Multi-component seismic data fusion and interpretation method

Abstract

The present application provides a multi-component seismic data fusion and interpretation method, including: obtaining original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume; performing multi-physical feature fusion processing on the preprocessed data volume to obtain a fused feature field; performing joint inversion processing of elastic parameters on the fused feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density; performing knowledge-graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, and the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithological boundary topological structure. Using this method can comprehensively utilize various seismic data information and improve the accuracy of geological interpretation under complex geological conditions.

Description

Technical Field

[0001] The present invention belongs to the field of geoinformation science, and particularly relates to a method for multi-component seismic data fusion and interpretation. Background Art

[0002] Traditional methods for interpreting seismic data mainly rely on single longitudinal wave (P-wave) data and are difficult to distinguish the anisotropic characteristics of complex geological bodies. For example, the geological structure in Shandong region is complex, with problems such as multi-phase fault systems, subtle reservoirs, and volcanic rock shielding. The existing technologies usually have the following problems in multi-component data fusion: on the one hand, the spatial registration error between shear wave (S-wave) and converted wave (PS-wave) data is large, resulting in low accuracy of the geographical basis; on the other hand, the multi-component feature fusion relies on artificial experience and lacks physical constraints, and the resolution of deep structure inversion is insufficient, affecting the accuracy of geological interpretation. Summary of the Invention

[0003] Based on this, it is necessary to provide a method for multi-component seismic data fusion and interpretation for the above technical problems, which can comprehensively utilize various seismic data information and improve the accuracy of geological interpretation under complex geological conditions.

[0004] In a first aspect, the present application provides a method for multi-component seismic data fusion and interpretation, including:

[0005] Obtaining original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume;

[0006] Performing multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field;

[0007] Performing elastic parameter joint inversion processing on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density;

[0008] Performing knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, where the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithological boundary topological structure.

[0009] Further, obtaining original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume includes:

[0010] Performing quantum computing optimization processing on the original multi-component seismic data to obtain sub-wavelength-aligned multi-component data, where the quantum computing optimization processing includes encoding spatial displacement parameters as quantum bit states and performing parameter search through quantum gate operations;

[0011] Performing polarization direction tensor decomposition processing on the S-wave data in the multi-component data to obtain S-wave data with eliminated polarization distortion;

[0012] Perform matched filtering on the PS wave data in the multi-component data to obtain the PS wave data after azimuth difference correction;

[0013] Perform spatio-temporal domain weighted fusion on the multi-component data, S wave data, and PS wave data to obtain a preprocessed data volume.

[0014] Furthermore, perform multi-physical feature fusion on the preprocessed data volume to obtain a fusion feature field, including:

[0015] Perform tensor integration on the P-wave amplitude envelope, S-wave polarization ellipticity, PS-wave phase gradient, and acoustic impedance parameters in the preprocessed data volume to obtain a four-dimensional feature tensor;

[0016] Perform cross-component attention weight calculation on the four-dimensional feature tensor to obtain a correlation weight matrix of P-wave and S-wave features;

[0017] Perform multi-scale convolutional neural network processing on the correlation weight matrix to obtain a fusion feature field with optimized resolution.

[0018] Furthermore, perform elastic parameter joint inversion on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field, including:

[0019] Construct an elastic wave full waveform inversion objective function based on the fusion feature field, and the objective function includes a data fitting term and a geological prior constraint term;

[0020] Perform alternating direction multiplier method solution on the objective function to obtain a three-dimensional physical property parameter field of P-wave velocity field, S-wave velocity field, and density field;

[0021] Perform shear wave splitting parameter calculation on the shear wave velocity field to obtain a fracture density field, and the fracture density field is used to correct the anisotropic characteristics in the three-dimensional physical property parameter field, where the calculation formula of the shear wave splitting parameter is as follows:

[0022] ;

[0023] where, is the shear wave splitting parameter, is the fast shear wave velocity, is the slow shear wave velocity, is the background shear wave velocity.

[0024] Furthermore, perform knowledge graph-driven semantic modeling on the three-dimensional physical property parameter field to obtain a geological interpretation model, including:

[0025] Use the following formula to perform radial basis function interpolation on the three-dimensional physical property parameter field to generate an implicit geological interface function:

[0026] ;

[0027] wherein, is an implicit geological interface function, is a weight coefficient, is the influence radius, is the central coordinate of the radial basis function;

[0028] Perform level set extraction processing on the implicit geological interface function to generate an initial geological interface model;

[0029] Perform confidence weighted correction processing with drilling data constraints on the initial geological interface model to generate a geological interpretation model after error correction.

[0030] Furthermore, perform quantum computing optimization processing on the original multi-component seismic data to obtain sub-wavelength aligned multi-component data, including:

[0031] Perform quantum bit superposition state mapping processing on the spatial displacement parameter to generate an initial quantum state;

[0032] Perform Hamiltonian-driven quantum evolution processing on the initial quantum state, where the Hamiltonian includes a data alignment error term and a spatial smoothing regularization term;

[0033] Perform projection measurement processing on the evolved quantum state to obtain an optimal spatial displacement parameter, and the optimal spatial displacement parameter is used to perform spatial displacement correction on the original multi-component seismic data to generate multi-component data.

