Seismic directional random noise suppression by radial‐trace time–frequency peak filtering using the Hurst exponent statistic

Random noise attenuation is a fundamental task in seismic data processing aimed at improving the signal‐to‐noise ratio (SNR) of seismic data, thereby improving the efficiency and accuracy of subsequent seismic data processing and interpretation. To this end, model‐based and data‐driven seismic data denoising methods have been widely applied, including f‐x deconvolution (FXDECON), K‐singular value decomposition (K‐SVD), feed‐forward denoising convolutional neural network (DnCNN), and U‐Net (an improved fully convolutional neural network structure), which have received widespread attention for their effectiveness in attenuating random noise. However, they often struggle with low‐SNR data and complex noise environments, leading to poor performance and leakage of effective signals. To address these issues, we propose a novel approach for random noise attenuation. This approach employs a multi‐model stacking structure, where the parameters governing this structure are optimized using a particle swarm optimizer. In the model‐based denoising method, we choose the FXDECON method, while in the data‐driven denoising method, we choose K‐SVD for shallow learning and U‐Net for deep learning as components of the multi‐model stacking structure. The optimal parameters for the multi‐model stacking structure are obtained using a particle swarm optimizer, guided by the proposed novel hybrid fitness function incorporating weighted signal‐to‐noise ratio, structural similarity, and correlation parameters. Finally, the effectiveness of the proposed method is verified with three synthetic and two real seismic datasets. The results demonstrate that the proposed method is effective in attenuating random noise and outperforms the benchmark methods in denoising both synthetic and real seismic data.This article is protected by copyright. All rights reserved

Section: Introductionmentioning

confidence: 99%

Multi‐model stacked structure based on particle swarm optimization for random noise attenuation of seismic data

Zhang,

Liao,

Luo

et al. 2024

“…With the development of information technology, more and more signal analysis and processing technologies have been applied to the study of time-frequency analysis, including the shorttime Fourier transform (STFT), continuous wavelet transform (CWT) and S transform. The methods described above are mainly used to depict seismic thin-layer (Huang et al, 2018;Wang et al, 2014). In order to take advantage of various methods, Stockwell et al (1996) adopted S transform with both STFT and CWT, and the time-frequency resolution could be adjusted adaptively (Deng et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

A seismic thin‐layer detection factor calculated by integrated S transform with non‐negative matrix factorization

Zhao,

Cao,

Yang

et al. 2024

Time–frequency analysis is one of the effective methods for seismic thin‐layer detection. Conventional time–frequency analysis technology for seismic thin‐layer detection is interfered by the energy of adjacent frequency signals, and there is information redundancy in the frequency‐domain analysis. Therefore, an S transform with improved window factor, which is based on the constrained non‐negative matrix factorization, is constructed to realize seismic thin‐layer detection. First, the seismic data is processed by the S transform of the improved window factor, and then we can obtain the frequency‐domain information with strong time–frequency focus by changing the adjustment factor and attenuation factor in the window function. Furthermore, the key frequency of the seismic data spectrum, which can also be called the key frequency characteristic factor, can be calculated by the non‐negative matrix factorization algorithm. Fortunately, the overthrust model shows a good correspondence between the key frequency characteristic factor and the thin‐layer interface. The field data example shows that this approach provides a new approach for thin‐layer detection.

Application of a cascading filter implemented using morphological filtering and time–frequency peak filtering for seismic signal enhancement

Liu

Zhou

2020

Inspired by the idea of the iterative time–frequency peak filtering, which applies time–frequency peak filtering several times to improve the ability of random noise reduction, this article proposes a new cascading filter implemented using mathematic morphological filtering and the time–frequency peak filtering, which we call here morphological time–frequency peak filtering for convenience. This new method will be used mainly for seismic signal enhancement and random noise reduction in which the advantages of the morphological algorithm in processing nonlinear signals and the time–frequency peak filtering in processing nonstationary signals are utilized. Structurally, the scheme of the proposed method adopts mathematic morphological operation to first preprocess the signal and then applies the time–frequency peak filtering method to ultimately extract the valid signal. Through experiments on synthetic seismic signals and field seismic data, this paper demonstrates that the morphological time–frequency peak filtering method is superior to the time–frequency peak filtering method and its iterative form in terms of valid signal enhancement and random noise reduction.