利用TM6数据反演陆地表面温度新算法研究

A Study on the New Algorithms for Retrieving Land Surface Temperature Based on TM6 Data

陆地表面温度(LST)反演一直是热红外遥感研究中的一大难题。虽然TM 6数据具有较高的空间分辨率(120 m),但由于只有一个热通道,要得到地表真实温度,原来需要利用辐射传输方程的方法,实时资料的缺乏限制了该方法的应用。因而由TM 6数据得到的通常都是星上亮度温度,而星上亮度温度与实际地表温度差距较大,因此,其反演的温度精度不高。而单窗算法和普适性单通道算法的提出为从TM 6数据较高精度地反演陆地表面温度提供了可能。分析和研究了这两个新的单通道温度反演算法,并针对北京市的实际情况,利用2005年5月6日的TM数据对北京市的陆地表面温度进行了反演,并用实地测量数据进行了比较验证。结果表明这两种温度反演算法都取得了较高的精度,它们的rm sd值分别为1.38°和2.18°。

Land surface temperature （LST） retrieval has been a key issue in the thermal infrared remote sensing research area. TM6 data has higher spatial resolution with a pixel size of 120 meters, and it is the only one thermal channel in a Landsat 5 TM scene. For this reason, the previous method of retrieving land surface temperature from TM6 data was based on radiative transfer equation, which seemed to be not practical due to the scarcity of in situ radiosounding data. Thus, in most cases, only at-satellite brightness temperature was obtained from TM6 data. However, there were large differences existed between satellite brightness temperature and the land surface temperature, which resulted in the not good precision of land surface temperature retrieval. While the approaches from the mono-window algorithm developed by Qin et al. and the generalized single-channel algorithm proposed by Jiménez-Munoz and Sobrino make it possible to retrieve land surface temperature from TM6 data with a higher precision. In this paper the two new algorithms were used to retrieve land surface temperature of Beijing city by performing tests on the TM6 data acquired on May 6 2005 respectively. The comparison results between the land surface temperature measured in situ and the retrieved by the algorithms have shown that the significant precisions from both the algorithms are obtained with the root mean square deviation （RMSD） values of 1.38 degrees and 2.18 degrees respectively.