[0034] Furthermore, performing confidence weighted correction processing with drilling data constraints on the initial geological interface model to generate a geological interpretation model after error correction includes:

[0035] Perform confidence calculation processing on the fault of the geological interpretation model after error correction based on fuzzy Petri net rules, specifically, determine the fault confidence based on the weighted minimum of the longitudinal wave velocity gradient confidence and the shear wave splitting parameter confidence;

[0036] Perform topological consistency verification processing on the geological interpretation model after error correction, and analyze the spatial geometric relationship between the formation dip angle and the fault strike through the verification to generate a final model that conforms to the formation contact relationship;

[0037] When the interface position error of the final model exceeds the preset threshold, trigger the adversarial generation network reconstruction processing, and the reconstruction processing generates a geological interface model consistent with the three-dimensional physical property parameter field gradient distribution.

[0038] In a second aspect, the present application also provides a multi-component seismic data fusion and interpretation device, characterized in that the device includes:

[0039] A data preprocessing module, configured to obtain original multi-component seismic data and perform spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume;

[0040] A data fusion module, configured to perform multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field;

[0041] A three-dimensional parameter generation module, configured to perform joint inversion processing of elastic parameters on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density;

[0042] A geological interpretation module, configured to perform knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, where the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithologic boundary topological structure.

[0043] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the above multi-component seismic data fusion and interpretation method is implemented.

[0044] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above multi-component seismic data fusion and interpretation method is implemented.

[0045] The above multi-component seismic data fusion and interpretation method performs spatial alignment and signal correction on the original multi-component seismic data through spatio-temporal registration and polarization correction processing, eliminates the phase difference and polarization distortion between different component data, and forms a preprocessed data volume with spatio-temporal consistency; based on multi-physical feature fusion processing, cross-component feature extraction and non-linear interaction are performed on the corrected P-wave, S-wave, and PS-wave data to construct a fusion feature field including multi-dimensional attributes such as amplitude, polarization, and phase, breaking through the representation limitation of a single data attribute; through joint inversion processing of elastic parameters, the three-dimensional distribution fields of longitudinal wave velocity, transverse wave velocity, and density parameters are synchronously solved to achieve coupled constraint inversion of multi-physical fields; and combined with knowledge graph-driven semantic modeling processing, the geological rule base is deeply fused with the gradient features of the physical property parameter field to generate a three-dimensional topological model including a fault system, a stratigraphic interface, and a lithologic boundary. This method improves the structural recognition accuracy and reservoir prediction reliability of complex geological bodies such as continental faulted basins through a cross-component data collaboration mechanism and a geological knowledge embedding strategy, and improves the accuracy of geological interpretation under complex geological conditions. Description of the Drawings

[0046] To more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the accompanying drawings required for the description of the embodiments or related technologies. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0047] Figure 1 It is a flowchart of a multi-component seismic data fusion and interpretation method provided by an embodiment of the present invention.

[0048] Figure 2 It is a schematic structural diagram of an interactive advertising system based on multi-modal perception and AI dynamic generation provided by an embodiment of the present invention. Detailed implementation manners

[0049] In order to make the purpose, technical solutions and advantages of the present application clearer, the following further details the present application in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] First, a brief introduction is made to the nouns involved in the embodiments of the present application.

[0051] Seismic data refers to various information related to earthquakes collected by seismic monitoring devices such as seismographs, including the waveforms, amplitudes, frequencies, propagation times, etc. of seismic waves. These data record the vibration conditions of the earth's internal medium during an earthquake, reflecting the characteristics of the earthquake's intensity, location, origin time, etc. It is an important basis for studying the laws of earthquake activities, the earth's internal structure, and for earthquake early warning, disaster assessment and other work.

[0052] A knowledge graph is a graph-based data structure composed of nodes (entities) and edges (relationships), used to describe concepts, entities and their mutual relationships in the objective world in a structured manner. It integrates and correlates knowledge in various fields, connecting the originally scattered information into an organic knowledge network. Through the knowledge graph, a computer can understand and process human knowledge, realizing various applications such as intelligent search, knowledge answering, recommendation systems, etc., providing people with more accurate and comprehensive information services, and helping people better understand and utilize knowledge.

[0053] Quantum computing optimization is a method that utilizes the principles of quantum mechanics to process and optimize problems. In the processing of original multi-component seismic data, it encodes spatial displacement parameters into qubit states, and by leveraging the properties of qubit superposition and entanglement, performs efficient searches on these parameters through quantum gate operations to find the optimal solution or an approximate optimal solution, thereby achieving the optimized processing of multi-component seismic data and obtaining multi-component data aligned at the sub-wavelength level. Compared with traditional computing methods, quantum computing optimization can process complex computing tasks more quickly and accurately.

[0054] Based on the above noun explanations, the implementation environment of the multi-component seismic data fusion and interpretation method provided by the embodiments of this application is described. Schematically, this implementation environment includes: a processor and a terminal. Among them, the processor and the terminal are connected through a network signal; the terminal is equipped with a data collector to obtain original multi-component seismic data; the processor includes but is not limited to a central processing unit (CPU), a graphics processing unit (GPU), an artificial intelligence chip, etc., which are not limited here.

[0055] Combined with the above noun explanations and implementation environment, the application scenarios of the embodiments of this application are described. The multi-component seismic data fusion and interpretation method provided in the embodiments of this application can be applied to the following scenarios including but not limited to:

[0056] Under complex fault geological conditions, such as in continental fault-depressed basins, due to the multi-stage development of the fault system in such areas, it is difficult for traditional methods to accurately identify small faults and fractures. Multi-component data fusion can combine the amplitude information of P-waves and the polarization characteristics of S-waves to improve the accuracy of depicting the spatial distribution of faults.

