Velocity-independent redatuming: A new approach to the near-surface problem in land seismic data processing

2002 ◽  
Vol 21 (8) ◽  
pp. 730-735 ◽  
Author(s):  
Panos G. Kelamis ◽  
Kevin E. Erickson ◽  
Dirk J. Verschuur ◽  
A. J. Berkhout
Keyword(s):  
Data Processing ◽  
Seismic Data ◽  
Seismic Data Processing ◽  
Near Surface ◽  
New Approach

scholarly journals Muting the noise cone in near‐surface reflection data: An example from southeastern Kansas

Geophysics ◽  
1998 ◽  
Vol 63 (4) ◽  
pp. 1332-1338 ◽  
Author(s):  
Gregory S. Baker ◽  
Don W. Steeples ◽  
Matt Drake
Keyword(s):  
Data Processing ◽  
Seismic Data ◽  
Hydrocarbon Exploration ◽  
Seismic Reflection Profile ◽  
Seismic Data Processing ◽  
Near Surface ◽  
Surface Reflection ◽  
Reflection Data ◽  
Data Volume ◽  
Total Data

A 300-m near‐surface seismic reflection profile was collected in southeastern Kansas to locate a fault(s) associated with a recognized stratigraphic offset on either side of a region of unexposed bedrock. A substantial increase in the S/N ratio of the final stacked section was achieved by muting all data arriving in time after the airwave. Methods of applying traditional seismic data processing techniques to near‐surface data (200 ms of data or less) often differ notably from hydrocarbon exploration‐scale processing (3–4 s of data or more). The example of noise cone muting used is contrary to normal exploration‐scale seismic data processing philosophy, which is to include all data containing signal. The noise cone mute applied to the data removed more than one‐third of the total data volume, some of which contains signal. In this case, however, the severe muting resulted in a higher S/N ratio in the final stacked section, even though some signal could be identified within the muted data. This example supports the suggestion that nontraditional techniques sometimes need to be considered when processing near‐surface seismic data.

ATTENUATION OF SHORT-PERIOD MULTIPLES IN SEISMIC DATA PROCESSING OF THE JEQUITINHONHA BASIN, BRAZIL

2014 ◽  
Vol 32 (3) ◽  
pp. 395 ◽  
Author(s):  
Silmara L.R. Oliveira ◽  
Rosângela Corrêa Maciel ◽  
Michelângelo G. da Silva ◽  
Milton José Porsani
Keyword(s):  
Data Processing ◽  
Seismic Data ◽  
Time Windows ◽  
Fixed Time ◽  
Seismic Data Processing ◽  
Multiple Reflections ◽  
Seismic Line ◽  
New Approach ◽  
Predictive Deconvolution ◽  
Short Period

