Seismic wave impedance inversion method and device and electronic equipment

Abstract

The invention provides a seismic wave impedance inversion method and device and electronic equipment, and relates to the technical field of seismic inversion, and the method comprises the steps: obtaining real seismic data of a to-be-processed work area, carrying out the discretization of the real seismic data, obtaining a plurality of pieces of single-channel seismic data, carrying out the processing of the plurality of pieces of single-channel seismic data through a target inversion model, and obtaining a plurality of pieces of single-channel seismic data; and obtaining a seismic wave impedance inversion result of the work area to be processed. Wherein the target inversion model is a model obtained by training the initial inversion model. Obviously, the method is substantially an inversion method based on data driving, so that dependence on an initial model is avoided. Besides, the encoding process of the encoder module and the decoding process of the decoder module in the target inversion model are both realized based on the mamba module, and the mamba module provides strong feature extraction and representation capabilities for the encoder, so that compared with the prior art, the method can ensure the accuracy of the seismic wave impedance inversion result.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic inversion, and particularly to a seismic wave impedance inversion method, device, and electronic device. Background Art

[0002] Seismic inversion requires extracting geophysical properties and spatial structures from seismic images. These inversions are crucial for evaluating geological features and reservoir properties, especially in subsurface environments. The P-wave impedance is a key elastic property reflecting reservoir characteristics and is an indispensable parameter in reservoir exploration and identification.

[0003] In the prior art, traditional inversion techniques based on models are often used, such as recursive inversion, trace integration, geostatistical inversion methods, etc. However, the above methods highly depend on the initial model and also require careful design of the regularization term. If the initial model is inaccurate, it may lead to deviation in the inversion results. In addition, problems such as frequency band mismatch, forward modeling approximation, and noise pollution make it difficult to obtain stable and high-resolution inversion results from field seismic data. In summary, the existing seismic wave impedance inversion methods have the technical problems of strong dependence on the initial model and poor accuracy of the inversion results. Summary of the Invention

[0004] The purpose of the present invention is to provide a seismic wave impedance inversion method, device, and electronic device to alleviate the technical problems of strong dependence on the initial model and poor accuracy of the inversion results existing in the existing seismic wave impedance inversion methods.

[0005] In a first aspect, the present invention provides a seismic wave impedance inversion method, including: obtaining real seismic data of a work area to be processed; discretizing the real seismic data to obtain a plurality of single-trace seismic data; processing the plurality of single-trace seismic data by using a target inversion model to obtain a seismic wave impedance inversion result of the work area to be processed; wherein, the target inversion model is a model obtained by training an initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

[0006] Optionally, the encoder module and the decoder module are the mamba encoder and the mamba decoder in the trained Masked Autoencoder (MAE); the training process of the MAE includes: obtaining a first training sample set; wherein, the first training sample set includes: a plurality of sample single-channel seismic data; performing a chunking process on the first target sample single-channel seismic data to obtain a discrete data volume; wherein, the first target sample single-channel seismic data represents any sample in the first training sample set; performing a random masking process on the discrete data volume to obtain a recovery index of the masked data, an unmasked data set, and a position marker of each unmasked data in the discrete data volume; using the mamba encoder to encode the unmasked data set to obtain latent features of the unmasked data set; constructing input data for the mamba decoder based on the latent features, the position marker of each unmasked data in the discrete data volume, and the recovery index of the masked data; using the mamba decoder to decode the input data to reconstruct the single-channel seismic data based on the decoding result to obtain a predicted value of the single-channel seismic data; calculating a first loss function value based on the first target sample single-channel seismic data and the predicted value of the single-channel seismic data; iteratively training the initial MAE based on the first loss function value until a preset end condition is reached to obtain the target MAE.

[0007] Optionally, calculating the first loss function value based on the first target sample single-channel seismic data and the predicted value of the single-channel seismic data includes: screening out corresponding predicted data from the predicted value of the single-channel seismic data based on the position of the masked data in the first target sample single-channel seismic data; calculating the Euclidean distance between the masked data and the predicted data; determining the Euclidean distance as the first loss function value.

[0008] Optionally, it further includes: obtaining a second training sample set; wherein, the second training sample set includes multiple groups of second training samples, and each group of second training samples includes: sample single-channel seismic data and corresponding true seismic wave impedance data; using the initial inversion model to process the second target sample single-channel seismic data to obtain corresponding predicted seismic wave impedance data; wherein, the second target sample single-channel seismic data represents any sample in the second training sample set; calculating a second loss function value based on the second target sample single-channel seismic data, the corresponding true seismic wave impedance data, and the predicted seismic wave impedance data; iteratively training the initial inversion model based on the second loss function value to continuously adjust the model parameters of the first projection module and the second projection module until a preset iteration termination condition is reached.

