Due to environmental and economic constraints on their acquisition, seismic data are always irregularly sampled and include bad or missing traces, which can cause problems for seismic data processing. Recently, many researchers have attempted to improve seismic data reconstruction using machine learning (ML) techniques, such as convolutional neural networks, which are inspired by computer vision and imaging processing. In this letter, we propose a novel approach for reconstructing missing traces in seismic data using ML techniques, especially recurrent neural network (RNN) algorithms. Instead of processing seismic data as an image, the proposed approach performs seismic trace interpolation using traces that are sequences of time-series data. More specifically, we adopt deep bidirectional long short-term memory (LSTM) for seismic trace interpolation and test models with and without skip connections. Field seismic data are used to demonstrate the effectiveness of the proposed approaches, and the deep bidirectional LSTM (DBiLSTM) with skip connections shows the best performance compared to cubic interpolation, minimum weighted norm interpolation (MWNI), and DBiLSTM without skip connection.

Due to the rapid development and spread of deep learning technologies, potential applications of artificial intelligence technology in the field of geophysical inversion are being explored. In this study, we applied a deep neural network (DNN) to reconstruct one-dimensional electrical resistivity structures from airborne electromagnetic (AEM) data for varying sensor heights. We used numerical models and their simulated AEM responses to train the DNN to be an inversion operator, and determined that it was possible to train the DNN without the use of stabilisers on the subsurface structures. We compared the quantitative performance of DNN and Gauss–Newton inversion of synthetic datasets, and demonstrated that DNN inversion reconstructed the subsurface structure more accurately, and within a significantly shorter period. We subsequently applied DNN inversion to a field dataset to quantify the effectiveness and applicability of the proposed method for real data. The results of the current study will open new avenues for real-time imaging of subsurface structures from AEM data.

Various geophysical data types have advantages for exploring the subsurface, and more reliable exploration can be realized through integration of such data. However, the imaging of physical properties based on deep learning (DL) techniques, which has received considerable attention because of its enormous potential, has generally been performed using only a single type of data. We have developed a cooperative inversion method based on supervised DL for salt delineation. Controlled-source electromagnetic (CSEM) data, which can effectively distinguish a salt body with high electrical resistivity from the surrounding medium, are used as data for cooperative inversion, with high-resolution information derived from seismic data used as the constraint. This approach can entrain seismic information into a fully convolutional network designed to invert CSEM data to reconstruct the resistivity distribution. The inversion network is trained using large synthetic data sets, including the seismic information derived from seismic data as well as CSEM data and resistivity models. A cooperative strategy for reasonable entrainment of seismic information into the inversion network is established based on analysis of the network and kernels of the convolutional layers. The performance of the proposed method is demonstrated through experiments on test data generated for resistivity models for complex salt structures. The trained cooperative inversion network shows improved salt delineation results compared to the independent inversion network, irrespective of noise levels added to the test data, due to restriction of the resistivity model to fit seismic information. Moreover, training with noise-added data decreased the effects of noise on prediction results, similar to the case of adversarial training. We develop the possibility of combining geophysical data with a constraint using DL-based techniques.

Deep-learning (DL) methods have recently been introduced for seismic signal processing. Using DL methods, many researchers have adopted these novel techniques in an attempt to construct a DL model for seismic data reconstruction. The performance of DL-based methods depends heavily on what is learned from the training data. We focus on constructing the DL model that well reflect the features of target data sets. The main goal is to integrate DL with an intuitive data analysis approach that compares similar patterns prior to the DL training stage. We have developed a two-sequential method consisting of two stages: (1) analyzing training and target data sets simultaneously for determining the target-informed training set and (2) training the DL model with this training data set to effectively interpolate the seismic data. Here, we introduce the convolutional autoencoder t-distributed stochastic neighbor embedding (CAE t-SNE) analysis that can provide the insight into the results of interpolation through the analysis of training and target data sets prior to DL model training. Our method was tested with synthetic and field data. Dense seismic gathers (e.g., common-shot gathers) were used as a labeled training data set, and relatively sparse seismic gathers (e.g., common-receiver gathers [CRGs]) were reconstructed in both cases. The reconstructed results and signal-to-noise ratios demonstrated that the training data can be efficiently selected using CAE t-SNE analysis, and the spatial aliasing of CRGs was successfully alleviated by the trained DL model with this training data, which contain target features. These results imply that the data analysis for selecting target-informed training set is very important for successful DL interpolation. In addition, our analysis method can also be applied to investigate the similarities between training and target data sets for other DL-based seismic data reconstruction tasks.

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 the underlying basis can be hardly interpreted. Subjectivity is therefore often introduced during the quality control process, and any interpretation of DL models can become an important source of information. To provide a such degree of interpretation and retain a higher level of human intervention, the development and application of explainable DL methods have been explored. To showcase the usefulness of such methods in the field of geoscience, we utilize a prototype-based neural network (NN) for the seismic facies classification problem. The "prototype” vectors, jointly learned to have the stereotypical qualities of a certain label, form a set of representative samples. The interpretable component thereby transforms "black boxes” into "grey boxes”. We demonstrate how prototypes can be used to explain NN methods by directly inspecting key functional components. We describe substantial explanations in three ways of examining 1) prototypes’ corresponding input-output pairs, 2) the values generated at the specific explainable layer, and 3) the numerical structure of specific shallow layers located between the interpretable latent prototype layer and an output layer. Most importantly, the series of interpretations shows how geophysical knowledge can be used to understand the actual function of the seismic facies classifier and therefore help the DL’s quality control process. The method is applicable to many geoscientific classification problems when in-depth interpretations of NN classifiers are required.

Diffraction images can be used for modeling reservoir heterogeneities at or below the seismic wavelength scale. However, the extraction of diffractions is challenging because their amplitude is weaker than that of overlapping reflections. Recently, deep-learning (DL) approaches have been used as a powerful tool for diffraction extraction. Most DL approaches use a classification algorithm that classifies pixels in the seismic data as diffraction, reflection, noise, or diffraction with reflection and takes whole values for the classified diffraction pixels. Thus, these DL methods cannot extract diffraction energy from pixels for which diffractions are masked by reflections. We have developed a DL-based diffraction extraction method that preserves the amplitude and phase characteristics of diffractions. Through the systematic generation of a training data set using synthetic modeling based on t-distributed stochastic neighbor embedding analysis, this technique extracts not only faint diffractions but also diffraction tails overlapped by strong reflection events. We also determine that the DL model pretrained with a basic synthetic data set can be applied to seismic field data through transfer learning. Because the diffractions extracted by our method preserve the amplitude and phase, they can be used for velocity model building and high-resolution diffraction imaging.