《Improving PolSAR Land Cover Classification With》.pdfVIP

848 IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, VOL. 5, NO. 3, JUNE 2012 Improving PolSAR Land Cover Classification With Radiometric Correction of the Coherency Matrix Donald K. Atwood, Member, IEEE, David Small, Member, IEEE, and Rüdiger Gens, Senior Member, IEEE Abstract— The brightness of a SAR image is affected by topog- Important information can be extracted from single polar- raphy due to varying projection between ground and image coor- ization SAR data, where signal brightness and texture can dinates. For polarimetric SAR (PolSAR) imagery being used for offer useful information on land cover. Howeve, the most purposes of land cover classification, this radiometric variability is shown to affect the outcome of a Wishart unsupervised classifi- significant gains have occurred with the recent availability cation in areas of moderate topography. The intent of this paper of polarimetric SAR (PolSAR) data from sensors such as the is to investigate the impact of applying a radiometric correction to -band ALOS PALSAR, -band Radarsat-2, and -band the PolSAR coherency matrix for a region of boreal forest in inte- TerraSAR-X. Within the constraints of frequency, incidence rior Alaska. The gamma naught radiometric correction estimates angle, and spatial resolution, quad-pol data yields a complete the local illuminated area at each grid point in the radar geom- characterization of the SAR microwave interaction with the etry. Then, each element of the coherency matrix is divided by the local area to produce a polarimetric product that is radiometrically target [1]. Incoherent