[0009] Optionally, based on the second target sample single-channel seismic data, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data, calculate the second loss function value, including: performing forward modeling on the seismic wave impedance prediction data using a preset forward modeling operator to obtain synthetic single-channel seismic data; calculating the residual between the synthetic single-channel seismic data and the second target sample single-channel seismic data to obtain a reconstruction loss; calculating the mean square error between the true seismic wave impedance data and the seismic wave impedance prediction data to obtain a well loss; and calculating the second loss function value based on the reconstruction loss and the well loss.

[0010] Optionally, the second loss function is expressed as: ; where both represent preset weight factors, represents the reconstruction loss, represents the well loss.

[0011] Optionally, a sparsity constraint is added to the hidden layer of the encoder module.

[0012] In a second aspect, the present invention provides a seismic wave impedance inversion device, including: a first acquisition module for acquiring true seismic data of a work area to be processed; a discretization module for discretizing the true seismic data to obtain a plurality of single-channel seismic data; and an inversion module for processing the plurality of single-channel seismic data using a target inversion model to obtain a seismic wave impedance inversion result of the work area to be processed; where the target inversion model is a model obtained by training an initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; and the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

[0013] In a third aspect, the present invention provides an electronic device, including a memory and a processor, where a computer program executable on the processor is stored on the memory, and when the processor executes the computer program, the seismic wave impedance inversion method according to any one of the foregoing embodiments is implemented.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium, where computer instructions are stored on the computer-readable storage medium, and when the computer instructions are executed by a processor, the seismic wave impedance inversion method according to any one of the foregoing embodiments is implemented.

[0015] The present invention provides a seismic wave impedance inversion method. After obtaining the real seismic data of the work area to be processed, the method first discretizes the real seismic data to obtain a plurality of single-channel seismic data, and then uses a target inversion model to process the plurality of single-channel seismic data to obtain the seismic wave impedance inversion result of the work area to be processed, wherein the target inversion model is a model obtained by training an initial inversion model. It can be seen that the essence of the present invention is a data-driven inversion method, which avoids the dependence on the initial model. In addition, the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module, and the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module. In view of the fact that the mamba module has higher efficiency and scalability compared with the traditional self-attention mechanism when processing long sequence data, and can also maintain good modeling ability, providing the encoder with strong feature extraction and representation ability. Therefore, compared with the prior art, the present invention can ensure the accuracy of the seismic wave impedance inversion result. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0017] Figure 1 It is a flowchart of a seismic wave impedance inversion method provided by an embodiment of the present invention; Figure 2 It is a network structure diagram of MAE provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the seismic wave impedance distribution of the Marmousi2 model; Figure 4 For Figure 3 It is a schematic diagram of the synthetic seismic data calculated; Figure 5 It is a schematic diagram of the prediction results of four seismic wave impedance prediction methods; Figure 6 It is a schematic diagram of the difference between the prediction results of four seismic wave impedance prediction methods for a single seismic trace and the true value; Figure 7 It is a functional module diagram of a seismic wave impedance inversion device provided by an embodiment of the present invention; Figure 8 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Components of the embodiments of the present invention generally described and illustrated in the figures herein can be arranged and designed in a variety of different configurations.

[0019] Therefore, the detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0020] The following will describe in detail some embodiments of the present invention with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0021] Embodiment 1 Figure 1 is a flowchart of a seismic wave impedance inversion method provided for an embodiment of the present invention. As Figure 1 shown, the method specifically includes the following steps: Step S102, obtain the real seismic data of the work area to be processed.

[0022] Step S104, discretize the real seismic data to obtain a plurality of single-channel seismic data.

[0023] Step S106, process the plurality of single-channel seismic data by using the target inversion model to obtain the seismic wave impedance inversion result of the work area to be processed.

[0024] Among them, the target inversion model is a model obtained by training the initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

[0025] Specifically, to avoid the dependence on the initial model in traditional model-driven inversion methods, the embodiments of the present invention apply a data-driven inversion method when performing seismic wave impedance inversion. That is, a target inversion model with single-channel seismic data as the input data and corresponding seismic wave impedance data as the output data is pre-trained. Considering that the input data of the target inversion model is single-channel seismic data and the single-channel seismic data is long-sequence data, to ensure the scalability of the model and the data processing efficiency, in the embodiments of the present invention, both the encoder module and the decoder module in the target inversion model are implemented based on the mamba module. The application of the mamba module can also avoid the problems of gradient disappearance and lack of effective parallelism in the target inversion model.

