地球物理方向频率合并的期刊文献

When multiple seismic surveys are acquired over the same area using different technologies that produce data with different frequency content, it may be beneficial to combine these data to produce a broader bandwidth volume. We propose a workflow for matching and blending seismic images obtained from shallow high-resolution seismic surveys and conventional surveys conducted over the same area. The workflow consists of three distinct steps: (a) balancing the amplitudes and frequency content of the two images by non-stationary smoothing of the high-resolution image; (b) estimating and removing variable time shifts between the two images; and (c) blending the two images together by least-squares inversion. The proposed workflow is applied successfully to images from the Gulf of Mexico.

Seismic blended source acquisition, also referred to as simultaneous source acquisition, is a cost‐effective technology that achieves a significant reduction in acquisition cycle time and increases seismic crew field productivity. The dispersed source array is a blended acquisition field technique that simultaneously employs sources emitting different types of sweeps (i.e. multi‐sweep), in terms of frequency bandwidth and length, which ultimately will result in a full broadband seismic data. In this paper, deblending of 3D multi‐sweep seismic blended data and the subsequent merging of the data volumes having different frequency bandwidths will be discussed. In specific data domains where the signal component is coherent, interference shots (i.e. blending noise) are randomly distributed in the data space according to its own shot firing time. Therefore, the deblending process, which separates interference shots from a signal component, becomes a noise attenuation problem. A sparse inversion methodology is applied in the frequency–wavenumber–wavenumber (f–kx–ky) domain to attenuate blending noise. By applying this deblending methodology to both dispersed source array's low‐ and mid‐high‐frequency bandwidths, we obtained high‐quality deblending results. For both frequency bandwidths of the deblended dispersed source array data, additional effort was made to combine the two datasets to a single broadband data volume. Consequently, deblending and merging of the dispersed source array blended data generated a broadband, deblended and well‐balanced seismic volume suitable for further processing and reservoir characterization applications.

The detection of subsurface hydrocarbon reservoirs often relies on the frequency content of 3D seismic data. Generally, it is best to use broadband seismic data for seismic interpretation, attribute mapping, and amplitude inversion. We describe an experiment in which a 3D seismic volume was made by combining different frequency bands from two overlapping 3D seismic surveys. This was done to increase the frequency bandwidth and associated resolution with respect to either individual survey.

… In this study, we aim to assess a workflow that combines two post-migration seismic data conditioning methods. Starting with a merged dataset exhibiting amplitude and frequency …

… merging is done in the frequency-space domain using simple weighting functions. We carried out numerical validation to merge … We proposed to perform the merging using seismic …

“Spectral Fusion” is a new tool designed to combine different seismic datasets covering the same surveyed area, with<br>different but overlapping bandwidths resulting for instance from various source/receiver depths for marine streamer data. By<br>combining data with overlapping spectra a larger bandwidth can be recovered, with the benefit of using all of the available<br>information in the overlapping frequency domain without requiring any wave shaping or predefined filters, and therefore<br>better preserving the individual phase and amplitude character of the input data.

During exploration, it is common to have access to multiple 3D post-stack seismic volumes, with different acquisition parameters, geometries and data quality that means conducting integrated interpretation and analyses of data between and across surveys can be challenging. Merging a 3D seismic survey would ideally require full re-processing using pre-stack data. However, given time, budgetary, data availability, or compute constraints, this is not always a practical solution. An alternative solution is to merge post-stack seismic volumes, which can provide a suitable substitute for many regional structural interpretation workflows. However, achieving a seamless post-stack seismic merge is challenging and often matching amplitudes, phase, and time shifts consistently cannot be done using bulk corrections across an entire survey due to lateral changes in geology. Seams at the boundaries between seismic surveys or changes in waveform characteristics will create challenges for interpretation, especially for the predictive tools being adopted for automated interpretation workflows. For example, a seam between two datasets is a large discontinuity that will lead to a potential false positive fault prediction. The seam may also prevent horizon interpretation tools from traversing between the two volumes. To reduce these challenges, we present an automated workflow for seamlessly merging two or more post-stack 3D seismic volumes that will enable easier interpretation workflows across regions where interpretations need to span multiple volumes. Traditionally, survey merging is achieved by applying seismic data interpolation/regularization, phase rotation, amplitude matches, and the optimal bulk time shift obtained to several seismic surveys to condition the different vintages and produce a seamless merged data set. This has been mainly done by designing a post-migration match filter which estimates wavelets in each time zone after isolating trace segments in the defined window and derived by least-squares time-domain with a smoothness constraint. However, after applying those estimated time, phase and amplitude adjustments, the surveys especially their spectra are matched in a loose sense. In case where the low frequency legacy data and high-resolution modern data are required to merge, the final spectrum is closer to low frequency spectrum of legacy data. The varying levels of resolution, acquisition patterns, and data quality in seismic surveys make their integration complex task. To overcome these challenges a mixture of selected deep-learning, optimization and classical machine-learning algorithms will be utilized to analyze and process data from multiple surveys. These algorithms will identify overlaps in the data and employ advanced techniques to match the spectra and phase of the seismic data, resulting in a seamless transition between different seismic surveys. Ultimately a complete merged 3D regional seismic survey will be generated. Various machine learning algorithms, including different types of neural networks will be utilized to analyze the data, detect gaps, and fill them with accurate data values.

… to frequency, phase, energy and time differences in the merge processing on the … merged seismic data, the proposed MMP significantly enhances the consistency of amplitude frequency…

Merging 3D seismic surveys into a seamless single 3D volume, either post-stack stage or pre-stack stage, is a challenging task on seismic data processing. This study describes some tips from coastline Abu Dhabi where we successfully managed merging two partially overlapping surveys during post-stack stage, one from transition zone (land and shallow marine) while the other one from offshore 3D OBC seismic survey, in order to understand subtle geological structure relationship among two areas. Since spatial sampling between two surveys are greatly diverse due to different orientations and grid sizes, a conjugate grid which is identical to a main cube had been applied over the subordinate one that enable the whole dataset to interpolate and process as a same grid. Then we deployed pre-conditioning steps over the subordinate dataset to minimize their quality differences where we particularly focused on residual noise, multiples, frequency contents and event timings. Lastly, a matching filter was designed and applied to the subordinate side to compensate residual amplitude, frequency and phase and produce a final dataset for structure interpretation. A single consolidated seamlessly merged volume was produced throughout the steps as described above along with well-to-seismic calibrations. Seismic interpretation was conducted over the main reservoir between two datasets/fields with a good degree of confidence. The present day structure separation and structure growth history were analyzed as a part of the structure interpretation. Moreover, this case study illustrates the add values of the seismic 3D merge from the aspect of regional structure restoration, and revealing structure relationship between the overlapping surveys. The final merging result outcome of this case study has an amenable structure continuity and seamless horizon mapping of the common reservoir target level between the two surveys. Additionally, the merged seismic cube did showcase a lateral zero phase wavelet stability of the existing wells that verified the reliability of the conducted post-stack seismic merging processing workflow. This case study has successfully demonstrated and summarized key technical tips that are recommended for merging datasets on post-stack domain in the future. However, pre-stack merge is also and still strongly recommended since static corrections, velocity pickings and imaging processing can be applied throughout the two surveys in one go while these cannot be fixed on post-stack merge. From data acquisition perspectives; enough overlap to reaching the full-fold rim of the overlapping surveys is highly recommended.