[0057] In the exploration of hidden reservoirs in volcanic rock shield areas, through the joint analysis of the phase characteristics of PS converted waves and the longitudinal wave attenuation attributes, it is possible to penetrate the high-speed volcanic rock cover layer and identify the top and bottom interfaces of the sand bodies below, improving the prediction accuracy of reservoir thickness in low signal-to-noise ratio areas.

[0058] In the exploration of salt dome structures, by using the tomography technology jointly constrained by multi-component velocity fields, it is possible to effectively solve the imaging distortion problem caused by sudden velocity changes at the salt body boundary and achieve the accurate restoration of the subsalt structure.

[0059] Schematically, the multi-component seismic data fusion and interpretation method provided by the embodiments of this application can also be applied to other application scenarios. Only examples are given here and the specific application scenarios are not limited.

[0060] In an exemplary embodiment, such as Figure 1As shown, a multi-component seismic data fusion and interpretation method is provided. In this embodiment, the algorithm is applied to the terminal in the foregoing implementation environment as an example. It can be understood that the system can also be applied to a server, or to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps 101 to 104:

[0061] Step 101: Obtain the original multi-component seismic data and perform spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume.

[0062] Specifically, the original multi-component seismic data can be obtained through a high-performance computing platform. Perform spatio-temporal registration processing on the original multi-component seismic data, calculate the time difference and spatial displacement of different component data using the cross-correlation algorithm, and correct the polarization direction distortion of the S wave through the polarization analysis algorithm to eliminate the signal phase inconsistency problem caused by the acquisition azimuth difference, and generate a preprocessed data volume with a unified spatio-temporal reference.

[0063] Step 102: Perform multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field.

[0064] Specifically, various physical features can be extracted from the preprocessed data volume, such as the amplitude, frequency, phase, etc. of seismic waves. Using analysis methods such as weighted average and principal component analysis, different seismic features are fused. For example, the amplitude feature and the frequency feature are combined according to a certain weight to generate a fusion feature field to comprehensively reflect the various characteristics of seismic data.

[0065] Step 103: Perform elastic parameter joint inversion processing on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density.

[0066] Exemplarily, geological data of the target area can be collected, such as information on known stratigraphic structures and rock type distributions, as geological constraint conditions. Substitute the fusion feature field data into the elastic parameter joint inversion algorithm, and combine with geological constraints to continuously iterate and calculate to solve parameters such as P-wave velocity, S-wave velocity, and density that can reflect the characteristics of underground media, and construct a three-dimensional physical property parameter field to present the distribution of the physical properties of underground media in three-dimensional space.

[0067] Step 104: Perform knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, and the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithological boundary topological structure.

[0068] Specifically, a knowledge graph containing geological domain knowledge can be constructed. The graph covers relevant concepts such as faults, strata, lithology, etc. and the relationships between them. The data of the three-dimensional physical property parameter field is associated and matched with the knowledge graph, and semantic modeling technology is used to judge the position of the fault system, identify the stratigraphic interface, and divide the topological structure of the lithology boundary according to the parameter characteristics, so as to generate a geological interpretation model.

[0069] The above multi-component seismic data fusion and interpretation method performs spatial alignment and signal correction on the original multi-component seismic data through spatio-temporal registration and polarization correction processing, eliminates the phase difference and polarization distortion between different component data, and forms a preprocessed data volume with spatio-temporal consistency; based on multi-physical feature fusion processing, cross-component feature extraction and non-linear interaction are performed on the corrected P-wave, S-wave and PS-wave data to construct a fusion feature field containing multi-dimensional attributes such as amplitude, polarization, and phase, breaking through the representation limitation of a single data attribute; through elastic parameter joint inversion processing, the three-dimensional distribution fields of the longitudinal wave velocity, transverse wave velocity, and density parameters are solved synchronously to realize the coupled constrained inversion of multiple physical fields; and combined with the semantic modeling processing driven by the knowledge graph, the geological rule base is deeply fused with the gradient features of the physical property parameter field to generate a three-dimensional topological model containing the fault system, stratigraphic interface and lithology boundary. This method improves the structural recognition accuracy and reservoir prediction reliability of complex geological bodies such as continental faulted basins through the cross-component data collaboration mechanism and geological knowledge embedding strategy, and improves the accuracy of geological interpretation under complex geological conditions.

[0070] In a possible embodiment, obtaining the original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain the preprocessed data volume may include:

[0071] Step 201, perform quantum computing optimization processing on the original multi-component seismic data to obtain multi-component data aligned at the sub-wavelength level, where the quantum computing optimization processing includes encoding the spatial displacement parameter as a quantum bit state and performing parameter search through quantum gate operations.

[0072] Specifically, the spatial displacement parameters (including the horizontal displacement amounts Δx, Δy and the vertical displacement amount Δz) in the original multi-component seismic data can be encoded as the superposition state of quantum bits, a quantum circuit for parameter search is constructed through quantum gate operations, and the objective function is optimized based on the quantum annealing algorithm. Measuring the quantum state to obtain the optimal displacement parameter, realizing the sub-wavelength level spatial alignment of P-wave, S-wave and PS-wave data, and improving the registration efficiency and accuracy of multi-component data in complex structural areas through the characteristics of quantum parallel computing.

[0073] Step 202, perform polarization direction tensor decomposition processing on the S-wave data in the multi-component data to obtain S-wave data with eliminated polarization distortion.

[0074] Specifically, a polarization direction tensor matrix can be calculated for the spatially aligned S-wave data. The principal polarization direction vector corresponding to the maximum eigenvalue can be extracted through eigenvalue decomposition, and a polarization correction operator can be constructed to perform direction reprojection on the original S-wave data, eliminating polarization distortion caused by formation anisotropy or acquisition azimuth deviation to ensure the reliability of shear wave splitting parameters in subsequent inversion.

