利用LANDSAT/TM热红外通道反演地表温度的三种方法比较

Comparison between three different methods of retrieving surface temperature from Landsat TM thermal infrared band

利用北京2004年1月28日、4月1日、4月17日、5月19日和7月6日过境的5景LANDSAT/TM影像和实测探空数据,分别运用大气辐射传输模型、覃志豪(2001)单窗算法和Jimёnez-Mu?oz&Sobrino(2003)单窗算法反演北京城区地表温度。通过对反演结果进行对比分析,结果表明,覃志豪(2001)单窗算法与基于探空数据的辐射传输方程法结果具有较好的一致性。提出在没有实时探空数据情况,对只有一个热红外通道的LANDSAT/TM数据源采用覃志豪(2001)单窗算法反演地表温度,精度是可以接受的。

Surface temperature has an important applying value in the research fields of climate change, vegetation ecology, environmental monitoring, urban heat-island effect, etc., and is also an important input parameter in the monitoring and RS models of surface flux （including sensible heat flux, latent heat flux and CO2 flux）, soil moisture content, yield estimation of crops, moisture stress of crops, etc. It is currently one of the means to derive spatial distribution of large-area surface temperature by applying satellite thermal infrared sensors, and how to retrieve the data of surface temperature from thermal infrared RS data is paid great attention to. In this paper, based on the atmospheric profile data and 5 scenes of Landsat TM thermal infrared band （TM6） images of Beijing, which were obtained on January 28, April 1, April 17, May 19 and July 6, 2004 separately, the corresponding data of surface temperature were retrieved with three different methods （thermal radiance transfer equation, mono-window algorithm carried out by Qin Zhihao in 2001, mono-window algorithm carried out by JimSnez-Munoz ＆ Sobrino in 2003）. The retrieved data are compared and analyzed. The results show that the data retrieved by thermal radiance transfer equation are highly similar to that retrieved by mono-window algorithm by Qin Zhihao. Therefore, when the in situ atmospheric profile data are unavailable, the mono-window algorithm by Qin Zhihao can be used to retrieve the data of surface temperature from one thermal infrared band of Landsat TM images, and the precision of the retrieved data of surface temperature is acceptable.