SIGNAL-TO-NOISE RATIO ENHANCEMENT IN MULTICHANNEL SEISMIC DATA VIA THE KARHUNEN-LOÉVE TRANSFORM*

The Karhunen-Loéve transform, which optimally extracts coherent information from multichannel input data in a least-squares sense, is used for two specific problems in seismic data processing. The first is the enhancement of stacked seismic sections by a reconstruction procedure which increases the signal-to-noise ratio by removing from the data that information which is incoherent trace-to-trace. The technique is demonstrated on synthetic data examples and works well on real data. The Karhunen-Loéve transform is useful for data compression for the transmission and storage of stacked seismic data. The second problem is the suppression of multiples in CMP or CDP gathers. After moveout correction with the velocity associated with the multiples, the gather is reconstructed using the Karhunen-Loéve procedure, and the information associated with the multiples omitted. Examples of this technique for synthetic and real data are presented.     

Pre-stack Kirchhoff depth migration local space-shift imaging condition: synthetic and data examples

Extracting accurate common image angle gathers from pre-stack depth migrations is important in the generation of any incremental uplift to the amplitude versus angle attributes and seismic inversions that can lead to significant impacts in exploration and development success. The commonly used Kirchhoff migration outputs surface common offset image gathers that require a transformation to angle gathers for amplitude versus angle analysis. The accuracy of this transformation is one of the factors that determine the robustness of the amplitude versus angle measurements. Here, we investigate the possibility of implementing an extended imaging condition, focusing on the space-lag condition, for generating subsurface reflection angle gathers within a Kirchhoff migration. The objective is to determine if exploiting the spatial local shift imaging condition can provide any increase in angle gather fidelity relative to the common offset image gathers. The same restrictions with a ray-based approach will apply using the extended imaging condition as both the offset and extended imaging condition method use travel times derived from solutions to an Eikonal equation. The aims are to offer an alternative ray-based method to generate subsurface angle gathers and to understand the impact on the amplitude versus angle response. To this end, the implementation of the space-shift imaging condition is discussed and results of three different data sets are presented. A layered three-dimensional model and a complex two-dimensional model are used to assess the space shift image gathers output from such a migration scheme and to evaluate the seismic attributes relative to the traditional surface offset common image gathers. The synthetic results show that the extended imaging condition clearly provides an uplift in the measured amplitude versus angle over the surface offset migration. The noise profile post-migration is also improved for the space-lag migration due to the double summation inside the migration. Finally, we show an example of a space-lag gather from deep marine data and compare the resultant angle gathers with those generated from an offset migration and a time-shift imaging condition Kirchhoff migration. The comparison of the real data with a well log shows that the space-lag result is a better match to the well compared to the time-lag extended imaging condition and the common offset Kirchhoff migration. Overall, the results from the synthetics and real data show that a Kirchhoff migration with an extended imaging condition is capable of generating subsurface angle gathers with an incremental improvement in amplitude versus angle fidelity and lower noise but comes at a higher computational cost.     

在成像空间中衰减多次波方法研究

在偏移后成像空间中的共成像点道集中可以对多次波进行衰减,对于给定的偏移速度模型,一次波与多次波在叠前偏移后的共成像点道集中具有不同的动校时差,这样我们就可以使用类似于偏移前衰减多次波的方法将一次波和多次波进行分离.本文在成像空间中应用抛物Radon变换分离多次波和有效波,由于每个共成像点道集都包含了复杂三维波场传播效应,所以本文方法具有处理三维数据和复杂地下构造的能力.相比于SRME以及传统Radon变换衰减多次波方法,本文方法能够在保持较小的计算量的同时,保证了衰减多次波的准确性.通过对模型数据试算和对实际数据的处理验证了本文方法在叠前时间偏移后衰减多次波的能力,并取得了很好的成像效果.     

Identification and classification of multiple reflections with self-organizing maps

Artificial neural networks can be used effectively to identify and classify multiple events in a seismic data set. We use a specialized neural network, a self-organizing map (SOM), that tries to establish rules for the characterization of the physical problem. Selected seismic data attributes from CMP gathers are used as input patterns, such that the SOM arranges the data to form clusters in an abstract space. We show with synthetic and real data how the SOM can identify and classify primaries and multiples, and how it can classify the various types of multiple corresponding to a certain generating mechanism in the subsurface.     

