Phase Ambiguity Resolution For Stacked Radar

Phase ambiguity resolution is one of the main problems in geodetic techniques involving electromagnetic phase measurements. For repeat-pass radar interferometry, the phase difference between two resolution cells in an interferogram usually reflects a topographic height difference, line-of-sight surface deformation, and atmospheric delay, or a combination of the three. Conventional techniques for two-dimensional phase ambiguity resolution, commonly referred to as `phase unwrapping', apply pathfollowing integration methods, using heuristic assumptions on the maximum phase gradient between adjacent resolution cells. If several SAR acquisitions are available, constraints can be introduced to estimate phase differences between two resolution cells, irrespective of the phase behavior of resolution cells in between these observations. This is advantageous, i.e., when areas of the data are very noisy and only a few coherent resolution cells can be identified. We can describe the interferometric phase observations as a linear function of topographic height (depending on the geometric satellite configuration) , surface deformation (dependent of the time interval between the observations), and the integer ambiguity parameters, regarding the atmospheric signal as a component of the stochastic model. Hence, the problem is to resolve the real-valued unknowns and the integer ambiguities. In this paper it is demonstrated how the integer leastsquares principle can be applied for this problem, and how a priori estimates on the success rates of the method can be given based on the satellite acquisition characteristics