[0026] It is known that in a model with an encoder-decoder structure, the output of the encoder needs to be projected before being input into the decoder. Therefore, a first projection module is set between the encoder module and the decoder module to map the output of the encoder module to the input space of the decoder module. In addition, after the decoder module decodes and outputs, its output result should also be mapped to the seismic wave impedance domain to obtain the seismic wave impedance inversion result. Therefore, a second projection module needs to be set after the decoder module to achieve the mapping between the above data domains. In an optional implementation manner, both the first projection module and the second projection module are implemented using fully connected layers.

[0027] Since the input data of the target inversion model is single-channel seismic data, after obtaining the real seismic data of the work area to be processed, the real seismic data needs to be discretized. It is known that the real seismic data is a set of seismic data recorded by multiple seismic channels. Therefore, the discretization process is to segment the real seismic data in units of seismic channels to obtain multiple single-channel seismic data. Then, the target inversion model is used to process the multiple single-channel seismic data to obtain the corresponding seismic wave impedance inversion results.

[0028] An embodiment of the present invention provides a seismic wave impedance inversion method. After obtaining the real seismic data of the work area to be processed, the method first discretizes the real seismic data to obtain multiple single-channel seismic data, and then uses a target inversion model to process the multiple single-channel seismic data to obtain the seismic wave impedance inversion result of the work area to be processed. The target inversion model is a model obtained by training an initial inversion model. It can be seen that the essence of the present invention is a data-driven inversion method, which avoids the dependence on the initial model. In addition, the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module. The encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module. In view of the fact that the mamba module has higher efficiency and scalability compared with the traditional self-attention mechanism when processing long sequence data, and can also maintain good modeling ability, providing the encoder with strong feature extraction and representation ability. Therefore, compared with the prior art, the present invention can ensure the accuracy of the seismic wave impedance inversion result.

[0029] In an optional embodiment, the encoder module and the decoder module are the mamba encoder and the mamba decoder in the trained masked autoencoder (MAE).

[0030] The training process of the MAE includes the following steps: Step S201, obtain a first training sample set; wherein, the first training sample set includes: multiple sample single-channel seismic data.

[0031] Step S202, perform block processing on the first target sample single-channel seismic data to obtain a discrete data volume; wherein, the first target sample single-channel seismic data represents any sample in the first training sample set.

[0032] Step S203, perform random masking processing on the discrete data volume to obtain the recovery index of the masked data, the set of unmasked data, and the position markers of each unmasked data in the discrete data volume.

[0033] Step S204, use the mamba encoder to encode the set of unmasked data to obtain the latent features of the set of unmasked data.

[0034] Step S205, based on the latent features, the position markers of each unmasked data in the discrete data volume, and the recovery index of the masked data, construct the input data of the mamba decoder.

[0035] Step S206, use the mamba decoder to decode the input data to reconstruct the single-channel seismic data based on the decoding result to obtain the predicted value of the single-channel seismic data.

[0036] Step S207: Calculate the first loss function value based on the first target single-channel seismic data and the predicted value of the single-channel seismic data.

[0037] Step S208: Iteratively train the initial MAE based on the first loss function value until a preset end condition is reached to obtain the target MAE.

[0038] Generally, to obtain the target inversion model, the model parameters of the encoder module, the first projection module, the decoder module, and the second projection module in the initial inversion model can be initialized first, and then a large number of samples are used for training to iteratively update the model parameters of the above four modules. The target inversion model obtained in this way can realize the inversion from single-channel seismic data to seismic wave impedance data. In the embodiments of the present invention, in order to enhance the understanding ability of the target inversion model for seismic signals and improve the inversion accuracy, different from the traditional model construction and training process, before constructing the initial inversion model, a masked autoencoder MAE using a mamba encoder and a mamba decoder needs to be pre-trained first. After the MAE reaches the preset training target, the trained mamba encoder and mamba decoder are used as the encoder module and the decoder module in the initial inversion model respectively, and then a large number of samples are used for training to continuously optimize the model parameters of the first projection module and the second projection module, thereby obtaining the target inversion model.

