In oilfield exploration and development, drilling noise creates significant interference, severely reducing the signal-to-noise ratio (SNR) of seismic data. Due to the complex characteristics of noise and signal, suppressing noise while recovering effective signals poses a challenge for denoising models. To address this, we propose a multiscale generator network with adaptive feature extraction for drilling noise suppression. The constructed network primarily consists of a dual-layer encoder-decoder structure. In the encoder, we designed an adaptive feature extraction module (AFEM) and a depthwise separable encoding module. The former utilizes deformable convolutions for adaptive extraction of effective features in seismic data, while the latter employs depthwise separable convolutions and an inverse bottleneck design to reduce computational complexity while maintaining effective feature extraction. The decoding module in the decoder uses two convolutional layers to reconstruct the seismic data with minimal computational cost. To prevent network degradation, residual connections are employed in both the encoding and decoding modules. The dual-layer structure extracts semantic information at different scales, preserving richer effective signal features and maximizing drilling noise suppression. Experimental results on both synthetic and field data demonstrate that the proposed method achieves higher-quality denoising results compared to denoising convolutional neural network (CNNS) (DnCNN), Unet, and MLGNet, while retaining the maximum amount of effective data.

Resolution of time-frequency based seismic attributes mainly relies on the time-frequency analysis tool. This study proposes an improved second-order synchroextracting wavelet transform (SSEWT) by optimizing the scale parameters and extraction scheme. Time–frequency computation on synthetic data shows a 5% improvement in efficiency. Then we apply the proposed transform to fluid mobility calculation on field data, yielding a 5.6% increase in computational efficiency and an 11.26% improvement in resolution, demonstrating its superior performance. Field data tests demonstrate that the proposed transform and the related fluid mobility result outperform conventional methods. Despite remaining computational challenges, the method offers significant advancements in reservoir characterization and fluid detection.

Acoustic impedance (AI) reflects the physical properties of subsurface materials and is a key parameter in reservoir evaluation and hydrocarbon exploration. Traditional poststack AI inversion methods typically assume a time-invariant wavelet, neglecting the nonstationary effects caused by attenuation and absorption during wave propagation. Therefore, performing attenuation compensation prior to inversion can effectively enhance seismic recording resolution and improve inversion accuracy. Nonetheless, attenuation compensation frequently encounters challenges such as incomplete correction and the amplification of noise. In this context, impedance inversion of nonstationary seismic records that accounts for attenuation effects can effectively avoid cumulative errors and noise amplification caused by attenuation compensation, thereby enhancing both the efficiency and accuracy of the inversion process. Existing nonstationary inversion methods, typically based on constant-Q models and reliant solely on time-domain information, struggle to handle realistic frequency-dependent Q variations. To address these limitations, we propose a poststack acoustic impedance inversion method in the time-frequency mixed domain, applicable to frequency-dependent Q attenuation. First, we derive a nonstationary convolution model in both time and frequency domains based on the frequency-dependent Q model. Then, we construct an impedance inversion objective function in the time-frequency mixed domain and solve it using the alternating direction method of multipliers (ADMM). The ADMM algorithm enables efficient iterative optimization by decomposing the objective into subproblems and solving them alternately. Tests on synthetic and field data demonstrate that the proposed method can effectively perform AI inversion from nonstationary seismic records with frequency-dependent Q attenuation. By incorporating frequency-domain information, it enhances the accuracy and resolution of the AI inversion.

Acoustic impedance (AI) inversion method based on the assumption of spatially invariant seismic wavelets has been widely applied in reservoir prediction. In reality, seismic wavelets are relatively stable but also exhibit subtle spatial variations, particularly when inconsistencies arise between seismic data and reservoir heterogeneity. To address this issue, we propose an alternate iterative inversion method of AI and wavelet based on sparse constraints. This method assumes that both the wavelet matrix and the reflection coefficients are sparse. Based on the Lp norm, we establish objective functions for wavelet estimation and AI inversion with sparse regularization constraints of different forms, ensuring the stability of the inversion results during the iterative process. In the inversion process, the statistical wavelet is used as the initial wavelet input to obtain the AI; then, the AI results are used to update the wavelet, and the two parameters are iteratively updated in this way. Through this alternating iteration, spatial variations of the wavelet are effectively captured, eliminating discrepancies in waveform and frequency characteristics of seismic data in the horizontal direction due to nonimpedance variations. Finally, the effectiveness of the proposed method is validated through model testing and application to actual data. The proposed method can invert impedance results that are more consistent with logging data, and the estimated wavelets synthesized seismogram show a better match with the well bypass. In particular, the proposed method is more applicable to the case where it is difficult to use a single wavelet to calibrate all wells simultaneously during the well-seismic calibration in the work area with multiple well data.