ABSTRACT. Short-period multiples attenuation is a difficult problem for shallow water marine seismic data processing. In the past few decades many filteringmethods have been developed to solve this problem and to improve the quality of seismic imaging. The Wiener-Levinson predictive deconvolution method is one of themost useful and well known filter methods used in the seismic data processing flow. It is a statistical approach to reduce redundancy along the time variable seismictrace, allowing us to both improve the time resolution and also attenuate multiple reflections of the seismic traces. One of the assumptions of the Wiener-Levinsonmethod is that the seismic wavelet is stationary along the entire seismic trace. However, this is not true for real seismic data and to bypass this limitation the methodis normally applied using fixed time windows, distributed along the seismic trace. The present study tested a new adaptive predictive deconvolution approach for theattenuation of short-period multiples. The new approach is based on a sliding window of fixed length that is shifted sample by sample along the entire seismic trace.At each position, a new filter is computed and applied. The implied systems of equations are solved by using a recursive Levinson-type algorithm. The main differencewith respect to the conventional Wiener-Levinson approach is that the filter is updated for each data sample along the trace and no assumption is imposed on the dataoutside the considered window. The new adaptive predictive deconvolution approach was tested using a seismic line of the Jequitinhonha Basin acquired by Petrobras.The results demonstrated that the new approach is very precise for the attenuation of short-period multiples, producing better results than the ones obtained fromthe conventional Wiener-Levinson predictive deconvolution approach. The results were obtained with filters of 25 coefficients, predictive distance of 5 samples andwindow length equal to 55 samples.Keywords: seismic processing, Jequitinhonha Basin, adaptive predictive deconvolution, multiple of attenuation, Wiener-Levinson deconvolution.RESUMO. A atenuação de reflexões múltiplas de curto período, presentes nos dados sísmicos adquiridos sobre lâmina d’água rasa, representa um grande problemado processamento de dados sísmicos marítimos. Nas últimas décadas, vários métodos de filtragem de dados sísmicos têm sido desenvolvidos com o propósito deatenuar reflexões múltiplas e melhorar a qualidade das seções sísmicas. O método de filtragem conhecido como deconvolução preditiva de Wiener-Levinson é bastante utilizado na indústria do petróleo. Ele permite melhorar a resolução temporal dos dados sísmicos e atenuar reflexões múltiplas, podendo ser visto como um método estatístico que remove a coerência temporal dos traços sísmicos. O método de Wiener-Levinson pressupõe que o pulso sísmico é estacionário, fato este que não ocorrenos dados sísmicos reais. Para contornar este problema, o método de Wiener-Levinson é normalmente aplicado utilizando-se janelas de tempo fixas, distribuídas ao longo do tempo de registro. No presente trabalho, empregamos um método de deconvolução preditiva adaptativa no qual as janelas de tempo deslizantes são deslocadas amostra a amostra ao longo de todo o traço sísmico. Os sistemas de equações são resolvidos com o algoritmo recursivo tipo-Levinson. Na deconvolução de Wiener-Levinson, com janelas de tempo fixa, os filtros são gerados e aplicados dentro de cada janela. Já na deconvolução preditiva adaptativa o algoritmo calcula um novo filtro a cada posição da janela deslizante. Para teste da nova abordagem utilizamos os dados sísmicos da Bacia de Jequitinhonha, cedidos pela Petrobras. Os melhores resultados foram obtidos com filtros de 25 coeficientes, distância de predição igual a 5 amostras e janela móvel de 55 amostras. Os resultados obtidos com a nova abordagem demonstram que a deconvolução preditiva adaptativa atua com eficácia na atenuação de múltiplas de curto período, apresentando resultados melhores queos gerados pelo método de deconvolução preditiva de Wiener-Levinson.Palavras-chave: processamento sísmico, Bacia do Jequitinhonha, deconvolução adaptativa, atenuação de múltiplas, deconvolução de Wiener-Levinson.

A hybrid learning-based framework for seismic denoising

2019 ◽  
Vol 38 (7) ◽  
pp. 542-549 ◽  
Author(s):  
Chengbo Li ◽  
Yu Zhang ◽  
Charles C. Mosher
Keyword(s):  
Data Processing ◽  
Seismic Data ◽  
Hybrid Learning ◽  
Environmental Noise ◽  
Great Promise ◽  
Seismic Data Processing ◽  
Weak Signals ◽  
Coherent Noise ◽  
Near Surface ◽  
Conventional Methods

Noise attenuation has been a long-standing problem in seismic data processing. It presents unique challenges on land due to a complex near surface coupled with unavoidable environmental noise sources. In many cases, weak signals are embedded in much stronger noise, which makes conventional methods less effective at extracting those signals. In addition, conventional methods may lack adaptability to various noise types and patterns. Machine learning has shown great promise in solving geophysical problems including seismic data processing and interpretation. Here, we propose a novel method that is applicable to attenuating both incoherent noise, such as environmental noise, and coherent noise, such as ground roll and scattered noise, under a unified learning-based framework. This framework takes advantage of conventional methods to build the initial models and then employs dictionary learning and sparse inversion to invert both signal and noise simultaneously. The proposed method augments conventional methods by leveraging learning to recover residual weak signals from strong noise. We have applied this hybrid learning-based method successfully to some of the most difficult data areas where conventional denoising methods underperformed. Synthetic and real data examples demonstrate the effectiveness of the method for various noise types.

Sign in / Sign up

Export Citation Format

Share Document