IEEE Geoscience and Remote Sensing Letters, Год журнала: 2024, Номер 21, С. 1 - 5

Опубликована: Янв. 1, 2024

Urban land surface temperature (ULST) is one of the core parameters in monitoring urban thermal environment, which has received extensive attention several study and application areas. Thermal infrared (TIR) remote sensing technology can efficiently observe large-scale radiance information a critical approach used to obtain ULST quickly. Traditional LST retrieval algorithms are conducted using classical transfer equation (RTE) based on assumption that flat, may be challenging hold for complex landscapes. Moreover, with improvement spatial resolution images, influence caused by geometric structure will more obvious. Various models have been proposed successfully applied TIR images tens meters resolutions, such as Landsat, ECOSTRESS, Gaofen-5. Current airborne sensors ultra-high (sub-meter). In this paper, ensemble learning method model (UHURT), new algorithm developed estimate directly from observed brightness temperature. The applies images. It end-to-end advantage not relying atmospheric or emissivity, known traditional algorithms, thus avoiding limitations due lack available input data. Validation results simulation dataset showed higher theoretical accuracy than split-window algorithm. As sky view factor (SVF) decreases, becomes pronounced, growing 0.149 K (SVF = 1.0) 1.085 0.25). image also indicated (RMSE 2.093 K) accurate those SW 2.490 K), correlation between resultant error building density lower, accurately reduce effect better.

Язык: Английский