学卫星哨兵2号(Sentinel-2)观测数据、研究区高程数据以及当地官方公布的受灾面积进行了结果验证分析。结果显示Sentinel-1比GF-3噪点更多且误提取了更多山区阴影,因而导致Sentinel提取洪水面积偏大,GF-3提取的洪水面积更接近于官方公布的受淹面积。本研究通过斯里兰卡的GF-3和Sentinel-1应用对比,显示了GF-3在洪涝灾害信息提取中具有较为准确的精度,对GF-3在洪水灾害动态监测的应用精度提供了应用实例。另外一方面,通过GF-3与Sentinel-1协同校正应用,可以得到更多的洪水过程信息,为防灾减灾领域提供可靠的数据支持。

In this paper,taking the flood disaster in Sri Lanka,which started on May 25,2017,as the research object,based on the multi-period data obtained by GF-3 and Sentinel-1 during and after the disaster,we used the optimal extraction method that according to the regional characteristics to extract the multi-period flood submergence range.At the same time,we validated and analyzed the results of Sentinel-2 observation data,the elevation data of the research area and the disaster area officially announced by the local authorities.The results show that Sentinel-1 has more noise points than GF-3,and more mountain shadows are extracted by mistake,which leads to larger flood area extracted by Sentinel-l,and the GF-3 extracted flood area is closer to the official flooded area.By comparing the application of GF-3 and Sentinel-1 in Sri Lanka,it is found that GF-3 has accurate accuracy in flood disaster information extraction.On the other hand,through the application of GF-3 and Sentinel-1,more flood process information can be obtained and reliable data support can be provided for disaster prevention and mitigation.