On the concept of the polarimetric matched filter in high resolution radar imaging

Speckle reduction has long been recognized as the main problem of coherent imaging and many processing techniques have been advanced to overcome it. The vast majority of these techniques, however, are of scalar nature simply because vector/matrix imaging data are so sparse and have become available only very recently. The authors have such data, which were taken with the NASA JPL/CV-990 dual-polarization L-band (1.225 GHz) SAR (synthetic-aperture radar) system. They investigate the potential of an exclusively polarimetric image filtering approach, i.e., filtering which takes full advantage of the polarized SAR matrix data provided on a pixel-by-pixel basis and which complements the existing scalar speckle reduction techniques.<>