Seismic data interpolation is a longstanding issue. Most current methods are only suitable for randomly missing cases. To deal with regularly cases, an antialiasing strategy should be included. However, seismic survey design using random distribution of shots and receivers always operationally challenging impractical. We have used deep-learning-based approaches interpolation, which could extract deeper features the training in nonlinear way by self-learning. It can also avoid linear events,...

Independent component analysis (ICA) proves to be effective in the removing ocular artifact from electroencephalogram recordings (EEG). While using ICA correction, a crucial step is correctly identify components among decomposed independent components. In most previous works, this of selecting was manually implemented, which time consuming and inconvenient when dealing with large amount EEG data. We present new method automatically selects eye blink based on pattern their scalp topographies,...

In general, we wish to interpret the most broadband data possible. However, do not always provide best insight for seismic attribute analysis. Obviously, spectral bands contaminated by noise should be eliminated. tuning gives rise with higher signal-to-noise ratios. To quantify geologic discontinuities in different scales, combined decomposition and coherence. Using decomposition, amplitudes corresponding a given scale discontinuity, as well some subtle features, which would otherwise buried...

Seismic inversion is a process of predicting high-resolution stratigraphic parameters from low-resolution seismic data. Traditional methods tend to impose human prior knowledge, such as sparsity, the modeling process. Nowadays, with development deep learning, idea by learning data has gained great attention in varieties research fields. As data-driven method, an artificial neural network (ANN) already been explored many researchers field inversion. Compared ANN, convolutional (CNN) stronger...

Deep learning has achieved great success in a variety of research fields and industrial applications. However, when applied to seismic inversion, the shortage labeled data severely influences performance deep learning-based methods. In order tackle this problem, we propose novel impedance inversion method based on cycle-consistent generative adversarial network (Cycle-GAN). The proposed Cycle-GAN model includes two subnets discriminative subnets. Three kinds loss, including estimation are...

The spectral decomposition technique plays an important role in reservoir characterization, for which the time-frequency distribution method is essential. deconvolutive short-time Fourier transform (DSTFT) achieves a superior resolution by applying 2D deconvolution operation on (STFT) spectrogram. For seismic decomposition, to reduce computation burden caused DSTFT, STFT spectrogram cropped into smaller area, includes positive frequencies fallen signal bandwidth only. In general, because...

Nowadays, convolutional neural network (CNN) has achieved great attention in varieties of research fields. Benefit from the strong learning ability CNN, knowledge contained labeled data set can be efficiently extracted. However, validity CNN is guaranteed by a sufficient number data. In field seismic impedance inversion, amount always limited. order to mitigate dependence on data, we propose solve inversion problem using 1D cycle-consistent generative adversarial (Cycle-GAN), which consists...

Deep learning has been widely adopted in seismic inversion. One of the major obstacles when adopting deep inversion is demand for labeled data sets. There are mainly two approaches to address this problem. generate massive numbers synthetic and then transfer trained model real data. The other introduce theoretical constraints reduce parameter spaces learning. In letter, we propose a physics-constrained impedance method based on Robinson convolution forward process provide process. Bilateral...

The short-time Fourier transform (STFT) spectrogram, which is the squared modulus of STFT, a smoothed version Wigner-Ville distribution (WVD). STFT spectrogram 2-D convolution signal WVD and window function WVD. In this letter, we propose deconvolutive (DSTFT) method, improves time-frequency resolution reduces cross-terms simultaneously by applying deconvolution operation on spectrogram. Compared to obtained proposed method shows clear improvement in resolution. Computer simulations are...

Characterization of seismic attenuation, quantified by medium quality factor <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Q</i> , is desirable in processing and interpretation. Some basic methods have been already proposed, such as the spectral ratio method, centroid frequency shift (CFS) peak (PFS) method. Each has advantages disadvantages. In this letter, we propose an improved method for estimation combining benefits CFS PFS together while...

I have developed an accelerated sparse time-invariant Radon transform (RT) in the mixed frequency-time domain based on iterative 2D model shrinkage time domain. denote it as SRTIS. In traditional RT domain, is modeled a inverse problem that solved by iteratively reweighted least-squares (IRLS) algorithm and forward RTs are implemented frequency this method, IRLS replaced shrinkage, i.e., sparsity of promoted some simple operations Synthetic real data demultiple examples using parabolic given...

When plane-wave destruction (PWD) is implemented by implicit finite differences, the local slope estimated an iterative algorithm. We propose analytical estimator of that based on convergence analysis Using estimator, we design a noniterative method to estimate slopes three-point PWD filter. Compared with estimation, proposed needs only one regularization step, which reduces computation time significantly. With directional decoupling filter, algorithm also applicable 3D estimation. present...

