一种基于指数函数的遥感水体信息提取方法

A method extract remote sensing information of water-body based on exponential function

MODIS数据具有高时相分辨率、高光谱分辨率和中等地面分辨率的特征,利用MODIS数据来监测和提取旱情严重地区的水体信息已经成为一个重要的研究方向.本文利用各类地物在MODIS各个波段上的波谱特征差异性进行水体提取,但与以往水体信息提取方法有创新之处在于:本文借助于非线性的指数函数对低反射的水体的反射率进行了拉伸而对高反射的非水体地物的反射率进行了压缩,从而提高了水体在短波红外上的反射率而抑制了非水体在红波段上的反射率,通过波段间的归一化处理进而构建出EXPWI(EXP Water Index)方法 .为了更进一步地提高水体信息提取的自动化程度,本文用ENVI/IDL编写了EXPWI模型.研究表明利用IDL编写的EXPWI法快速实现了中等分辨率遥感影像中水体信息的自动提取,能有效的抑制背景地物的干扰而清晰的识别出湖泊、河道和地表的微弱水体信息,而且能有效降低云对水体信息提取的干扰,更重要的是在抑制大气噪音对提取结果的干扰方面,EXPWI法的效果很好.

MODIS data have the characteristics of high spatial spectral resolution and medium ground surface resolution. The change study of water-bodies using MODIS data to monitor and extract has become an important direction in the remote sensing application. This paper uses all kinds of objects have different reflectances on spectrum characteristic, but this method is difference from normal methods, it use the exponential function to stretch the reflectance of water-bodies and compress the reflectance of non-water-bodies, so we can use the non-linear function to construct the EXPWI（EXP Water Index） . In order to improve the processing speed,we use ENVI/IDL to compile EXPWI method, The results show that it can automatically extract the medium remote sensing of water-bodies and identify the weak water-body information of lakes, rivers, earth surface and effective suppression of interference features in the background. In addition, taking the advantage of the ration computation, the EXPWI method can remove shadow noise from water information without using sophisticated procedures and reducing aerosological influence is one of the excellent results.