基于共聚焦点道集的叠前深度偏移

共聚焦点(CFP)偏移技术是一种基于等时原理,将Kirchhoff积分法的一步偏移分两步聚焦（即激发聚焦和检波聚焦）来完成的叠前地震成像方法.该方法借助于逆时聚集算子和共聚焦点道集来实现叠前偏移成像.基于共聚焦点道集的叠前深度偏移是把基于共炮集的深度偏移的算法引入到CFP技术上来,基于波场延拓的理论来实现偏移成像,该方法首先生成共聚焦点道集,然后基于面炮合成的理论合成聚焦震源,最终通过相关成像来实现叠前偏移成像.该方法选取较少的聚焦点就可以实现对于地下构造的偏移成像,和炮域波动方程偏移相比,其计算效率得到了提高.通过模型试算和实际资料的试处理,验证了该方法在实现叠前深度偏移成像上的正确性和有效性.     

Sub-basalt imaging problems and the application of Artificial Neural Networks

In many areas of the world, the presence of shallow high velocity, highly heterogeneous layers complicate seismic imaging of deeper reflectors. Of particular economic interest are areas where potentially hydrocarbon-bearing strata are obscured by layers of basalt. Basalt layers are highly reflective and heterogeneous. Using reflection seismic, top basalt is typified by a high-amplitude, coherent reflector with poor resolution of reflectors below the basalt, and even bottom basalt. Here, we present a new approach to the imaging problem using the pattern recognition abilities of a back-propagation Artificial Neural Network (ANN). ANNs are computational systems that attempt to mimic natural biological neural networks. They have the ability to recognize patterns and develop their own generalizations about a given data set. Back-propagation neural networks are trained on data sets for which the solution is known and tested on the data that are not previously presented to the ANN in order to validate the network result. We show that Artificial Neural Networks, due to their pattern recognition capabilities, can invert the medium statistics based on the seismic character. We produce statistically defined models involving a basalt analogous layer, and calculate full wavefield finite difference synthetic seismograms. We vary basalt layer thickness and source frequency to generate a synthetic model that produces seismic that is similar to real sub-basalt seismic, i.e. high amplitude top basalt reflector and the absence of base basalt and sub-basalt events. Using synthetic shot gathers, generated in a synthetic representation of the sub-basalt case, we can invert the velocity medium standard deviation by using an ANN. By inverting the velocity medium standard deviation, we successfully identified the transition from basalt to sub-basalt on the synthetic shot gathers. We also show that ANNs are capable of identifying the basalt to sub-basalt transition in the presence of incoherent noise. This is important for any future applications of this technique to the real-world seismic data, as this data is never completely noise-free. There is always a certain level of residual (noise remaining after initial noise filtering) environmental/ambient noise present on the recorded seismics, hence, neural network training with noise-free synthetic seismic is less than optimal.     

OFFSET CONTINUATION OF SEISMIC SECTIONS*

Geometrical acoustic and wave theory lead to a second-order partial differential equation that links seismic sections with different offsets. In this equation a time-shift term appears that corresponds to normal moveout; a second term, dependent on offset and time only, corrects the moveout of dipping events. The zero-offset stacked section can thus be obtained by continuing the section with maximum offset towards zero, and stacking along the way the other common-offset sections. Without the correction for dip moveout, the spatial resolution of the section is noticeably impaired, thus limiting the advantages that could be obtained with expensive migration procedures. Trade-offs exist between multiplicity of coverage, spatial resolution, and signal-to-noise; in some cases the spatial resolution on the surface can be doubled and the aliasing noise averaged out. Velocity analyses carried out on data continued to zero offset show a better resolution and improved discrimination against multiples. For instance, sea-floor multiples always appear at water velocity, so that their removal is simplified. This offset continuation can be carried out either in the time-space domain or in the time-wave number domain. The methods are applied both to synthetic and real data.     

