A method is presented which uses logarithmic statistics to detect and characterise class mixtures and targets in background clutter in synthetic aperture radar (SAR) images. Mixtures of ground cover types show up as extreme radar texture in statistical analysis of SAR images. Instead of modelling this as a spatially nonstationary radar cross section, this paper demonstrates how a mixture model analysis can be used to characterise the separate components and estimate their mixing proportions. 1 Theory 1.1 Mixture Model Let X be a real and positive random variate which represents a measurement obtained within a certain region of a SAR image. It is assumed that the region of interest is heterogeneous, and that X can be modelled with a twocomponent mixture model. This means that the observation X will be drawn from a distribution with probability density function (pdf) pX1(x) with probability π1, or from a distribution with pdf pX2(x) with probability π2. The pdfs are distinct, meaning that pX1(x) 6= pX2(x), and the mixing proportions obey
[1]
Torbjørn Eltoft,et al.
Estimation of the Equivalent Number of Looks in Polarimetric Synthetic Aperture Radar Imagery
,
2009,
IEEE Transactions on Geoscience and Remote Sensing.
[2]
Jean-Marie Nicolas.
1 - Introduction aux Statistiques de deuxième espèce : applications des Logs-moments et des Logs-cumulants à l'analyse des lois d'images radar
,
2002
.
[3]
Stian Normann Anfinsen,et al.
Statistical Analysis of Multilook Polarimetric Radar Images with the Mellin Transform
,
2010
.
[4]
Sylvia Frühwirth-Schnatter,et al.
Finite Mixture and Markov Switching Models
,
2006
.
[5]
H. White,et al.
On More Robust Estimation of Skewness and Kurtosis: Simulation and Application to the S&P500 Index
,
2003
.
[6]
Torbjørn Eltoft,et al.
Application of the Matrix-Variate Mellin Transform to Analysis of Polarimetric Radar Images
,
2011,
IEEE Transactions on Geoscience and Remote Sensing.
[7]
S. Quegan,et al.
Understanding Synthetic Aperture Radar Images
,
1998
.