Description of fluid properties, natural fractures, and stress conditions is significant for seismic characterization of unconventional fractured reservoirs. For the first two aspects, traditional methods to solid-liquid decoupling by low-frequency anisotropic Gassmann fluid substitution equation fall short in two aspects: 1) neglecting the combined effects of fractures and pore fluid, and 2) disregarding the influence of fracture inclination on the seismic anisotropy. To address the above two issues simultaneously, this article derives a novel PP-wave reflection coefficient incorporating fluid bulk modulus, vertical effective stress correlation parameter, fracture density, and coupled anisotropic fluid indicator (CFI) in the tilted transverse isotropy (TTI) medium. The derived TTI-saturated stiffnesses with CFI are revealed to provide the improved accuracy regarding physical property parameters (porosity, fluid fillings, and matrix mineral content) and fracture parameters (including fracture density and inclination). To invert the above key parameters from offset vector tile (OVT) domain seismic data, the stepwise seismic inversion strategy based on Lp quasi-norm sparsity constraints is employed. Synthesized azimuthal seismic data with varying signal-to-noise ratios (SNR) validate the feasibility and robustness of the proposed inversion method. Ultimately, the innovative method exhibits compelling effectiveness when applied to fractured shale gas-bearing reservoirs in the Sichuan Basin, China.

Presents corrections to the paper, (Corrections to “Seismic Inversion of Anisotropic and Isotropic Fluid Indicators and Fracture Density in the TTI Medium”).

Accurate first break picking is essential for tunnel microseismic monitoring. Here, we propose a self-supervised convolutional clustering picking (SCCP) method for automatically picking the first break of microseismic recordings. The time–frequency features are decomposed and reconstructed using accurate convolutional encoding and decoding under self-supervision. Then, the autoencoder output is unsupervisedly clustered into useful and invalid waveform sections employing the fuzzy$c$-means (FCMs) algorithm under long short-term memories, global attention, and self-attention constraints. Furthermore, the first point of the useful waveform is determined as the first break. Our results demonstrate that the proposed SCCP method outperforms the short-term average/long-term average (STA/LTA) and Akaike information criterion (AIC). Compared with PhaseNet, a supervised deep-learning method, the SCCP, produces similar performance without using human-labeled data. Practically, when the signal-to-noise ratio (SNR) is reduced to −6 dB, the average mean absolute error and standard deviation of the picking results remain at 1.12 and 9.19 ms, respectively.

During the multicomponent seismic exploration, the P- and S-waves can be recorded, enabling us to obtain more information about the subsurface media. In the interpretation and processing of most multicomponent seismic data, the accurate separation of P- and S-waves in forward simulation and wavefield extrapolation is essential, such as elastic reverse time migration and elastic full-waveform inversion. We decouple the elastic wavefield based on the first-order velocity-stress elastic wave decoupling equation. However, this method produces obvious separation artifacts due to the nonzero spatial derivative of the elastic parameters at the geological interfaces, which can affect subsequent seismic processing and interpretation. To address this issue, an adaptive median filtering method is proposed to remove these artifacts. First, calculate the velocity change rate. Utilizing a designed threshold to filter velocity change rates, we determine the location of the interface where the artifacts appear. Then, the window calculation function based on the wavelength of seismic is used to obtain the median filtering window used for processing. Finally, it applies adaptive median filtering to remove the separated artifacts while preserving the valid information of the wavefields. This method preserves the effective wavefields and effectively removes separation artifacts. Through numerical simulations, it is verified that the adaptive median filtering method can effectively and completely remove the separation artifacts occurring at all interfaces.

