Banca de DEFESA: EBERTON DA SILVA MARINHO

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE : EBERTON DA SILVA MARINHO
DATA : 22/08/2016
HORA: 10:00
LOCAL: Sala de seminários do DFTE
TÍTULO:

STATISTICAL ANALYSES OF COMPRESSIVE SENSING APPLIED TO DATA SEISMIC


PALAVRAS-CHAVES:

Seismic Data Reconstruction, Bayesian Compressive Sensing (BCS), $\ell_1$-MAGIC, StOMP, Wavelets, Kustosis, Entropy, Sparsity


PÁGINAS: 120
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Física
RESUMO:
The Compressive Sensing (CS) is an efficient data recovery processing technique and construction of signals from a lower sampling rate than required by the Nyquist-Shannon theorem. This technique allows a great reduction in data for signals that may be sparsely represented. The wavelet transform has been used to compress and represent many natural signals, including seismic, in a sparsely way. There are several signals reconstruction algorithms that use the CS method, for example: $\ell_1 $-MAGIC, Fast Bayesian Compressive Sensing (Fast BCS) and Stagewise Orthogonal Matching Pursuit (StOMP).
This thesis compares the recovery of seismic traces in a statistical perspective using different methods of CS, wavelet transforms and sampling rates. The correlation is measured between the relative error (RE) by the CS and recovery measurements: coefficient of variation, skewness, kurtosis and entropy of the original signal. There seems to be a correlation between the kurtosis and entropy of the signal with the reconstruction of RE by CS.
Also, it was analyzed the RE distribution in the CS. The $\ell_1 $-MAGIC had better results for sample rates up to 40%. Moreover, the RE distribution in $\$ell_1$-MAGIC had more normal, symmetrical and mesokurtic in Fast BCS histograms. However, for above 50% sample rates the Fast BCS showed a better performance in average RE.

MEMBROS DA BANCA:
Externo ao Programa - 1379465 - GILBERTO CORSO
Externo à Instituição - HUGO ALEXANDRE DANTAS DO NASCIMENTO - UFG
Interno - 2492756 - JOAO MEDEIROS DE ARAUJO
Presidente - 004.056.634-04 - LIACIR DOS SANTOS LUCENA - UFRN
Externo à Instituição - MARCOS VINICIUS CANDIDO HENRIQUES - UFERSA
Notícia cadastrada em: 09/08/2016 09:17
SIGAA | Superintendência de Tecnologia da Informação - (84) 3342 2210 | Copyright © 2006-2024 - UFRN - sigaa05-producao.info.ufrn.br.sigaa05-producao