应用微波遥感数据的东北地区地表温度反演  

Surface temperature inversion in Northeast China based on AMSR-E passive microwave data

作　　者：满浩然 臧淑英[1] 李苗[1] 张鑫[1] MAN Haoran;ZANG Shuying;LI Miao;ZHANG Xin(Heilongjiang Province Key Laboratory of Geographical Environment Monitoring and Spatial Information Service in Cold Regions,Harbin Normal University,Harbin 150025,China)

机构地区：[1]哈尔滨师范大学寒区地理环境监测与空间信息服务黑龙江省重点实验室,哈尔滨150025

出　　处：《测绘科学》2021年第3期124-132,共9页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41571199,41901072)。

摘　　要：针对无源微波遥感时间分辨率高可以克服云层影响获取地表温度的问题,该文应用AMSR-E微波亮度温度数据,分别选取了基于发射率估计的单通道反演法和多通道线性拟合法反演东北地区地表温度。在原有方法的基础上提出算法改进:对单通道反演法按照植被生长周期在生长季与非生长季分别建立发射率估计方程,探究各微波通道在每种地表覆被类型的反演能力并组合反演精度最高的通道,将微波极化差异指数作为表征发射率参数加入多通道拟合方程。结果显示,获取的地表温度剔除水体和冰雪无效像元后可用性达到100%,改进后的单通道反演法均方根误差由3.58~4.6降低至2.0~3.1,在75%的区域的误差小于2 K;多通道拟合法的最终均方根误差为2.6~3.5,同样有较高精度且只使用微波亮温数据就能获取地表温度。Passive microwave remote sensing with high time resolution can overcome the influence of clouds to obtain surface temperature,which is of great significance for studying the changes of active layers in cold regions.In this paper,AMSR-E microwave brightness and temperature data are used to select the single-channel inversion method based on emissivity estimation and the multi-channel linear fitting method to invert the surface temperature in the northeast region.Based on the original method,the algorithm is improved:the single-channel inversion method is based on the vegetation growth cycle to establish the emissivity estimation equation in the growing season and the non-growth season,to explore the inversion ability of each microwave channel in each surface cover type and combine the channels with the highest inversion accuracy,and add the microwave polarization difference index as the emissivity parameter to the multi-channel fitting equation.The results show that the obtained surface temperature is 100%usable after removing water bodies and ice and snow invalid pixels.The improved single-channel inversion method root mean square error is reduced from 3.58~4.6 to 2.0~3.1,and the error in 75%of the area is less than 2 K;the final root-mean-square error of the multi-channel fitting method is 2.6~3.5,which also has higher accuracy and can obtain the surface temperature only by using the microwave brightness temperature data.

关 键 词：被动微波遥感 微波亮温 地表温度反演 东北地区 

分 类 号：P407[天文地球—大气科学及气象学]