Improving the geocoding of persistent scatterers at facades using lattice topology information

Current applications of Synthetic Aperture Radar (SAR) data and methods like the Persistent Scatterer Interferometry (PSI) require very accurate three-dimensional coordinates of any imaged object to associate it to its equivalent in the real world. Transforming the radar system's inherent coordinates to Cartesian world coordinates, known as geocoding, is prone to errors on the one hand but crucial for the interpretation and further usage of the data. We present an approach to improve the geocoding by exploiting lattice topology information which is generated by grouping of Persistent Scatterer in the SAR geometry. In urban areas a SAR scene is mainly characterized by temporally stable and dominant scatterers, so called Persistent Scatterers (PS). These very characteristic reflectors are often induced by window corners or balconies at facades. Since those structures are most often regularly aligned, PS also appear in regular patterns. One novel application of PSI is to exploit this regularity and aggregate single PS to lattices in order to obtain more detailed knowledge about them. One possibility to unveil the physical nature of PS is to match them to their equivalent in oblique optical images. The matching can take place in the coordinate system of the optical image or in three-dimensional world coordinates. In both cases, a very accurate geocoding is required. In this paper, we show an approach for improving the geocoding by performing a linear adjustment which contributes to the goal of gaining a better understanding of Persistent Scatterers.