Effective number of looks for a multilook interferometric phase distribution

Interferometric phases have fluctuations that result from decorrelation between the complex images in an interferometric pair. Recently it has been proposed that a PDF derived for multilook polarimetric copolar-phase data can also be applied to interferometric SAR data. This PDF is parameterized by the number of looks and the complex correlation coefficient. Such a PDF allows the authors to determine phase error as a function of decorrelation. Actual SAR data usually have an affective number of looks less than the nominal number of looks. As a result, the assumption that samples are independent does not hold. Rederivation of the PDF to account for the nonindependence of samples represents a challenging, if not intractable, problem. Using χ²-squared goodness-of-fit tests, however, the authors show that an effective number of looks can be chosen in place of the nominal number of looks. The authors initially test this model using simulated data that has the same spectral characteristics as ERS-1 data. In addition to the phase fluctuations caused by decorrelation, actual data also have fluctuations caused by the topography and the imaging geometry. As a result, it is difficult to compare the phase statistics predicted by the PDF with actual data. There are a few cases, however, where these additional fluctuations are negligible and statistical comparisons can be made. They find that with an effective number of looks less than the nominal number, this PDF provides a good statistical model for multilook interferometric phases from both simulated and actual ERS-1 data