Unsupervised Classification of SAR Images Using Markov Random Fields and ${\cal G}_{I}^{0}$ Model

This letter deals with synthetic aperture radar (SAR) data classification in an unsupervised way. Many models have been proposed to fit SAR data (K, Weibull, Log-normal, etc.), but none of them are flexible enough to model all kinds of surfaces (particularly when there are urban areas present in the image). Our main contribution is the application of a statistical model G0 in a classification process which is shown to be able to model areas with different degrees of heterogeneity. The quality of the classification obtained by mixing this model and a Markovian segmentation is high. We use an iterative conditional estimation method to estimate the parameters of the proposed model.

[1]  A. Lopes,et al.  Statistical Distribution And Texture In Multilook And Complex Sar Images , 1990, 10th Annual International Symposium on Geoscience and Remote Sensing.

[2]  Pierre Lanchantin,et al.  Statistical image segmentation using triplet Markov fields , 2003, SPIE Remote Sensing.

[3]  A. A. Yushkevich,et al.  Strong Markov Processes , 1956 .

[4]  E. Jakeman,et al.  A model for non-Rayleigh sea echo , 1976 .

[5]  Gabriela Palacio,et al.  Markovian classification of SAR images using $\mathcal{G}^{0}_{I}$ model , 2009 .

[6]  L. Pizarro,et al.  M-estimators with asymmetric influence functions: the distribution case , 2006 .

[7]  Jean-Marie Nicolas,et al.  Unsupervised segmentation of SAR images using Triplet Markov fields and fisher noise distributions , 2007, 2007 IEEE International Geoscience and Remote Sensing Symposium.

[8]  Donald Geman,et al.  Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images , 1984 .

[9]  Nelson D. A. Mascarenhas,et al.  Pseudo-likelihood equations for Potts model on higher-order neighborhood systems: A quantitative approach for parameter estimation in image analysis , 2009 .

[10]  Henri Maître,et al.  A new statistical model for Markovian classification of urban areas in high-resolution SAR images , 2004, IEEE Transactions on Geoscience and Remote Sensing.

[11]  Gerhard Winkler,et al.  Image Analysis, Random Fields and Markov Chain Monte Carlo Methods: A Mathematical Introduction , 2002 .

[12]  Florence Tupin,et al.  Unsupervised classification of radar images using hidden Markov chains and hidden Markov random fields , 2003, IEEE Trans. Geosci. Remote. Sens..

[13]  Marta Mejail,et al.  Approximation of distributions for SAR Images: proposal, evaluation and practical consequences , 2001 .

[14]  Paolo Carnevali,et al.  Image Processing by Simulated Annealing , 1985, IBM J. Res. Dev..

[15]  C. Oliver.  Optimum texture estimators for SAR clutter , 1993 .

[16]  M. V. Menon.  Estimation of the Shape and Scale Parameters of the Weibull Distribution , 1963 .

[17]  A. C. Frery,et al.  Classification of SAR images using a general and tractable multiplicative model , 2003 .

[18]  C. Oliver.  A Model for Non-Rayleigh Scattering Statistics , 1984 .

[19]  R. Dana,et al.  The Impact of Strong Scintillation on Space Based Radar Design II: Noncoherent Detection , 1983, IEEE Transactions on Aerospace and Electronic Systems.