Taking signal and noise to new dimensions — Introduction

PreviousNext You have accessGEOPHYSICSVolume 80, Issue 6Taking signal and noise to new dimensions — IntroductionAuthors: Laurent DuvalSergey FomelMostafa NaghizadehMauricio SacchiLaurent DuvalIFP Energies Nouvelles, Rueil-Malmaison, France. E-mail: .Search for more papers by this authorEmail the author at [email protected], Sergey FomelThe University of Texas at Austin, John A. and Katherine G. Jackson School of Geosciences, Austin, Texas, USA. E-mail: .Search for more papers by this authorEmail the author at [email protected], Mostafa NaghizadehShell Canada Energy, Calgary, Alberta, Canada. E-mail: .Search for more papers by this authorEmail the author at [email protected], and Mauricio SacchiUniversity of Alberta, Department of Physics, Edmonton, Alberta, Canada. E-mail: .Search for more papers by this authorEmail the author at [email protected]https://doi.org/10.1190/2015-0921-SPSEINTRO.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail This special section covers present-day algorithms for seismic signal processing. In the last decade, important efforts were made in the development of signal processing methods that exploit the multidimensionality of seismic data. A stringent requirement for multidimensional seismic data processing is the elaboration of strategies to represent seismic data. This special section contains interesting examples of novel transformation for multidimensional seismic signal separation, seismic noise attenuation, and reconstruction. Examples are algorithms that use seislets, chirplets, local slant stacks, and apex shifted Radon transforms. In addition, this special section also offers results from a new family of algorithms that are inspired by the fields of compressive sensing, reduce-rank filtering, and dictionary learning. These algorithms can be used for data reconstruction, signal-to-noise-ratio enhancement and simultaneous source data processing.A special section on seismic signal processing cannot be complete without articles that tackle pervasive problems in seismic data exploration. As such, we also present papers on wavelet processing and sampling theory. We hope you enjoy reading this brew of papers as much as we have enjoyed assembling this special section.Chen and Fomel develop a novel approach to random noise attenuation that eliminates remaining coherent signal in the removed noise by using local signal and-noise orthogonalization. Local signal-and-noise orthogonalization can retrieve the leakage signal in the initial noise section of traditional denoising approaches by locally orthogonalizing signal-and-noise components, and thus can help reduce the loss of valuable information when applying any random-noise attenuation techniques.Last, the special section contains a new algorithm to solve the classical wavelet estimation problem.Liu et al. propose a new velocity-dependent (VD) formulation of the seislet transform, where the normal moveout equation serves as a bridge between local slope patterns and conventional moveout parameters in the common-midpoint domain. Synthetic and field-data examples show the effectiveness of the VD-seislet transform for eliminating strong random noise and the VD-seislet frame for separation of primaries and peg-leg multiples of different orders.Hu et al. propose a new approach to efficiently compress the local slant stacks by combing the estimation of multiple local slopes using structure tensor and matching pursuit decomposition. Several data examples and migration results show that this compression algorithm can get a high compression ratio while restoring most of the significant events and can be incorporated with a beam migration algorithm.Guitton and Claerbout propose a method for the deconvolution of nonminimum phase wavelets. It incorporates a sparseness criterion to make the polarity clearly evident and a log domain formulation to eliminate leg-jumps (output spike locations changing with time) with a Ricker-style regularization.Barros et al. propose to use the genetic global optimization algorithm differential evolution to estimate the parameters of the common-reflection surface stacking method. The authors also provide a convergence and a sensitivity analysis of the differential evolution, demonstrating its computational efficiency and the optimization method robustness with respect to the algorithm control parameters choice. The results obtained for a 2D real data set from Brazil indicate that the global strategy is a promising approach to the estimation of the common-reflection surface parameters.Ibrahim and Sacchi propose a fast transform that is equivalent to the apex shifted hyperbolic Radon transform using Stolt migration and demigration operators. This new transform for fast separation of simultaneous seismic sources by removing sources interference from common receiver gathers.Zhu et al. propose a novel seismic data denoising method based on a parametric dictionary learning scheme, which exploits the underlying sparse structure of the learned atoms over a base dictionary and significantly reduces the dictionary elements that need to be learned. This method achieves the best denoising performance and minimizes visual distortion.Wang and Xu present the time dispersion transforms, which can be applied to eliminate time dispersion artifacts introduced by finite difference calculation of time derivatives in both synthetic modeling and RTM imaging. This method works for both low- and high-order time finite-difference schemes, and both isotropic and anisotropic cases.Boßmann and Ma present an asymmetric Gaussian chirplet model and establish a dictionary-free variant of the orthogonal matching algorithm for sparse representation of seismic data. Unlike the Fourier transform that assumes that the seismic signals consist of plane waves, the proposed method assumes the seismic signal consists of nonstationary compressed plane waves, e.g., symmetric and asymmetric chirplets.Sternfels et al. introduce a novel convex optimization strategy enabling the simultaneous attenuation of random and erratic noise with interpolation for prestack seismic data, referred as joint low-rank and sparse inversion (JLRSI). The authors take a new look at the well-known Cadzow/singular spectrum analysis filtering and its robust and interpolation derivatives, using insights from recent developments in the field of compressive sensing to formulate the JLRSI problem and solve it thanks to state-of-the-art convex optimization algorithms.Li and Demanet propose a method for decomposing a seismic record into atomic events defined by a smooth phase and a smooth amplitude. An application to frequency extrapolation is shown for a synthetic shot record from the shallow Marmousi model.Abma et al. demonstrate simultaneous source acquisition and processing methods for land and marine seismic surveys. These surveys result in improved seismic images due to the improved source sampling while taking advantage of the improved efficiency of using multiple sources simultaneously.Kumar and Wason address the challenge of source separation for simultaneous towed-streamer acquisitions via two compressed sensing-based approaches — i.e., sparsity-promotion and rank-minimization. The applicability of the proposed method is tested on 2D field and realistic synthetic data sets; a comparison was made with the NMO-based median filtering approach.Wang et al. present a noise-robust method that accomplishes the detection and tracking of salt domes in postmigrated volumes using the gradient of textures and tensor-based subspace learning techniques. The experiments using the Netherlands offshore F3 block show that, in contrast to other state-of-the-art methods, the proposed method is more robust to noise, and can delineate salt dome boundaries with the highest similarities to those labeled by interpreters.Naghizadeh introduces a stochastic simulation technique to analyze the effectiveness of sparse Fourier reconstruction methods for multidimensional sampling functions. Also, double-weave 3D seismic acquisition design is proposed for sparse Fourier recovery of seismic records in shot and receiver domains.Naghizadeh investigates the compatibility of 3D double-weave acquisition designs with sparse Fourier reconstruction algorithms using simple modeled seismic data. Also subsurface fold distribution of double-weave acquisition was compared to random and traditional orthogonal acquisition layouts.FiguresReferencesRelatedDetailsCited by(Modi) (Indian Modi Government's Economic Development Policy and Implication for Cooperation between Korea and India )SSRN Electronic Journal, Vol. 8 Volume 80Issue 6Nov 2015Pages: 1ND-Z124ISSN (print):0016-8033 ISSN (online):1942-2156 publication data© 2015 Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 23 Oct 2015Published in print: 01 Nov 2015 CITATION INFORMATION Laurent Duval, Sergey Fomel, Mostafa Naghizadeh, and Mauricio Sacchi, (2015), "Taking signal and noise to new dimensions — Introduction," GEOPHYSICS 80: WDi-WDii. https://doi.org/10.1190/2015-0921-SPSEINTRO.1 Plain-Language Summary PDF Download Metrics Loading ...

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.