基于Sentinel-1数据的水体信息提取方法研究

Study on new method for water area information extraction based on Sentinel-1 data

SAR影像对于水体和地表形变具有很好的辨识性,因此常用来进行水体识别、土壤湿度反演和地表形变检测研究与应用。利用载有C波段合成孔径雷达的Sentinel-1卫星数据对大范围的水体信息进行识别,提出了SDWI(Sentinel-1 Dual-Polarized Water Index)水体信息提取方法。该方法受到NDVI和NDWI方法的启发,结合微波遥感中水体信息在影像中的特点,进一步研究了Sentinel-1双极化数据(VV和VH)之间水体信息提取的关系,以此关系达到增强水体特征的目的,同时消除土壤和植被的存在。分别以Sentinel-1A巢湖区域和Sentinel-1B鄱阳湖区域SAR影像为例来提取水体信息,实验结果表明该方法显著有效,但对影像中阴影的处理是未来研究的难点。

In SAR images,water area and surface deformation have good identification,so the images were often used for water area identification,soil moisture inversion and surface deformation detection. Sentinel-1 satellite data of C-band Synthetic Aperture Radar (SAR) was used to identify a wide range of water area information,the SDWI (Sentinel-1 Dual-Polarized Water Index) water area extraction method was proposed. This method was inspired by NDVI and NDWI methods. In combination with the characteristics of water area information of SAR image in microwave remote sensing,the relationship of water information extraction between Sentinel-1 bipolarization data (VV and VH) was further studied to enhance the characteristics of water area information and eliminate the image of soil and vegetation. Taking the SAR image data of Chaohu Lake region (Sentinel-1 A) and Poyang Lake region (Sentinel-1 B) were taken as examples to extract water area Information. The experimental results show that the SDWI method is effective and the water area information is well extracted. However,it is difficult to deal with shadow in image,which is worthy of research in the future.