基于空间内插的云下地表温度估计及精度分析

Estimation and Error Analysis of Land Surface Temperature Under the Cloud Based on Spatial Interpolation

云覆盖下地表温度的估计,是热红外遥感研究的一个前沿问题。地表温度具有较高的空间相关性,理论上可通过空间内插值方法来估计云覆盖下的地表温度,从而消除云对地表温度产品的应用限制。本文基于普通克里格插值方法和规则样条函数插值方法提出了云下地表温度的估计方法,并通过ArcGIS软件进行应用试验,分析了这一方法在不同云覆盖范围内的估计精度。实验结果表明,当云覆盖小于9km2时,城市地区的估计误差小于1℃,而农田地区为0.6℃。因此,本文提出的云覆盖下地表温度的估计方法在一定的云覆盖范围内是可用的。

Estimation of the land surface temperature in cloud-covered area is the frontier of thermal remote sensing research.Due to the land surface temperature has high spatial correlation,we can theoretically estimate the land surface temperature in cloud-covered area through spatial interpolation to eliminate the obstacle of cloud cover on applications of land surface temperature products.On the basis of the ordinary Kriging interpolation and completely regularized function interpolation,we develop an approach to estimate the land surface temperature in cloud-covered area,and use ArcGIS to perform the experiments to examine the accuracy of the approach for various scales of cloud cover.The results show that the accuracy is within 1℃ and 0.6℃ when the cloud covered area is less than 9km2 in city and in country,respectively.Therefore we can conclude that the proposed approach is applicable to estimate the land surface temperature in cloud-covered areas for practical applications.