LYLiuqing Yang

Liuqing Yang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Liuqing Yang has a total of 72 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • QBTransNet: Deep Learning Earthquake Distinction from Quarry Blasts Using Dilated Convolutional Transformer

    by Liuqing Yang, Yangkang Chen, Daniel Siervo, Katerine Vallejo, Alexandros Savvaidis

    The accurate distinction between earthquakes and non-earthquake events such as quarry blasts is crucial for the subsequent data analysis. However, signal characteristics of quarry blasts are similar to earthquake events, leading to unreliable and potentially erroneous manual identification, especially in the absence of source location information. In this article, we propose a reliable deep-learning-based framework to distinguish between earthquakes and quarry blasts. In the data preprocessing stage, we apply the continuous wavelet transform algorithm to the 60 s three-channel waveforms for time–frequency conversion. The proposed discrimination framework comprises a dilated convolutional transformer (DCT) and a capsule neural network (CapsNet). DCT combines the local perception capability of traditional convolutional neural networks, effectively extracting spatial features from multichannel scalograms. In addition, the multihead self-attention module in the transformer dynamically adjusts feature weights across different positions to adaptively focus on significant features, which is crucial for handling complex background noise and irrelevant information in earthquake and quarry blast signals. Then, the features extracted by DCT are transferred to the CapsNet for hierarchical feature representation. The dynamic routing mechanism in the CapsNet allows for flexible and adaptive feature propagation and integration between capsules, enabling precise distinction between earthquakes and quarry blasts. We use an artificial intelligence earthquake dataset recorded by the Texas Seismological Network to demonstrate the classification performance of the proposed network. Compared to state-of-the-art classification networks, the proposed method has higher reliability and satisfactory results, with the testing accuracy, precision, recall, and F1-score reaching 99.22%, 99.34%, 99.01%, and 99.18%, respectively. We also demonstrate the robustness of the proposed network through a real-time monitoring test.


  • Pyntfa: A Python Package for Nonstationary Time–Frequency Analysis

    by Yangkang Chen, Chao Li, Yunfeng Chen, Liuqing Yang, Alexandros Savvaidis

    Time–frequency analysis is a widely used signal-processing approach to analyze nonstationary seismological data. In time–frequency analysis, the 1D time-domain seismic data are transformed into a 2D space, where different seismic phases and noise are better distinguished according to the time–frequency difference of these components. Here, we present a Python package, Pyntfa, for performing nonstationary time–frequency analysis on seismological datasets. To characterize data nonstationarity, we apply nonstationary regression, where the Fourier basis functions are constructed to fit the input 1D seismic data. To characterize model nonstationarity, that is, spatially heterogenous regression coefficients, we apply nonstationary smoothing with a priori information. In the most standard and efficient case, a local smoothing with a constant radius will suffice to obtain a high-resolution time–frequency representation of the seismic data. An appropriate nonstationary model constraint will further enhance the resolution of the time–frequency map. The Pyntfa package is backed by C programs to run the core computations, which ensures the package’s production-level efficiency. A suite of diverse applications of the Pyntfa package is provided to demonstrate the potential of this package.


  • Retrieving weak earthquake signals from highly noisy distributed acoustic sensing data by moving-deep-learning filtering

    by Gui Chen, Yang Liu, Yuhang Sun, Liuqing Yang, Mi Zhang, Haoran Zhang

    Distributed acoustic sensing (DAS) technology using fiber-optic cables offers a cost-effective and high-density approach for recording earthquake signals, particularly compared to conventional receivers. However, DAS recordings are often significantly degraded by strong ambient noise. Therefore, enhancing the signal-to-noise ratio of DAS data is crucial. We develop a data-driven solution, i.e., the moving-deep-learning-filtering (MDLF) framework, which uses learnable filtering coefficients (weights and biases of the neural network) to effectively extract weak DAS signals from significant noise. This MDLF framework incorporates a designed convolutional attention neural network (CANN) optimized with a synthetic data set, and a hybrid loss function consisting of a data distribution loss and a signal structure loss. The trained CANN within MDLF operates on DAS patches extracted using overlapping moving windows, allowing for direct generalization to field DAS data. Experiments on the synthetic data set, as well as two field DAS data sets from the Frontier Observatory for Research in Geothermal Energy and the San Andreas Fault Observatory at Depth projects, demonstrate the effectiveness of our MDLF method, which outperforms benchmark approaches.