[0039] Multiple scattering and complex wavefield interference in seismic signals can cause spatial ambiguity, resulting in a large amount of redundant information in seismic data. This process is similar to the image blurring and information redundancy phenomena caused by light scattering and reflection in natural images. Due to this spatial redundancy, even if most of the seismic data is masked, the original data can still be recovered by the MAE.

[0040] In an optional implementation manner, a sparsity constraint is added to the hidden layer of the encoder module, that is, only part of the data is allowed to participate in the calculation. This mechanism gives priority to the most informative part of the input data while ignoring less important details, reducing redundant information, and thus realizing a more efficient and more concentrated encoding process. This selective activation ensures that only the most relevant features are processed, further improving the efficiency and interpretability of the model.

[0041] Generally, the MAE uses the ratio of the embedding dimension to the patch size to maintain a size similar to the original data. In the embodiments of the present invention, a higher ratio of the embedding dimension to the patch size is adopted to greatly increase the information capacity and expand the seismic representation space. Figure 2 For the network structure diagram of the MAE provided by the embodiments of the present invention, refer to Figure 2, when training the MAE, first, the single-channel seismic data of the first target sample needs to be segmented, that is, it is discretized into multiple amplitude data to obtain a discrete data volume. Next, a random masking operation is performed on the discrete data volume to mask some of the data. Thus, an unmasked data set, the position markers of each unmasked data in the discrete data volume, and the recovery index of the masked data are obtained. The recovery index of the masked data includes: the position of each masked data and the value after masking.

[0042] Next, the mamba encoder encodes the unmasked data set to capture and represent the complex features and structures inherent in the seismic data it processes, that is, to obtain its latent features. Then, by combining the position markers of each unmasked data in the discrete data volume and the recovery index of the masked data, the input data of the mamba decoder can be constructed. The mamba decoder decodes the data it receives and then maps the decoding result to the seismic data domain according to the requirements of the MAE output data to obtain the predicted value of the single-channel seismic data, that is, the reconstruction of the single-channel seismic data is completed.

[0043] To quantify the training effect, after obtaining the predicted value of the single-channel seismic data, according to the predicted value of the single-channel seismic data and its corresponding single-channel seismic data of the first target sample, the value of the first loss function is calculated. Furthermore, the initial MAE is iteratively trained using the value of the first loss function until a preset end condition is reached, and then the target MAE can be obtained. The embodiments of the present invention do not specifically limit the preset end condition, which can be that the training reaches a specified number of rounds, or that the value of the first loss function converges below a preset threshold.

[0044] In an optional implementation manner, in the above step S207, calculating the value of the first loss function based on the single-channel seismic data of the first target sample and the predicted value of the single-channel seismic data specifically includes the following steps: Step S2071, screening out the corresponding predicted data from the predicted value of the single-channel seismic data based on the positions of the masked data in the single-channel seismic data of the first target sample.

[0045] Step S2072, calculating the Euclidean distance between the masked data and the predicted data.

[0046] Step S2073, determining the Euclidean distance as the value of the first loss function.

[0047] Specifically, in order to enable MAE to focus on learning how to recover masked data, it is different from the calculation method of traditional loss functions: calculating the difference between the single-channel seismic data of the first target sample and the predicted value of the single-channel seismic data. In the embodiment of the present invention, when calculating the value of the first loss function, only the difference between the masked data in the single-channel seismic data of the first target sample and the corresponding predicted data in the predicted value of the single-channel seismic data is considered. Specifically, the Euclidean distance between the masked data and the predicted data is used as the value of the first loss function.

[0048] In an optional embodiment, the embodiment of the present invention further includes the following steps: Step S301, obtaining a second training sample set; wherein, the second training sample set includes multiple groups of second training samples, and each group of second training samples includes: sample single-channel seismic data and corresponding true seismic wave impedance data.

[0049] Step S302, processing the single-channel seismic data of the second target sample by using the initial inversion model to obtain corresponding seismic wave impedance prediction data; wherein, the single-channel seismic data of the second target sample represents any sample in the second training sample set.

[0050] Step S303, calculating the value of the second loss function based on the single-channel seismic data of the second target sample, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data.

[0051] Step S304, iteratively training the initial inversion model based on the value of the second loss function to continuously adjust the model parameters of the first projection module and the second projection module until a preset iteration termination condition is reached.

[0052] As can be seen from the above description, the model parameters of the encoder module and the decoder module in the initial inversion model have been fixed, that is, the model parameters of the two are no longer adjusted. When training the initial inversion model, the parameters to be optimized and adjusted only include: the model parameters of the first projection module and the model parameters of the second projection module.

