Effective random noise attenuation is critical for subsequent processing of seismic data, such as velocity analysis, migration, and inversion. Thus, the removal with an uncertainty level meaningful. Attenuating 3-D in a supervised way based on deep learning (DL) challenging because clean labels are difficult to obtain. Therefore, it necessary develop adaptive unsupervised-based method attenuation. In this article, we propose deep-denoising unsupervised (DDUL) network attenuate 2-D/3-D data....

We develop a deep learning framework based on image prior (DIP) and attention networks for 3-D seismic data enhancement. First, the noisy are divided into several overlapped patches. Second, DIP network has U-NET architecture, where input patches encoded to extract significant latent features, while decoder tries reconstruct using these extracted features. Besides, is used scale features from encoder decoder. Third, output of concatenated with that obtain high-order guide most information...

Abstract Distributed acoustic sensing (DAS) is an emerging technology that offers great potential in the high-resolution multi-scale seismic investigation due to its dense spatial coverage and cost-effectiveness. However, DAS data notoriously suffer from low signal-to-noise ratio (SNR) various types of strong noise, for example, high-frequency high-amplitude erratic vertical or horizontal noise. Here, we propose a novel denoising framework by cascading several individual methods are designed...

Local slope is an important attribute that can help distinguish seismic signals from noise. Based on optimal estimation, many filtering methods be designed to enhance the signal-to-noise ratio of noisy data. We present open-source MATLAB code package for local estimation and corresponding structural filtering. This includes 2D 3D examples with two main executable scripts related subfunctions. All files are in format. In each script, estimated based well-known plane-wave destruction...

ABSTRACT Earthquake forecasting is one of the most challenging tasks in field seismology that aims to save human life and mitigate catastrophic damages. We have designed a real-time earthquake framework forecast earthquakes tested it seismogenic regions southwestern China. The input data are features provided by multicomponent seismic monitoring system acoustic electromagnetic AI (AETA), which recorded using two types sensors per station: (EM) geo-acoustic (GA) sensors. target location...

Accurate separation of signal and noise constitutes a fundamental prerequisite for achieving high-resolution seismic imaging. A notable recent advancement in this domain is the Deep Image Prior (DIP), an unsupervised deep learning method leveraging neural networks (DNNs). The success approach lies adoption autoencoders that enables adaptive extraction high-fidelity data features. However, establishing optimal balance between suppression preservation remains non-trivial challenge DIP-based...

Iceland has long been a focal point of geophysical research due to its potential association with mantle plume. While earlier studies generally supported plume originating from the core-mantle boundary minimal lateral displacement during ascent, recent high-resolution tomographic images have indicated presence curved, ascending beneath Iceland. Consequently, detailed source region and morphology remain debated. In this study, we provide new constraints on by examining structure transition...

We have developed a new method for simultaneous denoising and reconstruction of 5D seismic data corrupted by random noise missing traces. Several algorithms been restoration based on rank-reduction (RR) methods. More recently, damping operator has introduced into the conventional truncated singular-value decomposition (TSVD) formula to further remove residual noise, presence which disturbs quality results. Despite success damped RR (DRR) when observed an extremely low signal-to-noise ratio...

The local orthogonalization (LO) approach has been broadly applied to attenuate random noise and deal with the signal-leakage problem induced using traditional denoising schemes. First, this removes applies LO weight (LOW) operator originally estimated signal. Then, signal leakage is predicted subsequently recovered from original component. Finally, denoised component are mixed output seismic However, limits when data include erratic noise. We find that shortcoming mainly caused by weakness...

Seismic denoising will inevitably cause signal leakage, which is seen as coherent energy in the removed noise profile. Most traditional methods either ignore leakage while maintaining clean denoised data or adjust parameters to decrease causing more residual data. The local signal-and-noise orthogonalization method can help retrieve leaked signals without bringing significant noise. However, when seismic contain erratic noise, does not work properly because of unstable inversion solving...

