数字农科院2.0

A robust framework for accurate land surface temperature retrieval: Integrating split-window into knowledge-guided machine learning approach

文献类型： 外文期刊

作者： Yuanliang Cheng;Hua Wu;Zhao Liang Li;Frank M. Göttsche;Xingxing Zhang;Xiujuan Li;Huanyu Zhang;Yitao Li

作者机构：

关键词： Knowledge-guided machine learning;Land surface temperature;Retrieval;Split-window;Validation

期刊名称： Remote Sensing of Environment

ISSN： 0034-4257

年卷期： 2025 年 318 卷

页码：

收录情况： SCIE（2025版） ; ; EI（2025版）

摘要： Land surface temperature (LST) is a crucial parameter of the surface-atmosphere system, driving the water and heat exchange between the surface and the atmosphere. However, existing LST retrieval methods are highly sensitive to input errors. This study proposed a robust framework for retrieving LST, termed SW-NN, which integrates the physics-based Split-Window (SW) algorithm with a data-driven Neural Network (NN). The framework comprises of two main components: (1) a NN model that estimates SW coefficients as functions of key parameters such as brightness temperature (BT), water vapor content (WVC), land surface emissivity (LSE), and viewing zenith angle (VZA); and (2) a SW model that applies these coefficients to compute LST based on physical principles. By embedding the SW algorithm into the NN's loss function, this integrated design ensures that physical relationships guide the training process. The training data for the framework were generated by simulating satellite BT under a broad range of atmospheric and surface conditions using a radiative transfer model. To address the challenge of input errors, the proposed framework incorporates Gaussian noise into the training data to simulate realistic uncertainties in BT, WVC, and LSE. Specifically, noise with standard deviations of 0.05 K, 10 % of the WVC value, and 0.01 was added to BT, WVC, and LSE, respectively. Simulation analysis on an independent test set demonstrates that the proposed framework achieves a theoretical root-mean-square error (RMSE) of 0.60 K under the noise strategy, outperforming standalone NN and SW models. Sensitivity analysis, conducted using the same noise strategy applied during training, indicates that input errors affect LST retrieval by approximately 0.20 K, significantly enhancing the model's generalization and robustness. The proposed framework was also applied to MODIS data to retrieve LST, which was directly validated against global measurements from fifteen sites. Additionally, the proposed framework was compared with the NN method, the generalized split-window (GSW) method (MOD11 LST), and the Temperature Emissivity Separation (TES) method (MOD21 LST). The results showed that the proposed framework achieved an RMSE of 1.99 K, outperforming the NN method (RMSE = 2.08 K) and the GSW method (RMSE = 2.52 K), and performing comparably to the TES method (RMSE = 2.03 K). Further analysis in arid areas, where LSE accuracy is relatively lower, showed that the proposed framework improved the RMSE to 1.94 K compared to MOD11 LST, which had an RMSE of 3.02 K, utilizing the same LSE inputs. The proposed framework leverages the SW model's mechanism and the NN model's nonlinear fitting capability. It also demonstrates high robustness against input error, particularly LSE error. In summary, the proposed framework achieves robust and accurate LST retrieval, offering interpretability and a significant improvement over existing methods designed for sensors with two thermal infrared channels, especially in challenging environmental conditions. The proposed framework is available at https://github.com/YL-Cheng-IGSNRR/SW-NN.

分类号：

  • 相关文献

[1]Retrieval of land surface temperature from FY3D MERSI-II based on re-fitting Split Window Algorithm. Zhang Dejun,Yang Shiqi,Sun Liang,Liu Xiaoran,Tang Shihao,Zhu Hao,Ye Qinyu,Zhang Xinyu. 2022

[2]An Algorithm for Retrieving Land Surface Temperatures Using VIIRS Data in Combination with Multi-Sensors. Xia, Lang,Mao, Kebiao,Ma, Ying,Zhao, Fen,Qin, Zhihao,Jiang, Lipeng,Shen, Xinyi. 2014

[3]An Operational Split-Window Algorithm for Land Surface Temperature Estimation From Chinese FY-3C VIRR Data. Li, Jia-Hao,Li, Zhao-Liang,Fan, Jinlong,Liu, Niantang. 2024

[4]Evaluating Spatial Representativeness Across Multiple Scales for a Comprehensive Ground Validation Network Using Landsat Land Surface Temperature Data and Random Forest. He, Xuanwei,Liu, Xiangyang,Ru, Chen,Deng, Xiangyi,Zhao, Ruoyi,Yu, Wenping. 2025

[5]Estimating the crop leaf area index using hyperspectral remote sensing. Liu Ke,Zhou Qing-bo,Wu Wen-bin,Tang Hua-jun,Liu Ke,Zhou Qing-bo,Wu Wen-bin,Tang Hua-jun,Wu Wen-bin,Xia Tian.

