Due to environmental and economic constraints on their acquisition, seismic data are always irregularly sampled include bad or missing traces, which can cause problems for processing. Recently, many researchers have attempted improve reconstruction using machine learning (ML) techniques, such as convolutional neural networks, inspired by computer vision imaging In this letter, we propose a novel approach reconstructing traces in ML especially recurrent network (RNN) algorithms. Instead of...

Due to the rapid development and spread of deep learning technologies, potential applications artificial intelligence technology in field geophysical inversion are being explored. In this study, we applied a neural network (DNN) reconstruct one-dimensional electrical resistivity structures from airborne electromagnetic (AEM) data for varying sensor heights. We used numerical models their simulated AEM responses train DNN be an operator, determined that it was possible without use stabilisers...

Deep-learning (DL) techniques have been proposed to solve geophysical seismic facies classification problems without introducing the subjectivity of human interpreters’ decisions. However, such DL algorithms are “black boxes” by nature, and underlying basis can be hardly interpreted. Subjectivity is therefore often introduced during quality control process, any interpretation models become an important source information. To provide a degree retain higher level intervention, development...

With recent advances in machine learning, convolutional neural networks (CNNs) have been successfully applied many fields, and several attempts made the field of geophysics. In this letter, we investigated mapping subsurface electrical resistivity distributions from electromagnetic (EM) data with CNNs. To begin imaging using CNNs, carried out precise delineation a salt structure, which is indispensable for identification offshore hydrocarbon reservoirs, towed streamer EM data. For training...

Various geophysical data types have advantages for exploring the subsurface, and more reliable exploration can be realized through integration of such data. However, imaging physical properties based on deep learning (DL) techniques, which has received considerable attention because its enormous potential, generally been performed using only a single type We developed cooperative inversion method supervised DL salt delineation. Controlled-source electromagnetic (CSEM) data, effectively...

Deep-learning (DL) methods have recently been introduced for seismic signal processing. Using DL methods, many researchers adopted these novel techniques in an attempt to construct a model data reconstruction. The performance of DL-based depends heavily on what is learned from the training data. We focus constructing that well reflect features target sets. main goal integrate with intuitive analysis approach compares similar patterns prior stage. developed two-sequential method consisting...

Deep learning (DL) methods are recently used as a powerful tool in seismic signal processing. Most of trace reconstruction governed by the superresolution based on convolutional neural network (CNN). The performances these kinds depend not only how training model is constructed but also what learned from data, especially field data application. In this study, we propose two sequences interpolation through t-SNE and U-Net to provide guide optimal organization sets successful missing traces....

First-break picking is an important step during processing of both passive and active seismic data. Many automated algorithms have been developed to detect first-break points in large volumes However, it remains difficult determine precise seismograms with low signal-to-noise ratios. Therefore, we present a new approach based on the differences between multi-window energy ratios (DERs) that minimizes effects noise. First, DER defined thresholding method detecting using DERs proposed....

Reservoir production monitoring using marine controlled-source electromagnetic (mCSEM) has been studied recently because it is sensitive to resistivity changes resulting from variations in hydrocarbon saturation. However, mCSEM for [Formula: see text] sequestration scarcely investigated, although the method advantageous injection and migration. To investigate feasibility of sequestration, we conducted numerical experiments representative models at a deep brine aquifer shallow sea. By...

Diffraction images can be used for modeling reservoir heterogeneities at or below the seismic wavelength scale. However, extraction of diffractions is challenging because their amplitude weaker than that overlapping reflections. Recently, deep-learning (DL) approaches have been as a powerful tool diffraction extraction. Most DL use classification algorithm classifies pixels in data diffraction, reflection, noise, with reflection and takes whole values classified pixels. Thus, these methods...

We present a series of processes for understanding and analysing controlled-source electromagnetic (CSEM) responses conductive permeable earth. To realize the CSEM response, new 3-D forward modelling algorithm based on an edge finite element method both electrically magnetically heterogeneities is developed. The shows highly accurate results in validation tests against semi-analytic solution stratified earth integral form scattered field. describe vector behaviour anomalous magnetic field...

AbstractTo simulate wave propagation in a tilted transversely isotropic (TTI) medium with tilting symmetry-axis of anisotropy, we develop 2D elastic forward modelling algorithm. In this algorithm, use the staggered-grid finite-difference method which has fourth-order accuracy space and second-order time. Since velocity-stress formulations are defined for staggered grids, include auxiliary grid points z-direction to meet free surface boundary conditions shear stress. Through comparisons...

Full-waveform inversion (FWI) provides a high-resolution velocity model, but carries high computational cost. Additionally, modern seismic acquisition, with dense sources and receivers, generates massive data, resulting in an even greater To reduce the burden of FWI, we have developed FWI algorithm using plane-wave data. Using this approach, gathers transformed from shot are used as input data inversion. Because number is generally far smaller than that common for same set, can significantly...

Quantitative facies classification is the key to linking seismic data lithology evaluate important reservoir properties. During past several years, size of volumes has piled up extent that it challenging for experts examine every volume classify facies. This motivated machine learning approach predicting in an efficient way. However, labeled (well data) limited by various constraints and very expensive obtain, whereas, there a plethora unlabeled (seismic data). Geophysicists are tasked...

Conventional interpretation of airborne electromagnetic data has been conducted by solving the inverse problem. However, with recent advances in machine learning (ML) techniques, a 1D deep neural network inversion that predicts resistivity model using multifrequency vertical magnetic fields and sensor height information at one location applied. Nevertheless, since final this approach relies on connecting models, ML low accuracy for estimation an isolated anomaly, as conventional inversion....

Facies classification refers to the of rock types and pore fluids using information obtained from well log data core samples. A range elastic properties provide main input for models. The are closely related water saturation, porosity, shale volume. In addition, if impedance inversion is performed, same can be surface seismic area, thus linking data. Machine learning (ML)-based facies has advantage minimizing subjectivity associated with human interpretations maximizing time efficiency....

An acoustic logging tool in inclined or horizontal boreholes may be placed apart from the center and produce additional complicated wavefields. We investigate effects of an off-centered on monopole, dipole, quadrupole logs due to off-centering tool. In recent tools, can obtained by adding subtracting responses at four monopole (pressure) receiver arrays right angles. examine array system for three directions eccentricity: inline direction dipole source (Off [Formula: see text] case),...

The petrophysical facies classification in the field of hydrocarbon exploration is one important tasks for reservoir characterization. To predict seismic area, deep learning has recently been applied. However, when applying machine (ML) to classification, there a problem that data available training are very limited. When using acquired under such limited conditions, as well log data, can be severe imbalance number samples because amount area interest relatively less than nonhydrocarbon...