[0075] Step 203: Perform matching filtering on the PS-wave data in the multi-component data to obtain the PS-wave data after azimuth difference correction.

[0076] Specifically, an azimuth-related matching filter can be designed based on the propagation path difference between the PS-wave and the P-wave. The phase distortion of the wave field at different incident angles can be compensated through convolution operation, and the filter coefficients can be optimized using the least squares algorithm to reduce the azimuth amplitude difference of the PS-wave data and improve the consistency of multi-component data fusion.

[0077] Step 204: Perform spatio-temporal domain weighted fusion on the multi-component data, S-wave data, and PS-wave data to obtain a preprocessed data volume.

[0078] Specifically, by performing weighted superposition fusion on the sub-wavelength level aligned P-wave data, S-wave data with polarization distortion eliminated, and azimuth-corrected PS-wave data in the spatio-temporal domain, a preprocessed data volume with a unified spatio-temporal reference is generated.

[0079] Furthermore, perform multi-physical feature fusion on the preprocessed data volume to obtain a fused feature field, which may include:

[0080] Step 301: Perform tensor integration on the P-wave amplitude envelope, S-wave polarization ellipticity, PS-wave phase gradient, and acoustic impedance parameters in the preprocessed data volume to obtain a four-dimensional feature tensor.

[0081] Exemplarily, the P-wave amplitude envelope is extracted from the preprocessed data volume to characterize the formation impedance interface, the S-wave polarization ellipticity is extracted to reflect the anisotropic characteristics, the PS-wave phase gradient is extracted to indicate the thin layer tuning effect, and the acoustic impedance parameter is extracted to correlate with the rock physical properties. The four types of parameters are aligned according to the spatial coordinates and stacked along the feature dimension. Specifically, a four-dimensional feature tensor with a dimension of (X, Y, Z, 4) can be constructed, where X and Y are spatial coordinates, Z is the time / depth axis, and 4 is the number of feature channels. Through multi-parameter joint characterization, cross-dimensional correlation of different physical properties is achieved, and data utilization rate is improved.

[0082] Step 302: Perform cross-component attention weight calculation on the four-dimensional feature tensor to obtain a correlation weight matrix of the P-wave and S-wave features.

[0083] Specifically, a cross-component attention mechanism can be designed. The P-wave amplitude envelope feature is used as the query vector, and the S-wave polarization ellipticity feature is used as the key vector. The attention weight is calculated through the following formula:

[0084] ;

[0085] where, is the query vector, is the key vector, is the scaling factor consistent with the feature dimension. The above steps enhance the collaborative response of P-wave and S-wave features at geological targets such as faults and fractures through adaptive weight allocation, improve the accuracy of fusion, and contribute to more accurately analyzing the anisotropy and other characteristics of the underground geological structure.

[0086] Step 303: Perform multi-scale convolutional neural network processing on the correlation weight matrix to obtain a fusion feature field with optimized resolution.

[0087] Specifically, a multi-scale convolutional neural network model can be built. The model includes convolutional layers with different-sized convolutional kernels. The correlation weight matrix is used as the input data and input into the first layer of the convolutional neural network. The convolutional kernels of different sizes slide on the correlation weight matrix to perform convolutional operations. Among them, small convolutional kernels focus on local detailed features, and large convolutional kernels focus on global features. Through convolutional operations, the information in the correlation weight matrix is extracted and fused. After being processed by multiple convolutional layers and possibly pooling layers, activation function layers, etc., the network outputs a new feature matrix, which is the fusion feature field with optimized resolution. In this process, the convolutional neural network automatically learns the relationship between P-wave and S-wave features at different scales, fuses and optimizes the features, and improves the reliability of the entire seismic data processing flow and the accuracy of geological interpretation.

[0088] Furthermore, based on geological constraints, elastic parameter joint inversion processing is performed on the fusion feature field to obtain a three-dimensional physical property parameter field, which may include:

[0089] Step 401: Construct an elastic wave full waveform inversion objective function based on the fusion feature field. The objective function includes a data fitting term and a geological prior constraint term.

[0090] Specifically, the data fitting term ensures the consistency between the inversion result and the actual seismic data, and can accurately reflect the seismic response generated by the underground geological structure; the geological prior constraint term uses existing geological knowledge to limit the range of the inversion result, avoids the multi-solution problem that does not conform to the geological actual situation, and improves the reliability and accuracy of the inversion result.

[0091] Step 402: Solve the objective function by the alternating direction method of multipliers to obtain a three-dimensional physical property parameter field of the P-wave velocity field, S-wave velocity field, and density field.

[0092] Specifically, the initial model parameters of the P-wave velocity field, S-wave velocity field and density field can be set according to preliminary geological knowledge or previous data, and the alternating direction method of multipliers (ADMM) is used to iteratively solve the objective function. In each iteration, the S-wave velocity field and the density field are fixed, and the partial derivative of the objective function with respect to the P-wave velocity field is obtained. The P-wave velocity field is updated according to the ADMM iterative formula, the P-wave velocity field and the density field are fixed, the S-wave velocity field is updated, the P-wave velocity field and the S-wave velocity field are fixed, and the density field is updated. This cycle is repeated until the objective function converges, that is, the objective function value changes less than a preset threshold after multiple iterations. At this time, a three-dimensional physical property parameter field consisting of the P-wave velocity field, the S-wave velocity field and the density field is obtained.

