喀斯特城市地表温度遥感反演算法比较  

Comparison of Land Surface Temperature Retrieval Algorithms in Karst City

作　　者：王恺宁 王修信[2,3] 黄凤荣[1] 罗涟玲[2] Wang Kaining;Wang Xiuxin;Huang Fengrong;Luo Lianling(School of Urban and Environmental Sciences,Liaoning Normal University,Dalian 116029,China;College of Computer Science and Information Technology,Guangxi Normal University,Guilin 541004,China;State Key Laboratory of Remote Sensing Science,Beijing Normal University,Beijing 100875,China)

机构地区：[1]辽宁师范大学城市与环境学院,辽宁大连116029 [2]广西师范大学计算机科学与信息工程学院,广西桂林541004 [3]北京师范大学遥感科学国家重点实验室,北京100875

出　　处：《遥感技术与应用》2018年第5期803-810,共8页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(41561008);广西自然科学基金项目(2014GXNSFAA118289);广西高校科学技术研究项目(2013LX020;KY2015LX007)资助

摘　　要：针对喀斯特城市快速扩展所引发的热环境问题,提出喀斯特山峰混合像元比辐射率估算方法,使Landsat 8遥感数据的地表温度反演算法适用于喀斯特城市,利用5种单通道算法和劈窗算法反演地表温度,分析反演精度和敏感性因子。结果表明:在我国南方喀斯特地区大气水分含量较高的情况下,单通道算法比劈窗算法精度更高,Jimenez单通道算法(JSC)和覃志豪单窗算法(QMW)更适用于喀斯特城市地表温度反演,反演值和实测值的误差在1.0℃内。反演地表温度的统计值以JSC算法与QMW算法相近,平均值的差值为0.26℃,标准差的差值为0.01℃,建筑和裸岩温度平均值的差值分别为0.43℃和0.54℃,高于水体和茂密植被;Jimenez劈窗算法与Rozenstein劈窗算法相近,平均值的差值为1.14℃,标准差的差值为0.19℃;Weng单通道算法在劈窗算法与JSC和QMW算法之间。各算法对比辐射率ε较敏感,ε每增加0.01,地表温度反演值误差增加0.4～0.7℃;除QMW算法反演值随近地面气温每增加1.0℃而引入近0.5℃误差外,各算法对近地面气温、大气总水分含量、大气透射率的敏感性相对较低。研究结果可为喀斯特城市热环境监测提供科学依据。Urban land covers have changed greatly with the rapid expansion of Guilin karst city in recent two decades.Some agriculture lands,forest and pools converted to buildings and roads,and some karst hills also entered urban district. However, the vegetation on some karst hills was destroyed and parts of limestone hill body were exposed. High air temperature was felt frequently in summer.Spatial distribution of land surface thermal field was affected by land cover changes directly. Land surface thermal field could be quantita- tively descripted with Land Surface Temperature（LST）.In order to analyze the impact of urban rapid ex- pansion on thermal environment in karst eity,I.STs were derived from Landsat 8 images and five retrieval algorithms with the proposal of the emissivity estimation method in the mixed pixels on karst hills. Then the derived LST results were compared with measurements so that the available retrieving algorithm for karst city was got. Finally, sensitive factors on LST were analyzed. Result shows that LSTs from Single- Channel （SC） algorithms are more accurate than those from Split-Window（SW） algorithms with high at- mospheric moisture content in karst district.The errors are within1.0℃ between LST measurements and retrieval results from Jimenez SC （JSC） and Qin Mono-Window （QMW）. LST statistics derived from JSC and QMW are close with average difference of 0.26℃ and standard deviation difference of 0.01 ℃.Average LST differences of building and bare rock are 0.43 ℃ and 0.54 ℃ respectively,higher than those of water body and dense vegetation. LST statistics from Jimenez SW（JSW） and Rozenstein SW（RSW） are close with average LST difference of 1.14 ℃ and standard deviation difference of 0.19 ℃. LST statistics from Weng SC arebetween those from SWs and those from JSC,QMW.As five algorithms show high sensitivity to emissivity,LST average will change 0.4-0.7 ℃ with 0.01 increment of emissivity.Five algorithms are relatively less sensitive to air temperature, tot

关 键 词：喀斯特城市 遥感 地表温度 反演算法 敏感性分析 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]