基于保幅拉东变换的多次波衰减

为在去除多次波时有效保护地震一次反射波数据的AVO现象,给后续反演、解释提供准确的地震数据,本文提出了一种基于保幅拉东变换的多次波衰减方法,该方法是对常规抛物拉东变换的修改,把常规的稀疏拉东变换在拉东域分成两部分：一部分用于模拟零偏移距处的反射波能量,增加的另一部分用于模拟反射波振幅的AVO特性.该方法不仅考虑了反射波同相轴的形状,还考虑了反射波同相轴振幅幅度的变化,从而可把反射波信息进行有效转换,进而有利于多次波的消除,更好地恢复有效波的能量.在把地震数据由时间域转换到拉东域时,本文采用了IRLS算法实现保幅拉东算子的反演.模型数据和实际地震道集的试算分析表明,与常规拉东变换相比,保幅拉东变换在去除多次波的同时可有效保护一次反射波的AVO现象.     

Kirchhoff叠前时间偏移角度道集

提出三维Kirchhoff叠前时间偏移角度域共像点道集的改进算法,克服传统角度求取算法局限,可计算相对倾斜地层法线入射角;与Kirchhoff直射线叠前时间偏移求角度算法相比,本文方法考虑射线弯曲效应,包含层速度,角度范围加大,更接近真实入射角;计算走时采取弯曲射线或者适应线性横向变速介质的非对称走时等算法,角度道集在大角度处得到拉平;采用相对保幅的权因子以及覆盖次数校正技术,有利于叠前AVA反演.模型测试结果表明：叠前时间偏移角度道集,相对CMP、CRP所转化角度道集,更准确反应AVA效应;实际三维数据测试表明本文方法可以提供品质优良的角度道集,适用于AVA分析、反演,提高叠前反演分辨率.     

共转换点道集S变换谱相似性的多波地震数据横波波场提取技术

横波速度动校正后的共转换点(CCP)道集内,同时刻的各道横波信号S变换(ST)谱与其叠加道ST谱具有相似关系.因此,可基于这种相似关系设计自适应滤波器来提取多波地震数据中的横波波场.首先对共中心点(CMP)道集应用纵波速度动校正并在各道减去叠加道来去除数据中的纵波波场;然后在CCP道集应用横波速度动校正,将地震道振幅水平调整至叠加道振幅水平并做S变换,以叠加道ST谱为参考对地震道ST谱进行自适应滤波,去除数据中的残余纵波和噪声;最后,将滤波结果的振幅水平恢复至滤波前振幅水平.理论和实际数据试算表明,本文方法可有效提取多波地震数据中的横波波场,为多波多分量横波数据处理提供新思路.     

Seismic ground-roll noise attenuation using deep learning

We propose to adopt a deep learning based framework using generative adversarial networks for ground-roll attenuation in land seismic data. Accounting for the non-stationary properties of seismic data and the associated ground-roll noise, we create training labels using local time–frequency transform and regularized non-stationary regression. The basic idea is to train the network using a few shot gathers such that the network can learn the weights associated with noise attenuation for the training shot gathers. We then apply the learned weights to test ground-roll attenuation on shot gathers, that are not a part of training input to obtain the desired signal. This approach gives results similar to local time–frequency transform and regularized non-stationary regression but at a significantly reduced computational cost. The proposed approach automates the ground-roll attenuation process without requiring any manual input in picking the parameters for each shot gather other than in the training data. Tests on field-data examples verify the effectiveness of the proposed approach.     

一阶多次波聚焦变换成像

将多次波转换成反射波并按传统反射波偏移算法成像,是多次波成像的一种方法.聚焦变换能准确的将多次波转换为纵向分辨率更高的新波场记录,其中一阶多次波转换为反射波.本文对聚焦变换提出了两点改进:1)提出局部聚焦变换,以减小存储量和计算量,增强该方法对检波点随炮点移动的采集数据的适应性;2)引入加权矩阵,理论上证明原始记录的炮点比检波点稀疏时,共检波点道集域的局部聚焦变换可以将多次波准确转换成炮点与检波点有相同采样频率的新波场记录.本文在第一个数值实验中对比了对包含反射波与多次波的原始记录做局部聚焦变换和直接对预测的多次波做局部聚焦变换两种方案,验证了第二种方案转换得到的波场记录信噪比更高且避免了第一个方案中切聚焦点这项比较繁杂的工作.第二个数值实验表明:在炮点采样较为稀疏时,该方法能有效的将一阶多次波转换成反射波;转换的反射波能提供更丰富的波场信息,成像结果更均衡、在局部有更高的信噪比,以及较高的纵向分辨率.     

