基于组合不同裂窗算法反演黄河流域地表温度研究

Land Surface Temperature Calculation in Yellow River Basin by Combination of Split Window Algorithms

以黄河流域为研究区域,利用MODIS遥感数据,根据下垫面的水分状况和土地覆盖类型对整个流域分别进行分区.在整个流域不分区、水分状况分区和土地覆盖分区双因子分区2种情形下,对比了7种常用的地表温度遥感反演裂窗算法的结果,在分析每种反演算法适用性的基础上,针对不同分区单元分别选择效果最好的算法进行组合来进行流域地表温度的反演.反演结果和MODIS的NASA的温度产品相比,效果更为理想.

This paper chooses the Yellow River basin as research region, compared and analyzed 7 commonly used split-window algorithims for Land Surface Temperature （LST） calculation with Moderate Resolution Imaging Spectrometer （MODIS）data. Considering the soil moisture and land use/cover heterogeneity, the whole basin was divided into small patched in terms of land use/cover types and acrid conditions. With the 7 split window algorithms, the LST was derived for the whole basin and all the small patched within the basin. With the NASA LST product as reference data, the LST results were evaluated. Based on the analysis of the performances of the seven algorithims, a best algorithim combination scheme was derived by adopting the algorithim with best performance for each patch. With this combination, LST map for the whole basin was produced and showed better effects than NASA LST product.