The trace interval in the common shot and receiver gathers is always inconsistent. inconsistency affects final performance of seismic data processing, reconstruction methods can enhance consistency. Unfortunately, most interpolation algorithms are suitable randomly missing cases, difficulty increases sharply regularly especially with big gaps. As deep learning (DL) has a strong self-learning ability nonlinear characterizations to avoid linear events, sparsity, low rank assumptions, we...

Full waveform inversion (FWI) is a powerful tool for estimating the underground velocity model. However, it computationally expensive and resulting models tend to be not accurate enough. Thus, improve efficiency accuracy of FWI, we propose super-resolution (SR) method based on deep learning enhance resolution seismic Since edge images model are also widely used in geophysics, multitask (MTL) network with hard parameter sharing applied perform SR its images. The proposed MTL dubbed M-RUDSR...

Accurate estimation of volumetric seismic dip is great significance for subsequent processing and interpretation works. Recently, with the development deep learning techniques, convolutional networks are also applied estimation. Compared traditional approaches, estimating dips not only more efficient but shows promise in accuracy robustness. However, if we take estimated by approaches as labels train on field data directly, robustness learned influenced due to error labels. An alternative...

The difficulty of 3D magnetic inversion is to use 2D anomaly data obtain susceptibility structure. contribution the underground medium decreases rapidly with increase depth, which leads rapid attenuation resolution depth. In this paper, artificial intelligence (AI) technology applied predict model corresponding anomaly. network built in paper uses method down-sampling encoder receptive field and realize feature extraction data. decoder, attention fusion modules are added fuse maps from...

Seismic image segmentation is important in geological interpretation. In recent years, numerous studies have leveraged texture features to analyze seismic images. However, traditional feature extraction methods are computationally intensive and cannot be updated through back-propagation. To address these challenges, we propose a model named Lightweight Segmentation Network based on Co-occurring Matrix (LSCM-Net). The overall architecture of LSCMNet employs an asymmetric encoder-decoder...

Singular value decomposition (SVD) is an efficient tool for the separation of signal and noise subspaces. When it used to process seismic images, SVD can enhance signal-to-noise ratio (SNR) horizontal events effectively. In this paper, adaptive filter proposed non-horizontal by detection image texture direction then alignment estimated dip through data rotation. The features derived from co-occurrence matrix are estimate direction. parameter adapted according stacking energy along detected...

This article proposes a new higher-order-statistics-based coherence-estimation algorithm, which we denote as HOSC. Unlike the traditional crosscorrelation-based C1 coherence sequentially estimates correlation in inline and crossline directions uses their geometric mean estimate at analysis point, our method exploits three seismic traces simultaneously to calculate 2D slice of normalized fourth-order moment with one zero-lag then searches for maximum point on estimate. To include more...

The time-invariant Radon transform (RT) is commonly used to regularize and interpolate sparsely sampled or irregularly acquired prestack seismic data. sparseness of the model significantly influences results regularization. We have developed an effective efficient method for regularization interpolation 2D as well 3D accelerated sparse RT in mixed frequency-time domain improve performance RT-based data This incorporated iterative shrinkage algorithm instead traditional iteratively reweighted...

The time-frequency analysis tools, which are very useful for anomaly identification, reservoir characterization, seismic data processing, and interpretation, widely used in discrete signal analysis. Among these methods, the generalized S transform (GST) is more flexible, because its analytical window can be self-adjusted according to local frequency components of selected signal, besides there exist another two adjustable parameters make it superior (ST). But amplitude-preserving ability a...

Reservoir prediction is a significant issue in seismic interpretation, and it difficult to reach tradeoff point for the reservoir accuracy spatial continuity. Nowadays, though numerous machine learning methods have been widely applied prediction, so few available well-logging labels are still major obstacle improving performance. Considering such critical factor, we propose semisupervised deep-learning framework, which closed-loop convolutional neural network (CNN). virtual used. The CNN,...

The technologies with kilovoltage (kV) and megavoltage (MV) imaging in the treatment room are now available for image‐guided radiation therapy to improve patient setup target localization accuracy. However, development of strategies efficiently effectively implement these remains challenging. This study proposed an aggregated technique on‐board CT reconstruction using combination kV MV beam projections data acquisition efficiency image quality. These were acquired at position a new device...

Adaptive multiple subtraction is a critical and challenging procedure for the widely used surface-related attenuation (SRMA) techniques. In this paper, I present an adaptive algorithm based on independent component analysis (ICA). The method expresses problem of as blind source separation (BSS) with two mixtures (the seismic data predicted multiple) or more sources (primaries multiples). By taking advantage sparse property data, adopts geometric ICA to recover mixing matrix linear...

Surface-related multiple attenuation (SRMA) can effectively remove multiples from seismic data when other (e.g., Radon transform) methods have difficulty. SRMA generally includes two steps: prediction (or modeling) and adaptive subtraction (AMS). In some cases, is the main challenge for success of SRMA. Adaptive often posed as a least-squares minimization problem that minimizes energy difference between original input traces modeled traces. The minimum-output-energy approach be implemented...