Recently, a research team from the Hefei Institutes of Physical Science of the Chinese Academy of Sciences developed a high-accuracy land surface temperature (LST) retrieval algorithm based on spaceborne full-spectrum imaging spectrometer data. The new method significantly improves the accuracy and robustness of LST retrieval under complex surface and atmospheric conditions.
The related research results were published in Science of Remote Sensing.
Land surface temperature is a key parameter for understanding surface–atmosphere energy exchange processes. However, conventional LST retrieval algorithms often suffer from limited accuracy when applied to surfaces with strong spectral contrast or under complex atmospheric conditions, restricting the operational application of high-resolution satellite data.
In this study, the research team proposed a Maximum Emissivity Approximation–based Optimized Smoothing for Temperature and Emissivity Separation (MEAOSTES) algorithm. By combining an improved split-window model with an optimized smoothing approach and introducing strategies for more reliable surface emissivity estimation, the method enables a more coordinated retrieval of land surface temperature and surface properties.
The new algorithm was compared with several widely used land surface temperature retrieval methods. The results show that it performs well over surfaces with strong spectral differences while maintaining stable accuracy over more uniform surfaces. Further analysis indicates that retrieval performance is mainly influenced by satellite radiance quality, with atmospheric conditions also playing an important role.
This study provides a reliable technical framework for the operational application of high-resolution land surface temperature retrieval from spaceborne imaging.
Precise fitted curve results of emissivity and surface-leaving brightness temperature for different surface types via the MEAOSTES. (a) asphalt roofing shingle；(b) dolomitic marble；(c) brown silty loam；(d)grassland (Image by HUANG Honglian)
LST retrieval results. (a) GF-5B true-color composite image；(b)MODISLST prooduct；(c) OSTES；(d) MEAOSTES (Image by HUANG Honglian)
