A self-attention enhanced encoder-decoder network for seismic data denoising

Seismic data interpolation and denoising play vital roles in obtaining complete and clean data in seismic data processing. Seismic data usually misses along various spatial axes and always mix with random noise. In order to obtain complete and clean seismic data, reconstruction technology can interpolate missing data and attenuate random noise. Nonsubsampled contourlet transform is an effective transform to obtain multi-scale and multi-direction sparse domain data for compression sensing interpolation and denoising. However, conventional iterative shrinkage/thresholding cannot handle ill-posed and ill-conditioned equations for solving linear inverse problem. We present a two-step iterative log thresholding method to overcome ill-posed and ill-conditioned problems and improve the convergence rate and solution accuracy, which can interpolate and denoise seismic data simultaneously in the nonsubsampled contourlet transform framework. First, we use nonsubsampled contourlet transform to convert the seismic missing data with random noise to sparse domain. Then, we apply two-step iterative log thresholding algorithm to interpolate and denoise data in sparse domain. The result of each iteration is based on the results of the previous two iterations, which can accelerate convergence rate. In addition, log thresholding can further improve convergence rate and solution accuracy. Finally, we use inverse nonsubsampled contourlet transform to obtain the interpolated and denoised seismic data. The new method can reconstruct the irregularly missing data and attenuate random noise to obtain complete and clean seismic data with high accuracy, which is crucial for seismic imaging and inversion. We demonstrate the applicability and effectiveness of this simultaneous interpolation and denoising technique with successful applications to both synthetic and field data examples.

Deconvolution is an essential step in seismic data processing. Sparse-spike deconvolution often is used to enhance the resolution of the seismic image by adding a model-driven regularization term. However, this method does not consider the features of the data, nor does it exactly describe the relationship between seismic data and the desired attribute (such as seismic reflectivity). We have developed a data and model dual-driven seismic deconvolution method based on error-constrained joint sparse representation (SR) using borehole measurement and surface seismic data. The combined features of the borehole reflectivity and the surface seismic data can be obtained through joint dictionary learning. With the help of the joint dictionary, the relationship between seismic waveforms and reflectivity is captured by the sparse coefficients. We construct the regularization term of deconvolution by alternately decomposing the error of the synthesized data via sparse reconstruction and the observed seismic data. Unlike model-driven methods, the constraint term of the new method can be established by the error-constrained SR. Based on this SR, the initial model of reflectivity is obtained to realize the sparse deconvolution of seismic data under the constraint of borehole data features. In general, this method is a data and model dual-driven deconvolution. Synthetic and field data tests demonstrate that this method can effectively improve the resolution and accuracy of deconvolution.

Seismic reflection data are used by geologists to identify the best site for oil and gas explorations. The raw seismic image cannot be directly used for explorations because the noises may hinder the primary reflections and result in misinterpretations. In the proposed research work, the denoising method with adaptive dictionary learning is attempted to attenuate the coherent and random noises which affect the seismic reflections. In this method, Karhunen Loeve Transform (KLT) is combined with K-means Singular Value Decomposition (KSVD) to improve the Signal to Noise Ratio (SNR) of the seismic data. The combination of fixed transform using KLT and learning-based dictionary using KSVD remove the redundant data while retaining the necessary data for further interpretations. KLT is applied on the whole seismic image to decorrelate the coefficients and retain the primary reflections. The horizontal events in the seismic image are preserved as they represent the eigen images with large energy. The KSVD is applied to the KLT resultant data, which denoises the patch of data while simultaneously updating the dictionary. The noise is eliminated using local sparsity of the image where mutually overlapping small image blocks are learned to yield the self-adaptive redundant dictionary. It is then used to obtain the sparse representation of the image blocks by eliminating noise. The algorithm is tested for both synthetic and field seismic images, and the results indicate a better reduction of random and coherent noise with the acceptable execution time compared to other denoising algorithms used. The KLT combined with KSVD method decorrelate the seismic data from noise by preserving the primary reflections and discarding the redundant noise.

One of the most crucial challenges in seismic data processing is the reduction of the noise in the data or improving the signal-to-noise ratio. In this study, the 1D undecimated discrete wavelet transform (UDWT) has been acquired to attenuate random noise and ground roll. Wavelet domain ground roll analysis (WDGA) is applied to find the ground roll energy in the wavelet domain. The WDGA will be a substitute method for thresholding in seismic data processing. To compare the effectiveness of the WDGA method, we apply the 1D double density discrete wavelet transform (DDDWT) using soft thresholding in the random noise reduction and ground roll attenuation processes. Seismic signals intersect with ground roll in the time and frequency domains. Random noise and ground roll have many undesirable effects on pre-stack seismic data, and result in an inaccurate velocity analysis for NMO correction. In this paper, the UDWT by using the WDGA technique and DDDWT (using the soft thresholding technique) and the regular Fourier based method as f–k transform will be used and compared for seismic denoising.

