Adaptive Noise Reduction of InSAR Data Based on Anisotropic Diffusion Models and Their Applications to Phase Unwrapping

摘要

In this paper, the relationship between noise and phase unwrapping is analyzed. Advanced global and local denoising techniques have been exploited to serve the key step in InSAR application-phase unwrapping. Phase noise, abrupt slope, layover and shadow are main factors to degrade the quality of phase patterns. Some noise reduction techniques are presented, and their comparisons are discussed. Two main phase unwrapping algorithm are especially described. We present the improved minimum norm models with nonlinear regularization techniques. Although, higher computation load has to be paid currently, high quality unwrapped phase insensitive to noises makes the model useful. The vector digital filtering has been employed to the interferogram images, which aim at reducing the residues. Numerical examples show significantly lower computation cost.