Modeling the statistics of high resolution SAR images

نویسندگان

  • Vladimir Krylov
  • Josiane Zerubia
  • Gabriele Moser
  • Sebastiano B. Serpico
چکیده

In the context of synthetic aperture radar (SAR) image processing a crucial problem is represented by the need to develop accurate models for the statistics of the pixel intensities. In the current research report, we address the problem of parametric probability density function (pdf) estimation for modeling the amplitude distribution for high resolution SAR images. Hitherto, several theoretical and heuristic models for the pdfs of SAR data have been proposed in the literature, most of them being highly e ective for some particular land-cover typologies. Thus, given some SAR image with no prior information about the typology, the choice of a single optimal SAR parametric pdf becomes a hard task. In this report, we develop an estimation algorithm addressing the problem of pdf selection by adopting a nite mixture model (FMM) for the amplitude pdf, by mixing components belonging to a given dictionary of SAR-speci c pdfs. The proposed method automatically integrates the procedures of selection of the optimal model for each component, of parameter estimation, along with the optimization of the number of components, by combining the Stochastic Expectation Maximization (SEM) iterative methodology and the recently proposed method-of-log-cumulants (MoLC) for parametric pdf estimation for non-negative random variables. Experimental results on several real COSMO-SkyMed and RAMSES sensor images are presented, showing the capabilities of the proposed method to accurately model the statistics of SAR amplitude data. Key-words: synthetic aperture radar (SAR) image, probability density function (pdf), parametric estimation, nite mixture models, stochastic expectation maximization (SEM). ∗ EPI Ariana, UR INRIA Sophia Antipolis Méditeranée, 2004, Route des Lucioles, B.P.93, FR-06902, Sophia Antipolis Cedex (France); Faculty of Computational Mathematics and Cybernetics, Lomonosov Moscow State University, 119991 Leninskie Gory, Moscow (Russia), e-mail: [email protected]. † Dept. of Biophysical and Electronic Engineering (DIBE), University of Genoa, Via Opera Pia 11a, I-16145, Genoa (Italy), e-mail: [email protected]. ‡ Dept. of Biophysical and Electronic Engineering (DIBE), University of Genoa, Via Opera Pia 11a, I-16145, Genoa (Italy), e-mail: [email protected]. § EPI Ariana, UR INRIA Sophia Antipolis Méditeranée, 2004, Route des Lucioles, B.P.93, FR-06902, Sophia Antipolis Cedex (France), e-mail: [email protected]. in ria -0 03 42 68 1, v er si on 2 30 J an 2 00 9 Modélisation des statistiques des images radar (RSO) haute résolution Résumé : En télédétection, un problème vital est le besoin de développer des modèles précis pour représenter les statistiques des intensités des images. Dans ce rapport de recherche, nous traitons le problème d'estimation de la densité de probabilité pour la modélisation d'amplitude d'une image haute résolution de type Radar à Synthèse d'Ouverture (RSO). Précédemment, plusieurs modèles théoriques et heuristiques ont été ultilisés pour représenter l'amplitude d'un signal du type RSO et ils ont montré leur e cacité pour certains types d'occupation du sol, rendant ainsi di cile le choix d'un seul modèle de densité de probabilité paramétrique. Dans ce rapport de recherche, nous introduisons un algorithme d'estimation fondé sur un modèle de mélange ni de densités de probabilité d'amplitude dont les composantes appartiennent à un dictionnaire spéci que. La mèthode proposée intègre, de façon automatique: les procédures de sélection d'un modèle optimal pour chaque composante, l'estimation des paramètres, l'optimisation du nombre de composantes. Pour ce faire, nous utilisons simultanément l'algorithme EM stochastique et la méthode des log-cumulants en vue de l'estimation de la densité de probabilité paramétrique. Des résultats expérimentaux sur plusieurs images RSO réelles (issues des capteurs COSMO-SkyMed et RAMSES) sont présentés montrant que la méthode proposée est su samment précise pour modéliser des statistiques d'images d'amplitude RSO. Mots-clés : image radar à synthèse d'ouverture (RSO), densité de probabilité, estimation paramétrique, modèles de mélange ni, EM stochastique (SEM). in ria -0 03 42 68 1, v er si on 2 30 J an 2 00 9 Modeling the statistics of high resolution SAR images 3

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The Extended Sub-look Analysis In Polarimetric SAR Data For Ship Detection

The monitoring of maritime areas with remote sensing is essential for security reasons and also for the conservation of environment. The synthetic aperture radar (SAR) can play an important role in this matter by considering the possibility of acquiring high-resolution images at nighttime and under cloud cover. Recently, the new approaches based on the sub-look analysis for preserving the infor...

متن کامل

A Multi-Variate Contour Detector for High-Resolution Polarimetric SAR Images

This article presents a new method for contour detection in high-resolution polarimetric SAR images. The method is based on multi-variate statistics of speckle in homogeneous regions in a SAR image and uses a hypothesis test for the difference in variance between two adjacent regions to find the contours. The detector is directly applied on the single-look complex polarimetric SAR image. A prep...

متن کامل

Very High Resolution Parametric and Non- Parametric Sartomography Methods for Monitoring Urban Areas Structures

Synthetic Aperture Radar (SAR) is the only way to evaluate deformation of the Earth’s surface from space on the order of centimeters and millimeters due to its coherent nature and short wavelengths. Hence, by this means the long term risk monitoring and security are performed as precisely as possible. Traditional SAR imaging delivers a projection of the 3-D object to the two dimensional (2-D) a...

متن کامل

An Improved Scheme for Parameter Esti- Mation of G0 Distribution Model in High- Resolution Sar Images

Statistical modeling of Synthetic Aperture Radar (SAR) images is of great importance for speckle noise filtering, target detection and classification, etc. Moreover, it can provide a comprehensive understanding of terrain electromagnetics scattering mechanism. Over the past three decades, many sophisticated models have been developed for SAR images, such as Rayleigh, Gamma, K and G etc. The G0 ...

متن کامل

A Model for High Resolution TerraSAR-X Images of Natural Scenes

In the last year, the TerraSAR-X sensor began to provide a great amount of data, thus increasing the possibility to extract valuable information from SAR images. Indeed, these high resolution images are increasingly more textured and present many details previously masked due to the low resolution of the available sensors. This certainly represents an advance for remote sensing applications, bu...

متن کامل

Introducing An Efficient Set of High Spatial Resolution Images of Urban Areas to Evaluate Building Detection Algorithms

The present work aims to introduce an efficient set of high spatial resolution (HSR) images in order to more fairly evaluate building detection algorithms. The introduced images are chosen from two recent HSR sensors (QuickBird and GeoEye-1) and based on several challenges of urban areas encountered in building detection such as diversity in building density, building dissociation, building sha...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2009