[0053] During the model training process, when the single-channel seismic data of the second target sample is input into the initial inversion model, the corresponding seismic wave impedance prediction data can be output. In order to improve the credibility and accuracy of the model prediction results, in the embodiment of the present invention, when calculating the value of the second loss function, not only the prediction accuracy of the seismic wave impedance is considered, but also the synthetic seismic data is calculated by using the seismic wave impedance prediction data, and further the difference between the synthetic seismic data and the corresponding single-channel seismic data of the second target sample is considered. Then, the initial inversion model is iteratively trained by using the value of the second loss function until a preset iteration termination condition is reached. Optionally, the preset iteration termination condition is that the correlation coefficient between the synthetic seismic data and the corresponding single-channel seismic data of the second target sample converges to a specified threshold or above.

[0054] In an alternative embodiment, in step S303, based on the second target sample single-channel seismic data, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data, the second loss function value is calculated, which specifically includes the following steps: In step S3031, the seismic wave impedance prediction data is forward-processed using a preset forward operator to obtain synthetic single-channel seismic data.

[0055] In step S3032, the residual between the synthetic single-channel seismic data and the second target sample single-channel seismic data is calculated to obtain a reconstruction loss.

[0056] In step S3033, the mean square error between the true seismic wave impedance data and the seismic wave impedance prediction data is calculated to obtain a well loss.

[0057] In step S3034, based on the reconstruction loss and the well loss, the second loss function value is calculated.

[0058] In the embodiment of the present invention, the arithmetic formula for the synthetic single-channel seismic data is: , where represents the synthetic single-channel seismic data, represents the wavelet matrix, represents the differential matrix, represents the column vector composed of the seismic wave impedance prediction data, that is, the preset forward operator is .

[0059] In an alternative embodiment, the second loss function is expressed as: ; where both represent preset weight factors, represents the reconstruction loss, represents the well loss.

[0060] In summary, the embodiment of the present invention is a seismic wave impedance inversion method combining MAE and the mamba module. The MAE component adopts an asymmetric encoder-decoder architecture. The encoder processes the visible subset of the data block, and the lightweight decoder reconstructs the original input from the latent representation. This design minimizes the computational complexity and optimizes the accuracy. To effectively capture global features, the target inversion model adopts the mamba encoder and mamba decoder in MAE, and uses its self-attention mechanism to identify the global correlations in the seismic data. This combined method enables the embodiment of the present invention to have higher accuracy and computational efficiency in the seismic wave impedance inversion task.

[0061] The present invention uses the open-source data of the Marmousi2 model for experimental verification. Figure 3 is a schematic diagram of the seismic wave impedance distribution of the Marmousi2 model, as shown in Figure 3As shown, in the Marmousi2 model, the impedance values show significant changes around the fault. The Marmousi2 model is specifically designed to replicate the geological environment related to continental drift by combining various typical geological features, such as the temporal and spatial dimensions and the significant velocity changes in the fault structure. The dataset contains 13,601 record traces, each record trace consisting of 2,801 sampling points with a time interval of 1 ms. This model is used to calculate the reflection coefficient, which is convolved with a 30 Hz Ricker wavelet to generate synthetic seismic data as shown in Figure 4 shown.

[0062] In this model, 5,000 record traces are randomly selected as sample data, with 90% assigned to the training set and 10% assigned to the validation set. The validation set is used to evaluate the network performance every 100 time steps, facilitating the adjustment of model parameters to minimize the validation loss. The network parameters are updated using the Adam optimizer, with the weight decay set to 10^-8 and the learning rate to 0.01. Training is carried out using the PyTorch framework. For comparison, three widely used deep learning methods, RNN, LSTM, and Transformer, are evaluated. To ensure the consistency of the comparative experiments, the parameters and training processes of all models are similar.