  • Robust Bidirectional Q-Compensated Denoising for Seismic Data With Adaptive Structural Regularization

    by Chao Li, Guochang Liu, Liuqing Yang, Sergey Fomel, Yangkang Chen

    Seismic data is commonly contaminated by different types of noise with varying amplitudes, increasing the challenges of retrieving effective signals from strong background noise. Such issues become even worse when attenuation effects are considered because the amplitude of seismic waves will dissipate after propagation, making seismic signals easier to be inundated by noise, especially at deeper positions. To implement data enhancement and attenuation compensation without noise amplification, we propose a robust framework in a blind manner to reconstruct attenuation-compensated seismic data with a higher signal-to-noise ratio (SNR), namely, structural-oriented blind Q-compensated denoising (SBQD) method. Unlike classic Q-compensated denoising methods, the proposed SBQD does not require wavelet as a prior and can iteratively estimate wavelet and reflectivity series simultaneously, and the final stationary seismic data can be obtained by convolving the estimated wavelet with the reflectivity series. Moreover, guided by local adaptive structural regularization, the proposed SBQD can provide superior results with higher accuracy and fidelity and remove those noise-related artifacts during denoising and compensation. Compared with conventional two-step methods (e.g., denoising and attenuation compensation), the proposed SBQD can implement denoising and compensation simultaneously, which avoids introducing compensation-related errors (e.g., noise amplification) and effectively preserves useful signals. Synthetic and field examples are used to validate the robustness of the proposed SBQD method on noise removal and attenuation compensation for seismic data.


  • HCTNet: Robust prestack seismic inversion using a hybrid convolutional transformer

    by Liuqing Yang, Sergey Fomel, Shoudong Wang, Wenjin Li, Jinyu Meng, Chao Li, Yangkang Chen

    Accurately estimating subsurface elastic parameters is essential to characterize oil and gas reservoirs. Prestack amplitude-variation-with-angle (AVA) inversion methods are used to estimate subsurface elastic parameters. The traditional physical model-based linear AVA inversion methods are challenging for accurately calculating the velocity and density of the prestack data with large incidence angles. Data-driven deep-learning networks have received much attention because of their ability to construct nonlinear relationships between a source and target domains. For prestack seismic inversion, the limitation of wells affects the inversion performance of elastic parameters. Here, we construct a hybrid neural network with a 1D convolutional neural network (CNN) and transformer for prestack seismic inversion. We use a hybrid dilated convolutional block with a residual connection mechanism (RHDC) to initially extract features from prestack angle-domain gathers. RHDC has a larger receptive field than traditional convolutional layers, which allows the extraction of larger-scale features. Our method is a prestack elastic parameter simultaneous inversion network based on a multitask learning strategy. We include a seismic data reconstruction branch to improve the inversion performance, which is equivalent to a regularization method that focuses on the overall structure and consistency of the seismic data when inverting elastic parameters. We simulate synthetic data based on the feature distribution of the field data. For field data inversion, the network is pretrained using synthetic data with limited labels, and then the well-trained model is applied to train on field data, reducing reliance on wells. We use a hybrid loss function in the training stage and tune the loss weights across the three inversion branches to balance the inversion performance. Compared with two state-of-the-art deep-learning networks, i.e., a residual CNN and a hybrid-seismic-prior-guided neural network, our method has better horizontal continuity and noise robustness on synthetic and field data examples.


  • Salt3DNet: A Self-Supervised Learning Framework for 3-D Salt Segmentation

    by Liuqing Yang, Sergey Fomel, Shoudong Wang, Xiaohong Chen, Omar M. Saad, Yangkang Chen

    Salt body segmentation is a critical part of structural interpretation and oil and gas exploration for subsalt reservoirs. Existing automatic salt body segmentation techniques mostly use supervised learning strategies. It is challenging to generate a large number of labels by manual labeling, especially for 3-D salt bodies. Here, we propose a self-supervised learning (SSL) framework called Salt3DNet, for 3-D salt body segmentation. This framework is divided into two stages: pretraining and fine-tuning of downstream tasks. In the pretraining stage, we use the Barlow twins (BTs) method to pretrain the encoder and reduce redundancy in a contrastive learning manner to learn high-level data representations. In the fine-tuning stage, we construct two encoders to reconstruct 3-D seismic data and segment salt bodies in a multitask collaborative learning way. The encoder and decoder are composed of the 3-D fully convolutional DenseNet and soft attention mechanism, where the latter represents the selective kernel block (SKB) with multiple kernels of different sizes. Salt3DNet calculates the correlation matrix of features from different perspectives in the pretraining stage and makes it close to the identity matrix to obtain a more prosperous feature representation. Then, Salt3DNet uses a limited number of labeled samples for training. According to the evaluation metrics, the proposed network has demonstrated promising salt segmentation performance in 3-D SEG advanced modeling (SEAM) synthetic data and $F3$ block real seismic data. In addition, the proposed network is demonstrated to have higher prediction accuracy than state-of-the-art salt segmentation frameworks through ablation experiments.