Viscoelastic attenuation occurs when seismic waves propagate through subsurface media, significantly reducing the signal-to-noise ratio (SNR) and resolution, which severely impacts the application and interpretation of seismic records. Inverse Q filtering (IQF) compensates for this attenuation via the method of wavefield continuation. The current IQF algorithms use the constant Q model, which assumes a frequency-independent Q. However, some field data and laboratory measurements indicate that Q is a function of frequency and reveal the limitations in the traditional IQF. In this article, we propose inverse frequency-dependent Q filtering (IFQF) based on the power-law Q model. In this method, the phase compensation is unconditionally stabilized, whereas the amplitude compensation is numerically unstable. Thus, stabilization is achieved by incorporating a stabilization factor. The experimental results on synthetic and field data show that the conventional IQF ignores the frequency dependence of the Q value, potentially leading to overcompensation. IFQF effectively compensates for phase dispersion and amplitude attenuation induced by the frequency-dependent Q attenuation, which significantly improves the resolution of seismic recordings.

During the propagation of seismic waves in the Earth, the influence of Earth absorption leads to energy attenuation and phase distortion. Traditional seismic acoustic impedance (AI) inversion typically utilizes a constant wavelet to construct a wavelet library for AI inversion, resulting in outcomes that struggle to effectively portray the spatial variation characteristics of underground reservoirs. While using an inverse Q filter for energy compensation before AI inversion can enhance precision, accurately extracting the Q factor before the inverse Q filter is complex. Moreover, the inverse Q filter often amplifies noise interference, impacting the accuracy of the inversion. Therefore, we propose an alternating iterative synchronous inversion method for AI and the quality factor. Compared with the conventional method, the impedance information and attenuation parameters of seismic data can be obtained directly, which improves the efficiency and accuracy of inversion. The method can be divided into two steps. Firstly, we construct the L1-2-norm regularization of the AI inversion objective function, which is solved using the difference of convex algorithm (DCA) and the alternating direction method of multipliers (ADMM). Subsequently, the Markov Chain Monte Carlo (MCMC) method is employed to solve the nonstationary forward equation, with the fitting of the centroid frequency of seismic data and synthetic seismic records. Through continuous iteration of AI and the quality factor until the error is below the specified threshold or reaches the designated number of iterations, stable AI and average Q factors are obtained. Numerical simulations demonstrate the reliability of the proposed method, showcasing higher accuracy compared to stable wavelet inversion and constant Q inversion. Real data applications validate the method’s effectiveness, providing more dependable results for reservoir fluid identification. The obtained average Q factors can be employed for seismic data processing and interpretation.

The brittleness index (BI) is an essential parameter for evaluating rock brittleness, which can effectively guide artificial hydraulic fracturing for shale gas exploitation. However, some conventional BIs often fall short when applied to shale reservoirs with low permeability, low porosity, and high total organic carbon (TOC) content; we proposed a novel BI more sensitive to brittle rock fractions, porosity, and organic matter content, which was corroborated via the constructed petrophysical modeling method for shale reservoirs. Considering the disadvantage of accumulating error in the indirect calculation of elastic parameters, we further derived a new PP-wave reflection coefficient approximation equation incorporating the novel BI, Young’s modulus, and density, named the BYD reflection coefficient approximation equation. Rigorous validation against the exact Zoeppritz equation by four classes of amplitude versus offset (AVO) seismic response models underlines its accuracy and reliability. Furthermore, we enhanced the seismic inversion algorithm using the reweighted$L_{P}$(RLP) quasinorm, surpassing the sparsity of the$L_{P}$quasinorm. The nonlinear optimization problem was solved by the alternating direction method of multipliers (ADMM) and the$L_{P}$quasinorm iterative shrinkage thresholding algorithm (ISTA). Three sets of model tests with different noises have demonstrated the rationality and robustness of the inversion method. Finally, the 3-D seismic inversion of the novel BI parameter applied to a gas-bearing shale reservoir in Sichuan Basin, China, confirmed its effectiveness in characterizing brittleness distribution, facilitating optimal shale gas exploitation.

As a common seismic facies visualization analysis method, self-organizing map (SOM) projects the waveform or seismic attribute vectors into a two-dimensional topological plane in a nonlinear way, which can effectively and efficiently discover the topological structure of a dataset. SOM does not need to set the number of classes in prior and has friendly visualization characteristics and excellent generalization, which are conducive to seismic facies interpretation using unlabeled data. However, due to the competitive learning used in SOM and the imbalance of data distribution in real world, the samples from majority classes are expanded on the topological plane and the minority classes are compressed. As a result, the plane cannot accurately describe the global structure of data distribution. To improve the visualization precision by modifying the topological relationship of the prototype vectors of SOM, we utilize uniform manifold approximation and projection (UMAP), a novel manifold learning technique for dimension reduction, to correct the prototype vectors generated by SOM. By combining the advantages of SOM and UMAP in the representation of data topological structure, the global relationship between seismic data samples can be properly established, and the internal relative spatial structure of majority class samples can be retained as much as possible, resulting in a more reliable classification. Meanwhile, the framework maintains the advantages of SOM in visualization. In the modeling tests and real data experiments, we have demonstrated the effectiveness and rationality of UMAP-SOM on the spatial structure representation of three-dimensional seismic data.

