Speckle statistics and SAR processor evaluation 
Philippe MARTIN 
E.N.S.T.L[.M.A. 
6, Avenue de Claviéres 
30107 -ALES CEDEX 
FRANCE 
Abstract 
Synthetic Aperture Radar images are primarily affected by speckle 
noise due to the coherent interference of the radar wave and the surface 
scatterers. This paper reviews the classical theory of speckle statistics 
(Rayleigh speckle statistics), derives higher order statistical properties of the 
SAR raw and processed signals and shows their relationships to the surface 
reflectivity. The applications of this model are then developed. It is shown how 
this model can be applied : 
- to the estimation of the parameters of SAR sensors, and therefore 
to the design of optimized processor ; 
- to the evaluation of the quality of the processed image. 
and how it can contribute to speckle noise reduction. 
Introduction 
The speckle effect is a property of all coherent imaging systems, when 
the radiation (microwave or laser light) is scattered from a surface whose 
roughness is on the order of the wavelength. The effect is caused by the 
interference of the signals returns from distributed scatterers on the surface. 
454 
  
  
  
of 
th 
p! 
al 
re 
be 
in 
qi 
ac 
e» 
di 
CC 
af