[0063] Figure 5 The prediction results of four seismic wave impedance prediction methods are shown. Figure 5 The (a)–(d) views in Figure 5 show the seismic wave impedance prediction results of RNN, LSTM, Transformer, and MAE-Mamba (the method provided by the embodiment of the present invention), respectively. The RNN model shows the overall structural characteristics of the lithologic layers, but there are significant fluctuations in some specific areas. The LSTM model effectively retains the long-term memory of the sequence during the cyclic process through the gating mechanism, showing better longitudinal continuity. However, due to the complexity of the model, it is prone to deformation in complex lithologic environments, resulting in unstable results. The Transformer can better capture long-range dependencies through its self-attention mechanism, thus obtaining better global performance when processing long sequences. Its inversion results are relatively consistent with the actual impedance, but there are still deficiencies in some specific features. MAE-Mamba effectively combines the advantages of RNN and Transformer, thereby improving the overall prediction accuracy and superior horizontal continuity, especially in areas with complex structures. Figure 5 As can be seen from

[0064] To more thoroughly examine the prediction performance, the 8500th trace was extracted from the impedance profile. This seismic trace is located in a complex lithology environment of the Marmousi2 model. Figure 6 It is a schematic diagram of the difference between the prediction results of four seismic wave impedance prediction methods for a single seismic trace and the true value. By Figure 6 It can be seen that the method of the present invention has a good identification effect in the lithology change area, and the fluctuation range meets the expectations. It ensures a higher consistency with the ground truth impedance. In addition, the method of the present invention shows obvious advantages in global inversion, and the overall trend of the curve is clear.

[0065] The following Table 1 summarizes the time required for training and prediction of different method networks, indicating that MAE-Mamba has higher computational efficiency compared to other methods involved in the experiment. Table 2 gives the R 2 between the prediction results and the ground truth of the experiments conducted, Pearson correlation coefficient (PCC), and mean square error (MSE). According to the data in Table 2, the MAE-Mamba method is superior to other methods.

[0066] Table 1 Time required for training and prediction using Marmousi2

[0067] Table 2 R 2 between the impedance prediction value and the true value using Marmousi2, PCC, and MSE

[0068] In addition, a series of ablation experiments were conducted to verify the effectiveness of the method of the present invention. The experiments included disabling and adjusting the parameters of the MAE component to evaluate their impact on the Mamba block and the Transformer block. The performance metrics included Pearson correlation coefficient (PCC), R-squared value, and mean square error (MSE). The following Table 3 shows the performance metric results of different experiments, indicating that the model achieves the best results when the masking ratio is 0.75. In addition, the comparative analysis with the model excluding the MAE component provides additional evidence to support the effectiveness of the method proposed in the embodiments of the present invention.

[0069] Table 3 R 2 between the impedance prediction value and the true value of different ablation experiments, PCC, and MSE

[0070] Embodiment 2 The embodiment of the present invention also provides a seismic wave impedance inversion device, which is mainly used to execute the seismic wave impedance inversion method provided in the above Embodiment 1. The following is a specific introduction to the seismic wave impedance inversion device provided by the embodiment of the present invention.

[0071] Figure 7 This is a functional module diagram of a seismic wave impedance inversion device provided by an embodiment of the present invention. As Figure 7 shown, the device mainly includes: a first acquisition module 10, a discretization module 20, and an inversion module 30, where: The first acquisition module 10 is used to acquire the real seismic data of the work area to be processed.

[0072] The discretization module 20 is used to discretize the real seismic data to obtain a plurality of single-trace seismic data.

[0073] The inversion module 30 is used to process the plurality of single-trace seismic data by using a target inversion model to obtain the seismic wave impedance inversion result of the work area to be processed.

[0074] Among them, the target inversion model is a model obtained by training an initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

[0075] An embodiment of the present invention provides a seismic wave impedance inversion device. After the device acquires the real seismic data of the work area to be processed, it first discretizes the real seismic data to obtain a plurality of single-trace seismic data, and then uses the target inversion model to process the plurality of single-trace seismic data to obtain the seismic wave impedance inversion result of the work area to be processed, where the target inversion model is a model obtained by training an initial inversion model. It can be seen that the device of the present invention actually executes a data-driven inversion method, avoiding the dependence on the initial model. In addition, the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module, and the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module. In view of the fact that the mamba module has higher efficiency and scalability compared with the traditional self-attention mechanism when processing long sequence data, and can also maintain good modeling ability, providing strong feature extraction and representation ability for the encoder. Therefore, compared with the prior art, the embodiment of the present invention can ensure the accuracy of the seismic wave impedance inversion result.

[0076] Optionally, the encoder module and the decoder module are the mamba encoder and the mamba decoder in a trained masked autoencoder MAE; the device further includes: The second acquisition module is used to acquire the first training sample set; wherein, the first training sample set includes: a plurality of sample single-channel seismic data.

[0077] The chunking module is used to chunk the first target sample single-channel seismic data to obtain a discrete data volume; wherein, the first target sample single-channel seismic data represents any sample in the first training sample set.

