Land surface temperature estimation from remote sensing data - a case study in Kun Ming city

Land surface temperature is an important indicator of the earth's environmental analysis. Land surface emissivity is one of the key factors of remote sensing retrieving land surface temperature. In this paper, using the Landsat5 TM6 band data acquired in March 24, 2009 and taking Kunming city as the study area, we use NDVI threshold method and empirical formula model to calculate the land surface emissivity of Kunming city. In this paper, we apply universal single-channel algorithms to invert the surface temperature of Kunming. The results showed that: The results of land surface emissivity were calculated by NDVI threshold and empirical formula model are significantly related with MODIS surface emissivity products. The real temperature and the surface temperature inverted by two surface emissivity are little difference and they are in the accuracy of the control range. By empirical formula model calculated surface emissivity of the surface temperature anomalies in urban areas, exceeding the real temperature 2 °C - 4 °C, does not accurately reflect the surface temperature. And the surface temperature inverted by the result of NDVI threshold is closer to the real surface temperature. This method could reflect the temperature under different land covers more accurately. Therefore, land cover such good in Kunming city area, with NDVI threshold method calculated surface emissivity to invert surface temperature is more appropriately.