Random noise attenuation is an important step in seismic data processing. In this paper, we propose a novel denoising approach based on compressive sensing and the curvelet transform. We formulate the random noise attenuation problem as an L 1 norm regularized optimization problem. We propose to use the curvelet transform as the sparse transform in the optimization problem to regularize the sparse coefficients in order to separate signal and noise and to use the gradient projection for sparse reconstruction (GPSR) algorithm to solve the formulated optimization problem with an easy implementation and a fast convergence. We tested the performance of our proposed approach on both synthetic and field seismic data. Numerical results show that the proposed approach can effectively suppress the distortion near the edge of seismic events during the noise attenuation process and has high computational efficiency compared with the traditional curvelet thresholding and iterative soft thresholding based denoising methods. Besides, compared with f-x deconvolution, the proposed denoising method is capable of eliminating the random noise more effectively while preserving more useful signals.

Data interpolation is a very important step in seismic data processing. At present, there are many methods based on deep learning to solve the problem of seismic data interpolation, but few studies adopt the idea of multi-scale residuals combined with attention mechanism. In this paper, a seismic data interpolation algorithm based on multi-scale residual attention network is proposed. Multi-scale convolution is applied to residual structure for the first time, and the information of seismic data is extracted adaptively by cascade of multi-scale residual blocks. Then, the output of each residual block is used as hierarchical feature for global feature fusion. In the stage of residual mapping, the residual channel attention block is introduced to promote the flow of information features and make the channel attention mechanism focus on the high frequency and edge information of seismic data, thus enhancing the ability of network representation. A lot of numerical experiments show that the algorithm have short computation time and can achieve good reconstruction results.

Time‐lapse seismic data processing is an important technique for observing subsurface changes over time. The conventional time‐lapse seismic exploration has been conducted using a large‐scale exploration system. However, for efficient monitoring of shallow subsurface, time‐lapse monitoring based on the small‐scale exploration system is required. Small‐scale exploration system using a sparker source offers high vertical resolution and cost efficiency, but it faces challenges, such as inconsistent waveforms of sparker sources, inaccurate positioning information and a low signal‐to‐noise ratio. Therefore, this study proposes a data processing workflow to preserve the signal and enhance the repeatability of small‐scale time‐lapse seismic data acquired using a sparker source. The proposed workflow has three stages: pre‐stack, post‐stack and machine learning–based data processing. Conventional seismic data processing methods were applied to enhance the quality of the sparker seismic data during the pre‐stack data processing stage. In the post‐stack processing stage, the positions and energy correction were performed, and the machine learning–based data processing stage attenuated random noise and applied a matched filter. The data processing was performed using only the seismic signals recorded near the seafloor, and the results confirmed the improvement in the repeatability of the entire seismic profile, including that of the target area. According to the repeatability quantification results, the predictability increased and the normalized root mean square decreased during data processing, indicating improved repeatability. In particular, the repeatability of the data was greatly improved through vertical correction, energy correction and matched filtering approaches. The processing results demonstrate that the data processing method proposed in this study can effectively enhance the repeatability of high‐resolution time‐lapse seismic data. Consequently, this approach could contribute to a more accurate understanding of temporal changes in subsurface structure and material properties.

As a kind of supergene geological phenomenon, ground fissure has brought great inconvenience to human life. In addition, it also has a close relationship with earthquake. However, it is very difficult to ascertain the extension depth of ground fissure since its concealment and uncertainty. In this paper, 3D seismic exploration is used to detect ground fissure in Shanxi Province of China. Specific parameters for seismic data acquisition, processing and interpretation are analysed. Firstly, seismic data acquisition method and its corresponding parameters are discussed. Small dose explosive sources and high frequency geophones are used. Small trace interval and appropriate fold are also adopted. Secondly, seismic data processing is processed from shot record to seismic profile. Multi-domain loop iteration de-noising is used to get high signal-to-noise ratio data. Accurate near surface model, interactive iteration and residual static correction are used to eliminate the impact of low velocity zone and the static correction problem. Large common middle point bin and small velocity analysis interval are used for high accuracy velocity spectrum analysis. The mute parameter of stretching distortion and the migration aperture are researched for shallow ground fissure detection. Thirdly, seismic data interpretation is processed to get ground fissure distribution. Fault enhanced filter is used to improve the signal-to-noise ratio effectively and the chimney cube is used to identify ground fissure automatically. Thus, the specific 3D seismic exploration method used in this paper is suitable for ground fissure detection. Cited as : Shi, S., Liu, Z., Feng, J., Feng, G., Li, M. Using 3D seismic exploration to detect ground fissure. Advances in Geo-Energy Research, 2020, 4(1): 13-19, doi: 10.26804/ager.2020.01.02

High-amplitude noise could interfere with useful seismic signals, affecting our ability in processing and interpreting seismic data. Thus, attenuating seismic noise is an important task in seismic processing. Total variation (TV) has played an important role in many steps of seismic data processing but always neglected the seismic structural information. Directional TV (DTV), however, considers the structural direction of seismic events but tends to cause the staircasing effect on seismic records based on the first-order formulation. Here, we develop a high-order DTV (HDTV) method for seismic denoising. It considers the local structural direction of the seismic data and calculates the higher order derivatives of seismic images to avoid the staircasing effect. We design several synthetic models that are contaminated by various types of random noise to test the denoising ability. The denoising performance of our new method is compared with the first-order DTV, conventional high-order TV, and TV regularization methods from two aspects, i.e., the signal-to-noise ratio and the effective signal leakage degree. Then, the advantages of the proposed method are further validated via several field seismic data sets.

