利用空间插值法估算云覆盖像元地表温度的可行性研究  

Estimating of Land Surface Temperature under the Cloud Cover with Spatial Interpolation

作　　者：张军[1] 覃志豪[1,2] 刘梅[1] 涂丽丽[1] 周义[1] 杨强[1] 

机构地区：[1]南京大学国际地球系统科学研究所,江苏南京210093 [2]中国农业科学院农业资源与农业区划研究所,北京100081

出　　处：《地理与地理信息科学》2011年第6期45-49,F0003,共6页Geography and Geo-Information Science

基　　金：国家973计划项目(2010CB951500)课题4(2010CB951504)

摘　　要：云覆盖是热红外遥感应用和地表温度遥感定量反演的重要障碍。实验选取河北衡水地区Landsat TM影像,借助GIS的空间分析功能,探索运用反距离加权(IDW)和普通克里金插值(OK)方法近似估算3种典型地表类型(水面、城镇、农田)下的云覆盖像元地表温度的可行性。实验结果显示:当云覆盖区域达15 000个像元(约等于13.5km2)时,两种空间插值方法对城镇的云覆盖像元地表温度估算误差都为1.6°C左右,而农田和水面的估算误差在1°C左右(且水面估算误差小于农田),这与3种典型地表类型的地表温度空间异质性密切相关。因此,当Landsat TM图像中有少块云覆盖而发生地表温度值缺失时,利用空间插值法估算云覆盖像元的地表温度是切实可行的。The existing of the cloud is one of the main factors which are restricted the applications of RS image.In order to resolve the problem,an exploring study on LST estimation with TM data under cloudy cover using the ordinary kriging and inverse distance weighting under the spatial analysis function of geographic information system in Hengshui City,Hebei was tested in this paper.The TM image was divided into three typical land use（city,farm,water）.The results showed that：when the cloud cover area was within 15 000 pixels,the estimating error of the cloud covering land surface temperature in the city was below 1.6°C,in the farm and water were below 1°C,and the whole estimation error trend was water farm city,which was relation with the heterogeneity of land surface temperature of the three typical land patterns.Therefore we can see that the spatial interpolation is a very effective method to estimate the land surface temperature under the cloud cover pixel in the small scope.

关 键 词：云覆盖像元 地表温度 空间插值 估算误差 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]