… low-frequency phase shift, applied to the seismic relative impedances, in the search for the best spectral merge. The background models are specified by a low-cut corner frequency and …

… for seismic reservoir characterization, the amplitude-related seismic attributes are of greater importance. Therefore amplitude-spectrum matching is critical to survey merge processing. …

Recent advancements in seismic first-arrival picking have been driven by U-Net-based networks, yet their performance often degrades in low-SNR environments and complex subsurface conditions, where noise and interfering waves (e.g., reflections) obscure weak arrivals. Conventional approaches, relying predominantly on time-domain features, suffer from insufficient representation, limiting their accuracy. To address these challenges, we propose a time–frequency fusion network (TFFNet), a novel U-Net architecture that effectively integrates time- and frequency-domain features to enhance picking performance. At its core is the TFFConv module, which employs parallel branches within a time–frequency attention fusion (TFAF) framework. The frequency branch leverages a Haar-discrete-wavelet-transform (DWT)-based directional decomposition, augmented by learnable weights and frequency-domain attention (FDA), to suppress noise adaptively and amplify first-arrival signals. The time branch employs patch attention to extract task-relevant features. These branches are optimally combined via the power-averaged channel fusion (PACF) module. Comprehensive experiments on three low-SNR field datasets (Lalor, Qaidam, and Xiong’an) demonstrate TFFNet’s efficacy, outperforming four state-of-the-art methods (U-Net benchmark, STU-Net, DSU-Net, and U-Net3+). TFFNet achieves hit rate within 1-pixel error (HR@1px) of 90.9%, 79.6%, and 56.0% on Lalor, Qaidam, and Xiong’an, respectively, with improvements of up to 23.2% on the challenging Xiong’an dataset. These results underscore TFFNet’s potential for precise seismic first-arrival picking in noisy and complex geological settings. Codes are available at https://github.com/WHOAREYOUXSC/TFFNet

… seismic frequency-decomposed attribute fusion involves the analysis of seismic wave responses across various frequency … multiple seismic attributes derived from diverse frequency …

… More importantly, the GPR and seismic detection have similar imaging theory. The data were acquired at the Campus of Zhejiang University, China, with the purpose to identify the …

Abstract Resolution of seismic data is critical to reservoir prediction and the near surface underground structure. Low resolution of seismic data is caused by many factors, such as the surface environment, shielding layer and sand body. To widen the frequency bandwidth and improve the resolution of the post-stack seismic reflection profiles, we analyze the characteristics of amplitude-spectrum of the post-stack profiles and design frequency expanding inverse filter based on the logging data for data fusion, including the loss of longitudinal frequency information and the variation of transverse frequency. This method creates broadband synthetic seismogram by using a wide bandwidth seismic wavelet and a series of reflection coefficients calculated from acoustic logging and density logging data. By analyzing the synthetic seismogram of widening frequency band with borehole seismic traces, the loss regularity of vertical frequency signal generated by low-pass filter of formation is restored. Additionally, to compare borehole seismic traces with arbitrary cross-well trace, the filtering effect about transverse variation of formation is obtained. The successful application of the inverse filter in terms of data fusion to acquire a new profile with better resolution called the seismic-logging profile, demonstrated that the proposed data fusion method has the advantage of being efficient in compensating the high frequency components of seismic section.

… seismic attributes from the "optimized" frequency of seismic … frequency" is confirmed first based on the seismic frequency … the "optimized frequency". Besides, a novel strategy named "…

… -order differential fusion method in frequency domain to broaden the seismic bandwidth. … high frequency and suppress low frequency of signals, namely it has the attribute of frequency …

… propose an adaptive multi-frequency and multi-scale (… frequency hierarchical adaptive refinement module (MFAR) in the decoder. DCEW decomposes feature maps into multi-frequency …

… (2011), we demonstrate that taking into account a real seismic source s(t) with a limited frequency bandwidth, the WR suffers from distortions: |$\tilde{R}\left[\xi ,p\right]\left(t,a\right) = W\…

Abstract Defining the boundaries, thicknesses and sedimentary facies of fluvial reservoirs (sand bodies) is critical for predicting hydrocarbon volumes, designing schemes for petroleum exploration and development and improving oil recovery. Most reservoirs contain thick and thin sand bodies at the same intervals, while the amplitude values of seismic data usually highlight sand bodies near the 1/4 wavelength for the tuning phenomena. Hence, the application of spectral decomposition to seismic attributes and the combination of multiple frequency-decomposed (spectral-decomposed) seismic attributes have gained increasing attention for the readjustment of tuning thickness to predict sand bodies of various thicknesses. However, the popular method of red-green-blue blending is a simple linear combination of three frequency-decomposed seismic attributes that qualitatively analyzes the sand thickness without well-log interpretation. This research proposes machine learning fusion as a new nonlinear method for fusing high-, middle-, and low-frequency seismic attributes. This method uses machine learning to link well-log interpretation and multiple-frequency seismic attributes for the quantitative prediction of sand thickness, which is important for development work in a mature field. Test results of the conceptual model and the real case indicate that the predicted sand thickness after fusing multiple frequency-decomposed seismic attributes is approximately in line with the actual thickness (correlations between 80 and 90%). Combined with the coherence attribute and the red-green-blue blending results, the distributions and histories of sedimentary facies are analyzed based on the predicted sand thickness and well data. The results suggest that the proposed method can effectively readjust the tuning thickness and improve the resolution of seismic interpretation. This method is a potentially effective technique to characterize the sand thickness and sedimentary facies in other fields using a similar geological setting and dataset.

Abstract Understanding the hierarchical architectural elements of fluvial sand bodies is important for planning their development strategy and to enhance oil recovery. Red-Green-Blue (RGB) blending of multiple seismic attributes and forwarding seismic modelling are commonly used in the analysis of compound sand bodies. However, RGB blending of multiple seismic attributes can only qualitatively describe the boundaries and thickness of sand bodies. The forward seismic modelling techniques previously documented in the literature are not effective when depicting the geometry of, and stacking relationships between, sand bodies (i.e., reservoir compartmentalisation). Hence, we propose in this work a new workflow that combines fused spectral-decomposition seismic attributes (SDSAs) and forwarding seismic modelling to quantitatively predict sand thickness, and to characterise stacking relationships between sand bodies. First, we employ a Support Vector Machine (SVM) algorithm to fuse high, middle, and low frequency components (attributes) of seismic data so as to quantitatively predict the thickness of sand bodies. Second, we define the seismic waveform response patterns corresponding to the typical conceptual stacked sand bodies. With the constraints of waveform patterns and predicted sand thickness (fused SDSAs), the geometry and stacking relationships of the sand bodies are characterised by forward seismic modelling. To illustrate the effectiveness of our proposed workflow, we apply it to the Neogene Minghuazhen Formation (Nm) of the QHD 32–6 oil field, Bohai Bay Basin, China. We define five architectural elements of a meandering fluvial reservoir by analysing the hierarchy of sand bodies using our workflow. The predicted sand bodies in this workflow were further proven by horizontal drilling and production data.

