| Summary: | This thesis presents the theoretical basis and system model for interferometric synthetic aperture radar, INSAR, from spaceborne platforms. The model describes how the images are formed and what relevant ground parameters affect the signal. There are several useful ground parameters in spaceborne INSAR data. These are scene coherence and backscattering, ground topography, small surface movements and atmospheric artifacts, which may be used for different kinds of applications. This thesis covers four applications, land-cover characterization (using backscattering and coherence), digital height map generation, forest parameter retrieval (using coherence) and measurements of small sea ice movements. For a correct assessment of the accuracy of each application, the parameters have to be estimated correctly from the INSAR data. It is shown that although some parameter estimates are biased and noisy, they are still useful. Land-cover characterization is possible up to 75% accuracy for four land-cover classes for three different test areas. Height maps may have an accuracy down to 3 m for three-day repeat-pass INSAR and slightly higher for longer time intervals over a forested test area. The heights of the trees also add to the height measured over the forest, leading to a "forest height bias" which can vary widely for different INSAR images. Forest parameter retrieval is possible for the case of stem volume measurements with an accuracy down to 30 m3/ha. The small-scale rheology (small movements) in Baltic low-saline sea ice was possible to evaluate and measure with an unprecedented accuracy with INSAR images.
|
|---|