A Modification to Goldstein algorithm for TerraSAR-X interferometic phase filter

In Goldstein algorithm, the filter parameter is a pre-chosen parameter and affects the quality of interferometric phase (IP) filtering in InSAR processing if the parameter improperly. In this paper we present a novel adaptive filtering parameter construct method for the TerraSAR-X IP filter. Analyzed the relationship between the coherence and the phase standard deviation by the probability density function of the IP, the coherence can be used as a standard of IP´s quality. We use sine function of absolute value of complex coherence to modify the existing algorithm. The new filtering parameter can avoid the over-filtering or un-filtering. The improved algorithm is successfully used in both simulated data and TerraSAR-X datasets filtering. The experiment results showed this method can filter the noise effectively and also preserve phase resolution and fringe information of the IP.