基于Landsat系列数据地表温度反演算法对比分析——以齐齐哈尔市辖区为例

Algorithms Comparison of Land Surface Temperature Retrieval from Landsat Series Data:A Case Study in Qiqihar,China

Landsat热红外系列数据是地表温度反演的一项重要数据源。以齐齐哈尔市辖区为研究区域,基于2002、2008和2016年Landsat TM/ETM+/TIRS系列数据,分别采用单窗算法(MW算法)、单通道算法(SC算法)和辐射传输方程法(RTE算法)进行地表温度反演及对比分析,并利用MODIS地表温度产品对反演结果进行精度验证。结果表明:(1)基于Landsat系列数据,3种算法反演得到的地表温度的空间分布状况一致,总体上市区地表温度较高,水体区域温度最低;(2)基于ETM+数据,SC和RTE算法结果一致性较好,其中SC算法精度最高,MW算法在不同地物覆被区误差均较大;(3)MW算法基于TM数据反演精度最高,RTE算法次之,SC算法较差;(4)基于Landsat 8TIRS数据,SC算法精度最高,RTE算法误差较大。

Land Surface Temperature（LST）is considered to be one of the significant indicators of urban en- vironment analysis.Landsat thermal infrared series data is an important data source for retrieving surface temperature.In this paper,the thermal infrared band of the Landsat data in 2002,2008 and 2016 were used to retrieve LST by three different algorithms in municipal area of Qiqihar,China. These algorithms were the Mono-Window algorithm（MW algorithm）,the Single Channel algorithm（SC algorithm） and the Radiation Transport Equation method（RTE algorithm）.And the results of the retrieval were compared to each other and verified by MODIS surface temperature products.The LST distribution maps were accomplished according to the retrieval results.The results showed that： （1）The spatial distribution of the LST obtained by the retrieval of the Landsat series by the three algorithms is consistent,and the LSTof the urban center is higher and thetemperature of water is the lowest; （2）Based on ETM＋ data,the consistency between SC and RTE algorithm results is good,among which the SC algorithm has the highest precision, and the MW algorithm has large errors in different land cover areas- （3）The retrieval results by MW algorithm based on the TM data has the highest accuracy,RTE algorithm results is second,and the LST form SC algorithm is less consistent with the corresponding MODIS temperature products; （4）Based on the Landsat 8 TIRS da- ta,the SC algorithm has the highest accuracy and the RTE algorithm has a large error.