1 August 1992 Comparison of two retrieval methods for ground reflectance
Author Affiliations +
Abstract
A comparison of two retrieval methods is presented to calculate ground reflectance from Landsat Thematic Mapper satellite data. The first method is based on a modified two-stream approximation to simulate the radiative transfer above an inhomogeneous surface. The atmosphere is parametrized by the optical depth and the single scattering albedo. The theory of fractal geometry is employed to compute the structure measures of a scene, from which the ground variability is estimated. By a linear regression, the ground variability can be related to the atmospheric optical depth. The independent second method is based on model ATCOR (including LOWTRAN-7). Here, a priori knowledge is used (shape of spectral reflectance curve for vegetation, water, bare soil) to determine the unknown atmospheric parameters like optical depth and type of aerosol (single scattering albedo). The adjacency effect, which describes the influence of atmospheric crosstalk in modifying the radiances of adjacent fields of different reflectance, is taken into account by both procedures. Typically, deviations between both methods are up to 2% in reflectance for low to medium reflection (< 30%) targets and up to 4% for high reflectance (> 40%) targets of Landsat imagery. In view of the independent approaches, this level of agreement in retrieved ground reflectance is fairly good. The new method is particularly valuable if no a-priori knowledge is available and if the scene has a large dynamic range of spatial frequencies.
© (1992) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Carmen Tornow, Carmen Tornow, Rudolf Richter, Rudolf Richter, } "Comparison of two retrieval methods for ground reflectance", Proc. SPIE 1688, Atmospheric Propagation and Remote Sensing, (1 August 1992); doi: 10.1117/12.137887; https://doi.org/10.1117/12.137887
PROCEEDINGS
12 PAGES


SHARE
Back to Top