  • Interpretable Unsupervised Learning Framework for Multidimensional Erratic and Random Noise Attenuation

    by Liuqing Yang, Sergey Fomel, Shoudong Wang, Xiaohong Chen, Yaoguang Sun, Yangkang Chen

    Coherent and incoherent noise in seismic data inevitably reduces the quality of subsequent processing, e.g., migration and inversion. Different from random noise, erratic noise follows the non-Gaussian distribution and has high amplitude, which is a challenge to the conventional denoising frameworks based on deep learning (DL). In this study, we propose unsupervised learning (UL)-based multibranch attention network (MANet) to attenuate the erratic and random noise in 2-D and 3-D seismic data. MANet can adaptively attenuate noise in multidimensional seismic data without the need to manually generate labels to train the network. MANet integrates global features of waveforms extracted from multiple branches in a weighted way to enhance attention to significant features, thus obtaining a global and comprehensive representation of weights. To enhance the migration ability of shallow-level-to-deep-level features, we add some skip connections in the corresponding encoder and decoder. We use a robust mean-Huber loss function that is less sensitive to outliers to improve the denoising performance of erratic noise. We apply the proposed network for both 2-D and 3-D synthetic and field data. The denoising results demonstrate that the proposed method has better signal preservation and noise attenuation abilities compared with the conventional denoising methods and the state-of-the-art UL framework. We improve the interpretability of the network by visualizing the weight matrices and different encoders. Besides, the visualization schemes proposed in this article can be applied to more research, such as geological event interpretation, geological resource detection, and surface morphology analysis.


  • SelfMixed: Self-supervised mixed noise attenuation for distributed acoustic sensing data

    by Zitai Xu, Bangyu Wu, Yisi Luo, Liuqing Yang, Yangkang Chen

    Distributed acoustic sensing (DAS) is an emerging data acquisition technique known for its high sensing density, cost effectiveness, and environmental friendliness, making it a technology with significant future application potential in many fields. However, DAS signals are often contaminated by various types of noise, such as high-frequency, high-amplitude erratic, and horizontal noise, making their processing challenging. Therefore, it is crucial to leverage the physical characteristics of these diverse types of noise in DAS data and effectively attenuate them. In this work, we develop SelfMixed, a novel self-supervised learning method for mixed noise suppression of DAS data. We fully exploit the physical characteristics of different types of noise in DAS data and introduce a physical characteristic-based training strategy. Specifically, we use the ℓ2 norm to characterize random noise, the ℓ1 norm for erratic noise, and horizontal smoothness and vertical nonsmoothness for horizontal noise. In addition, we use a blind-spot-based training strategy for DAS denoising, relying solely on observed noisy data. To more effectively attenuate horizontal noise, we also introduce a Fourier transform-based parameterization method. By combining self-supervised deep priors with the physical characteristics of mixed DAS noise, our method effectively attenuates complex mixed noise in field DAS data. Extensive experiments on synthetic and field data from various geographic scenarios validate the superiority of SelfMixed over seven state-of-the-art DAS denoising approaches.


  • Deep learning with soft attention mechanism for small-scale ground roll attenuation

    by Liuqing Yang, Sergey Fomel, Shoudong Wang, Xiaohong Chen, Yangkang Chen

    Ground roll is a type of coherent noise with low frequency, low velocity, and high amplitude, which masks useful signals and decreases the quality of subsequent seismic data processing. It is a challenge for traditional signal processing methods to separate useful signals effectively when the ground roll and useful reflected signals overlap seriously in the low-frequency band. We develop a supervised-learning-based framework with soft attention residual learning mechanisms for suppressing the ground roll noise. To reduce the cost of manual labeling, the 2D patching technique is used to segment large-scale seismic data into a large number of small-scale patches for training. Our network includes a multibranch attention block that uses multiple branches with different kernel sizes to extract waveform features at different scales from input noisy patches. Then, we use the soft attention mechanism to select and fuse the feature maps of different branches. Our network can achieve encouraging ground roll attenuation performance by using a small number of training samples, which is demonstrated by synthetic and field data examples. Compared with one traditional method and two advanced deep-learning frameworks, our network has better abilities in preserving low-frequency useful signals and removing ground roll.