Atmospheric Correction
Monitoring the earth’s surface from satellites is hampered by the atmosphere. Atmospheric absorption and scattering processes influence the transport of solar radiation along its path from the sun to the surface of the earth, as well as from the surface of the earth to the sensor.
In order to obtain precise information about the earth’s surface and be able to compare optical remote sensing data sets spatially and temporally, the influence of the atmosphere must be corrected. This correction is fundamental whenever biophysical and geophysical parameters like leaf area index, proportion of photosynthetically active radiation, land use, emissivity or land surface temperature are being operationally derived in order to model and analyze the geosphere-biosphere-atmosphere system, as at DFD.
For a number of satellite-borne instruments DFD provides highly precise correction for the disruptive influences of clouds, aerosols and trace gases. Two new methodologies have been developed both of which enjoy patent protection, ATCOR and DurchBlick.
ATCOR calculates the reflectivity and, as needed, the surface emissivity and temperature, whereby aerosol content and water vapor column content are derived from the recorded scene, assuming the required spectral channels are available. Topographical effects (degree of slope and orientation) are taken into account, hazy areas are corrected, and the effect of thin/semithick cirrus clouds can be eliminated in hyperspectral data. In addition, areas with cloud shadow can be interactively corrected. A detailed description is available on the Internet (see the DLR-ATCOR link at left). The version for satellite sensors is limited to instruments with high spatial resolution and ignores the influence of variations in the viewing angle within one scene. The version for aerial sensors takes into account the influence of the viewing angle. For push broom hyperspectral sensors the central wavelength of a channel varies within one detector row from pixel to pixel (the so-called “smile” effect), and this instrument-specific effect must be taken into account in the atmospheric correction.
DurchBlick was developed in order to obtain a description of the state of the atmosphere which is as up-to-date and precise as possible above a satellite scene requiring atmospheric correction. For this purpose, the spatial and temporal remote sensing products relevant for the atmosphere (ozone, water vapor, clouds, aerosols) are automatically compiled and taken into account in the correction procedure.
The figure shows a Landsat-TM mosaic composed of scenes recorded on different dates. The scene on the left has not been corrected and the edges where the scenes overlap is especially obvious. Comparing the reflectivity in both parts of the image, for example to study land use, would thus lead to misinterpretations.
The scene on the right is more homogenous; the ground reflectivity has been reconstituted, and precise quantitative analysis of the entire area is now possible, thanks to atmospheric correction.
Downloads