… seismic signals with narrow frequency bands and obtain more refined seismic data,this paper proposes a seismic data … fusion in time domain, which automatically obtains the weighted …

… Therefore, the wide-frequency and high-density 3D seismic data acquisition was deployed … narrow azimuth 3D seismic acquisition direction, and the quality of seismic data and the …

… the responses of seismic bodies in different frequency bands. In order to make effective use of seismic frequency information and reasonably display the dominant frequency of each …

… to realize the effective fusion of multi-directional seismic information, increase the seismic illumination from different directions, improve the extension of frequency band, and improve …

The application of artificial intelligence in seismic impedance inversion makes the prediction of stratigraphic information more efficient. Semi-supervised framework for impedance inversion is the latest breakthrough method in this field. However, the 1-D semi-supervised methods now in use are unable to extract the spatiotemporal properties of the data solely through the network itself. Moreover, the initial model, a critical input for this method, is typically derived through extrapolation and interpolation of well log data. This can lead to significant errors, especially when the well data are sparse and the subsurface structures are complex. Well data only provide information for a limited section of the reservoir, thereby making it challenging to capture the overall behavior accurately. As a result, the creation of an accurate initial model is often fraught with errors. A more desirable approach is to use seismic attribute-guided methods, such as neural networks, which incorporate both seismic and well log data, leading to a more accurate low-frequency model with lateral variations. In this article, we develop a semi-supervised domain adversarial and spatial fusion (DASF) inversion framework. This method uses a 1-D convolutional neural network (CNN)-based global spatiotemporal analysis module and a 2-D CNN-based local spatiotemporal analysis module to complete the inversion and forward task simultaneously. Multiple spatiotemporal characteristics from two submodules can be successfully fused using an adaptive fusion approach. In this network, the step of extracting the initial model is incorporated into the learning process. Moreover, we adopt adversarial learning in the impedance domain to guide the training process, thereby reducing the network’s dependence on labels. The experiments on the synthetic and field dataset show that the proposed method can efficiently improve the prediction accuracy of the inversion results compared with conventional methods. Meanwhile, the local spatiotemporal analysis module can be used to create a more trustworthy initial model that incorporates the characteristic of seismic and well-logging data.

… Time-frequency analysis is vital in seismic data processing, and the … This paper traces the evolution of time-frequency … representation of seismic signals, thereby exposing their complex …

Seismic amplitude variation with offset (AVO) inversion is well-known as a popular and pragmatic tool used for the prediction of elastic parameters in the geosciences. Low frequencies missing from conventional seismic data are conventionally recovered from other geophysical information, such as well-log data, for estimating the absolute rock properties, which results in biased inversion results in cases of complex heterogeneous geologic targets or plays with sparse well-log data, such as marine or deep stratum. Broadband seismic data bring new opportunities to estimate the low-frequency components of the elastic parameters without well-log data. We have developed a novel AVO inversion approach with the Bayesian inference for broadband seismic data. The low-frequency components of the elastic parameters are initially estimated with the proposed broadband AVO inversion approach with the Bayesian inference in the complex frequency domain because seismic inversion in the complex frequency domain is helpful to recover the long-wavelength structures of the elastic models. Gaussian and Cauchy probability distribution density functions are used for the likelihood function and the prior information of model parameters, respectively. The maximum a posteriori probability solution is resolved to estimate the low-frequency components of the elastic parameters in the complex frequency domain. Furthermore, with those low-frequency components as initial models and constraints, the conventional AVO inversion method with the Bayesian inference in the time domain is further implemented to estimate the final absolute elastic parameters. Synthetic and field data examples demonstrate that the proposed AVO inversion in the complex frequency domain is able to predict the low-frequency components of elastic parameters well, and that those low-frequency components set a good foundation for the final estimation of the absolute elastic parameters.

… Abstract—The paper addresses marine broadband seismic data acquisition … group of broadband seismic solutions comprises various algorithms for special processing of seismic data …

In June 2022, an innovative seismic survey was conducted in Blötberget, central Sweden, to evaluate the effectiveness of employing both a broadband seismic source and broadband receivers for mineral exploration in a challenging hardrock setting. The Blötberget mine hosts high‐quality iron oxides, predominantly magnetite and hematite, sometimes enriched with apatite. These deposits comprise 10–50 m thick sheet‐like horizons with a moderate eastward dip ( ∼$\sim$ 45°) along an NNE‐trending zone. The survey employed a combination of co‐located micro‐electromechanical sensors, three‐component recorders, surface and borehole distributed acoustic sensing, along with a 77‐kN broadband seismic vibrator operating with 2–200 Hz linear sweeps. A tailored processing workflow was applied to preserve the broadband nature of the recorded data, and a one‐dimensional velocity model was derived from the borehole distributed acoustic sensing data for migration and time‐to‐depth conversion purposes. Compared to the previous seismic surveys, the resulting seismic cross section reveals several well‐defined reflections with improved resolution. Notably, a reflection intersecting the main deposits at a depth of approximately 1200 m exhibits a distinct polarity reversal relative to the reflection from the mineralization, providing further evidence for its interpretation as originating from a fault zone. Shallow reflections align with geological boundaries and partially coincide with weak magnetic anomalies. Additional reflections were revealed underneath the known mineralization on both sides of the fault zone and may suggest the presence of potential additional resources. The delineation of these reflections and the fault zone is critical for future mine planning and development in the region. This case study underscores the potential of broadband data in achieving high‐resolution subsurface imaging in hardrock environment and its pivotal role in mineral resource assessment processes.

… broadband seismic networks have been developed, and the retrieved data have been used for seismic … P and S receiver function analysis of seafloor borehole broadband seismic data. J…

The recent development of techniques to extend the bandwidth of marine towed-streamer surveys has significantly changed the marine seismic landscape. In fact, it has coined the new category of “broadband seismic,” now synonymous with the marine towed-streamer market. The bandwidth challenge for marine towed-streamer seismic is well documented and is related to mitigating, or completely removing, the interference pattern from the interaction of the upgoing primary wave and its surface reflection (i.e., its ghost) at the source and receiver side. The interference results in the ghost notches in the amplitude spectrum which bound the useful bandwidth of the data at the high and low ends of the spectrum.

… bandwidth contained on the seismic data. Ghost alters bandwidth to … broadband seismic data, lots of attempts are used, both on the acquisition and on the processing of seismic data. …

Seismic exploration employing wide-azimuth, broadband, high-density data (i.e. double-width single-height; “double-width” means wide-azimuth and broadband; “single-height” means high-density data, generally referring to small-bin-size data (less than 10 × 10 m); and double-width single-height is an abbreviation) enables more complete wave field information to be recorded, reduces aliasing, and produces abundant low-frequency information, which is conducive to broadband processing and the anisotropic study of seismic data. Based on the acquired double-width single-height seismic data, in this paper, we analyse the wave field characteristics, signal-to-noise ratio and frequency of seismic data. We also design a procedure for processing double-width single-height seismic data. The key techniques in the proposed procedure focus on the high-resolution amplitude-preserved characteristics of double-width single-height seismic data obtained with a low-frequency vibroseis sweep. Faults and sand bodies are characterised and described by using the final imaging results, which embody the advantages of double-width single-height seismic exploration.

