Yinghe Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yinghe Wu has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Shulin Pan
Southwest Petroleum University
7 shared publications - Guojie Song
School of Sciences, Southwest Petroleum University, Chengdu 610500, China
5 shared publications - Jingyi Chen
Department of Geosciences, Seismic Anisotropy Group, The University of Tulsa, Tulsa, OK, USA
3 shared publications - Shengbo Yi
School of Geoscience and Technology, Southwest Petroleum University, Chengdu, China
3 shared publications - Dawei Liu
Xi’an Jiaotong University
2 shared publications Yaojie Chen
Southwest Petroleum University, School of Geoscience and Technology 1 , Chengdu
2 shared publications
Latest Publications
- 1 January 2026
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 91 (1) , WA123-WA135
Rapid retrieval and classification of passive-source body wave events using a convolutional self-attention encoder: Application to gas storage monitoring
Passive-source seismic interferometry (SI) demonstrates significant potential in geophysical monitoring owing to its low cost and nondestructive attributes. However, uneven noise source distributions, pervasive surface waves, and other sources of noise often obscure body wave reflections, leading to very low signal-to-noise ratios in reconstructed signals. Additionally, long-term data acquisition generates large, high-noise data sets, increasing processing costs. Thus, rapidly retrieving body wave-dominated segments is essential for preprocessing. In this study, a convolutional self-attention encoder (CSE) was introduced to address these challenges. This method adopted multimodal inputs consisting of noise segments and their corresponding frequency attributes, enhancing overall performance. An attention mechanism was integrated to construct a binary classifier based on the CSE, enhancing contextual understanding. These steps helped the CSE to better learn features from the data sets, thereby improving generalization ability and prediction accuracy. Accordingly, the network was updated through incremental learning on small data sets, enabling continuous prediction and gradual adaptation to new data. This enabled us to rapidly generate large, high-quality training data sets and to train models with high performance. To further improve prediction data quality, the intermediate features extracted from the deep convolutional layers corresponding to the retrieved body wave events were extracted, and these features were clustered to perform quality classification of the retrieved data. Synthetic simulation data and field data from Northwest China validated that the proposed method retrieves and classifies body wave reflections with high accuracy and efficiency, providing robust data for subsequent monitoring and enhancing imaging results. The proposed method provides efficient and reliable data preprocessing for the application of passive-source SI in energy transition scenarios, such as subsurface fluid migration monitoring and carbon dioxide geological sequestration, and facilitates the development of low-carbon exploration and green transformation in the energy industry.
- 1 January 2026
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 91 (1) , WA123-WA135
Rapid retrieval and classification of passive-source body wave events using a convolutional self-attention encoder: Application to gas storage monitoring
Passive-source seismic interferometry (SI) demonstrates significant potential in geophysical monitoring owing to its low cost and nondestructive attributes. However, uneven noise source distributions, pervasive surface waves, and other sources of noise often obscure body wave reflections, leading to very low signal-to-noise ratios in reconstructed signals. Additionally, long-term data acquisition generates large, high-noise data sets, increasing processing costs. Thus, rapidly retrieving body wave-dominated segments is essential for preprocessing. In this study, a convolutional self-attention encoder (CSE) was introduced to address these challenges. This method adopted multimodal inputs consisting of noise segments and their corresponding frequency attributes, enhancing overall performance. An attention mechanism was integrated to construct a binary classifier based on the CSE, enhancing contextual understanding. These steps helped the CSE to better learn features from the data sets, thereby improving generalization ability and prediction accuracy. Accordingly, the network was updated through incremental learning on small data sets, enabling continuous prediction and gradual adaptation to new data. This enabled us to rapidly generate large, high-quality training data sets and to train models with high performance. To further improve prediction data quality, the intermediate features extracted from the deep convolutional layers corresponding to the retrieved body wave events were extracted, and these features were clustered to perform quality classification of the retrieved data. Synthetic simulation data and field data from Northwest China validated that the proposed method retrieves and classifies body wave reflections with high accuracy and efficiency, providing robust data for subsequent monitoring and enhancing imaging results. The proposed method provides efficient and reliable data preprocessing for the application of passive-source SI in energy transition scenarios, such as subsurface fluid migration monitoring and carbon dioxide geological sequestration, and facilitates the development of low-carbon exploration and green transformation in the energy industry.