Missing traces affect the result of subsequent steps, such as migration and amplitude versus offset (AVO) analysis, which harms the understanding of the subsurface structure and hydrocarbon exploration. Framelet transform can sparsely represent seismic data and it can describe data in detail. Compared with the commonly used$l_{1}$norm,$l_{p}$quasi-norm has higher sparsity. In this letter, we establish a new subject with$l_{p}$quasi-norm and framelet transform to reconstruct the seismic record. Instead of a conventional solver, we apply the alternating direction method of multiplier (ADMM) to solve the problem. Both synthetic test and field application prove that our proposed method not only gets a good result with high signal-noise-ratio (SNR) but also costs much less time than the conventional method. This indicates that the interplay of framelet transform and$l_{p}$quasi-norm can do a good job in seismic data reconstruction.

Improving the quality of microseismic recordings is a critical step in the microseismic data processing. We introduce a wavelet-weighted online dictionary learning (WWODL) denoising strategy for noisy microseismic recordings. We develop an adaptive parameters estimation approach for tunable$Q$-factor wavelet transform (TQWT), which provides accurate periodic and nonstationary subband information from microseismic data for online dictionary learning (ODL). A sliding time window is employed to divide the obtained subbands into a series of patches of equal length, which are then assembled into a matrix and fed into the ODL. The subband kurtosis information is weighted to the constraint function of the ODL, further enhancing the sparse coding ability for each subband. Shortening the oscillation duration, an improved ODL is developed with a faster convergence speed in calculating sparse coefficients. Our results confirm that the WWODL can suppress high-frequency, low-frequency, and shared-bandwidth noises and has a minimal impact on the first arrival. The time consumption and the signal-to-noise ratio (SNR) of the WWODL are on average 1/8.675 and 56.82% higher than ODL, respectively.

Sparse low-rank denoising methods are widely applied for seismic random noise attenuation. Due to the poor structural sparsity and poor low rank of the traditional method, random noises, effective signal loss, and poor continuity still exist. To overcome these barriers, a multitrace seismic random noise simultaneous attenuation method in the time–frequency using the Lp-variation and${\gamma }$norm constraint is proposed. This approach uses Lp-variation regularization to describe the structural sparsity of seismic data in the time–frequency domain. The structural sparsity can obtain the structural similarity of adjacent traces in the time–frequency domain. This similarity can improve the continuity of events and can further suppress low-amplitude random noise. Besides, the approach utilizes the${\gamma }$norm to constrain the low rank of seismic data in the time–frequency domain. The${\gamma }$norm can obtain more low-rank information than the nuclear norm. More low-rank information can improve the overall suppression effect of random noise. The Lp-variation and${\gamma }$norm constraints are used to construct the objective function. The alternating direction method of multipliers algorithm, the difference of convex programming, and singular value decomposition are utilized to obtain the attenuation algorithm. Both synthetic and field data tests prove that the proposed method has better denoising and effective signal protection ability.

Seismic wave acquisition is usually disturbed by natural noise and instrument noise. As the seismic wave propagates, the filtering effect of the Earth and its various layers will result in energy attenuation and velocity dispersion; these phenomena weaken the seismic time series amplitudes and distort the seismic phase data. In traditional processing methods, noise removal precedes the data compensation process, which attempts to retrieve information that was originally lost due to signal attenuation. If denoising is performed, weak seismic signals are often removed, resulting in the loss of useful signal data that cannot be recovered. Here, we propose a sparse regularization strategy based on dictionary learning. This inversion method performs the denoising and data compensation tasks in parallel. By applying this method to both synthetic and real datasets, we demonstrate that this technique effectively compensates and denoises the seismic data and improves both the data resolution and the signal-to-noise ratio of the seismic records of interest.

