基于通用分裂窗算法和Landsat-8数据的地表温度反演研究  

Land Surface Temperature Retrieval from Landsat-8 Data with the Generalized Split Window Algorithm

作　　者：李珊珊[1] 蒋耿明[1] Li $hanshan,Jiang Gengming(Key Laboratory of Information Science for Electromagnetic Waves, Ministry of Education ,Fudan University ,Shanghai 200433 ,China)

机构地区：[1]复旦大学电磁波信息科学教育部重点实验室,上海200433

出　　处：《遥感技术与应用》2018年第2期284-295,共12页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(41271012;61331020)

摘　　要：基于通用分裂窗算法和Landsat-8数据开展地表温度(Land Surface Temperature,LST)反演研究。首先,使用MODTRAN和SeeBor V5数据库开展辐射传输模拟,建立LST与Landsat-8热红外传感器(Thermal InfraRed Sensor,TIRS)第10(10.9μm)和11波段(12.0μm)观测亮温、地表比辐射率(Land Surface Emissivities,LSEs)、观测角度和总可降水量(Total Precipitable Water,TPW)之间的关系。其次,对模拟数据进行分组,用多元线性回归求解算法系数,并进行敏感性分析。然后,从经过交叉辐射定标的晴空Landsat-8数据反演得到LST,其中LSEs从Landsat-8陆地成像仪(Operational Land Imager,OLI)数据估算,TPW从欧洲中尺度天气预报中心再分析数据提取。最后,用Terra卫星中分辨率成像光谱仪的LST/比辐射率产品(MOD11_L2 V5)对反演结果进行验证。该算法能够较精确地从Landsat-8数据反演LST,主要误差源自LSEs和TPW的不确定性,对LSEs和TPW纠正前后的误差分别为-0.64±0.81K和0.10±0.68K。This work addresses the LST retrieval from Landsat-8 data with the generalized split-window algorithm.Firstly,radiative transfer modeling experiment is conducted using MODTRAN 4.0,fed with SeeBor V5 atmospheric profile database to build a data set of LST related to brightness temperatures in the bands 10 and 11 of Thermal Infrared Sensor（TIRS）on Landsat-8,Land Surface Emissivities（LSEs）,viewing zenith angle and Total Precipitable Water（TPW）.Secondly,based on the modeling data set,the unknown coefficients of the generalized split-window algorithm are obtained,and the algorithm sensitivity is analyzed.Then,LSTs are derived from the inter-calibrated and clear sky Landsat-8 data with the generalized split-window algorithm,in which LSEs are estimated from Landsat-8 Operational Land Imager（OLI）data,and TPWs are extracted from the European Centre for Medium-range Weather Forecasts（ECMWF）reanalysis data.Finally,the results are validated with the Moderate resolution Imaging Spectroradiometer（MODIS）LST/LSE product（MOD11_L2 V5）.The results show that the generalized split window algorithm developed in this work can accurately retrieve LST from the Landsat-8 data,and the error is mainly come from the uncertainty of LSEs and TPW.Before and after correction of LSEs and TPW,the LST errors in this work are,respectively,-0.64 ± 0.81 Kand 0.10 ± 0.68 Kagainst the MOD11_L2 V5 product.

关 键 词：Landsat-8数据 辐射传输模拟实验 地表比辐射率 通用分裂窗算法 地表温度反演 交叉精度验证 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]