高分一号卫星遥感影像提取冰川堰塞湖水体方法比较

The Comparison of Methods for Extracting Water Body from GlacierBarrier Lake by GF-1 Satellite Remote Sensing Images

快速准确地从遥感影像提取冰川堰塞湖水体信息,是研究冰坝遥感监测与应急监测的核心热点问题。以GF-1卫星遥感影像为主要数据源,采用归一化差分水体指数(NDWI)、改进阴影水体指数(ENDWI)和面向对象(SVM)对克亚吉尔冰川堰塞湖水体进行定量提取。比较分析3种水体判识方法,3种方法均可以提取完整的水体边界,并且抑制了90%以上的非水体信息。NDWI法和ENDWI法可应用于GF-1地表水体提取,能够满足冰坝水体监测与应急监测需求,但面向对象法最适宜GF-1影像的水体信息准确提取。

The rapid and accurate extraction of water body information of the glacier barrier lake from the remote sensing image is the core hot spot of the research on the remote sensing monitoring and emergency monitoring of the ice dam.Taking GF-1 satellite remote sensing image as the main data source,the quantitative extraction of water body information of Keyajir glacier barrier lake was completed by adopting normalized difference water body index(NDWI),improved shadow water body index(ENDWI),and object-oriented method(SVM).Three methods of water extraction were compared and analyzed,the results show that the three methods can extract the complete water boundary and suppress more than 90%of the non-water information.NDWI method and ENDWI method can be applied to GF-1 surface water extraction,which can meet the requirements of ice dam water monitoring and emergency monitoring,but object-oriented method is most suitable for accurate extraction of GF-1 image water information.