In marine seismic data acquisition, varying the source depth along a sail line gives diversity in sequential shot gather frequency spectra. Undesired alterations of the frequency spectra are created by the source ghost and by air-gun bubble oscillations. By deliberately varying the source depth along a sail line, it is possible to obtain a seismic data set that will have energy more evenly distributed within the main frequency band of the source output. This is obtained when data acquired with different source depths are stacked in imaging. We formulated a simple inverse problem that seeks to find the optimal distribution of source depths over a sequential series of shots that shape the amplitude spectrum of the final image into a desired shape. We assumed that the data are receiver-side deghosted, that designature could be applied to each shot gather, and that the shot gathers could be redatumed to a common datum prior to imaging.

Seismic inversion is an important approach in parameters estimation in the fields of geosciences. The low-frequency component of the model parameter plays an important role in seismic inversion. The emergence of broadband seismic data acquisition and processing technologies is pushing the attention of the low frequency to a new level. With the review of the research status of the estimation of the low-frequency models in history and the low-frequency component contained in the complex frequency domain, a novel broadband seismic Bayesian inversion approach in the complex frequency domain is proposed to implement the estimation of the low-frequency component of the model parameter. The proposed approach makes full use of the advantage of broadband seismic data and the low-frequency component of the damped wave fields in the complex frequency domain. The kernel function of the proposed inversion approach is built with Bayesian inference. Synthetic examples demonstrate the feasibility and robustness of the proposed inversion approach in the estimation of the low-frequency component of the model parameter. A field data example verifies the feasibility and reasonability of the proposed inversion approach in application. Finally, the estimated low-frequency component of the model parameter is utilized as the initial model for the suggested seismic Bayesian inversion method in time domain. Model and field data examples further verify the effectiveness and superiority of the proposed inversion method in the final estimation of the model parameter by comparing with the conventional inversion approach.

Seismic processing and interpretation techniques provide important tools for oil and gas exploration in the Songliao Basin in eastern China, which is dominated by terrestrial facies. In the Songliao Basin, a large number of thin-sand reservoirs are widely distributed and they are the primary targets of potential oil and gas exploration and exploitation. An important part of exploration in the Songliao Basin is to accurately describe the distribution of these thin-sand belts and the sand-body shapes. However, the thickness of these thin-sand reservoirs is generally below the resolution of conventional seismic processing. Most reservoirs are thin interbeds of sand and mudstones with strong vertical and lateral variations. This makes it difficult to accurately predict the vertical and horizontal distribution of the thin-sand bodies using conventional seismic processing and interpretation methods. In addition, these lithologic traps are difficult to identify due to the complex controlling factors and distribution characteristics and the strong concealment. These challenges motivate us to improve the seismic data quality to help delineate thin-sand reservoirs. We have used the broadband, wide-azimuth, and high-density integrated seismic exploration technique to help delineate thin reservoirs. We first use field single-point excitation and single-point receiver acquisition to obtain seismic data with wide frequency bands, wide azimuth angles, and high folds, which contain rich geologic information. Next, we perform near-surface Q compensation, viscoelastic prestack time migration, seismic attributes, and seismic waveform indication inversion on the newly acquired seismic data. The 3D case study indicates the benefits of improving the imaging of thin-sand body and the accuracy of inversion and reservoir characterization using our method.

… In this paper, we present five seismic data sets from the SBB and CBB, in the area around … Using our multi-frequency data set, we have also investigated qualitatively attenuation (…

… After establishing the correspondence between acoustic impedance data and seismic data, we designated wells “15_9-F-12” and “15_9-F-11 A" as blind wells for validation in the deep …

… the multifrequency data into a single, improved dataset. Examples are then provided, based on data obtained both in the laboratory under controlled conditions and in field data from …

… Acquisition of seismic data using different sources along the … The great potential of multi-frequency seismic data in … which makes use of multi-frequency OBS data and leads to a …

… with single-frequency or multifrequency spectral data. They both work well in crosshole or VSP … , multi-frequency data are necessary in the three inversion schemes for surface seismic …

… High-frequency seismic data (>60 Hz) can delineate the preservation sites of ore bodies or mineralized alteration zones. This model may offer significant potential for guiding deep …

… to unstable convergence and reliance on labeled data. To overcome these limitations, this … structure, eliminating the need for labeled data and avoiding gradient imbalance caused by …

We present a 2D multi-offset, multi-frequency synthetic GPR data set specifically designed to evaluate and test processing, analysis and inversion techniques. The data set replicates realistic subsurface conditions at four sections separated by 2 m. We modeled four multi-offset GPR profiles at 50, 100 and 200 MHz frequencies using realistic wavelets. The data set provides a robust framework for validating advanced GPR algorithms and techniques such as pre-stack depth migration, amplitude versus offset analysis and full waveform inversion. Extensive technical validation ensures data reproducibility and affordability. The standardized, realistic synthetic data set can be used as a reliable benchmark for developing and testing new algorithms and methods, thereby advancing the understanding of subsurface imaging and real-world data interpretation.

It remains controversial whether the interaction between a mantle plume and a craton destabilizes or reinforces the craton. The Tarim basin, with a craton core, a Permian Large Igneous Province, and internal deformation, is an ideal place to investigate this interaction. Here, we construct high‐resolution S‐wave velocity structures down to 15 km in depth using multi‐frequency receiver functions from two temporary seismic arrays that largely cover the Tarim Basin. Our results reveal a strong velocity‐increasing discontinuity across the basin and several large‐scale high‐Vs anomalies. The discontinuity is flat at about 3.5 km depth in the majority of eastern Basin but is uplifted and folded to ∼3 km depth around the Bachu Uplift in the central‐western basin and depressed to more than 6 km depth in the northwestern and southwestern basin. The high‐Vs anomalies, with an average Vs of ∼3.4 km/s, are concentrated under this discontinuity around the Bachu Uplift. Analysis with drilling data, experimental rock‐physics data and previous geophysical observations indicates that the discontinuity corresponds to the top of early Permian strata, and the high‐Vs anomalies are the magmatic intrusions from the early Permian mantle plume. There is strong deformation around the Bachu Uplift formed during Cenozoic Indian‐Eurasian collision, exhibiting a strong spatial correlation with the Permian magmatic intrusions. This suggests that the western Tarim Craton, compared to the east, may be weakened in strength by the Permian mantle plume and exhibits more localized Cenozoic deformation.

… Figure 5 shows the low frequency, middle frequency and high frequency seismic … seismic data. And then, based on the 3-D AVAF seismic data volume, we performed the Multi-frequency …

In solving for frequency-domain seismic wavefields, traditional numerical methods, such as frequency-domain finite difference (FDFD), require multiple computations for each fixed frequency. To overcome this issue, we choose to incorporate frequency distribution into the input of the physics-informed neural network (PINN) to predict multifrequency wavefields with a single training session. However, for VanillaPINN, simply increasing the input dimensions without embedding the sinusoidal feature of seismic waves into the network will lead to convergence difficulty. To address this problem, we develop a seismic wavefield simulation method based on GaborPINN to simulate multifrequency wavefields with improved training convergence. Furthermore, we propose to use the Halton sequence to collect the input coordinates, and the wavefield results demonstrate that the Halton sequence, with its more uniform distribution, achieves higher accuracy than the conventional random sampling. We demonstrate the superiority of the proposed method over the VanillaPINN and the Fourier feature PINN on a section of the Sigsbee2A model.