[0093] Step 403, the shear wave splitting parameter is calculated and processed on the shear wave velocity field to obtain a fracture density field. The fracture density field is used to correct the anisotropic characteristics in the three-dimensional physical property parameter field. The calculation formula of the shear wave splitting parameter is as follows:

[0094] ;

[0095] in, is the shear wave splitting parameter, is the fast shear wave speed, is the slow shear wave velocity, is the background shear wave velocity.

[0096] Specifically, the fracture density field can be obtained by calculating the shear wave splitting parameters to quantitatively describe the development of underground fractures. Fractures are the key cause of the anisotropy of underground media. By correcting the three-dimensional physical parameter field with the help of the fracture density field, the differences in physical properties of underground media in different directions can be accurately reflected, and the anisotropic characteristics of underground geological structures can be more finely characterized, providing high-quality geological models for geological interpretation, oil and gas exploration and other work. The above method supplements and improves the anisotropic information of the three-dimensional physical parameter field, making the physical parameter field more complete and accurate, making up for the shortcomings of the traditional physical parameter field in the description of anisotropy, and improving the overall quality of seismic data processing and geological analysis processes.

[0097] Furthermore, the 3D physical parameter field is processed by semantic modeling driven by knowledge graph to obtain a geological interpretation model, which may include:

[0098] Step 501, using the following formula, radial basis function interpolation processing is performed on the three-dimensional physical property parameter field to generate an implicit geological interface function:

[0099] ;

[0100] in, is an implicit geological interface function, is a weight coefficient, is the influence radius, is the central coordinate of the radial basis function.

[0101] Specifically, radial basis function interpolation can flexibly interpolate according to the spatial distribution of physical property parameters, effectively capturing the complex morphology of geological interfaces; by adjusting the weight coefficient and influence radius, it can adapt to the interface characteristics under different geological conditions. Compared with traditional interpolation methods, the above method can better handle non-uniform data and accurately interpolate in areas where physical property parameters change drastically, providing a more accurate basis for subsequent extraction of geological interfaces and helping to discover hidden geological structures and interfaces.

[0102] Step 502: Perform level set extraction processing on the implicit geological interface function to generate an initial geological interface model.

[0103] Specifically, level set extraction processing has strong topological self-adaptability and can automatically handle complex topological changes such as the merging and splitting of geological interfaces. Compared with other methods that directly construct interfaces based on data points, the geological interfaces generated by level set extraction are smoother and more continuous, conforming to the natural morphology of geological bodies and providing a more reliable basic model for subsequent geological interpretation.

[0104] Step 503: Perform confidence-weighted correction processing with drilling data constraints on the initial geological interface model to generate a geological interpretation model after error correction.

[0105] Specifically, through confidence-weighted correction, the high-precision advantage of drilling data is fully utilized, effectively correcting the errors in the initial geological interface model and improving the accuracy and practicality of the model.

[0106] Furthermore, quantum computing optimization processing is performed on the original multi-component seismic data to obtain sub-wavelength-aligned multi-component data, which may include:

[0107] Step 601: Perform quantum bit superposition state mapping processing on the spatial displacement parameters to generate an initial quantum state.

[0108] Specifically, by utilizing the superposition property of quantum bits, it is possible to simultaneously encode the possible values of multiple spatial displacement parameters in a single quantum state. Compared with traditional computing methods, it improves the information storage and processing efficiency; traditional methods may need to sequentially traverse each parameter value, while the quantum bit superposition state can simultaneously process a large number of parameter combinations, providing a rich parallel processing basis for subsequent quantum evolution operations, shortening the parameter search time, and improving the overall processing speed.

[0109] Step 602: Perform Hamiltonian-driven quantum evolution processing on the initial quantum state. The Hamiltonian includes a data alignment error term and a spatial smoothing regularization term.

[0110] Specifically, the data alignment error term in the Hamiltonian directly guides the quantum state to evolve in a direction that meets the data alignment requirements, ensuring that the finally obtained spatial displacement parameters can effectively correct the spatial displacement of the original multi-component seismic data and improve the data alignment accuracy. The spatial smoothing regularization term, starting from the actual geological situation, ensures the rationality of the spatial displacement parameters and avoids abnormal parameters that do not conform to geological laws. Based on the quantum evolution method of the Hamiltonian, by making full use of the principles of quantum mechanics, it can efficiently search for the optimal solution in a complex high-dimensional parameter space. Compared with traditional optimization algorithms, it can converge to a better parameter combination more quickly, improving the accuracy and efficiency of data processing.

[0111] Step 603: Perform projection measurement processing on the evolved quantum state to obtain the optimal spatial displacement parameters. The optimal spatial displacement parameters are used to perform spatial displacement correction on the original multi-component seismic data to generate multi-component data.

[0112] Specifically, perform projection measurement on the evolved quantum state, obtain the classical measurement values of each qubit, decode to obtain the optimal spatial displacement parameters, and apply the optimal spatial displacement parameters to the three-dimensional spatial resampling of the original P-wave and S-wave data to eliminate the spatial misalignment caused by acquisition geometry differences, generate multi-component data, and improve the quality of multi-component seismic data.

[0113] Furthermore, the confidence-weighted correction processing of the initial geological interface model constrained by drilling data to generate an error-corrected geological interpretation model may include:

[0114] Step 701: Perform confidence calculation processing on the faults of the error-corrected geological interpretation model based on fuzzy Petri net rules. Specifically, determine the fault confidence based on the weighted minimum value of the longitudinal wave velocity gradient confidence and the shear wave splitting parameter confidence.

