Summary: | This masters thesis is a part of the nordic collaboration project NorSEN, whose goal is to improve the environmental surveillance in the regions north of the Arctic circle. This implies the build-up of a data-acquisition network that can be used for tracing environmental changes over the years. The data from these sites can also be used as a reference for satellite image validation and to improve the atmospheric correction procedures of satellite imagery. Atmospheric correction is the process where an image taken at satellite level is recreated as it would have looked like on the ground. The signal is distorted along its way through the atmosphere by the means of absorption and scattering. Swedens contribution to this project is a raft with on board spectrometers placed on the lake Torneträsk near Abisko in the summer of 2005. The lake is surrounded by mountains which are often covered with clouds that changes the light conditions. This poses a problem for the atmospheric correction since these types of situations are more difficult to model. The lake is also often exposed to rough weather with strong winds, leading to high waves. When this occurs, it is much harder to calculate the reflectance from the data. Severe rainfalls can also lead to sediments in the lake which changes the lakes reflectance. One part of the project is therefore to evaluate if Torneträsk is a good reference surface. This part will however not be covered by the report. The masters thesis has been devoted to the evaluation and changing of the satellite signal simulation program 6S, which is also able to apply atmospheric correction to known satellite signals. Pictures from the instrument MERIS on board ESAs Earth observation satellite ENVISAT has been available within the project. Some assumptions in 6S that has proved to be a problem is that of horizontal homogeneity and the plane-parallel assumption. Since the clouds and mountains need to be modeled, an inhomogeneous model must be implemented. This can be done by introducing a horizontal ...
|