Abstract Amplitude variation with offset and azimuth (AVOAz) inversion is an effective tool to invert the azimuthal seismic reflection data for fracture weaknesses, which play an important role in seismic characterization of fractures. Conventional AVOAz inversion is implemented in the time domain, which simultaneously uses full-frequency components of the seismic data causing an unstable and inaccurate inversion result in the case of noisy seismic data. Here we propose a novel multiscale frequency-domain seismic inversion approach to improve resolution and accuracy of fracture weakness estimates. Based on new derived frequency-domain expression of azimuthal seismic amplitude difference for a horizontal transversely isotropic (HTI) medium, we build the kernel function incorporated with the Fourier operator, which avoids the Fourier inverse transform in the frequency-domain inversion. Low frequency information in seismic data is often missing or contaminated with noises, thus we establish an inversion objective function combining low frequency regularization term and Cauchy sparse regularization term in the Bayesian framework. We adopt the multiscale strategy and perform inversion successively from low to high frequency group utilizing a limited number of frequency components with a high signal-to-noise (S/N) ratio. In addition, we compare the multiscale frequency-domain inversion approach, the frequency-domain simultaneous inversion approach, and the time-domain inversion approach using synthetic data and field data. Synthetic example demonstrates that the cross-correlation coefficients (CCs) between the true fracture weaknesses and the inversion results obtained by multiscale frequency-domain inversion approach are above 0.83 even with a S/N of 2. Compared to the time-domain inversion approach and the frequency-domain simultaneous inversion approach, the inversion results with multiscale frequency-domain inversion approach have higher resolution and accuracy. Field data example further validates the feasibility and superiority of the proposed multiscale frequency-domain inversion approach.

… narrow-beam echosounding data (Parasound). The data allow spatial correlation between a … An integrated interpretation of the multi-frequency data set provides insight into the regional …

Continuous seismic profiling is a geophysical method widely used in shallow-water geological and geotechnical investigations. Although other geophysical techniques, e.g., electric, electromagnetic, and potential methods, can also assist in investigating these environments, they do not produce adequate data to technically support engineering projects from a quantitative perspective. A guide for selecting the technique or techniques that can provide different projects with optimum data is still lacking. Wrong procedures can still be found in this professional and research field in Brazil and elsewhere as regards the selection of acoustic sources that produce the best results or solutions for different underwater projects. This article aims at contributing to the discussion on the performance of seismic sources in shallow-water surveys. It concludes that the best final product is obtained by using multi-frequency acoustics systems simultaneously. Operating multiple seismic sources at the same time can yield both good resolution and good penetration, thereby meeting all the needs of any given underwater engineering project, e.g., dams, ports, pipelines, bridges, tunnels, offshore wind farms, basic geology and archeology surveys, dredging projects, and investigations of silting processes in rivers and water reservoirs.

Backscatter mosaics based on a multi-frequency multibeam echosounder survey in the continental shelf setting of the North Sea were compared. The uncalibrated backscatter data were recorded with frequencies of 200, 400 and 600 kHz. The results showed that the seafloor appears mostly featureless in acoustic backscatter mosaics derived from 600 kHz data. The same area surveyed with 200 kHz reveals numerous backscatter anomalies with diameters of 10–70 m deviating between −2 dB and +4 dB from the background sediment. Backscatter anomalies were further subdivided based on their frequency-specific texture and were attributed to bioturbation within the sediment and the presence of polychaetes on the seafloor. While low frequencies show the highest overall contrast between different seafloor types, a consideration of all frequencies permits an improved interpretation of subtle seafloor features.

Cold seeps on Opouawe Bank, situated in around 1000 m water depth on the Hikurangi Margin offshore North Island. New Zealand, were investigated using multibeam bathymetry, 75 and 410 kHz sidescan sonar imagery, and 2–8 kHz Chirp sediment echosounder data. Towed video camera observations allowed ground-truthing the various geoacoustic data. At least eleven different seep locations displaying a range of seep activity were identified in the study area. The study area consists of an elongated, northward-widening ridge that is part of the accretionary Hikurangi Margin and is well separated from direct terrigenous input by margin channels surrounding the ridge. The geoacoustic signature of individual cold-seep sites ranged from smooth areas with slightly elevated backscatter intensity resulting from high gas content or the presence of near-surface gas hydrates, to rough areas with widespread patches of carbonates at the seafloor. Five cold seeps also show indications for active gas emissions in the form of acoustic plumes in the water column. Repeated sidescan sonar imagery of the plumes indicates they are highly variable in intensity and direction in the water column, probably reflecting the control of gas emission by tides and currents. Although gas emission appears strongly focused in the Wairarapa area, the actual extents of the cold seep structures are much wider in the subsurface as is shown by sediment echosounder profiles, where large gas fronts were observed.

Knowledge about the stochastic nature of heterogeneity in subsurface hydraulic properties is critical for aquifer characterization and the corresponding prediction of groundwater flow and contaminant transport. Whereas the vertical correlation structure of the heterogeneity is often well constrained by borehole information, the lateral correlation structure is generally unknown because the spacing between boreholes is too large to allow for its meaningful inference. There is, however, evidence to suggest that information on the lateral correlation structure may be extracted from the correlation statistics of the subsurface reflectivity structure imaged by surface‐based ground‐penetrating radar measurements. To date, case studies involving this approach have been limited to 2D profiles acquired at a single antenna centre frequency in areas with limited complementary information. As a result, the practical reliability of this methodology has been difficult to assess. Here, we extend previous work to 3D and consider reflection ground‐penetrating radar data acquired using two antenna centre frequencies at the extensively explored and well‐constrained Boise Hydrogeophysical Research Site. We find that the results obtained using the two ground‐penetrating radar frequencies are consistent with each other, as well as with information from a number of other studies at the Boise Hydrogeophysical Research Site. In addition, contrary to previous 2D work, our results indicate that the surface‐based reflection ground‐penetrating radar data are not only sensitive to the aspect ratio of the underlying heterogeneity, but also, albeit to a lesser extent, to the so‐called Hurst number, which is a key parameter characterizing the local variability of the fine‐scale structure.

Abstract Ground-penetrating radar (GPR) is a non-destructive technique that utilizes high-frequency electromagnetic waves to detect and locate subsurface objects and interfaces. Resolution can be improved with an increase of the source frequency; however, this will lead to decreased investigation depth due to the stronger attenuation of the high frequency signal. Thus, there is a trade-off between the resolution and the investigation depth in the single central frequency-based GPR system. To obtain both high resolution and deep penetrating ability simultaneously, we applied multi-frequency GPR data fusion with three algorithms: time-domain fusion with weights/without, frequency-domain fusion with weights. The fusion effect is qualitatively and quantitatively evaluated and compared by the fused radar profile and Laplacian operator, which is a second derivative gradient operator and commonly used in image edge detection by detecting the zero-crossings of image intensity. The results of the numerical simulation and field data showed that the fused profile was able not only to retain the high resolution in the shallow area from the high-frequency antenna but also take advantage of the significant investigation depth of the low-frequency antenna by merging the multi-frequency data into a single profile. Thus, the fused GPR data has the ability to create a single profile with more information and detailed characteristics.