[0115] Exemplarily, the fuzzy Petri net rule is a rule system that combines the graphical modeling ability of the Petri net with the advantage of fuzzy logic in processing uncertain information. Based on elements such as places and transitions in the Petri net, it represents the degree of uncertainty of information by assigning fuzzy values to places, and transitions are triggered according to fuzzy rules. Specifically, in the geological field, for example, when dealing with the calculation of fault confidence, fuzzy information such as the confidence of the longitudinal wave velocity gradient and the confidence of the shear wave splitting parameter can be used as inputs. By setting appropriate fuzzy rules such as "when the confidence of the longitudinal wave velocity gradient is high and the confidence of the shear wave splitting parameter is high, the fault confidence is high; when one of the confidences is low, the fault confidence takes the weighted minimum of the two", the result is output through fuzzy inference operations, thereby effectively dealing with the uncertainty and fuzziness commonly existing in geological data and making the analysis and judgment of geological phenomena more in line with the actual complex situation.

[0116] Step 702: Perform a topological consistency verification process on the geologic interpretation model after error correction. After passing the verification, analyze the spatial geometric relationship between the formation dip angle and the fault strike to generate a final model that conforms to the formation contact relationship.

[0117] Specifically, the topological consistency verification ensures that the geologic interpretation model conforms to geological laws in terms of the overall structure. There are inherent spatial geometric constraint relationships between formations and faults in geological structures. Models that do not conform to this relationship cannot accurately reflect the underground geological structure. Through topological consistency verification, possible unreasonable structures in the model can be discovered, avoiding incorrect interpretations of geological structures caused by incorrect model construction. The generated final model that conforms to the formation contact relationship has higher reliability in geologic interpretation and subsequent geological research.

[0118] Step 703: When the interface position error of the final model exceeds the preset threshold, trigger the adversarial generative network reconstruction process. The reconstruction process generates a geological interface model that is consistent with the gradient distribution of the three-dimensional physical property parameter field.

[0119] Specifically, the adversarial generative network reconstruction process provides an effective method to solve the problem of excessive interface position error in the geologic interpretation model. Traditional methods often have poor effects when dealing with models with large errors, while GAN can make full use of the gradient distribution information of the three-dimensional physical property parameter field to generate a geological interface model that is consistent with the characteristics of the original data. Through the adversarial training of the generator and the discriminator, the model is continuously optimized, so that the reconstructed model not only meets the geological physical laws but also improves the accuracy of the interface position. It improves the accuracy of the geologic interpretation model and enhances the adaptability of the model to complex geological conditions, providing a more accurate geological model for geological exploration and research.

[0120] In summary, a multi-component seismic data fusion and interpretation method provided by an embodiment of the present application uses a quantum optimization algorithm to achieve sub-wavelength spatial registration and polarization correction, generating a preprocessed data volume with high consistency; fuses multi-physical features such as P-wave amplitude, S-wave polarization, and PS-wave phase through a cross-component attention mechanism to construct a fusion feature field with enhanced representation; further combines geological prior constraints to perform joint inversion of elastic parameters on the fusion feature field, synchronously solving the three-dimensional distribution fields of longitudinal wave velocity, transverse wave velocity, and density parameters, and correcting anisotropic features based on shear wave splitting parameters; through knowledge graph-driven semantic modeling, deeply fuses the physical property parameter field with the geological rule library, and uses radial basis function interpolation, level set extraction, and confidence-weighted correction to generate a three-dimensional geological interpretation model, accurately depicting the topological structures of fault systems, stratigraphic interfaces, and lithologic boundaries. The above technical solutions can solve the problems of insufficient data utilization and strong interpretive non-uniqueness under complex geological conditions, and can comprehensively utilize various seismic data information to improve the accuracy of geological interpretation under complex geological conditions.

[0121] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indications of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps is not strictly restricted by order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps in other steps.

[0122] Based on the same inventive concept, an embodiment of the present application also provides a multi-component seismic data fusion and interpretation device for implementing the above-mentioned multi-component seismic data fusion and interpretation method. The solution for solving problems provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the multi-component seismic data fusion and interpretation device provided below can refer to the limitations on the multi-component seismic data fusion and interpretation method in the above text, and will not be repeated here.

[0123] In an exemplary embodiment, as Figure 2 shown, a multi-component seismic data fusion and interpretation device 20 is provided, including:

[0124] A data preprocessing module 21, configured to obtain original multi-component seismic data and perform spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume.

[0125] The data fusion module 22 is used to perform multi - physical feature fusion processing on the pre - processed data volume to obtain a fused feature field.

[0126] The three - dimensional parameter generation module 23 is used to perform elastic parameter joint inversion processing on the fused feature field based on geological constraints to obtain a three - dimensional physical property parameter field; the three - dimensional physical property parameter field includes P - wave velocity, S - wave velocity, and density.

[0127] The geological interpretation module 24 is used to perform knowledge - graph - driven semantic modeling processing on the three - dimensional physical property parameter field to obtain a geological interpretation model, and the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithological boundary topological structure.

[0128] Further, the data pre - processing module 21 may include:

[0129] The quantum computing optimization unit 211 is used to perform quantum computing optimization processing on the original multi - component seismic data to obtain sub - wavelength - level aligned multi - component data, where the quantum computing optimization processing includes encoding spatial displacement parameters as qubit states and performing parameter search through quantum gate operations.

[0130] The S - wave distortion elimination unit 212 is used to perform polarization direction tensor decomposition processing on the S - wave data in the multi - component data to obtain S - wave data with eliminated polarization distortion.