[0078] The masking module is used to perform random masking on the discrete data volume to obtain the recovery index of the masked data, the unmasked data set, and the position markers of each unmasked data in the discrete data volume.

[0079] The encoding module is used to encode the unmasked data set using a mamba encoder to obtain the latent features of the unmasked data set.

[0080] The construction module is used to construct the input data of the mamba decoder based on the latent features, the position markers of each unmasked data in the discrete data volume, and the recovery index of the masked data.

[0081] The decoding and reconstruction module is used to decode the input data using a mamba decoder to reconstruct the single-channel seismic data based on the decoding result to obtain the predicted value of the single-channel seismic data.

[0082] The first calculation module is used to calculate the value of the first loss function based on the first target sample single-channel seismic data and the predicted value of the single-channel seismic data.

[0083] The first training module is used to iteratively train the initial MAE based on the value of the first loss function until a preset end condition is reached to obtain the target MAE.

[0084] Optionally, the first calculation module is specifically used for: Screen out the corresponding predicted data from the predicted value of the single-channel seismic data based on the positions of the masked data in the first target sample single-channel seismic data.

[0085] Calculate the Euclidean distance between the masked data and the predicted data.

[0086] Determine the Euclidean distance as the value of the first loss function.

[0087] Optionally, the device further includes: The third acquisition module is used to acquire the second training sample set; wherein, the second training sample set includes multiple groups of second training samples, and each group of second training samples includes: sample single-channel seismic data and corresponding true seismic wave impedance data.

[0088] A processing module, configured to process the single-channel seismic data of the second target sample by using an initial inversion model to obtain corresponding seismic wave impedance prediction data; wherein, the single-channel seismic data of the second target sample represents any sample in the second training sample set.

[0089] A second calculation module, configured to calculate a second loss function value based on the single-channel seismic data of the second target sample, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data.

[0090] A second training module, configured to iteratively train the initial inversion model based on the second loss function value to continuously adjust the model parameters of the first projection module and the second projection module until a preset iteration termination condition is reached.

[0091] Optionally, the second calculation module is specifically configured to: Perform forward modeling on the seismic wave impedance prediction data by using a preset forward operator to obtain synthetic single-channel seismic data.

[0092] Calculate the residual between the synthetic single-channel seismic data and the single-channel seismic data of the second target sample to obtain a reconstruction loss.

[0093] Calculate the mean square error between the true seismic wave impedance data and the seismic wave impedance prediction data to obtain a well loss.

[0094] Calculate the second loss function value based on the reconstruction loss and the well loss.

[0095] Optionally, the second loss function is expressed as: ; wherein, both represent preset weight factors, represents the reconstruction loss, represents the well loss.

[0096] Optionally, a sparsity constraint is added to the hidden layer of the encoder module.

[0097] Embodiment 3 Refer to Figure 8 , an embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62, and a communication interface 63, and the processor 60, the communication interface 63, and the memory 61 are connected through the bus 62; the processor 60 is configured to execute an executable module stored in the memory 61, such as a computer program.

[0098] Among them, the memory 61 may include high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk memory. The communication connection between this system network element and at least one other network element is realized through at least one communication interface 63 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0099] The bus 62 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of easy representation, Figure 8 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0100] Among them, the memory 61 is used to store programs. After receiving an execution instruction, the processor 60 executes the program. The method executed by the device defined by any embodiment of the foregoing embodiments of the present invention can be applied to or implemented by the processor 60.

[0101] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor 60 or the instructions in the form of software. The above-mentioned processor 60 can be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it can also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by the combination of the hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, etc. This storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and combines its hardware to complete the steps of the above method.

[0102] A computer program product for an earthquake wave impedance inversion method, device and electronic device provided by an embodiment of the present invention includes a computer-readable storage medium storing non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For specific implementation, reference can be made to the method embodiments and will not be elaborated herein.

[0103] In addition, in each embodiment of the present invention, each functional unit may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0104] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk, or an optical disc that can store program code.

[0105] It should be noted that similar reference numerals and letters indicate similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0106] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the present invention is usually placed during use. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0107] In addition, terms such as "horizontal", "vertical", "hanging", etc. do not mean that the components are required to be absolutely horizontal or hanging, but can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but can be slightly inclined.

