L1-norm regularization based SAR phase unwrapping: initial result

Abstract

Synthetic aperture radar interferometry (InSAR) exploits two complex-valued images of the same target area to extract interferogram. However, due to the periodicity of trigonometric functions, the interferometric phase is only measured modulo 2π and wrapped between [-π,π). Therefore, phase unwrapping is necessary in InSAR processing for absolute terrain phase. The phase unwrapping method has a great influence on the accuracy of phase recovery. The phase difference between wrapped phase and unwrapped phase, usually called phase ambiguity cycle, can be estimated as an integer multiple of . The phase unwrapping problem is essentially the problem of estimating the phase ambiguity cycle. Since the phase ambiguity cycle is equal to zero in most cases, its estimation can be regarded as a sparse signal recovery problem. In this paper, an L₁-norm regularization-based SAR phase unwrapping method is introduced to estimate the phase ambiguity cycle and obtain the high-quality absolute phase. In the proposed method, we firstly construct the phase unwrapping model based on the relationship between wrapped and absolute phase. Then the phase ambiguity cycle will be recovered by solving an L₁-norm regularization problem. Experimental results based on simulated and real data validate the proposed method.