Complex Wishart distribution-based change detection with polarimetric TerraSAR-X imagery

Frank Thonfeld, Allan Aasbjerg Nielsen, Henning Skriver, Knut Conradsen, Morton John Canty

AbstractIn this contribution, we present a change detection method based on the complex Wishart distribution. A maximum likelihood estimate for the complex covariance matrix of the distribution can be realized. In case that the components of the matrix are not independent and identically distributed, the so-called equivalent number of looks (ENL) may be used to parameterize the distribution. For two co-registered, look-averaged polarimetric images acquired at times t1 and t2, a per-pixel, generalized likelihood ratio test for equal covariance matrices has a critical region, assumed to be the same for both images. The associated asymptotic probability of obtaining a smaller value of the test statistic is given in the maximum likelihood estimate for the complex covariance matrix, so that the decision threshold can be chosen for any desired confidence level. In this contribution we illustrate a change detection procedure applicable to single, dual and quad polarimetric TerraSAR-X images and examine its sensitivity to the ENL parameter.
TypeConference paper [With referee]
ConferenceTerraSAR-X Science Team Meeting
Year2013    Month June
PublisherDLR Oberpfaffenhofen, Germany
Electronic version(s)[pdf]
BibTeX data [bibtex]
IMM Group(s)Image Analysis & Computer Graphics, Geoinformatics