- 24 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-10
Physically Guided High-Resolution Acoustic Impedance Inversion Based on Hybrid Networks
Seismic acoustic impedance inversion is essential for reservoir prediction and characterization. In recent years, deep learning has shown immense potential as a data-driven approach in seismic data processing, inversion, and interpretation. As a data-driven method, deep learning-based seismic inversion results better when sufficient labeled data are provided. Overfitting and poor generalization often occur when labels are insufficient. Due to the lack of labeled data in seismic inversion problems, the difficulty of inversion increases, leading to unstable and poor generalization of prediction results. To partially address this issue, we propose a constrained seismic inversion strategy. Since seismic records are time series, we exploit the CNN and Bi-LSTM network structures that are more applicable to time series. We combine the physical model and the initial model as constraints to improve the network stability and generalization ability, and impose sparse constraints on the reflection coefficient to further improve the prediction accuracy. The network structure transformation improves the efficiency and stability of the training process. Through numerical experiments and real data tests, it is proved that the proposed method improves the vertical resolution and geological reliability, providing a more stable and efficient method for seismic inversion under conditions of limited labeled data. The overall performance improved by 2% through comparative analysis.
- 24 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-10
Physically Guided High-Resolution Acoustic Impedance Inversion Based on Hybrid Networks
Seismic acoustic impedance inversion is essential for reservoir prediction and characterization. In recent years, deep learning has shown immense potential as a data-driven approach in seismic data processing, inversion, and interpretation. As a data-driven method, deep learning-based seismic inversion results better when sufficient labeled data are provided. Overfitting and poor generalization often occur when labels are insufficient. Due to the lack of labeled data in seismic inversion problems, the difficulty of inversion increases, leading to unstable and poor generalization of prediction results. To partially address this issue, we propose a constrained seismic inversion strategy. Since seismic records are time series, we exploit the CNN and Bi-LSTM network structures that are more applicable to time series. We combine the physical model and the initial model as constraints to improve the network stability and generalization ability, and impose sparse constraints on the reflection coefficient to further improve the prediction accuracy. The network structure transformation improves the efficiency and stability of the training process. Through numerical experiments and real data tests, it is proved that the proposed method improves the vertical resolution and geological reliability, providing a more stable and efficient method for seismic inversion under conditions of limited labeled data. The overall performance improved by 2% through comparative analysis.
- 11 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-12
Unsupervised-Learning Stable Inverse Q Filtering for Seismic Resolution Enhancement
Affected by near-surface absorption, seismic wave energy attenuation and phase distortion greatly reduce the resolution and signal-to-noise ratio (SNR) of seismic data, causing changes in seismic attributes much greater than other factors. Inverse Q filtering is a common method to compensate for these undesirable effects. To overcome the drawbacks of the traditional inverse Q filtering, such as the difficulty of parameter selection and the instability of wave amplitude compensation, we propose a new unsupervised inverse Q filtering method in a deep learning (DL) framework, using a forward attenuation operator based on the seismic wave attenuation theory to drive the network. The filtering strategy does not require actual training labels and avoids the numerical instability of the amplitude compensation. First, we design a hybrid CNN-BiLSTM-Attention model for multivariate time series prediction, and then take the data to be compensated as input for the DL network and the compensated data as output. The output is then attenuated using a forward attenuation operator constructed from the near-surface Q model. After, the error between the attenuated data and the original input data is transmitted back to the DL network to modify the network output, and the error is minimized by optimizing the network parameters to generate the final compensation result. In the entire prediction process, there is no need to produce un-attenuated data labels, which achieves the effect of unsupervised learning. The results with synthetic and field data demonstrate that the unsupervised method can effectively and stably compensate for seismic signals. Compared to the classical inverse Q filtering, the proposed method improves the resolution and SNR of seismic records.
