Land surface temperature retrieval based on TM8 thermal infrared data and its application

The land surface temperature （LST）is a key factor widely applied in agricultural drought monitoring, crop yield estimation model and it has an important application in drought remote sensing monitoring. In this paper, we proposed a new split-window algorithm flowchart and derived new parameters to retrieve LST from Landsat 8 satellite according to operation land imager (OLI)and the thermal infrared sensor (TIRS)band characteristics. The atmospheric transmittance and surface emissivity were computed and analyzed in detail. Selecting MODIS temperature product and meteorological stations in the same region as reference, this paper analyzed the result and evaluated the accuracy. The results showed that RMSE was 0.3845 from LST scatter and LST error ranged from -0.6℃ to 0.6℃between retrieval and product. Analyzing the statistical values from meteorological stations, the mean temperatures of product stations and retrieved by algorithm and observations from meteorological stations were 21.61℃, 21.19℃and 21.38℃respectively. RMSD on the high temperature region from meteorological stations was 0.742and the error was less than 1℃. The method of applying example can quickly and efficiently retrieve LST parameters for agricultural drought monitoring, which got better retrieval LST and improved the accuracy agricultural drought monitoring model.