[0131] The PS - wave correction unit 213 is used to perform matching filtering processing on the PS - wave data in the multi - component data to obtain azimuth - difference - corrected PS - wave data.

[0132] The data fusion unit 214 is used to perform spatio - temporal domain weighted fusion processing on the multi - component data, S - wave data, and PS - wave data to obtain a pre - processed data volume.

[0133] Further, the data fusion module 22 may include:

[0134] The tensor integration unit 221 is used to perform tensor integration processing on the P - wave amplitude envelope, S - wave polarization ellipticity, PS - wave phase gradient, and acoustic impedance parameters in the pre - processed data volume to obtain a four - dimensional feature tensor.

[0135] The attention weight calculation unit 222 is used to perform cross - component attention weight calculation processing on the four - dimensional feature tensor to obtain a correlation weight matrix of P - wave and S - wave features.

[0136] The convolutional neural network unit 223 is used to perform multi - scale convolutional neural network processing on the correlation weight matrix to obtain a fused feature field with optimized resolution.

[0137] Further, the three - dimensional parameter generation module 23 may include:

[0138] The constraint inversion unit 231 is used to construct an elastic wave full waveform inversion objective function based on the fused feature field, and the objective function includes a data fitting term and a geological prior constraint term.

[0139] The ADMM unit 232 is used to solve the objective function by the alternating direction method of multipliers to obtain the three-dimensional physical property parameter fields of the P-wave velocity field, S-wave velocity field, and density field.

[0140] The S-wave splitting parameter calculation unit 233 is used to calculate the S-wave splitting parameters for the S-wave velocity field to obtain the fracture density field, and the fracture density field is used to correct the anisotropic characteristics in the three-dimensional physical property parameter fields. The calculation formula for the S-wave splitting parameter is as follows:

[0141] ;

[0142] Where is the S-wave splitting parameter, is the fast S-wave velocity, is the slow S-wave velocity, is the background S-wave velocity.

[0143] Furthermore, the geological interpretation module 24 may include:

[0144] The interpolation unit 241 is used to perform radial basis function interpolation on the three-dimensional physical property parameter fields using the following formula to generate an implicit geological interface function:

[0145] ;

[0146] Where is the implicit geological interface function, is the weight coefficient, is the influence radius.

[0147] The initial geological interface generation unit 242 is used to perform level set extraction on the implicit geological interface function to generate an initial geological interface model.

[0148] The error correction unit 243 is used to perform confidence weighted correction on the initial geological interface model with drilling data constraints to generate an error-corrected geological interpretation model.

[0149] Furthermore, the quantum computing optimization unit 211 may include:

[0150] The initial quantum state sub-unit 2111 is used to perform quantum bit superposition state mapping on the spatial displacement parameters to generate an initial quantum state.

[0151] The quantum evolution sub-unit 2112 is used to perform Hamiltonian-driven quantum evolution on the initial quantum state, and the Hamiltonian includes a data alignment error term and a spatial smoothing regularization term.

[0152] A projection measurement subunit 2113 is configured to perform projection measurement processing on the evolved quantum state to obtain an optimal spatial displacement parameter, and the optimal spatial displacement parameter is used to perform spatial displacement correction on the original multi-component seismic data to generate multi-component data.

[0153] Further, the error correction unit 243 may include:

[0154] A fuzzy Petri net rule subunit 2431 is configured to perform confidence calculation processing on the fault of the geologic interpretation model after error correction based on fuzzy Petri net rules. Specifically, the fault confidence is determined based on the weighted minimum value of the longitudinal wave velocity gradient confidence and the shear wave splitting parameter confidence.

[0155] A topology check subunit 2432 is configured to perform topology consistency verification processing on the geologic interpretation model after error correction. By analyzing the spatial geometric relationship between the formation dip angle and the fault strike, a final model that conforms to the formation contact relationship is generated.

[0156] An adversarial generation network subunit 2433 is configured to trigger adversarial generation network reconstruction processing when the interface position error of the final model exceeds a preset threshold. The reconstruction processing generates a geologic interface model that is consistent with the three-dimensional physical property parameter field gradient distribution.

[0157] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of a multi-component seismic data fusion and interpretation method as described above are implemented.

[0158] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0159] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial descriptions of the method embodiments. The device embodiments described above are only illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0160] The above-described embodiments merely represent several implementation manners of the embodiments of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A multi-component seismic data fusion and interpretation method, characterized in that The method includes: Obtaining original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume; Performing multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field; Performing joint inversion processing of elastic parameters on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density; Performing knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, where the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithological boundary topological structure; Among them, the step of obtaining original multi-component seismic data and performing spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume includes: Performing quantum computing optimization processing on the original multi-component seismic data to obtain multi-component data aligned at the sub-wavelength level, where the quantum computing optimization processing includes encoding spatial displacement parameters as qubit states and performing parameter search through quantum gate operations; Performing polarization direction tensor decomposition processing on the S-wave data in the multi-component data to obtain S-wave data with polarization distortion eliminated; Performing matching filtering processing on the PS-wave data in the multi-component data to obtain PS-wave data with azimuth difference corrected; Performing spatio-temporal domain weighted fusion processing on the multi-component data, the S-wave data, and the PS-wave data to obtain the preprocessed data volume; The step of performing joint inversion processing of elastic parameters on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field includes: Constructing an elastic wave full waveform inversion objective function based on the fusion feature field, where the objective function includes a data fitting term and a geological prior constraint term; Performing alternating direction multiplier method solution processing on the objective function to obtain the three-dimensional physical property parameter field of the P-wave velocity field, S-wave velocity field, and density field; Performing shear wave splitting parameter calculation processing on the shear wave velocity field to obtain a fracture density field, where the fracture density field is used to correct the anisotropic characteristics in the three-dimensional physical property parameter field, and the calculation formula for the shear wave splitting parameter is as follows: ； wherein, is the shear wave splitting parameter, is the fast shear wave velocity, is the slow shear wave velocity, is the background shear wave velocity.