- 11 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-12
Unsupervised-Learning Stable Inverse Q Filtering for Seismic Resolution Enhancement
Affected by near-surface absorption, seismic wave energy attenuation and phase distortion greatly reduce the resolution and signal-to-noise ratio (SNR) of seismic data, causing changes in seismic attributes much greater than other factors. Inverse Q filtering is a common method to compensate for these undesirable effects. To overcome the drawbacks of the traditional inverse Q filtering, such as the difficulty of parameter selection and the instability of wave amplitude compensation, we propose a new unsupervised inverse Q filtering method in a deep learning (DL) framework, using a forward attenuation operator based on the seismic wave attenuation theory to drive the network. The filtering strategy does not require actual training labels and avoids the numerical instability of the amplitude compensation. First, we design a hybrid CNN-BiLSTM-Attention model for multivariate time series prediction, and then take the data to be compensated as input for the DL network and the compensated data as output. The output is then attenuated using a forward attenuation operator constructed from the near-surface Q model. After, the error between the attenuated data and the original input data is transmitted back to the DL network to modify the network output, and the error is minimized by optimizing the network parameters to generate the final compensation result. In the entire prediction process, there is no need to produce un-attenuated data labels, which achieves the effect of unsupervised learning. The results with synthetic and field data demonstrate that the unsupervised method can effectively and stably compensate for seismic signals. Compared to the classical inverse Q filtering, the proposed method improves the resolution and SNR of seismic records.
- 4 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
An Automatic Screening Method for the Passive Surface-Wave Imaging Based on the F–K Domain Energy Characteristics
Due to low cost and nondestructive characteristics, the passive surface-wave imaging has shown great potentials in urban near-surface exploration. However, the imaging methods are facing with many challenges in practical applications, such as uneven noise source distribution and complex site environment, which will seriously affect the dispersion imaging quality of surface waves, and result in the failure of retrieving accurate dispersion curve and inaccurate inversion. Therefore, data screening is required to improve the accuracy of passive surface-wave dispersion imaging. This process is usually completed manually, which is time-consuming for processing the large data sets. To solve this problem, we propose an automatic data screening method for the near-surface passive surface-wave imaging. Based on the distribution characteristics of the surface-wave energy of noise data in the F-K domain, this proposed method uses the least-squares technique to fit the quadratic distribution of energy, and sorts the noise time segments according to the defined correlation coefficients and bandwidth coefficients. Thus, we can automatically detect the noise time segments with high signal-to-noise ratio (SNR) without manual intervention. In order to verify the effectiveness of this proposed method, both synthetic data and field data are used in this study. The results show that this automatic data screening method significantly improves the accuracy of the passive surface-wave dispersion imaging, effectively expands the surface-wave energy band, and realizes rapid and automatic data screening.
- 4 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-12
An Automatic Screening Method for the Passive Surface-Wave Imaging Based on the F–K Domain Energy Characteristics
Due to low cost and nondestructive characteristics, the passive surface-wave imaging has shown great potentials in urban near-surface exploration. However, the imaging methods are facing with many challenges in practical applications, such as uneven noise source distribution and complex site environment, which will seriously affect the dispersion imaging quality of surface waves, and result in the failure of retrieving accurate dispersion curve and inaccurate inversion. Therefore, data screening is required to improve the accuracy of passive surface-wave dispersion imaging. This process is usually completed manually, which is time-consuming for processing the large data sets. To solve this problem, we propose an automatic data screening method for the near-surface passive surface-wave imaging. Based on the distribution characteristics of the surface-wave energy of noise data in the F-K domain, this proposed method uses the least-squares technique to fit the quadratic distribution of energy, and sorts the noise time segments according to the defined correlation coefficients and bandwidth coefficients. Thus, we can automatically detect the noise time segments with high signal-to-noise ratio (SNR) without manual intervention. In order to verify the effectiveness of this proposed method, both synthetic data and field data are used in this study. The results show that this automatic data screening method significantly improves the accuracy of the passive surface-wave dispersion imaging, effectively expands the surface-wave energy band, and realizes rapid and automatic data screening.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 20 (1545598X) , 1-5
