Geometric features of LOS data derived by SAR PSI technologies and the three-dimensional data fusion

Abstract

At present, the satellite SAR persistent scatterer interferometry can already estimate surface changes with a near to 1 mm theoretical precision limit. However, the ascending and descending acquisitions of available SAR services cannot provide three-dimensional changes routinely, though the slow deformation processes are basically three-dimensional (3D). In this paper the geometric features of ascending and descending SAR data and possible fusion with geodetic data are summarised. All the geometric equations are introduced, which are necessary to derive the two characteristic changes in the observation plain defined by ascending and descending unit vectors pointing to SAR satellite positions. The unambiguously derivable characteristic changes can be transformed into vertical and east changes, but they may be biased by possible north displacement. The geometric features of symmetric and asymmetric acquisitions are also investigated. Monte-Carlo simulation is used to investigate the precision of two estimated components. It is experienced that the precisions are not sensitive to one degree standard deviations of positional angles. The Gauss–Markov model of least square adjustment method is used to derive only the statistical properties of reasonable data fusion which can contribute to the 3D applications. Although complementary satellites, which are already proposed in the literature, could provide precise autonomous solutions, in the practise GNSS and levelling data can be used for direct data fusion. Whereas, even errorless levelled high changes cannot contribute to the proper estimation of northern components, GNSS derived changes are the best candidates, which can be interpolated or measured directly. Moreover, these two techniques can properly compensate the weaknesses of each other. The interferometric SAR techniques are not sensitive enough to the north changes, but can contribute to the precision of height estimation, which are the weakest components of the GNSS technique. This statement is valid if the standard deviations of combined data are comparable. For test computations the geometric parameters of available Sentinel-1A images are used, which cover the area of the Széchenyi István Geophysical Observatory, where experimental integrated geodetic benchmark is located combining ascending and descending backscatterers with the possibility of the GNSS, gravimetric and traditional geodetic measurements, as well.