Noise and missing traces usually influence the quality of multidimensional seismic data. Therefore, it is necessary to estimate useful signal from its noisy observation. The damped rank-reduction (DRR) method has emerged as an effective reconstruct matrix incomplete observations. However, higher noise level larger ratio traces, weaker DRR operator becomes. Consequently, estimated low-rank (LR) includes a significant amount residual that influences following processing steps. we focus on...

Abstract New sensing techniques like the nodal geophones and distributed acoustic enable a spatial sampling ratio that was unprecedentedly high in earthquake seismology. The much higher of seismic wavefields is close to level exploration seismology calls for unified processing approach multichannel data regardless research interest, example, oil gas oriented or earthquake-study oriented. Here, we present first Python package benefits both communities, is, seismology, called Pyseistr....

Abstract We present a multifunctional open-source package—pyekfmm for eikonal-based travel-time calculation in 2D and 3D heterogeneous anisotropic media based on the well-documented fast marching method. Different from existing packages, pyekfmm package offers seamless compilation of backbone C programs Python environment through state-of-the-art pip installation. As result, users can use different scientific purposes with convenience enabled by interfaces efficiency offered programs. The...

Most traditional seismic denoising algorithms will cause damage to useful signals, which are visible from the removed noise profiles and known as signal leakage. The local signal-and-noise orthogonalization method is an effective for retrieving leaked signals noise. Retrieving while rejecting compromised by smoothing radius parameter in method. It not convenient adjust because it a global parameter, whereas data highly variable locally. To retrieve adaptively, we have adopted new...

Diffractions in the seismic data are associated with small-scale subsurface structures, thus their separation and imaging helpful characterizing underground discontinuities a high resolution that cannot be reached by traditional reflection methods. Traditional slope-based diffraction methods strongly affected accuracy stability of slope estimation methods, e.g., plane-wave destruction (PWD) method. When local is not properly estimated, separated waves suffer from mixture between energy due...

5D data is the original recorded form in 3D seismic acquisition, which includes sufficient information from all five dimensions. However, environmental and economic logistic difficulties often severely impact acquisition geometry, leading to raw with missing traces strong contaminating random noise. This deficiency causes troubles subsequent processing. Thus, an efficient interpolation denoising method required recover useful signals. Unfortunately, practical applications of many existing...

Diffraction separation plays a critical role in obtaining high-resolution images of subsurface diffractors by eliminating the diffractions’ contribution from reflections. It is usually based on difference spatial coherency between seismic reflection and diffraction energy time domain. Methods local slope differences, e.g., plane-wave destruction (PWD) method, tend to separate point line simultaneously. This paper applies rank-based known as localized rank-reduction (LRR) distinguish...

The development of the distributed acoustic sensing (DAS) technique enables us to record seismic data at a significantly improved spatial sampling rate meter scales, which offers new opportunities for high-resolution subsurface imaging. However, DAS recordings are often characterized by low signal-to-noise ratio (S/N) due presence noise, degrading reliability imaging and interpretation. Current noise reduction methods remain insufficient in simultaneously preserving weak signals eliminating...

Recently, studies on multidimensional seismic data interpolation through rank-constrained matrix or tensor completion have led to many effective methods, with satisfactory results. Despite the success of e.g., damped rank reduction (DRR), and high-order orthogonal iteration (HOOI), strong noise highly decimated traces could still make reconstruction results not acceptable. In this article, we find that implementing only one constraint solve recovery problem is sufficient. Therefore, consider...

Abstract The SS precursors have been extensively utilized in mapping the mantle transition zone (MTZ). However, their applications are often challenged by weak phases that arise from small impedance contrasts of discontinuities, noise contamination, and localized thermal/compositional heterogeneities. We develop a new data processing workflow for more reliable MTZ imaging adopting recently proposed robust damped rank‐reduction (RDRR) method exploration seismology. This exploits signal...