Noisy Supervised Deep Learning for Remote Sensing Image Segmentation Using Electronic Maps
Deep learning has made substantial progress in remote sensing image segmentation tasks. It usually requires a large number of high-quality annotation maps (i.e., clean labels), which are labor-intensive. In this paper, we propose to use abundant electronic maps (i.e., noisy labels) to supplement a small number of clean labels to solve the problem of massive label production. In addition, we propose a multi-stage noise supervised framework (NSDI) to prevent noise from deteriorating the performance of deep model. NSDI consists of clean training , weight initialization , and hybrid training stages. In the clean training stage, we train a segmentation model using a small amount of clean labels and remote sensing images to compute the confusion probability matrix. In the weight initialization stage, we use the confusion probability matrix as well as the prediction probability to calculate the label reliability of the electronic map. In the hybrid training stage, an adaptive weighted loss function based on cross-entropy is used to dynamically update the label reliability. Then we train model further using electronic maps with the support of the adaptive weighted loss function and label reliability. Experiments were undertaken on 2,656 images of 512 × 512 pixels. Ablation studies show that NSDI improves the model robustness as well as the segmentation quality.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 20 (1545598X) , 1-5
Noisy Supervised Deep Learning for Remote Sensing Image Segmentation Using Electronic Maps
Deep learning has made substantial progress in remote sensing image segmentation tasks. It usually requires a large number of high-quality annotation maps (i.e., clean labels), which are labor-intensive. In this paper, we propose to use abundant electronic maps (i.e., noisy labels) to supplement a small number of clean labels to solve the problem of massive label production. In addition, we propose a multi-stage noise supervised framework (NSDI) to prevent noise from deteriorating the performance of deep model. NSDI consists of clean training , weight initialization , and hybrid training stages. In the clean training stage, we train a segmentation model using a small amount of clean labels and remote sensing images to compute the confusion probability matrix. In the weight initialization stage, we use the confusion probability matrix as well as the prediction probability to calculate the label reliability of the electronic map. In the hybrid training stage, an adaptive weighted loss function based on cross-entropy is used to dynamically update the label reliability. Then we train model further using electronic maps with the support of the adaptive weighted loss function and label reliability. Experiments were undertaken on 2,656 images of 512 × 512 pixels. Ablation studies show that NSDI improves the model robustness as well as the segmentation quality.
- 5 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-13
An Unsupervised Inversion Method for Seismic Brittleness Parameters Driven by the Physical Equation
Brittleness is an important parameter characterizing the fracturing properties of shale reservoir, which can be predicted by the pre-stack seismic inversion. In order to overcome the low efficiency and ill-posed problems of the traditional pre-stack brittleness inversion, we propose a new unsupervised deep learning (DL) inversion method for seismic brittleness parameters based on the physical equation. This method integrates DL framework and the physical equation, and provides a DL inversion strategy without actual labels. We first input the original seismic data into the Fastformer network, and use the low-frequency model as the physical constraint to predict the brittle parameters. Then, the prediction results of brittleness parameters are sent to the forward modeling module (a linear approximation equation) to calculate the synthetic seismic data. Next, the error between the calculated seismic data and original seismic data is used to update the network prediction results. The network parameters are iteratively optimized to minimize the error, and the brittle prediction parameters are finally output. In the whole training process, it is not necessary to use the real brittle parameters as the labels. Through this method, the effect of approximate unsupervised learning is obtained. Finally, we apply the proposed method to the synthetic data and field data, and compared with the results inverted by the traditional L1 method. The experimental results show that the proposed method has higher inversion accuracy and efficiency than the traditional L1 method, which has a great potential in the practical application.