… of the RGB blend generated from the original seismic data. The … data intersecting the package. The dominant effect on the colour blend is variation in bed thickness and the RGB blend …

Despite routine demand from petroleum explorationists and field developers, interpreting (inverting) seismic data for reservoir thickness from acoustic impedance (AI) or lithology volume requires a high-quality, unbiased well database and the special skills of elite geophysicists. I have developed a new method, based on linear combination and color blending of multiple-frequency panels, to estimate AI and thickness without the strict implementation of complex mathematics and extensive well control. Aimed at readjusting the thin-bed tuning effect in a formation of normal thickness range (up to [Formula: see text]; [Formula: see text] = dominant wavelength), a linear combination of three frequency panels from [Formula: see text] data would lead to a reasonable visual match between a sandstone (shale) body and its seismic event, should the combined amplitude spectrum roughly match the AI spectrum. A red-green-blue blending of frequency panels further extends the interpretive benefits by illustrating the thickness in color, adding a sense of thickness cyclicity on the vertical view and that of sandstone thickness map on stratal-slice view. Tests using a simple wedge model and a complex, geologically realistic multi-thin-bed model demonstrate that the proposed workflow may achieve decent geometry (thickness) estimation and reasonably high correlation ([Formula: see text]) for AI prediction with minimal or no well control. The results are similar to colored inversion in the fast-track principle, with improved stability and less error (at least in this study). More complex procedures — such as linear regression and model-based inversion — may lead to minor to moderate improvement with adequate well control. An application to a field data set confirmed the value of the methods in high-resolution reservoir-thickness imaging, with a strong potential for stratigraphically oriented studies, such as seismic chronostratigraphy, sequence stratigraphy, and seismic sedimentology.

ABSTRACT Acquisition of incomplete data, i.e., blended, sparsely sampled, and narrowband data, allows for cost-effective and efficient field seismic operations. This strategy becomes technically acceptable, provided that a satisfactory recovery of the complete data, i.e., deblended, well-sampled, and broadband data, is attainable. Hence, we explore a machine-learning approach that simultaneously performs suppression of blending noise, reconstruction of missing traces, and extrapolation of low frequencies. We have applied a deep convolutional neural network in the framework of supervised learning in which we train a network using pairs of incomplete-complete data sets. Incomplete data, which are never used for training and use different subsurface properties and acquisition scenarios, are subsequently fed into the trained network to predict complete data. We develop matrix representations indicating the contributions of different acquisition strategies to reducing the field operational effort. We also determine that the simultaneous implementation of source blending, sparse geometry, and band limitation leads to a significant data compression where the size of the incomplete data in the frequency-space domain is much smaller than the size of the complete data. This reduction is indicative of survey cost and duration that our acquisition strategy can save. Synthetic and field data examples demonstrate the applicability of the proposed approach. Despite the reduced amount of information available in the incomplete data, the results obtained from the numerical and field data cases clearly show that the machine-learning scheme effectively performs deblending, trace reconstruction, and low-frequency extrapolation in a simultaneous fashion. It is noteworthy that no discernible difference in prediction errors between extrapolated frequencies and preexisting frequencies is observed. The approach potentially allows seismic data to be acquired in a significantly compressed manner while subsequently recovering data of satisfactory quality.

… This study, hence, explores a ML based scheme that aims at optimal data recovery from seismic data acquired in a blended, sparsely-sampled and narrowband manner. Synthetic and …

Blended source arrays are historically configured with equal source units, such as broadband vibrators (land) and broadband air-gun arrays (marine). I refer to this concept as homogeneous blending. I have proposed to extend the blending concept to inhomogeneous blending, meaning that a blended source array consists of different source units. More specifically, I proposed to replace in blended acquisition the traditional broadband sources by narrowband versions — imagine coded single air guns with different volumes or coded single narrowband vibrators with different central frequencies — together representing a dispersed source array (DSA). Similar to what we see in today’s audio systems, the DSA concept allows the design of dedicated narrowband source elements that do not suffer from the low versus high frequency compromise. In addition, the DSA concept opens the possibility to use source depths and spatial sampling intervals that are optimum for the low-, mid-, and high-frequency sources (multiscale shooting grids). DSAs are considered to be an important step in robotizing the seismic acquisition process.

… explain how we synthesize a blended distance separated simultaneous sweeping data set. We … , αs ,ω the angular frequency and the set of frequencies we invert. We define one source …

Seismic acquisition is a trade-off between economy and quality. In conventional acquisition the time intervals between successive records are large enough to avoid interference in time. To obtain an efficient survey, the spatial source sampling is therefore often (too) large. However, in blending, or simultaneous acquisition, temporal overlap between shot records is allowed. This additional degree of freedom in survey design significantly improves the quality or the economics or both. Deblending is the procedure of recovering the data as if they were acquired in the conventional, unblended way. A simple least-squares procedure, however, does not remove the interference due to other sources, or blending noise. Fortunately, the character of this noise is different in different domains, e.g., it is coherent in the common source domain, but incoherent in the common receiver domain. This property is used to obtain a considerable improvement. We propose to estimate the blending noise and subtract it from the blended data. The estimate does not need to be perfect because our procedure is iterative. Starting with the least-squares deblended data, the estimate of the blending noise is obtained via the following steps: sort the data to a domain where the blending noise is incoherent; apply a noise suppression filter; apply a threshold to remove the remaining noise, ending up with (part of) the signal; compute an estimate of the blending noise from this signal. At each iteration, the threshold can be lowered and more of the signal is recovered. Promising results were obtained with a simple implementation of this method for both impulsive and vibratory sources. Undoubtedly, in the future algorithms will be developed for the direct processing of blended data. However, currently a high-quality deblending procedure is an important step allowing the application of contemporary processing flows.

… in the frequency domain and discuss the matrices for one frequency slice of a data volume. … 2 we see the effect of the blending operator on a seismic experiment: the blended experiment …

In blended seismic acquisition, or simultaneous source seismic acquisition, source encoding is essential at the acquisition stage to allow for separation of the blended sources at the processing stage. In land seismic surveys, the vibroseis sources may be encoded with near-orthogonal sweeps for blending. In marine seismic surveys, the sweep type of source encoding is difficult because the main source type in marine seismic exploration is the air-gun array, which has an impulsive character. Another issue in marine streamer seismic data acquisition is that the spatial source sampling is generally coarse. This hinders the deblending performance of algorithms based on the random time delay blending code that inherently requires a dense source sampling because they exploit the signal coherency in the common-receiver domain. We have developed an alternative source code called shot repetition that exploits the impulsive character of the marine seismic source in blending. This source code consists of repeated spikes of ones and can be realized physically by activating a broadband impulsive source more than once at (nearly) the same location. Optimization of the shot-repetition type of blending code was done to improve the deblending performance. As a result of using shot repetition, the deblending process can be carried out in individual shot gathers. Therefore, our method has no need for a regular dense source sampling: It can cope with irregular sparse source sampling; it can help with real-time data quality control. In addition, the use of shot repetition is beneficial for reducing the background noise in the deblended data. We determine the feasibility of our method on numerical examples.