Relevance of the work. The paper considers challenging problems related with outlining of intervals with oil and gas presence in the mature Khylly field by use of latest 3D seismic survey techniques in order to gain larger crude resources base. The purpose of this research is to discover the most promising intervals of target horizons with relatively high reservoir properties outlined by 3D seismic data. The subjects of research are 3D seismic survey data, downhole seismic survey – Vertical Seismic Profiling (VSP) and well logging diagrams. The object of research is the Khylly deposit. The paper describes in brief geological and geophysical characteristics, stratigraphic and lithological features of rocks making the section. It is noted that despite repeated surveys by use of various geological and geophysical techniques, the field setting is not thoroughly studied and it has been covered by 3D seismic survey in 2012. Research results. 3D seismic survey applied across Khylly area is resulted in drawing of 4 structural maps for III and I horizons of Productive Series (PS), Akchagyl and Lower Absheron suites. Taking into account the relevance of structural planes of various stratigraphic levels and III horizon of PS being one of the major reference horizons the paper gives description of structural map drawn for this horizon. The detailed velocity model is designed based on VSP data with wide use of velocity analysis data. It has been made clear that Khylly area has block structure and each block has been described in detail. Based on acquired data it has been recommended to drill exploratory well R-1. Conclusion. Processing and interpretation of seismic data are aimed at solving some geological problems; the main task was to obtain results that ensure the study of the geological structure in the seismic survey area, including tracing of seismic horizons, faults and outlining the areas and section intervals which may be of interest due to possible oil and gas presence. VSP data acquired in well 2012 and velocity analysis made it possible to design velocity model of the section under the study, with the use of which the temporary 3D cube was transformed into a depth cube. The quality of seismic data is good and made it possible to solve the tasks set for this research.

The application of the seismic reflection method to detect stratigraphic changes away from well locations is beginning to gain acceptance among Reservoir Engineers. Unfortunately one of the limitations of this method is the inherent low resolution of seismic data as compared to well log data. Several approaches have been proposed to produce well log-like or pseudo log traces from seismic reflection data. Although there is no improvement in the resolution, the pseudo log traces provide a more convenient way of visually corelating well data to seismic data. A new stratigraphic extrapolation method designed to overcome some of the drawbacks of low resolution is presented. This method consists of relating a model trace (initially derived from well information) to an actual seismic trace by varying the parameters of the model trace within well defined constraints until best match is achieved between them. Extrapolation of the stratigraphy away from the well is performed from trace to trace, the starting model for each trace being determined either from the final model for the previous trace or from an independently postulated model. An interactive computer program is employed for this purpose which allows the incorporation of additional information to the initial model. In this way the final stratigraphic model of the reservoir is the one which best matches the seismic data. Several examples showing the application of the method will be discussed.

Ground roll attenuation is an important step in seismic data processing. We implement a procedure to eliminate ground roll from seismic data that is based on eigen-images. We use singular-value decomposition (SVD) to distinguish between signal and noise within narrow frequency bands of the ground-roll cone at near offsets in the data. The estimate of the energy of the signal comes from an uncontaminated portion of the seismic records. The method is able to adapt to the changing character of the noise and signal about the seismic survey. The modeled noise is then subtracted from the data while preserving the reflections.

High Resolution, Simultaneous VSP and Land Seismic Acquisition

Summary The random noise in seismic data seriously limits the extraction of information related to underground geological structure and lithology parameters. How to remove such noise sufficiently while keeping the fidelity of original seismic data is an important issue in the seismic data processing. In the poststack seismic profile, the most obvious feature is the structural feature and non-stationary feature. Inspired by it, we propose an improved non-local means (NLM) method with adaptive search window (ASWNLM). This method includes determining the direction of the search window by a correlation algorithm and the height of the search window by the instantaneous center frequency. Tests on synthetic data and real post-stack data prove our method can remove random noise as much as possible while preserving signal detail well even without a proper filtering parameter h.

Radon filters are often used for removal of multiple reflections from normal move-out-corrected seismic reflection data. In the conventional Radon transform, integration surfaces are hyperbolic rather than linear. This specific hyperbolic surface is equivalent to a parabola in terms of computational expense, but more accurately distinguishes multiples from primary reflections. The forward transform separates seismic arrivals by their differences in travel time move-out. Multiples can be suppressed by an inverse transform of the seismic data. Examples show that multiples are effectively attenuated in pre-stack and stacked seismograms. Based on the parabolic Radon transform, a new method is utilized for missing offset restoration, resampling and regularization of pre-stack individual common depth point (CDP) gathers. The method is also valid for resampling spatially aliased seismic data. Restoration of missing offsets and trace interpolation is an interesting and important problem in seismic data processing. Here we present an application of Radon transform on a multichannel seismic data set from the western continental margin of India (WCMI), which shows remarkable signal enhancement.