GAO Siyuan, HUANG Shifeng, SUN Yayong and WANG Hui
Abstract: Synthetic Aperture Radar is with observation characteristics of full time, all day, real time and wide coverage, which has unique advantage in fast monitoring of landslide disaster. Therefore, based on the serious landslide disaster in Maoxian County, Sichuan Province on June 24, 2017, this paper conducts the application research of landslide range extraction based on the Sentinel-1 SAR data in the disaster. First,by analyzing the radar backscattering mechanism of forest area and landslide area,the radar ratio index method is proposed based on VV polarization and VH polarization. Then,combining the two-phase Sentinel-1 SAR images before and during the disaster,the features of backscattering intensity change are investigated,and the rationality of extracting landslide mass by the radar ratio index method is verified. Finally, for the Sentinel-1 SAR data in the disaster, the radar ratio index method is adopted for the rapid extraction of landslide range and the monitoring of landslide disaster. The results show that the Maoxian landslide area is 1.3 km2 and the inclined sliding distance is about 2378m by performing the rapid extraction of Sentinel-1 SAR data. The landslide caused the serious damage to Xinmo village and the houses were completely destructed. The research shows that the radar ratio index can significantly characterize the image features of the landslide mass. This method presents great potential and advantages in landslide extraction and monitoring by Sentinel-1 SAR.
Keywords: Sentinel-1; landslide; Maoxian County; ratio index
