Title :
Spatially Variant Restoration for Polarimetric Synthetic Aperture Radar Imagery
Author :
Lee, Ken Yoong ; Bretschneider, Timo
Author_Institution :
Sch. of Comput. Eng., Nanyang Technol. Univ., Singapore
fDate :
July 31 2006-Aug. 4 2006
Abstract :
A. spatially variant speckle filter is proposed for multi-look polarimetric synthetic aperture radar (POLSAR) imagery. The central idea is that the filtering is applied only to homogeneous areas based on the scattering properties, while for detected edges, lines and point-like textural features, the original complex covariance matrix of these features is restored in order to preserve the actual features. The capabilities of the proposed filter were examined using nine-look NASA/JPL POLSAR C-band data. Based on the obtained results, the proposed filter showed a promising performance in speckle removal and radiometric preservation. Moreover, the point-like textural features as well as structural features (i.e. edges and lines) were well-retained in the filtered outputs.
Keywords :
covariance matrices; edge detection; geophysical techniques; image restoration; image texture; radar polarimetry; speckle; synthetic aperture radar; POLSAR imagery; covariance matrix; polarimetric synthetic aperture radar imagery; radiometric preservation; spatially variant restoration; speckle removal; textural features; Covariance matrix; Filtering; Filters; Image edge detection; Image restoration; NASA; Polarimetric synthetic aperture radar; Radar detection; Radar scattering; Speckle;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 2006. IGARSS 2006. IEEE International Conference on
Conference_Location :
Denver, CO
Print_ISBN :
0-7803-9510-7
DOI :
10.1109/IGARSS.2006.456