Abstract: Landsat TM data has a thermal band (TM6) to monitor the thermal radiance and surface temperature of the earth. For a long time, retrieval of surface temperature from TM6 data is done through the method called atmospheric correction, which requires the estimate of atmospheric thermal radiance and absorption. The computation is complicated and the accuracy is strongly subjected to the estimate, hence has few applications in real world. In this paper, a new method with simple computation and high accuracy is developed on the basis of thermal radiance transfer equation of the ground. The impacts of both atmosphere and the emitted ground on the thermal radiance transfer of the ground are directly involved into the proposed algorithm, which requires 3 essential parameters for the retrieval: emissivity, transmittance and atmospheric average temperature. Validation indicates that the accuracy of the algorithm is high up to <0.4℃ for the estimate of the essential parameters with no errors and <1.1℃ for the estimate moderate errors. Since the method is applicable for the remote sensing data with only one thermal band, it has been termed as mono window algorithm.

Key words: land surface temperature, thermal radiance, Landsat TM, mono window algorithm, brightness temperature