基于Landsat 8遥感数据的天津市地表温度反演

以Landsat 8遥感数据为数据源,进行天津市地表温度反演研究。首先采用单通道算法反演地表温度,并利用均值标准差法进行温度分级。然后建立不同温度等级面积比例的估算模型。再通过随机样点,从不同温度等级和土地覆盖类型2个角度,分别建立并比较不同类样点的地表温度与各指数的拟合模型。结果表明:次高温区域面积比例与人口密度、人均GDP都具有较高的决定系数;地表温度与NDVI、BAEM的二元线性回归决定系数高于地表温度与单一指数的决定系数;将样点分类后,低温点与MNDBI的决定系数高于其他温度等级样点,水域和植被样点与各种指数的决定系数高于其他地物类型样点。

Land Surface Temperature Inversion in Tianjin City based on Landsat 8 Remote Sensing Data

In this paper, Landsat 8 remote sensing data was used as data source to study the surface temperature inversion in Tianjin. The land surface temperature was retrieved using the single-channel algorithm. A mean standard deviation method was used to grade temperature. The estimating models of the area proportion of different temperature-level were built respectively. Then, random sampling points were chosen and classified from two perspectives including temperature levels and land cover types. Land surface temperature of each type sampling points was fitted respectively with each index, and fitting models were compared. The results showed that:The area proportion of sub-high temperature areas had relatively high determination coefficient with population density and per capita GDP. The binary linear regression of surface temperature fitting with NDVI and BAEM had a higher determination coefficient than that for the surface temperature fitting with a single index. After classification of sample points, the determination coefficient between low temperature samples and MNDBI was higher than those for other temperature-level sampling points. The determination coefficient between water area samples and each kind of indexes and the determination coefficient between vegetation samples and each kind of indexes were both higher than those for other land-cover type sampling points.