- 5 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-13
An Unsupervised Inversion Method for Seismic Brittleness Parameters Driven by the Physical Equation
Brittleness is an important parameter characterizing the fracturing properties of shale reservoir, which can be predicted by the pre-stack seismic inversion. In order to overcome the low efficiency and ill-posed problems of the traditional pre-stack brittleness inversion, we propose a new unsupervised deep learning (DL) inversion method for seismic brittleness parameters based on the physical equation. This method integrates DL framework and the physical equation, and provides a DL inversion strategy without actual labels. We first input the original seismic data into the Fastformer network, and use the low-frequency model as the physical constraint to predict the brittle parameters. Then, the prediction results of brittleness parameters are sent to the forward modeling module (a linear approximation equation) to calculate the synthetic seismic data. Next, the error between the calculated seismic data and original seismic data is used to update the network prediction results. The network parameters are iteratively optimized to minimize the error, and the brittle prediction parameters are finally output. In the whole training process, it is not necessary to use the real brittle parameters as the labels. Through this method, the effect of approximate unsupervised learning is obtained. Finally, we apply the proposed method to the synthetic data and field data, and compared with the results inverted by the traditional L1 method. The experimental results show that the proposed method has higher inversion accuracy and efficiency than the traditional L1 method, which has a great potential in the practical application.
- Open Access
Research on Rock Physics Modeling Methods for Fractured Shale Reservoirs
The Sichuan Basin is a significant region for exploration and development of shale gas in China, and it is essential to clarify the impact of deep shale gas reservoir parameters on cost-effective development at scale to ensure national energy security. Rock physics modeling is a significant means of communicating the physical and elastic parameters of rocks. A rock physics modeling method applicable to fractured shale gas reservoirs is proposed for the current situation of complex fluid relationships in shale gas reservoirs and unclear characteristics of gas identification seismic response. In this paper, based on the Self Consistent Approximation (SCA) model and the differential effective medium (DEM) model, the anisotropic source of shale is used as a starting point to add bound water, kerogen, clay, and brittle minerals, the Schoenberg linear slip theory is used to add fracture disturbance effects, and then the Brown–Korringa model is used to perform fluid replacement under anisotropic conditions. Finally a rock physics model applicable to fractured shale gas reservoirs is obtained, and the established rock physics model is used for analysis of elastic parameters, Thomsen parameters, and fracture weakness parameters. Rock physics tests were performed on shale in southern Sichuan as an example. The experimental results show that the model established by the process can accurately invert the longitudinal and transverse wave velocities of the shale, which can provide a conceptual basis for the study of fractured shale gas reservoirs.
Open AccessResearch on Rock Physics Modeling Methods for Fractured Shale Reservoirs
The Sichuan Basin is a significant region for exploration and development of shale gas in China, and it is essential to clarify the impact of deep shale gas reservoir parameters on cost-effective development at scale to ensure national energy security. Rock physics modeling is a significant means of communicating the physical and elastic parameters of rocks. A rock physics modeling method applicable to fractured shale gas reservoirs is proposed for the current situation of complex fluid relationships in shale gas reservoirs and unclear characteristics of gas identification seismic response. In this paper, based on the Self Consistent Approximation (SCA) model and the differential effective medium (DEM) model, the anisotropic source of shale is used as a starting point to add bound water, kerogen, clay, and brittle minerals, the Schoenberg linear slip theory is used to add fracture disturbance effects, and then the Brown–Korringa model is used to perform fluid replacement under anisotropic conditions. Finally a rock physics model applicable to fractured shale gas reservoirs is obtained, and the established rock physics model is used for analysis of elastic parameters, Thomsen parameters, and fracture weakness parameters. Rock physics tests were performed on shale in southern Sichuan as an example. The experimental results show that the model established by the process can accurately invert the longitudinal and transverse wave velocities of the shale, which can provide a conceptual basis for the study of fractured shale gas reservoirs.