A compound method for random noise attenuation

Random noise attenuation, preserving the events and weak features by improving signal‐to‐noise ratio and resolution of seismic data are the most important issues in geophysics. To achieve this objective, we proposed a novel seismic random noise attenuation method by building a compound algorithm. The proposed method combines sparsity prior regularization based on shearlet transform and anisotropic variational regularization. The anisotropic variational regularization which is based on the linear combination of weighted anisotropic total variation and anisotropic second‐order total variation attenuates noises while preserving the events of seismic data and it effectively avoids the fine‐scale artefacts due to shearlets from the restored seismic data. The proposed method is formulated as a convex optimization problem and the split Bregman iteration is applied to solve the optimization problem. To verify the effectiveness of the proposed method, we test it on several synthetic seismic datasets and real datasets. Compared with three methods (the linear combination of weighted anisotropic total variation and anisotropic second‐order total variation, shearlets and shearlet‐based weighted anisotropic total variation), the numerical experiments indicate that the proposed method attenuates random noises while alleviating artefact and preserving events and features of seismic data. The obtained result also confirms that the proposed method improves the signal‐to‐noise ratio.     

基于偏移成像道集的剩余静校正方法

针对陆上地震资料处理的静校正问题,提出了一种基于偏移成像道集的剩余静校正方法.与传统的由动校正后的CMP道集中拾取剩余时差不同,本文基于偏移成像道集求取剩余时差,避免了复杂情况下同相轴归位不准确导致的剩余时差拾取误差.通过生成随炮点和检波点位置变化的偏移道集,实现了由偏移道集中直接拾取炮、检点的地表一致性剩余时差;该炮、检点偏移道集只在指定的局部时窗生成,并不增加大的计算量.二维和三维实际数据测试表明了该方法的有效性和实用性.     

Attenuation of marine seismic interference noise employing a customized U-Net

Marine seismic interference noise occurs when energy from nearby marine seismic source vessels is recorded during a seismic survey. Such noise tends to be well preserved over large distances and causes coherent artefacts in the recorded data. Over the years, the industry has developed various denoising techniques for seismic interference removal, but although well performing, they are still time-consuming in use. Machine-learning-based processing represents an alternative approach, which may significantly improve the computational efficiency. In the case of conventional images, autoencoders are frequently employed for denoising purposes. However, due to the special characteristics of seismic data as well as the noise, autoencoders failed in the case of marine seismic interference noise. We, therefore, propose the use of a customized U-Net design with element-wise summation as part of the skip-connection blocks to handle the vanishing gradient problem and to ensure information fusion between high- and low-level features. To secure a realistic study, only seismic field data were employed, including 25,000 training examples. The customized U-Net was found to perform well, leaving only minor residuals, except for the case when seismic interference noise comes from the side. We further demonstrate that such noise can be treated by slightly increasing the depth of our network. Although our customized U-Net does not outperform a standard commercial algorithm in quality, it can (after proper training) read and process one single shot gather in approximately 0.02 s. This is significantly faster than any existing industry denoising algorithm. In addition, the proposed network processes shot gathers in a sequential order, which is an advantage compared with industry algorithms that typically require a multi-shot input to break the coherency of the noise.     

叠前三参数非高斯反演方法研究

针对地球物理反演中广泛采用的"噪声高斯分布假设",本文研究了叠前地震资料中噪声的非高斯分布特征,提出了针对非高斯噪声的地震叠前非高斯反演概念和思想,构造了能同时压制高斯和非高斯噪声的混合范数作为反演目标函数,采用改进的Powell算法进行求解,有效地抑制了叠前地震资料中的高斯和非高斯混合噪声.模型试算和实际地震数据的反演结果验证了方法的正确性和算法的可靠性.     