We used a novel iterative estimation scheme for separation of blended seismic data from simultaneous sources. The scheme is based on an augmented estimation problem that can be solved by iteratively constraining the deblended data using shaping regularization in the seislet domain. We formulated the forward modeling operator in the common-receiver domain, in which two sources were assumed to be blended using a random time-shift dithering approach. The nonlinear shaping-regularization framework offered some freedom in designing a shaping operator to constrain the model in an underdetermined inverse problem. We designed the backward operator and the shaping operator for the shaping-regularization framework. The backward operator can be optimally chosen as half of the identity operator in the two-source case, and the shaping operator can be chosen as coherency-promoting operator. The high performance deblending effect of the iterative framework was tested on three numerically blended synthetic data sets and one numerically blended field data set. Compared with alternative f-k domain thresholding and f-x predictive filtering, seislet-domain soft thresholding exhibits the most robust behavior.

Blended acquisition, which allows multiple sources almost simultaneously fired, has become an effective way for accelerating seismic data acquisition. In order to use conventional processing methods for imaging, deblending is necessary for this special acquisition. Convolutional neural network‐based deblending methods provide a novel end‐to‐end framework for source separation. We proposed a field‐data‐based augmentation method that uses shuffled deblending noise as the features to be learned and take the inaccurate labels as the output of the network. Synthetic data experiments show that a network trained on data set with the proposed data augmentation method has higher accuracy for deblending even if the labelled data are noisy. Besides, 2D discrete wavelet transform, which has the advantage of multiscale decomposition and dimensionality reduction, is introduced to accelerate the computation of the network. The data augmentation method for data set generation and the computational speedup method for network training/predicting are also applied to field data. The results from synthetic and field data all confirm the performance of our methods.

Summary Deblending algorithms based on current blended acquisition design often require a dense source sampling to ensure a high-quality result. However, in practice the spatial source sampling is usually too coarse. In this abstract, we discuss an alternative approach to blending in marine seismic, called shot repetition, which can overcome this requirement. Shot repetition refers to activating a broadband source more than once at the same location. We extend the general forward model of source blending to include the case of shot repetition. By exploiting the repetitive shots acquired at the same location, deblending can be implemented in the common-shot domain, and therefore our method has no restrictions on source sampling. We applied the method to numerically blended field data and obtained satisfactory results.

… Moore [18] used the Radon transform to separate the blended data in the frequency domain and … where Pblend denotes the blended data, P is the seismic data generated by separate …

… data. Pi bl is the separated data from blended seismic data at ith shooting Si sin. Based on Гi k of equation (2), the blending data separation problem can be expressed by the following …

We present a simultaneous multifrequency inversion approach for seismic data interpretation. This algorithm inverts all frequency data components simultaneously. A data-weighting scheme balances the contributions from different frequency data components so the inversion process does not become dominated by high-frequency data components, which produce a velocity image with many artifacts. A Gauss-Newton minimization approach achieves a high convergence rate and an accurate reconstructed velocity image. By introducing a modified adjoint formulation, we can calculate the Jacobian matrix efficiently, allowing the material properties in the perfectly matched layers (PMLs) to be updated automatically during the inversion process. This feature ensures the correct behavior of the inversion and implies that the algorithm is appropriate for realistic applications where a priori information of the background medium is unavailable. Two different regularization schemes, an L2-norm and a weighted L2-norm function, are used in this algorithm for smooth profiles and profiles with sharp boundaries, respectively. The regularization parameter is determined automatically and adaptively by the so-called multiplicative regularization technique. To test the algorithm, we implement the inversion to reconstruct the Marmousi velocity model using synthetic data generated by the finite-difference time-domain code. These numerical simulation results indicate that this inversion algorithm is robust in terms of starting model and noise suppression. Under some circumstances, it is more robust than a traditional sequential inversion approach.

Pre‐stack seismic inversion, which estimates subsurface elastic parameters (Vp, Vs and ρ) from amplitude‐variation‐with‐offset (AVO) data, remains challenged by ill‐posedness and limited well control. Although deep learning (DL) offers data‐driven non‐linear mapping, its generalization is often constrained by scarce labels and physical inconsistency. This study introduces a semi‐supervised DL framework that integrates a time–frequency joint convolutional neural network (TF‐CNN) with physics‐guided constraints. The framework uniquely combines (1) a time–frequency hybrid attention mechanism that adaptively weights seismic features across domains; (2) exact Zoeppritz equation forward modelling to generate physically consistent pseudo‐labels for semi‐supervised learning and (3) a low‐frequency structural prior to stabilize long‐wavelength components. Validated on synthetic (Marmousi2) and field datasets, the method demonstrates enhanced accuracy, improved noise robustness and tighter uncertainty quantification compared to conventional AVO inversion and standard convolutional neural network (CNN) baselines. Quantitative metrics confirm significant gains in parameter accuracy and uncertainty reliability. The proposed approach effectively mitigates ill‐posedness through multi‐constraint integration, offering a practical and interpretable tool for reservoir characterization in label‐scarce settings.

… This work provides a new paradigm for AI-driven pre-stack seismic inversion by unifying multi-scale feature representation learning. We will open-source the code on https://github.com/…

… frequency model based on wells are processed by frequencydivision fusion,and a multiscale fusion model … data;(b)main frequency and frequency band analysis of seismic data. …

This paper discusses the joint inversion of seismic and electromagnetic data in geophysical exploration. It recognizes seismic exploration as a primary method, especially through full waveform inversion (FWI) that enables high-resolution subsurface velocity estimation. However, it acknowledges the severe non-uniqueness problem associated with inverting subsurface velocities using solely seismic data. The abstract introduces electromagnetic exploration as an effective supplement due to its expansive detection range, cost-effectiveness, sensitivity to subsurface anomalies, and provision of deep resistivity information. By jointly inverting electromagnetic and seismic data, complementary information from both datasets is integrated, ameliorating the non-uniqueness issue and enhancing the accuracy and reliability of the inversion results. However, existing joint inversion techniques encounter hurdles such as complex objective function design, difficulty in achieving convergence, and inadequate coupling between seismic and electromagnetic data. To surmount these challenges, we propose a Multi-scale Feature Fusion Network (MFF Net) grounded in joint learning. This approach harnesses the robust nonlinear fitting capacities of neural networks to optimize the objective function efficiently. Furthermore, coupling between seismic and electromagnetic data is established across multiple sampling scales. By leveraging the frequency band complementarity of pre-stack seismic and electromagnetic data and considering the low-frequency constraints from electromagnetic data alongside mid-to-high frequency constraints from seismic data, high-resolution resistivity and velocity models are inverted. This ultimately boosts the precision and reliability of the inversion results. We validate the efficacy of this approach using synthetic data, demonstrating the effective reconstruction of the underground salt dome structure, even in the absence of low-frequency data.

With the continuous development of complex marine hydrocarbon reservoirs, broadband seismic data have shown growing advantages in revealing abundant stratigraphic information. Affected by acquisition conditions and stratigraphic attenuation, the acquired seismic data commonly suffer from narrow bandwidth, and conventional broadband processing techniques are incapable of optimizing the overall frequency band. This study proposes a coordinated high- and low-frequency broadband compensation method based on adaptive weight fusion to effectively extend the frequency bandwidth of seismic data. Firstly, wavefield separation is used to suppress ghost reflections, compensate low-frequency effective signals, and restore the continuity of the low-frequency spectrum. Then, based on the spectrum extrapolation method of maximum entropy spectrum estimation, a spectrum prediction model is established to achieve the continuation of high-frequency effective signals. Finally, in combination with the signal-to-noise ratio of each frequency band, the adaptive weight fusion algorithm is applied for weighted summation. The acquired broadband seismic data feature a continuous spectrum and balanced energy, greatly improving seismic imaging quality. Comparative results obtained using conventional processing methods verify that the proposed method can significantly improve stratigraphic continuity and wave group characteristics.

