An Analytic Solution Method for Retrieving Land Surface Temperature from Remotely Sensed Thermal Infrared Imagery

This paper investigates land surface temperature (LST) retrieval method based on multiband remotely sensed thermal infrared data. In search of quantitative remote sensing retrieval methods for the inversion model of the radiance transfer equation, a quantitative retrieval analytical solution method is proposed, which adds a constraint of the difference between the LST and the effective mean atmospheric temperature based on prior knowledge, and introduces prior knowledge for the second time in deriving an analytical solution of the retrieved LST. The verification tests with both simulated data and real MODIS data show that this analytical solution method is relatively superior to both the split-window algorithm and the iterative retrieval method for the radiance transfer equation in terms of retrieval precision, stability and efficiency; moreover, the new method reveals more general applicability theoretically, and it may be applied to comprehensive processing of thermal infrared imagery recorded by different sensors.