[0108] In the description of the present invention, it should also be noted that unless otherwise clearly specified and defined, the terms "set", "install", "connected", "connected to" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0109] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An earthquake wave impedance inversion method, characterized in that, Including: Obtain the real seismic data of the work area to be processed; Discretize the real seismic data to obtain multiple single-trace seismic data; Process the multiple single-trace seismic data by using the target inversion model to obtain the seismic wave impedance inversion result of the work area to be processed; Wherein, the target inversion model is a model obtained by training an initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

2. The seismic wave impedance inversion method according to claim 1, characterized in that The encoder module and the decoder module are the mamba encoder and mamba decoder in the trained masked autoencoder (MAE); The training process of the MAE includes: Obtain a first training sample set; wherein, the first training sample set includes: multiple sample single-trace seismic data; Perform block processing on the first target sample single-trace seismic data to obtain a discrete data volume; wherein, the first target sample single-trace seismic data represents any sample in the first training sample set; Perform random masking processing on the discrete data volume to obtain the recovery index of the masked data, the set of unmasked data, and the position markers of each unmasked data in the discrete data volume; Use the mamba encoder to encode the set of unmasked data to obtain the latent features of the set of unmasked data; Based on the latent features, the position markers of each unmasked data in the discrete data volume, and the recovery index of the masked data, construct the input data of the mamba decoder; Use the mamba decoder to decode the input data to reconstruct the single-trace seismic data based on the decoding result, and obtain the predicted value of the single-trace seismic data; Based on the first target sample single-trace seismic data and the predicted value of the single-trace seismic data, calculate the value of the first loss function; Perform iterative training on the initial MAE based on the value of the first loss function until a preset end condition is reached to obtain the target MAE.

3. The seismic wave impedance inversion method according to claim 2, wherein Based on the first target sample single-trace seismic data and the predicted value of the single-trace seismic data, calculating the value of the first loss function includes: Screen out the corresponding predicted data from the predicted value of the single-trace seismic data based on the position of the masked data in the first target sample single-trace seismic data; Calculate the Euclidean distance between the masked data and the predicted data; Determine the Euclidean distance as the value of the first loss function.

4. The seismic wave impedance inversion method according to claim 1, wherein It further includes: Obtain a second training sample set; wherein, the second training sample set includes multiple groups of second training samples, and each group of the second training samples includes: sample single-trace seismic data and corresponding real seismic wave impedance data; Process the second target sample single-trace seismic data by using the initial inversion model to obtain the corresponding predicted seismic wave impedance data; wherein, the second target sample single-trace seismic data represents any sample in the second training sample set; Calculate a second loss function value based on the second target sample single-channel seismic data, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data; Iteratively train the initial inversion model based on the second loss function value to continuously adjust the model parameters of the first projection module and the second projection module until a preset iteration termination condition is reached.

5. The seismic wave impedance inversion method according to claim 4, characterized in that Calculating a second loss function value based on the second target sample single-channel seismic data, the corresponding true seismic wave impedance data, and the seismic wave impedance prediction data includes: Performing forward modeling on the seismic wave impedance prediction data using a preset forward operator to obtain synthetic single-channel seismic data; Calculating the residual between the synthetic single-channel seismic data and the second target sample single-channel seismic data to obtain a reconstruction loss; Calculating the mean square error between the true seismic wave impedance data and the seismic wave impedance prediction data to obtain a well loss; Calculating the second loss function value based on the reconstruction loss and the well loss.

6. The seismic wave impedance inversion method according to claim 5, wherein The second loss function is expressed as: ; where both represent preset weight factors, represents the reconstruction loss, represents the well loss.

7. The seismic wave impedance inversion method according to claim 1, characterized in that Add a sparsity constraint to the hidden layer of the encoder module.

8. An earthquake wave impedance inversion device, characterized in that Including: A first acquisition module for acquiring true seismic data of a work area to be processed; A discretization module for discretizing the true seismic data to obtain a plurality of single-channel seismic data; An inversion module for processing the plurality of single-channel seismic data using a target inversion model to obtain a seismic wave impedance inversion result of the work area to be processed; Wherein, the target inversion model is a model obtained by training an initial inversion model; the target inversion model includes: an encoder module, a first projection module, a decoder module, and a second projection module; the encoding process of the encoder module and the decoding process of the decoder module are both implemented based on the mamba module; the first projection module is used to map the output of the encoder module to the input space of the decoder module; the second projection module is used to map the output of the decoder module to the seismic wave impedance domain.

9. An electronic device, comprising a memory and a processor, wherein a computer program capable of running on the processor is stored on the memory, characterized in that, When the processor executes the computer program, it implements the seismic wave impedance inversion method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by the processor, they implement the seismic wave impedance inversion method according to any one of claims 1 to 7.

Images