MODIS数据在几种典型地表发射率光谱特性中的适用性分析

Applicability analysis of MODIS data on several typical surface emissivity spectrum characteristics

发射率是识别地物的重要参数,受季节和下垫面的影响发射率波谱特性曲线呈一定变化规律.以沙漠、土壤、冻土、湖泊以及渤海海域为研究区域,利用MOD021KM反演发射率,并与MOD09产品所得发射率在可见光与近红外波段进行对比及交叉验证,探究两种算法在地表识别中的适用性.结果表明:在乌兰布和沙漠,二者相关系数达到0.8900,表现最好,且均方根误差RMSE值为0.0512,是最低的;祁连山冻土区、大同土林、太湖1、太湖2、辽东湾、渤海中部、渤海湾以及莱州湾相关系数R2分别为0.8646、0.6480、0.4282、0.3970、0.3320、0.1265、0.1446、0.1703,RMSE分别为0.0563、0.0707、0.2150、0.2140、0.1187、01481、0.2149、0.1150.总体来看,两种算法反演所得发射率在沙漠、土壤、冻土区适用性较好,能够用来识别典型地表,在湖面和海面适应性则相对较低,可以用来定性分析发射率的空间分布.

As an important parameter to determine the ground characteristics,the emissivity is affected by the season and the underlying surface,and its characteristic curve shows a certain change law.In this paper,we use desert,soil,frozen ground,lakes and the Bohai Sea as the study areas,and use MOD021KM inverse emissivity to compare and cross-validate with the MOD09 product in the visible and near-infrared bands to investigate the applicability of the two algorithms for surface identification.The results show that in the Ulan Buh Desert,the best performance was achieved with a correlation coefficient of 0.8900 and the lowest RMSE value of 0.0512.The correlation coefficients R2 for Qilian Mountain permafrost region,Datong soil forest,Taihu Lake 1,Taihu Lake 2,Liaodong Bay,the center area of Bohai,Bohai Bay and Laizhou Bay are:0.8646,0.6480,0.4282,0.3970,0.3320,0.1265,0.1446,0.1703,and RMSE are:0.0563,0.0707,0.2150,0.2140,0.1187,0.1481,0.2149,0.1150.In general,the emissivity obtained from the above two algorithms is more applicable in desert,soil and frozen ground and can be used to identify typical surfaces,while it is less applicable in lake and sea surfaces and can only be used to qualitatively analyze the spatial distribution of emissivity.