基于单程波偏移算子的地表相关多次波成像

在常规地震资料处理中,多次反射波被视为噪声并从地震数据中去除,以免在之后的地震资料解释中造成误解.而事实上,多次波也是地震信号,是照明波场的一部分,能够对地下构造成像的精度做出贡献.本文分析了多次波在传统单程波叠前深度偏移中产生构造假象的机制和表现,为实现基于单程波偏移算子的多次波成像,修改了单程波叠前深度偏移的边界条件,即将输入的震源波场用包含多次波的记录来替代,输入的记录波场用预测出的表层相关多次波来替代,实现了基于单程波偏移算子的地表相关多次波成像,并从理论上给出了其成像依据.通过基于二范式最小能量差原则求取的匹配因子,将多次波成像结果与一次波成像结果进行匹配叠加,应用多次波成像来弥补一次波成像的不足.简单模型验证了基于单程波偏移算子的多次波成像方法的有效性,最后对Sigsbee2B模型进行了一次波与多次波联合成像试算,盐边界高陡构造成像质量得到了明显改善.     

Interferometric interpolation of sparse marine data

We present the theory and numerical results for interferometrically interpolating 2D and 3D marine surface seismic profiles data. For the interpolation of seismic data we use the combination of a recorded Green's function and a model‐based Green's function for a water‐layer model. Synthetic (2D and 3D) and field (2D) results show that the seismic data with sparse receiver intervals can be accurately interpolated to smaller intervals using multiples in the data. An up‐ and downgoing separation of both recorded and model‐based Green's functions can help in minimizing artefacts in a virtual shot gather. If the up‐ and downgoing separation is not possible, noticeable artefacts will be generated in the virtual shot gather. As a partial remedy we iteratively use a non‐stationary 1D multi‐channel matching filter with the interpolated data. Results suggest that a sparse marine seismic survey can yield more information about reflectors if traces are interpolated by interferometry. Comparing our results to those of f‐k interpolation shows that the synthetic example gives comparable results while the field example shows better interpolation quality for the interferometric method.     

Deep neural network application for 4D seismic inversion to changes in pressure and saturation: Optimizing the use of synthetic training datasets

In this work, we tackle the challenge of quantitative estimation of reservoir dynamic property variations during a period of production, directly from four-dimensional seismic data in the amplitude domain. We employ a deep neural network to invert four-dimensional seismic amplitude maps to the simultaneous changes in pressure, water and gas saturations. The method is applied to a real field data case, where, as is common in such applications, the data measured at the wells are insufficient for properly training deep neural networks, thus, the network is trained on synthetic data. Training on synthetic data offers much freedom in designing a training dataset, therefore, it is important to understand the impact of the data distribution on the inversion results. To define the best way to construct a synthetic training dataset, we perform a study on four different approaches to populating the training set making remarks on data sizes, network generality and the impact of physics-based constraints. Using the results of a reservoir simulation model to populate our training datasets, we demonstrate the benefits of restricting training samples to fluid flow consistent combinations in the dynamic reservoir property domain. With this the network learns the physical correlations present in the training set, incorporating this information into the inference process, which allows it to make inferences on properties to which the seismic data are most uncertain. Additionally, we demonstrate the importance of applying regularization techniques such as adding noise to the synthetic data for training and show a possibility of estimating uncertainties in the inversion results by training multiple networks.     

局部倾角约束最小二乘偏移方法研究

随着石油勘探难度的进一步加大,地震数据往往存在采样不规则、地震道缺失等现象,如果不对其进行处理,会对后续的地震成像产生影响,引入成像噪音.针对这一问题,一般是通过地震道插值或数据规则化对叠前数据进行处理,然后采用常规的偏移方法进行成像,本文则是将地震成像看作最小二乘反演问题,在共成像点道集引入平滑算子,在共偏移距/角度道集引入平面波构造算子(PWC)进行约束,通过预条件共轭梯度法使得反偏移后数据与输入数据之间的误差达到最小,最终得到信噪比更高、振幅属性更为可靠的成像结果.理论模型和实际资料处理表明,本文方法不仅可以有效压制数据不规则对成像产生的噪音,而且具有更高的成像精度.