Sparse acoustic impedance inversion methods are widely used in seismic processing and interpretation. Due to the insufficiency of the sparse representation and absence of time-frequency domain prior information in traditional sparse inversion methods, the inversion results have low accuracy in pinch-out points and thin layers. To overcome these barriers, a multi-trace acoustic impedance inversion method based on arctangent total variation and hybrid domain constraints is proposed. Different from the traditional L1 norm sparsity constraint, the arctangent total variation sparsity constraint can obtain sparser information, both in vertical and horizontal directions. This sparser information obtains more accurate boundaries and better continuity, which can reduce the error of inversion result in pinch-out points and thin layers. In addition, the hybrid domain constraint is utilized to introduce time-frequency domain prior information in the inversion. The time-frequency domain information can further improve the accuracy of the inversion. The alternating direction method of the multiplier algorithm, the Sylvester equation, and the proximity operator are utilized to solve the multi-constraint inversion problem. Numerous experimental results show that the proposed method can improve the accuracy of the inversion, especially in pinch-out points and thin layers.

Traditional seismic inversion has the problems of low lateral resolution and poor anti-noise performance. To address these challenges, the advantages of inversion methods in time and frequency domains are combined; the Lp norm that can preserve more sparsity information is introduced into seismic inversion. Meanwhile, multi-trace simultaneous inversion has become the mainstream of inversion methods, which can effectively improve the quality of pre-stack seismic inversion. However, traditional multi-trace simultaneous inversion methods have high computational costs and are difficult to obtain the 3D result. Ensuring spatial continuity of inversion results is a challenging task. To address this issue, a 3D global pre-stack inversion method in the time-frequency mixed domain is proposed. The proposed method adds a frequency domain fidelity constraint the objective function of traditional time domain inversion to improve the resolution of the inversion results. To improve the computational efficiency of multi-trace simultaneous inversion, the anisotropic total variation regularization method under Lp norm constraint is employed. In addition, the 2D Sylvester equation is extended to 3D space to enable 3D global inversion. The method not only improves computational efficiency, but also ensures spatial continuity of the results. The stability and feasibility of the proposed method have been demonstrated through its application on both the 3D overthrust model and field pre-stack seismic data.

Missing traces are a common problem in seismic data acquisition, which can affect the quality of subsequent processing and interpretation. Therefore, seismic data interpolation is an essential step to recover the missing information. Recently, deep learning has emerged as a powerful tool for seismic data interpolation, especially generative adversarial networks (GAN). GAN can generate realistic data by learning from existing samples. In this paper, we propose an improved GAN for seismic data interpolation. The generator is set as U-Net, which could extract more features from the input data via skip connections. For the discriminator, we add a spectral normalization layer to preserve the information content of the discriminator’s weights. The Wasserstein loss function is used to stabilize the training process. With those changes, the improved GAN outperforms the traditional GAN. Both synthetic and field data tests demonstrate its effectiveness. Our proposed network can intelligently interpolate seismic data with high signal-to-noise ratio and enhance the efficiency of seismic data processing and analysis.

Sparse impedance inversion is widely applied for hydrocarbon prediction. Due to the low sparseness of traditional sparse constraints, pseudolayer and low-resolution problems still exist. To overcome this barrier, an impedance inversion method based on the reweighted L1-norm sparse constraint is proposed. Different from the traditional L1-norm constraint, which considers the location information of impedance boundaries, the reweighted L1-norm uses the amplitude information of impedance boundaries. The amplitude information can improve the sparseness, which helps inversion obtain more precise boundaries of impedance and weaken the pseudolayer phenomenon. Both the reweighted L1-norm constraint and initial model constraint are used to construct the objective function. The alternating direction method of multipliers (ADMM) algorithm is utilized to obtain the inversion algorithm. Both synthetic and field data tests prove that the impedance boundary of the proposed method is more accurate, and the pseudolayer phenomenon is weakened.

During the secondary seismic exploration in the maturing oil-field, due to the complex ground conditions, well-pump noise strongly reduced the signal-to-noise ratio (SNR) and affected the subsequent application of seismic data. At present, deep learning is mainly used in suppress random noise and ground-roll noise. Due to the obvious difference between the features of well-pump noise with random noise and ground-roll, the existing deep learning seismic denoising method which only extracts the features on one scale is not suitable for the suppression of well-pump noise. Therefore, we adopted the coarse-to-fine denoise strategy and presented a new method using multi-layer generator network (MLGnet) to suppress well-pump coherent noise. In proposed method, the network mainly consists of multi-layer encoder-decoder architecture, which could combine different layer feature information to obtain accurate denoised results. Meanwhile, we split the noisy seismic data into multiple patches in each layer, which could effectively expand the receiving range of denoising network and extract more useful features from well-pump noise. In this way, the proposed network can effectively utilize the multi-scale semantic information to suppress well-pump noise from seismic data. Experimental results on synthetic data and field data illustrated that our approach could obtain high-quality denoised results and retain the valid data to the greatest extent, compared with DnCNN and GAN.

