A Method Combining Multilook Processing and Non Linear Filter for Phase Noise Reduction in Interferometric SAR

摘要

In this paper, a method combining mutilook processing and non linear filter for phase reduction in interferometric SAR(INSAR) is presented. Multilook processing is a common method to reduce the phase noise in INSAR and the quality of interferogram increases with the number of looks. But the improvement is not dramatic because multilook approach brings a loss of spatial resolution and the number of looks is restricted to the critical number which is proposed in this paper. Three nonlinear filters, which suppress phase noise effectively, are analyzed, including conditioned neighbor averaging, modified median filtering and modified mode filtering. Then a new practical method, which combines multilook average and non linear filter to reduce noise in the wrapped phase map, is presented to suppress the noise in INSAR and the experimental results by processing ERS-1/2 data are used to confirm that the method is more effective.