[6]LAND SURFACE EMISSIVITY ESTIMATION FROM SATELLITE DATA WITH MACHINE LEARNING. Xiu Juan Li,Hua Wu,Zhao Liang Li,Yong Gang Qian,Si Bo Duan. 2021

[7]A normal form for synchronous land surface temperature and emissivity retrieval using deep learning coupled physical and statistical methods. Han Wang,Kebiao Mao,Jiancheng Shi,Sayed M. Bateni,Dorjsuren Altantuya,Bayarsaikhan Sainbuyan,Yuhai Bao. 2024

[8]An Improved Spatial Downscaling Procedure for TRMM 3B43 Precipitation Product Using Geographically Weighted Regression. Chen, Cheng,Zhao, Shuhe,Chen, Cheng,Zhao, Shuhe,Chen, Cheng,Zhao, Shuhe,Duan, Zheng,Qin, Zhihao. 2015

[9]Comparison of Six Methylation Methods for Fatty Acid Determination in Yak Bone Using Gas Chromatography. Jia, Wei,Wu, Ting,Yu, Qunli,Jia, Wei,Liu, Wenyuan,Mi, Si,Zhang, Chunhui,Li, Xia,Wu, Ting,Yu, Qunli,Liu, Wenyuan,Yu, Qunli.

[10]Simultaneous Determination of 70 Pesticide Residues in Coffees by Gas Chromatography-Mass Spectrometry. Yang, Xin,Wang, Jing,Yang, Xin,Zhang, Hua,Xu, Dechang,Qiu, Jianwei,Wang, ZhenYu,Cao, WeiQiang. 2010

[11]Comparisons of MODIS LAI products and LAI estimates derived from Landsat TM. Yang, Peng,Chen, Zhongxin,Zhou, Qingbo,Zha, Yan,Wu, Wenbin,Shibasaki, Ryosuke. 2006

[12]REMOTE SENSING PRODUCTS VALIDATION ACTIVITY AND OBSERVATION NETWORK IN CHINA. Jin, Rui,Li, Xin,Che, Tao,Jin, Rui,Li, Xin,Che, Tao,Ma, Mingguo,Ge, Yong,Xiao, Qing,Zhao, Kai,Xin, Xiaoping. 2016

[13]Multiresidue method for determination of 88 pesticides in berry fruits using solid-phase extraction and gas chromatography-mass spectrometry: Determination of 88 pesticides in berries using SPE and GC-MS. Yang, X.,Zhang, H.,Zhang, Y. C.,Dong, A. J.,Zhao, H. T.,Cui, J.,Liu, Y.,Yang, X.,Wang, J.,Sun, C. H..

[14]Assessment of the MODIS LAI Product Using Ground Measurement Data and HJ-1A/1B Imagery in the Meadow Steppe of Hulunber, China. Li, Zhenwang,Tang, Huan,Xin, Xiaoping,Zhang, Baohui,Wang, Dongliang. 2014

[15]Validation of a Crop Yield and CO2 Fixation Model over Southeast Asia by Carbon Partitioning in Grain Plants. Kaneko, Daijiro,Yang, Peng,Kumakura, Toshiro. 2011

[16]Multiscale Validation of the 8-day MOD16 Evapotranspiration Product Using Flux Data Collected in China. Tang, Ronglin,Li, Zhao-Liang,Wu, Hua,Tang, Bo-Hui,Shao, Kun,Li, Zhao-Liang,Zhang, Li,Zhou, Guoqing. 2015

[17]Equations to predict methane emissions from cows fed at maintenance energy level in pasture-based systems. Stergiadis, Sokratis,Zou, Caixia,Chen, Xianjiang,Wills, David,Yan, Tianhai,Stergiadis, Sokratis,Zou, Caixia,Chen, Xianjiang,Allen, Michelle.

[18]VALIDATION OF MODIS FAPAR PRODUCTS IN HULUNBER GRASSLAND OF CHINA. Li, Gang,Xin, Xiaoping,Zhang, Hongbin,Li, Gang,Xin, Xiaoping,Zhang, Hongbin,Wang, Daolong,Zhang, Hua,Xin, Xiaoping,Zhang, Hongbin,Liu, Shimin. 2010

[19]Developing a photosynthetic sterility model to estimate CO2 fixation through the crop yield in Asia with the aid of MODIS data. Kaneko, Daijiro,Yeh, P. J. -F.,Kumakura, Toshiro,Yang, Peng. 2010

[20]Testing the modified Rothamsted Carbon Model for paddy soils against the results from long-term experiments in southern China. Xu, Minggang,Jiang, Guiying,Shirato, Yasuhito,Yagasaki, Yasumi,Huang, Qinghai,Li, Zuzhang,Nie, Jun,Shi, Xiaojun. 2013

作者其他论文 更多>>