Identifying useful microseismic events is one of the key steps in monitoring tunnel rockbursts. Here, we propose a strong noise-tolerance deep learning (SNTDL) network for the automatic classification of noisy microseismic events. The training set, validation set, and test set of the SNTDL network consist of 27,989 unfiltered microseismic recordings. First, to comprehensively characterize the microseismic events, we extract 10 weakly correlated features of the microseismic recordings as the input of the SNTDL network. Then, the skip connection and concatenation structure are added to this network, which can further enhances its generalization ability. Additionally, the SNTDL, AlexNet, Inception, Visual Geometry Group, and ResNet are compared using the synthetic microseismic recordings with different signal-noise ratios. The results demonstrate that the SNTDL network has a higher accuracy and stronger noise-tolerance capability than the other approaches. Application to a dataset collected from a different construction environment confirms that the SNTDL network can still achieve an accurate classification result, which further verifies that the proposed network has a reliable generalization performance.

Seismic data interpolation and denoising play vital roles in obtaining complete and clean data in seismic data processing. Seismic data usually misses along various spatial axes and always mix with random noise. In order to obtain complete and clean seismic data, reconstruction technology can interpolate missing data and attenuate random noise. Nonsubsampled contourlet transform is an effective transform to obtain multi-scale and multi-direction sparse domain data for compression sensing interpolation and denoising. However, conventional iterative shrinkage/thresholding cannot handle ill-posed and ill-conditioned equations for solving linear inverse problem. We present a two-step iterative log thresholding method to overcome ill-posed and ill-conditioned problems and improve the convergence rate and solution accuracy, which can interpolate and denoise seismic data simultaneously in the nonsubsampled contourlet transform framework. First, we use nonsubsampled contourlet transform to convert the seismic missing data with random noise to sparse domain. Then, we apply two-step iterative log thresholding algorithm to interpolate and denoise data in sparse domain. The result of each iteration is based on the results of the previous two iterations, which can accelerate convergence rate. In addition, log thresholding can further improve convergence rate and solution accuracy. Finally, we use inverse nonsubsampled contourlet transform to obtain the interpolated and denoised seismic data. The new method can reconstruct the irregularly missing data and attenuate random noise to obtain complete and clean seismic data with high accuracy, which is crucial for seismic imaging and inversion. We demonstrate the applicability and effectiveness of this simultaneous interpolation and denoising technique with successful applications to both synthetic and field data examples.

The inversion of reservoir elastic parameters and fracture parameters is of great significance to oil and gas production. In shale reservoirs with inclined fractures, using the reflection coefficient equation of vertical fractures or horizontal fractures under the vertical transverse isotropy (VTI) background has certain limitations. For this reason, based on the linear-slip model, this article establishes the approximate stiffness matrix of the monoclinic medium with a set of inclined fractures under the background of VTI and combines the Born scattering theory to further derive the PP-wave linear reflection coefficient equation of the monoclinic medium. The equation includes parameters, such as Young’s modulus, Poisson’s ratio, density, fracture weaknesses, and bedding weaknesses. Then, a rock physics model that comprehensively considers horizontal bedding and inclined fractures is established, and the influence of the inclination of the fractures on the reflection coefficient is analyzed. Finally, taking the advantage of Bayesian theory, the fracture weakness inversion method based on the azimuth amplitude difference is established, and the inversion accuracy is improved by adding Cauchy constraints and smoothing model regularization terms. Then, take the fracture weaknesses as inputs to invert the parameters, such as Young’s modulus, Poisson’s ratio, density, and bedding weaknesses. The inversion results of theoretical data show that the method can invert the fracture information well, and it can also be effectively applied to low-to-medium signal-to-noise ratio (SNR) data. The inversion results of theoretical and actual data prove that the inversion equation derived in this article can be used to obtain more accurate elastic and fracture parameters in fractured shale reservoirs.