2. The method according to claim 1, characterized in that, The step of performing multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field includes: Performing tensor integration processing on the P-wave amplitude envelope, S-wave polarization ellipticity, PS-wave phase gradient, and acoustic impedance parameters in the preprocessed data volume to obtain a four-dimensional feature tensor; Performing cross-component attention weight calculation processing on the four-dimensional feature tensor to obtain a correlation weight matrix of P-wave and S-wave features; Performing multi-scale convolutional neural network processing on the correlation weight matrix to obtain a fusion feature field with optimized resolution.

3. The method according to claim 1, wherein The step of performing knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model includes: Using the following formula to perform radial basis function interpolation processing on the three-dimensional physical property parameter field to generate an implicit geological interface function: ； Among them, is the implicit geological interface function, is the weight coefficient, is the influence radius, is the central coordinate of the radial basis function; Performing level set extraction processing on the implicit geological interface function to generate an initial geological interface model; Performing confidence weighted correction processing with drilling data constraints on the initial geological interface model to generate a geological interpretation model after error correction.

4. The method according to claim 1, wherein Performing quantum computing optimization processing on the original multi-component seismic data to obtain sub-wavelength level aligned multi-component data, including: Performing quantum bit superposition state mapping processing on the spatial displacement parameters to generate an initial quantum state; Performing Hamiltonian-driven quantum evolution processing on the initial quantum state, where the Hamiltonian includes a data alignment error term and a spatial smoothing regularization term; Performing projection measurement processing on the evolved quantum state to obtain optimal spatial displacement parameters, and the optimal spatial displacement parameters are used to perform spatial displacement correction on the original multi-component seismic data to generate the multi-component data.

5. The method according to claim 3, characterized in that, Performing confidence weighted correction processing with drilling data constraints on the initial geological interface model to generate an error-corrected geological interpretation model, including: Performing confidence calculation processing on the fault of the error-corrected geological interpretation model based on fuzzy Petri net rules, specifically, determining the fault confidence based on the weighted minimum value of the longitudinal wave velocity gradient confidence and the shear wave splitting parameter confidence; Performing topological consistency verification processing on the error-corrected geological interpretation model, and the verification generates a final model that conforms to the stratigraphic contact relationship by analyzing the spatial geometric relationship between the formation dip angle and the fault strike; When the interface position error of the final model exceeds a preset threshold, triggering adversarial generation network reconstruction processing, and the reconstruction processing generates a geological interface model consistent with the three-dimensional physical property parameter field gradient distribution.

6. A multi-component seismic data fusion and interpretation device, characterized in that The device is used to implement the method described in claim 1, including: A data preprocessing module, configured to obtain original multi-component seismic data and perform spatio-temporal registration and polarization correction processing to obtain a preprocessed data volume; A data fusion module, configured to perform multi-physical feature fusion processing on the preprocessed data volume to obtain a fusion feature field; A three-dimensional parameter generation module, configured to perform joint inversion processing of elastic parameters on the fusion feature field based on geological constraints to obtain a three-dimensional physical property parameter field; the three-dimensional physical property parameter field includes P-wave velocity, S-wave velocity, and density; A geological interpretation module, configured to perform knowledge graph-driven semantic modeling processing on the three-dimensional physical property parameter field to obtain a geological interpretation model, and the geological interpretation model includes at least one of a fault system, a stratigraphic interface, and a lithology boundary topological structure; Among them, the data preprocessing module includes: A quantum computing optimization unit, configured to perform quantum computing optimization processing on the original multi-component seismic data to obtain sub-wavelength level aligned multi-component data, where the quantum computing optimization processing includes encoding spatial displacement parameters into quantum bit states and performing parameter search through quantum gate operations; An S-wave distortion elimination unit, configured to perform polarization direction tensor decomposition processing on the S-wave data in the multi-component data to obtain S-wave data with eliminated polarization distortion; A PS-wave correction unit, configured to perform matching filtering processing on the PS-wave data in the multi-component data to obtain azimuth difference corrected PS-wave data; A data fusion unit, configured to perform spatio-temporal domain weighted fusion processing on the multi-component data, the S-wave data, and the PS-wave data to obtain the preprocessed data volume; The three-dimensional parameter generation module includes: A constraint inversion unit, configured to construct an elastic wave full waveform inversion objective function based on the fused feature field, where the objective function includes a data fitting term and a geological prior constraint term; An ADMM unit, configured to perform an alternating direction method of multipliers solution process on the objective function to obtain the three-dimensional physical property parameter fields of the P-wave velocity field, S-wave velocity field, and density field; An S-wave splitting parameter calculation unit, configured to perform an S-wave splitting parameter calculation process on the S-wave velocity field to obtain a fracture density field, where the fracture density field is used to correct the anisotropic characteristics in the three-dimensional physical property parameter field, and the calculation formula of the S-wave splitting parameter is as follows: ; Among them, is the shear wave splitting parameter, is the fast shear wave velocity, is the slow shear wave velocity, is the background shear wave velocity.

7. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the method according to any one of claims 1 to 5 is implemented.

Images