基于MODIS数据的水体识别指数方法的比较研究

Comparison and research on the different index models used in water extraction based on remote sensing data of MODIS

从典型地物光谱曲线分析入手,MODIS数据的第7、2、1波段对植被、陆地、水体区分明显。本文对长三角太湖流域地区,采用MODIS数据以五种常水体识别进行水体信息提取,然后分别从目视解译效果、水体区分度的计算和提取的水体面积三方面来对这五种指数方法进行比较和评价。结果表明,在应用MODIS数据进行水体识别时,比值植被指数(RVI,Ratio Vegetation Index)和混和水体指数(CIWI,Combined Index of Normalized Difference Water Index)模型的目视解译效果较好,且CIWI指数具有水陆区分度较大和水体面积提取精度较高的优点。

According to the discussion and analysis of spectrum profile of typical ground objects, it shows that the 7th, 2nd and 1st bands of MODIS data is useful for the distinction of vegetation,land and water. This paper gives application and comparison among five different index models used in water extraction in the MODIS region data of Taihu River. In the end, we interpret images by eyes and calculate the division degree and area of water body in order to carry out the comparison and evaluation of the five indexes. The result shows that,based on the MODIS data, the index RVI（ Ratio Vegetation Index） and CIWI （Combined Index of Normalized Difference Water Index） have the better image interpretation,and CIWI also has superiority to others in the relative precision in DD and area calculation.