[ABE-L] (sem assunto)

'Francisco Cysneiros' via abe-l@ime.usp.br abe-l em ime.usp.br
Seg Maio 18 13:39:03 -03 2026


Ciclo de Seminários do PPGE/2026.1

Temos o prazer de anunciar que, no dia  20 de maio de 2026,
o seminário será ministrado pelo  Prof. Dr. Willams Batista Ferreira da
Silva - DE/UFPE




Título: Estimation of the Equivalent Number of Looks on Correlated
Synthetic Aperture Radar Data


Data e horário: Quarta-feira, 20/05/2026, às 14h (GMT-03)

Local: Auditório Prof. Ruy Luis Gomes, sala 347 (*Auditório do DE*),

Departamento de Estatística - CCEN – UFPE


Palestrante : Prof. Dr. Willams Batista Ferreira da Silva - DE/UFPE



Abstract:This work addresses the challenge of accurately estimating the
Equivalent Number of Looks (ENL) in multilook Synthetic Aperture Radar
(SAR) imagery when pixel intensities exhibit spatial correlation. The ENL
is critical as it governs the effective signal-to-noise ratio in SAR data.
Traditional methods of estimating the ENL do not account for spatial
dependence. To address this limitation, the effects of spatial correlation
on ENL estimation are examined, and two novel ENL estimators are proposed
based on the two-dimensional generalized gamma autoregressive moving
average (2D G$\Gamma$-ARMA) model, which explicitly accounts for spatial
dependence. Fundamental properties of the 2D G$\Gamma$-ARMA process are
derived, including its Fisher information matrix and $s{\text{th}}$
non-central moment. The asymptotic normality of the proposed estimators is
also established. Evidence from extensive Monte Carlo simulations indicates
that the proposed estimators outperform traditional approaches,
particularly in spatially correlated scenarios. Two multilook SAR images
from San Francisco Bay and Flevoland are analyzed. Notably, the proposed
estimators effectively address the persistent over- and underestimation
problems that arise from ignoring correlation. Our results highlight the
importance of modeling spatial dependence for reliable ENL estimation in
practical SAR applications.

-- 
Francisco José A. Cysneiros
Full professor
Departament of Statistics - UFPE
Co-Director of Department of Statistics, UFPE 2005-2009
Co-Director of Graduate Studies, Statistics, UFPE 2009-2013
Director of Graduate Studies, Statistics, UFPE 2013-2017
Co-Director of Graduate Studies, Statistics, UFPE 2026-2028
Editor in Chief of Brazilian Journal of Probability and Statistics
My URLs:
www.de.ufpe.br/~cysneiros  <http://www.de.ufpe.br/~cysneiros>
http://150.161.44.6/cysneiros

Lattes:

*ResearchedID:*
*Publons:*
http://lattes.cnpq.br/1313497098151734
http://www.researcherid.com/rid/G-6333-2012
https://publons.com/a/1540091
ORCID: http://orcid.org/0000-0001-6757-6969


Description: Enter a Description
<https://www.researcherid.com/ProfileGet.action?focus=description>
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20260518/0f298f17/attachment.htm>


Mais detalhes sobre a lista de discussão abe