Incorporating Surface Emissivity into a Thermal Atmospheric Correction
Incorporating Surface Emissivity into a Thermal Atmospheric Correction
复制标题
将表面发射率纳入热大气校正
DOI:
--
复制
发表时间:
2002
期刊:
影响因子:
--
通讯作者:
R. Gillies
中科院分区:
文献类型:
--
作者:
N. Brunsell;R. Gillies
as NDVI threshold methods. The issue of incorporating surface emissivity into a thermal Previous research has focused on the incorporation of atmospheric correction of thermal remotely sensed data is emissivity assuming an atmospheric correction has already addressed. The Normalized Difference Vegetation Index (NDVI) been performed. However, the literature is lacking in a clear is derived using atmospherically corrected surface reflectance description of a methodology on how to perform an atmos- values, which is subsequently used to estimate the percent of pheric correction in thethermal portion of the spectrum for sat- vegetation cover. Surface emissivity is approximated by a linear ellite remote sensing with emissivity incorporated as part of interpolation between a minimum bare soil emissivity and a the algorithm. Accurate determination of radiometric surface maximum vegetation value of emissivity. An application of temperatures requires information concerning the state of the the method to an image over the Southern Great Plains 1997 atmosphere at the time of satellite overpass. (SGP97) Hydrology Experiment for the Advanced Very High An alternate method that is widely applied for atmos- Resolution Radiometer (AVHRR) band 4 demonstrates pheric corrections of temperature is the split window (e.g., temperature corrections up to 8C, with a mean correction of McMillan (1975) and Waltonet al. (1990)). This method has the 3.7C. The temperatures within the fully vegetated pixels show benefit of notrequiring prior knowledge ofatmospheric condi- good agreement with air temperature measurements at the tions (Sobrino et al., 1994) as the transmittance effects are esti- time of satellite overpass. mated as a function of brightness temperatures in band 5 and band 4 based on many assumptions. However, split window