An iterative technique to estimate baseline parameters in SAR interferometry

Baseline, the shortest distance between two satellite trajectories, is an important parameter in Synthetic Aperture Radar Interferometry as it has direct impact on the accuracy of Digital Elevation Models and surface displacement measurements due to land subsidence, earthquake etc. Exact and simplified models for the estimation of baseline parameters for a flat earth as well as a curved earth were developed. The results are compared with the existing simplified models and found to be in good agreement. The baseline parameters are estimated using a Newton-Raphson method followed by a least square technique by simulating 90 ground control points which were well distributed all over the satellite scene (heights - phases). The interferogram phases were corrupted with various types and levels of phase noise. An iterative technique for the improvement of baseline estimation by minimizing the phase noise is illustrated with examples. A transformation function is developed to reduce the phase noise of the rest of the image for generating a DEM.