利用Sentinel-2影像进行洱海水体水域提取的指数方法分析研究

Analysis of Index Method for Extracting Water Area of Erhai Lake Using Sentinel-2 Image

针对高原复杂地形中湖泊水体的快速精准提取问题,基于Sentinel-2数据,以云南高原湖泊洱海及周边区域为研究对象,分别使用归一化差异水体指数(NDWI)、改进的归一化差异水体指数(MNDWI)、阴影水体指数(SWI)、自动水体提取指数(AWEIsh)和基于线性判别分析水体指数(WI2015)提取水体信息;后利用同期Geogle Earth影像目视解译结果进行精度验证。结果表明,5种指数均能成功提取研究区内大部分水体;SWI法总体提取精度最高,达93.2%,NDWI法和WI2015法次之,MNDWI法和AWEIsh法总体提取精度最低。

To address the problem of rapid and accurate extraction of lake water bodies in complex plateau terrain, the study is based on Sentinel-2 MSI data, taking Erhai Lake and its surrounding areas on the plateau in Yunnan as the research object. The normalized difference water index(NDWI),the modified normalized difference water index(MNDWI),the shadow water index(SWI),the automatic water extraction index(AWEIsh) and the water index(WI2015) were used respectively to extract the distribution information of water bodies in the plateau. The accuracy is verified by using the results of visual interpretation of the same period of Geogle earth images. The results show that all five water body indices can successfully extract most of the water bodies in the study area. The SWI method has the highest overall extraction accuracy of 93.2%,followed by the NDWI method and WI2015method. The MNDWI method and AWEIsh method have the lowest accuracy of all extractions.