Speckle reduction and edge detection for TerraSAR-X single-look dual-polarization imagery

Preliminary work on speckle reduction and edge detection of TerraSAR-X single-look dual-polarization (HH and VV polarizations) spotlight data is reported. A speckle reduction approach is introduced based on a beta-distributed sample squared radius. The performance of the speckle reduction approach was benchmarked against the boxcar filter and assessed with respect to the degree of speckle reduction, radiometric preservation as well as image feature retention. For detecting edges in TerraSAR-X single-look dual-polarization speckled imagery, a constant false alarm rate (CFAR) detector is presented in this paper based on the Wilks' lambda. The applicability of the edge detector for TerraSAR-X single-look single-polarization intensity data is also discussed.