Geophysical exploration is developing from qualitative seismic imaging to quantitative imaging, and broadband acoustic impedance is the core. Directly estimating broadband impedance using full-waveform inversion is a strong nonlinear problem. It is difficult to obtain a reliable result in practice. We proposed an alternative way: estimate background velocity, density, and broadband reflectivity first and then fuse them to be the broadband impedance by information fusion. This article studies the method of fusing band-limited reflectivity and background impedance. Due to the observation with band-limited seismic wavelet, only band-limited reflectivity can be obtained even after a lot of processing. The band-limited reflectivity can lead to oscillation error in impedance. Different from the conventional poststack impedance inversion, this article introduces an iterative process without the need of wavelet extraction. Start from the broadband reflectivity that has been subjected to fidelity imaging, least-squares migration, and magnitude calibration. In order to reduce the oscillation error, reflectivity is sparsely promoted such that the reflection coefficients from large to small are gradually fused with background impedance. Reflectivity and impedance are mutually constrained and iteratively updated, and lateral continuity is incorporated. Numerical experiment and 3-D field data application demonstrate the effectiveness of the method. Impedance shows higher interpretability than reflectivity.

… High-precision seismic wave imaging, providing broadband reflection coefficient imaging … of the key technical challenges for high-resolution seismic imaging.Based on this, it is crucial to …

… it into an information fusion problem that reconstructs the broadband wave impedance model by … of quantitative broadband reflection coefficient first depends on fidelity seismic imaging, …

… of high-quality broadband seismic data with improvements in … On the other hand, traditional seismic imaging technologies … fusion is an effective way to achieve high-resolution imaging. …

… seismic data and high-resolution data (eg the result of convolution between reflectivity and one broadband … operators that transform narrowband seismic data into broadband high-…

… Next, the broadband synthetic record and the well-side seismic … seismic data, as well as multi-attribute fusion, and ultimately realizing high-precision fault recognition based on the fusion …

… reflection seismic imaging strategy based on deployment of spatially sparse surface seismic … technologies including Fusion's ThinMAN(trademark) broadband spectral inversion. Seal …

—The paper focuses on the wide-azimuth, broadband, and high-density (WBH) seismic data application methodology, which was used to complete a more detailed structural interpretation of the K oilfield, to identify a number of low-relief structures in its periclinal parts, and to detect the potential residual oil zones (ROZ) in the oilfield. The obtained wide-azimuth, high-density field data and the results of broadband inversion are the main factors that increase the degree of ROZ prediction. A comprehensive analysis has shown that potential ROZ are arched faulted-nose structures in the periclinal parts of the oilfield, low-relief anticlines in the periclinal parts of the oilfield, and lithologic pinchout zones. Technical support has been provided for identifying ROZ in the given oilfield, and the basis has been laid for predicting the residual oil distribution in analogous oilfields with high productivity and high water cut that are at the middle and late stages of development.

We develop an inversion procedure for deriving multi‐scale velocity models with waveform inversions of earthquake and ambient noise data at multi‐frequency bands recorded by regional and dense sensor configurations. The method is applied for the area around the 2019 Ridgecrest earthquake rupture zones, utilizing data recorded by regional stations and dense 2D and 1D arrays with station spacings of ∼5 km and ∼100 m, respectively. Starting with regional Vp, Vs models and locations of Ridgecrest aftershocks, the velocity models and event locations are improved iteratively by inversions of waveforms recorded by regional stations and the 2D array, using a minimum spectral element size of ∼600 m. Waveforms from local events recorded by dense 1D arrays across the M7.1 rupture zone with frequencies of up to 10 Hz are used to resolve small‐scale features of the rupture zone and shallow crust with a local spectral element size of 80 m. The refined models provide self‐consistent descriptions of the rupture zone and the shallow crust embedded in the regional structures. The results reveal pronounced low Vs and high Vp/Vs in the M6.4 and M7.1 rupture zones coinciding with concentrations of seismicity, and also around the Garlock fault and in several local basins. We also observe clear velocity contrasts across the Garlock fault with polarity reversals along strike and with depth. The obtained multi‐scale velocity models can be used to improve derivations of earthquake source properties, simulations of dynamic ruptures and ground motions, and the understanding of fault and tectonic processes in the region.

This paper presents an f–x domain Transformer network for multi‑frequency ground‐penetrating radar data fusion, which processes single‑sided amplitude spectra to preserve complementary spectral features while avoiding distortions common in time‑domain approaches. The method introduces full‑range two‐dimensional sinc interpolation for calibration, an energy‑driven frequency‑window selection strategy and an encoder‑only Transformer adapted for spectral inputs. Experiments on synthetic and field data show that the proposed fusion yields higher spectral entropy and reduces training time by 70%–90% on CPU and ∼40% on GPU, offering a physically consistent and computationally efficient solution for enhanced subsurface interpretation in engineering and geophysical surveys.

Abstract High frequency GPR signals offer high resolution while low frequency GPR signals offer greater depth of penetration. Effective fusion of multiple frequencies can combine the advantages of both. In addition, GPR attribute analysis can improve subsurface imaging, but a single attribute can only partly highlight details of different physical and geometrical properties of subsurface potential targets. In order to overcome these challenges, we implement an advanced multi-frequency and multi-attribute GPR data fusion approach based on 2-D wavelet transform utilizing a dynamic fusion weight scheme derived from edge detection algorithm, which is tested on data from a small glacier in the north-eastern Alps by 250 & 500 MHz central frequency antennas. Besides, information entropy and spatial frequency are developed as quantitative evaluation parameters to analyze the fusion outcomes. The results demonstrate that the proposed approach can enhance the efficiency and scope of GPR data interpretation in an automatic and objective way.

… The multi-channel frequency-domain feature fusion mechanism enriches the feature representation of the data, providing an effective new approach for GPR data inversion. …

GPR systems with a single central frequency suffer limitations due to the unavoidable trade-off between resolution and penetration depth that multi frequency equipments can overcome. We propose a new semi-supervised multi-frequency merging algorithm based on Deep Learning and specifically on Bi-Directional Long-Short Term Memory to automatically merge varying numbers of data sets at different frequencies. The proposed methodology is tested on synthetic and field data, to evaluate performances and robustness. The proposed merging algorithm can manage the complementarity of information at different central frequencies, properly merging different types of data. Results show not only a smooth transition in time, but, even more important, a remarkable broadening of the bandwidth thus increasing the overall resolution. Our approach is not limited to specific frequency components or geological setting but can be potentially exploited to merge any type of dataset with different spectral components.

… GPR data and subsequent fusion of signals in the frequency domain. … frequencies in the frequency domain. To obtain the preferred fusion algorithm for multi‐frequency data‐level fusion, …