- Open Access
- 25 December 2022
- Published by MDPI AG in Applied Sciences
- Vol. 13 (1) , 246
Direct Inversion Method of Brittleness Parameters Based on Reweighted Lp-Norm
Brittleness is an important factor that indicates shale properties, as well as fracturability, and it can be well-represented using the elasticity parameter Eρ. Seismic inversion allows direct access to the Eρ ofbrittleness parameters. Seismic inversion is a typical ill-posed problem that has an enormous multiplicity of solutions. In order to invert reservoir brittleness parameters more stably and reliably, a direct inversion method for determining brittleness parameters based on a reweighted Lp-norm is proposed, and the reweighted Lp method is introduced to brittleness parameter inversion for the first time. The alternating direction method of multipliers (ADMM) is used to establish the inversion structure and to optimize the objective function in blocks, which effectively improves the convergence speed. We first introduce a reweighted Lp method and establish a pre-stack inversion objective function based on the reweighted Lp method. Then, theoretical simulation data are applied to compare the inversion outcomes of the new method with those of the traditional method, and the effect of the method in this manuscript is verified. Finally, the feasibility of this method is further verified using actual data for experimental analysis. Through an analysis of the experimental results, we find that this method can be well-applied to seismic pre-stack inversion calculation and provides a new direct inversion method for the determination of brittleness parameters for exploration geophysics.
- 25 December 2022
- Published by MDPI AG in Applied Sciences
- Vol. 13 (1) , 246
Open AccessDirect Inversion Method of Brittleness Parameters Based on Reweighted Lp-Norm
Brittleness is an important factor that indicates shale properties, as well as fracturability, and it can be well-represented using the elasticity parameter Eρ. Seismic inversion allows direct access to the Eρ ofbrittleness parameters. Seismic inversion is a typical ill-posed problem that has an enormous multiplicity of solutions. In order to invert reservoir brittleness parameters more stably and reliably, a direct inversion method for determining brittleness parameters based on a reweighted Lp-norm is proposed, and the reweighted Lp method is introduced to brittleness parameter inversion for the first time. The alternating direction method of multipliers (ADMM) is used to establish the inversion structure and to optimize the objective function in blocks, which effectively improves the convergence speed. We first introduce a reweighted Lp method and establish a pre-stack inversion objective function based on the reweighted Lp method. Then, theoretical simulation data are applied to compare the inversion outcomes of the new method with those of the traditional method, and the effect of the method in this manuscript is verified. Finally, the feasibility of this method is further verified using actual data for experimental analysis. Through an analysis of the experimental results, we find that this method can be well-applied to seismic pre-stack inversion calculation and provides a new direct inversion method for the determination of brittleness parameters for exploration geophysics.
- 29 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
A Surface-Wave Inversion Method Based on FHLV Loss Function in LSTM
Surface-wave analysis methods have been widely applied to construct near-surface shear-wave velocity structures. Whether it is an active source or passive source, the near-surface shear-wave velocity structure is obtained by inverting the surface-wave dispersion curve. In order to solve the problems of low inversion efficiency and poor inversion results in traditional surface-wave exploration, we have studied the surface-wave inversion methods based on deep learning technology. In this study, we propose a long short-term memory (LSTM) surface-wave inversion method based on the first height last velocity (FHLV) loss function. The core of our proposed method is the FHLV loss function consisting of two parts: a speed loss and a thickness loss, which improves the overall prediction accuracy through optimizing the learning process of the thickness parameter by the network. To verify the accuracy of the proposed LSTM surface-wave inversion method based on the FHLV loss function, experiments are conducted on both synthetic and real datasets. The results show that our proposed method can efficiently and accurately invert the near-surface shear-wave velocity structure.
- 29 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 19 (1545598X) , 1-5
A Surface-Wave Inversion Method Based on FHLV Loss Function in LSTM
Surface-wave analysis methods have been widely applied to construct near-surface shear-wave velocity structures. Whether it is an active source or passive source, the near-surface shear-wave velocity structure is obtained by inverting the surface-wave dispersion curve. In order to solve the problems of low inversion efficiency and poor inversion results in traditional surface-wave exploration, we have studied the surface-wave inversion methods based on deep learning technology. In this study, we propose a long short-term memory (LSTM) surface-wave inversion method based on the first height last velocity (FHLV) loss function. The core of our proposed method is the FHLV loss function consisting of two parts: a speed loss and a thickness loss, which improves the overall prediction accuracy through optimizing the learning process of the thickness parameter by the network. To verify the accuracy of the proposed LSTM surface-wave inversion method based on the FHLV loss function, experiments are conducted on both synthetic and real datasets. The results show that our proposed method can efficiently and accurately invert the near-surface shear-wave velocity structure.