You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.
 
 Earthnet Home  
Introduction
Sessions and Session Summaries
Deadlines
First Annoucement
Scientific Committee
ESA Specific Links
Organising Committee
Programme
Round Table Discussion Questions
Conference Photos
List of Participants
Proceedings
All papers


 

Model based SAR polarimetric speckle noise filter

Carlos López-Martínez (1)

(1) Technical University of Catalonia (UPC), Jordi Girona, 1-3, Edifici D3-118, E-08940, Barcelona, Spain

Abstract

Speckle noise is the most important problem for a correct interpretation of polarimetric SAR data. Polarimetric SAR data are described by the complex hermitian covariance matrix. For the backscattering case, this matrix has 3 by 3 elements. The diagonal elements contain intensity information, where speckle noise follows a multiplicative noise model. Up to now, speckle was assumed also multiplicative within the off-diagonal elements. The authors have recently proved that speckle noise presents a multiplicative as well as an additive behaviour within the off-diagonal elements, by presenting a complete speckle noise model for the covariance matrix. Noise reduction techniques that are based on assuming only a multiplicative speckle noise model for all the covariance matrix elements shall reduce speckle in a non-optimum form. Based on the new speckle noise model, the authors shall present in this paper a new technique to optimally reduce speckle noise for all the covariance matrix elements. Results of this new polarimetric speckle noise reduction technique shall be presented by employing real polarimetric SAR data. The paper shall be divided in two main parts. The first part shall be devoted to present the theoretical background as well as the implementation of the polarimetric speckle noise reduction technique. The second part shall show different results of applying the speckle noise reduction technique to real polarimetric SAR data.

 

Full paper

 

  Higher level                 Last modified: