A multi-modal geophysical seismic exploration method and system

Abstract

The application discloses a kind of multi-state geophysical seismic exploration method and system, it is related to geophysical exploration technical field, and ground and underground wave field data are collected by distributed array;Effective reflection signal is separated using time-frequency analysis;Detect environmental change and utilize convolutional neural network to identify interference type and adjustment requirement;Synchronous adjustment instruction of excitation source and receiving array is generated and issued;Fusion multi-platform space-time alignment data, high-resolution underground structure image is reconstructed using reverse time migration imaging.The application realizes the dynamic optimization of detection parameter and multi-platform cooperation, and improves the exploration precision and efficiency under complex geological conditions.

Description

Technical Field

[0001] This invention belongs to the field of geophysical exploration technology, and in particular relates to a multi-mode geophysical seismic exploration method and system. Background Technology

[0002] Seismic exploration, as a core technology in geophysical exploration, plays an irreplaceable role in energy exploration, engineering construction, and disaster early warning. This technology aims to infer subsurface geological structures by collecting and analyzing the propagation response of artificially generated seismic waves in the subsurface medium. Traditional seismic exploration methods typically rely on pre-set fixed parameter schemes, achieving reliable results in relatively simple and stable geological environments. However, in practical operations, especially in areas with complex surface conditions and variable geological structures, the limitations of such static methods are becoming increasingly apparent.

[0003] Specifically, the core deficiency of existing technologies lies in their lack of dynamic adaptability to the detection environment. When there is strong, time-varying vibration interference on the surface, or when there are abrupt changes in the underground geological structure, acquisition systems based on fixed parameters cannot adjust their operating strategies in real time. For example, in urban environments or near-shore areas, continuous ground traffic and human activities generate high-frequency noise, masking effective deep reflection signals; while in complex tectonic areas, drastic changes in the strata can lead to a shift in the dominant frequency of reflected waves and abnormal attenuation of amplitude. Traditional methods struggle to identify these changing characteristics online, and even more so, they cannot optimize the frequency, energy, or geometric layout of the excitation source or the receiving array in real time, resulting in low signal-to-noise ratio and weak effective signals in the acquired raw data, directly affecting the accuracy of subsequent imaging.

[0004] Furthermore, as exploration missions demand higher efficiency and wider coverage, multi-platform, distributed collaborative operation modes are gradually becoming the trend. However, existing technologies have significant shortcomings in multi-platform collaboration. Each detection platform often operates independently, lacking a unified, high-precision time synchronization and real-time command coordination mechanism. This makes it difficult to accurately align observation data from different spatial locations in time and space, preventing the formation of high-quality joint datasets. In practical applications, such as in joint land-sea or large-area 3D exploration, this lack of collaboration can lead to artifacts at data stitching points, reduced overall imaging resolution, and failure to meet the needs of high-precision resource assessment and engineering decision-making. Solving the challenge of dynamic adaptation and efficient multi-platform collaboration in detection systems is key to improving seismic exploration results in complex environments. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a multi-mode geophysical seismic exploration method and system to resolve the issues present in the prior art.

[0006] Firstly, to achieve the above objectives, the present invention provides a multi-mode geophysical seismic exploration method, comprising the following steps:

[0007] Raw time-series data containing surface vibration and subsurface wave field response were collected using a distributed sensor array.

[0008] The original time-series data is subjected to time-frequency decomposition to separate the effective reflection signal components and obtain a purified reflection wavefield dataset.

[0009] Based on the purified reflected wave field dataset, detect whether the detection environment has undergone significant changes;

[0010] When the significant changes are detected, a convolutional neural network model is used to identify the dominant disturbance type and the geological stratum adjustment requirements, and to generate targeted detection parameter optimization instructions.

[0011] Based on the targeted detection parameter optimization instructions, generate and issue instructions for adjusting the excitation source frequency and receiver array geometry synchronization to each detection platform;

[0012] Based on the new round of wavefield response data collected by each detection platform after adjustment, a spatiotemporally aligned joint observation dataset is generated by fusion.

[0013] The spatiotemporally aligned joint observation dataset is subjected to reverse time migration imaging to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image.

[0014] Output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

[0015] Secondly, the present invention also provides a multi-mode geophysical seismic exploration system for implementing a multi-mode geophysical seismic exploration method, the system comprising:

[0016] The data acquisition module is used to acquire raw time-series data containing surface vibration and subsurface wave field response through a distributed sensor array;

[0017] The data processing module is used to perform time-frequency decomposition processing on the original time-series data to separate the effective reflection signal components and obtain a purified reflection wave field dataset.

[0018] The environmental detection module is used to detect whether the detection environment has undergone significant changes based on the purified reflected wave field dataset;

[0019] The decision optimization module is used to identify the dominant disturbance type and geological stratum adjustment needs using a convolutional neural network model when the significant change is detected, and to generate targeted detection parameter optimization instructions.

[0020] The collaborative control module is used to generate and send excitation source frequency and receiver array geometric synchronization adjustment instructions to each detection platform based on the targeted detection parameter optimization instructions;

[0021] The data fusion module is used to fuse the new round of wavefield response data collected by each detection platform after adjustment to generate a spatiotemporally aligned joint observation dataset.

[0022] The imaging reconstruction module is used to perform reverse time migration imaging processing on the spatiotemporally aligned joint observation dataset to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image.

[0023] The results output module is used to output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

[0024] Thirdly, the present invention also provides a computer terminal device, comprising:

[0025] One or more processors;

[0026] A memory, coupled to the processor, for storing one or more programs;

[0027] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the polymorphic geophysical seismic exploration method in the first aspect described above.

[0028] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the polymorphic geophysical seismic exploration method in the first aspect described above.

[0029] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the polymorphic geophysical seismic exploration method described in the first aspect above.

[0030] Compared with the prior art, the present invention has the following advantages and technical effects:

[0031] This invention provides a multi-modal geophysical seismic exploration method and system, which effectively improves the accuracy and efficiency of seismic exploration in complex environments. By analyzing acquired data in real time and dynamically determining environmental changes, the exploration system can adaptively adjust excitation and reception parameters, thereby effectively suppressing interference and enhancing effective signals. Convolutional neural networks are used to intelligently identify interference types and geological stratigraphic requirements, achieving precise and automated optimization of detection parameters. High-precision synchronous control and spatiotemporal alignment fusion technology ensure data consistency across multiple platforms, forming a high-quality joint observation dataset. Reverse-time migration imaging processing based on this dataset can reconstruct clearer and more accurate subsurface velocity models and reflective interface images, ultimately outputting a high-resolution three-dimensional geological structure description, providing more reliable technical support for resource assessment and engineering decision-making. Attached Figure Description

[0032] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0033] Figure 1 This is an overall flowchart of the multi-state geophysical seismic exploration method according to an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram of the time-frequency decomposition and purification process of reflected wave field data according to an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of multi-platform synchronous adjustment and joint data imaging in an embodiment of the present invention. Detailed Implementation

[0036] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0037] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0038] Example 1

[0039] Figure 1This paper demonstrates the complete process of the multi-mode geophysical seismic exploration method of the present invention. It illustrates the coherent steps and logical relationships from the acquisition of raw time-series data through a distributed sensor array, through time-frequency analysis and signal purification, environmental change detection, intelligent decision-making and parameter optimization based on neural networks, synchronous issuance of instructions and status updates across multiple platforms, data fusion to form a joint observation dataset, and geological structure reconstruction by reverse time migration imaging, until the final output of a three-dimensional subsurface structure description.

[0040] like Figure 1 As shown, this embodiment provides a multi-mode geophysical seismic exploration method, including:

[0041] Raw time-series data containing surface vibration and subsurface wave field response were collected using a distributed sensor array.

[0042] The original time-series data is subjected to time-frequency decomposition to separate the effective reflection signal components and obtain a purified reflection wavefield dataset.

[0043] Based on the purified reflected wave field dataset, detect whether the detection environment has undergone significant changes;

[0044] When the significant changes are detected, a convolutional neural network model is used to identify the dominant disturbance type and the geological stratum adjustment requirements, and to generate targeted detection parameter optimization instructions.

[0045] Based on the targeted detection parameter optimization instructions, generate and issue instructions for adjusting the excitation source frequency and receiver array geometry synchronization to each detection platform;

[0046] Based on the new round of wavefield response data collected by each detection platform after adjustment, a spatiotemporally aligned joint observation dataset is generated by fusion.

[0047] The spatiotemporally aligned joint observation dataset is subjected to reverse time migration imaging to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image.

[0048] Output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

[0049] S101, through a pre-established multi-field information acquisition framework, synchronously acquires earthquake motion data, electromagnetic signals and gravity change data, and uses standardized processing methods to unify the format and filter noise of various types of data to obtain a pre-processed multi-source dataset.

[0050] S102. Based on the pre-processed multi-source dataset, an information association mechanism is constructed. Using a feature mapping method based on physical constraints, the features of earthquake motion data, electromagnetic signals and gravity change data are cross-compared to determine the set of potential consistency features among the data.

[0051] S103, for the consistency feature set, adopts a multi-field information fusion algorithm to weight and integrate seismic motion data, electromagnetic signals and gravity change data to generate a comprehensive feature matrix for subsequent preliminary characterization of geological structure;

[0052] S104: By integrating the feature matrix, a physical constraint model is constructed, prior knowledge of deep underground exploration is introduced, contradictory features in the matrix are corrected, and it is determined whether the corrected features conform to geological laws, thus obtaining the optimized feature description.

[0053] S105. Based on the optimized feature description, data mapping is performed on the geological structure prediction module, and a hierarchical analysis method is used to divide the deep underground region into structural sections to determine the structural distribution characteristics of each level.

[0054] S106. Based on the structural distribution characteristics and the reservoir property judgment rules, if the vibration data indicates the existence of the reservoir but the electromagnetic signal shows an anomaly, then gravity change data is introduced for auxiliary verification to obtain the final reservoir property judgment result.

[0055] S107. Based on the final reservoir property determination results, generate a resource distribution characterization image, mark key areas in the image, and output a comprehensive geological image suitable for resource exploration.

[0056] As one implementation method in this embodiment, the process of acquiring raw time-series data through a distributed sensor array includes:

[0057] A clock synchronization protocol is used to synchronously acquire multi-channel analog signals and generate multi-channel synchronous digital waveform data.

[0058] The active signal range is extracted based on the amplitude change rate of the multi-channel synchronous digital waveform data;

[0059] Based on the frequency distribution characteristics of the active signal range, it is divided into surface vibration components and subsurface wave field response components.

[0060] The divided components are subjected to spectral aliasing analysis and wavefront snapshot inversion to generate a feature signal sequence with physical attribute annotations.

[0061] Step S101: Through the pre-established multi-field information acquisition framework, earthquake motion data, electromagnetic signals and gravity change data are acquired synchronously. Standardized processing methods are used to unify the format and filter noise of various types of data to obtain a pre-processed multi-source dataset.

[0062] The system receives seismic amplitude sequences, electromagnetic frequency fluctuation sequences, and gravity anomaly change sequences, aligns them according to high-precision timestamps, and generates a time-synchronized raw multidimensional data stream. It then parses the data frame header protocol of the time-synchronized raw multidimensional data stream and performs interpolation resampling to construct a standardized multi-field data matrix with a unified format. Finally, it identifies the noise regions to be processed in the standardized multi-field data matrix and applies an adaptive notch filter for amplitude attenuation, obtaining a clean multi-physical quantity sequence with a high signal-to-noise ratio, and outputs the pre-processed multi-source dataset.

[0063] Specifically, after receiving the seismic amplitude sequence, electromagnetic frequency fluctuation sequence, and gravity anomaly change sequence, the sequence is first aligned based on a high-precision timestamp.

[0064] Specifically, the timestamps recorded by the three sensors are mapped onto the same reference time axis.

[0065] In one possible implementation, the data frame header protocol of the original time-synchronized multidimensional data stream is parsed and resampled using interpolation. The data frame header typically contains information such as sequence identifier, sampling rate marker, channel number, and parity bit. After parsing, cases where the original sampling rates are 100Hz, 50Hz, and 1Hz are identified. Preferably, spline interpolation is used to resample all sequences to a uniform sampling rate of 200Hz. The standardized multi-field data matrix constructed in this way has each row corresponding to the same time moment, and the three columns store the seismic acceleration value, electromagnetic field intensity component, and gravity residual value, forming a matrix with a consistent format, which facilitates subsequent batch processing.

[0066] It should be noted that identifying the noise region to be processed in the standardized multi-field data matrix and applying an adaptive notch filter are key steps.

[0067] Understandably, the multi-source dataset output after the above series of processing steps has its physical quantities strictly synchronized in time, has a uniform sampling rate, and its noise is specifically suppressed. This makes it easier for the subsequent intelligent recognition algorithm to capture the weak multi-field coupling characteristics of earthquake precursors, and the sensitivity and reliability of the overall monitoring system are significantly improved.

[0068] As one implementation method in this embodiment, the process of performing time-frequency decomposition processing on the original time-series data includes:

[0069] The generalized S-transform algorithm is used to map the original time series data into a two-dimensional time-frequency distribution matrix;

[0070] The local spectral energy density is calculated based on the two-dimensional time-frequency distribution matrix, and the characteristics of effective reflected signals are distinguished by singular value decomposition.

[0071] A denoising time-frequency mask is constructed and subjected to a Hadamard product operation with the two-dimensional time-frequency distribution matrix to retain effective reflected signal components.

[0072] Perform the inverse generalized S-transform operation on the retained effective reflected signal components.

[0073] Figure 2 The process of time-frequency decomposition and purification of reflected wave field data is described in detail. The diagram shows the core steps and data morphology changes of the original time-series data being converted into a two-dimensional time-frequency distribution matrix through generalized S-transform, then separating the effective signal components by calculating the local spectral energy density, constructing a denoising time-frequency mask and performing Hadamard product operation, and finally obtaining the purified reflected wave field dataset through inverse transformation.

[0074] like Figure 2 As shown, in step S102, based on the pre-processed multi-source dataset, an information association mechanism is constructed. Using a feature mapping method based on physical constraints, the features of earthquake vibration data, electromagnetic signals, and gravity change data are cross-compared to determine the potential consistency feature set among the data.

[0075] Seismic motion waveform data, electromagnetic spectrum signals, and gravity anomaly values ​​are acquired to generate a spatiotemporally synchronized data matrix. This spatiotemporally synchronized data matrix is ​​then input into a geophysical constraint model to obtain a pure feature vector set. The pure feature vector set is processed using a physical constraint feature mapping function to construct a standardized feature projection matrix. The geometric distance between the standardized feature projection matrices is calculated. If the geometric distance meets certain conditions, a multi-source feature cross-comparison index list is generated. Based on the multi-source feature cross-comparison index list, physically homogeneous feature elements are extracted to determine the potential consistency feature set among seismic motion data, electromagnetic signals, and gravity variation data.

[0076] For example, when processing earthquake-related multi-source data, one can start by examining the spatiotemporal characteristics of the data to explore how to achieve effective data integration and analysis. This research focuses on the topic of acquiring earthquake motion data, electromagnetic signals, and gravity variation data, and generating spatiotemporally aligned sequences based on acquisition timestamps and spatial coordinate grids.

[0077] Understandably, spatiotemporal alignment is a fundamental step in multi-source data processing. Suppose that in a certain earthquake monitoring operation, seismic motion data is recorded at 100 sampling points per second, electromagnetic signal acquisition frequency is 50 times per second, and gravity change data is collected once per minute. By aligning the timestamps, all three are unified onto a time grid of once per second. Simultaneously, combined with spatial coordinates, the data is mapped to a 10 km by 10 km grid area, forming a unified spatiotemporal sequence. This alignment method helps reduce data bias in subsequent analysis.

[0078] For example, in the step of reconstructing a comprehensive feature matrix from a weighted feature map, the feature values ​​in the map can be rearranged according to a spatial grid to form a two-dimensional matrix. Each element of the matrix represents a comprehensive physical property value for a specific region. Assuming the comprehensive value of a certain grid point is 0.9, it indicates that there may be significant changes in the subsurface medium in that region. This matrix form facilitates subsequent analysis.

[0079] As one implementation method in this embodiment, the process of detecting whether the detection environment has undergone significant changes based on the purified reflected wave field dataset includes:

[0080] The time-frequency distribution matrix of the purified reflected wave field dataset was obtained by decomposing it using the short-time Fourier transform algorithm.

[0081] Extract the instantaneous main frequency value based on the time-frequency distribution matrix and calculate the main frequency offset distribution map;

[0082] Calculate the amplitude attenuation rate data based on the main frequency offset distribution map;

[0083] The main frequency offset distribution map and amplitude attenuation rate data are used to construct an environmental difference feature vector, and a judgment is made based on whether its modulus value exceeds a preset anomaly judgment threshold.

[0084] Step S103: For the consistency feature set, a multi-field information fusion algorithm is used to weight and integrate the seismic motion data, electromagnetic signals and gravity change data to generate a comprehensive feature matrix for subsequent preliminary characterization of geological structures.

[0085] Seismic motion data, electromagnetic signals, and gravity variation data are acquired, and a spatiotemporal aligned sequence is generated based on the acquisition timestamps and spatial coordinate grids. Features are extracted from the spatiotemporal aligned sequence to obtain a normalized numerical set, which is then used to generate a cross-correlation matrix. If the correlation values ​​in the cross-correlation matrix are greater than a preset threshold, a fusion weight vector is constructed. This fusion weight vector is then used in conjunction with the normalized numerical set to obtain a weighted feature map. The weighted feature map is reconstructed to construct a comprehensive feature matrix, which is then analyzed to define the physical property boundaries of the subsurface medium, thus achieving a preliminary characterization of the geological structure.

[0086] Specifically, constructing a comprehensive feature matrix reflecting the correlation of underground media properties is achieved by mapping multi-source deep-penetration data to a high-dimensional vector space. For example...

[0087] In one embodiment, the acquired seismic wave propagation time series, resistivity profile data, and crustal density perturbation values ​​are first spatially gridded, projecting the observation points of different physical quantities onto the same three-dimensional coordinate system to form the original data tensor. Subsequently, principal component extraction and standardization are performed on each physical quantity to obtain a set of feature vectors with consistent dimensions. These vector sets are then concatenated according to predefined attribute association rules to form the initial comprehensive feature matrix.

[0088] It should be noted that this mapping method can better preserve the implicit physical coupling relationship between different observation methods. Inputting the comprehensive feature matrix into the physical constraint model and using prior knowledge to perform numerical iterative correction of contradictory features is a key step in ensuring the physical rationality of the matrix.

[0089] Specifically, physical constraint models typically embed empirical rock physics formulas and known stratigraphic sequence patterns.

[0090] For example, if a location in the matrix exhibits both high wave velocity and low resistivity characteristics, while the geological background of the area should be dense carbonate rock, the model will identify the contradiction and initiate an iterative adjustment process.

[0091] In one possible implementation, the model first calculates the Mahalanobis distance between the current feature vector and the prior distribution. If the distance exceeds the confidence interval, a smoothing correction is applied to the contradictory components using the principle of minimum perturbation. After multiple iterations, contradictory features are gradually suppressed, generating a corrected feature set. This correction significantly improves the reliability of subsequent interpretations. Determining whether the corrected feature set conforms to geological laws primarily relies on consistency checks based on the regional tectonic evolution history and drilled lithological sequences.

[0092] For example, if the corrected features show a combination of low density and high wave velocity in a certain depth segment, and this segment is located in the core of a known paleo-uplift, then it is considered to conform to the pattern; conversely, if high density is accompanied by extremely low wave velocity without any evidence of fracture, then it is considered not to conform, and it is necessary to return to the previous step to re-optimize the weights.

[0093] Preferably, the consistency judgment can also incorporate a small amount of known well point data as anchor points to further constrain the model output. If the judgment conforms to geological laws, then dimensionality reduction and reconstruction are performed to obtain an optimized feature description.

[0094] In one embodiment, a supervised linear discriminant analysis method is used to project high-dimensional features into a low-dimensional space, while maximizing the separability of different lithological categories.

[0095] For example, the original matrix may contain hundreds of dimensions of information. After dimensionality reduction, about 10 main dimensions are retained, which not only preserves the differences in medium properties, but also greatly reduces noise interference and computational complexity.

[0096] For example, after this step, the previously scattered feature points exhibit a clear clustered structure on the two-dimensional plane, with one cluster corresponding to clastic reservoirs and the other to igneous intrusions. This clear separation provides high-quality input for subsequent reservoir prediction and structural interpretation.

[0097] Understandably, by combining the above-mentioned physical constraint iteration with dimensionality reduction and reconstruction, not only is the inherent inconsistency of multi-source data eliminated, but the accuracy and interpretability of the description of underground media properties are also significantly improved, laying a solid foundation for fine geological modeling.

[0098] Step S104: By integrating the feature matrix, a physical constraint model is constructed, prior knowledge of deep underground exploration is introduced, contradictory features in the matrix are corrected, and it is determined whether the corrected features conform to geological laws, thus obtaining the optimized feature description.

[0099] A comprehensive feature matrix reflecting the correlation of subsurface media properties is constructed. This matrix is ​​obtained by mapping acquired deep-penetration data to a vector space. The comprehensive feature matrix is ​​then input into a physical constraint model. Prior knowledge fused within the physical constraint model is used to numerically iteratively correct contradictory features in the comprehensive feature matrix, generating a corrected feature set. It is then determined whether the corrected feature set conforms to geological laws. If it does, dimensionality reduction and reconstruction are performed to obtain an optimized feature description.

[0100] For example, when acquiring and inverting microgravity variation data covering areas of signal conflict, gravity anomalies in the area can be collected using microgravity measurement equipment, allowing the derivation of the subsurface residual density distribution. Suppose that microgravity data indicates the presence of a low-density anomaly within the signal conflict area; combined with geological background analysis, this might correspond to a low-density reservoir or porous structure. This inversion process provides crucial information for subsequent spatial matching, reducing the risk of misjudgment due to data conflicts.

[0101] For example, when outputting the final reservoir property determination result based on the physical property validation set, a multi-parameter weighted analysis can be used to comprehensively assess the likelihood of reservoir existence. Assuming that in the above case, all indicators in the validation set point to reservoir characteristics, such as low density and high porosity, the area is ultimately determined to be a potential oil and gas reservoir. This multi-data fusion approach can effectively improve the reliability of the determination, provide a reliable basis for subsequent exploration and development, and reduce exploration risks.

[0102] For example, from the perspective of expanding the scope of the study, if the microgravity data in areas of signal conflict is insufficient to support the conclusions, additional geomagnetic or well logging data can be introduced as supplementary verification methods. Suppose that in a certain exploration, the microgravity data resolution is low, but the geomagnetic data reveals anomalous signals consistent with reservoir characteristics. This multi-source data combination approach can further enrich the content of the physical property verification set and improve the comprehensiveness of the final judgment. This method is particularly suitable for exploration tasks in complex geological environments.

[0103] As one implementation method in this embodiment, the process of generating and issuing synchronization adjustment instructions to each detection platform includes:

[0104] Acquire and process environmental data to obtain spectral response characteristics;

[0105] The excitation source frequency value is calculated based on the spectral response characteristics, and the geometric topology of the receiver array is generated by combining the platform spatial distribution data.

[0106] Based on the aforementioned geometric topology, transmission delay data is calculated to correct synchronization timing deviations;

[0107] The corrected synchronization timing deviation and excitation source frequency value are integrated to generate an adjustment instruction sequence, which is then encapsulated into an adjustment instruction set.

[0108] Figure 3 The key points illustrate the collaborative mechanism of multi-platform synchronous adjustment and joint data imaging. The diagram shows the entire process of interaction and data flow, including generating an instruction set containing frequency adjustment and geometric redistribution information according to optimization instructions, issuing and driving each platform to update its status synchronously through the communication link, transmitting data collected by each platform back to the central node for spatiotemporal alignment and fusion, and performing reverse time migration imaging based on the joint dataset to obtain high-resolution geological images.

[0109] like Figure 3 As shown, in step S105, based on the optimized feature description, data mapping is performed on the geological structure prediction module, and a hierarchical analysis method is used to divide the deep underground region into structural layers to determine the structural distribution characteristics of each layer.

[0110] The optimized feature description is obtained, and a mapping feature vector is generated through a preset mapping rule library. The mapping feature vector is input into the geological structure prediction module to obtain a layered impedance sequence. The layered impedance sequence is subjected to deep discretization processing to generate a depth slice image. The gradient value of the depth slice image is calculated, and the positions corresponding to the gradient values ​​exceeding a preset threshold are connected to form the stratigraphic boundary coordinates. The sedimentary facies classification result is obtained by matching the sedimentary boundary coordinates with a sedimentary facies classification library. Based on the sedimentary facies classification result, the structural distribution characteristics of each layer in the deep underground region are determined.

[0111] Specifically, after obtaining the final reservoir property determination result, an initial resource distribution characterization image can be generated based on the result.

[0112] Specifically, the image uses color gradients to reflect the spatial distribution of reservoir physical parameters, for example, marking high-porosity areas with warm colors and low-porosity areas with cool colors, thus visually presenting potential resource-rich areas.

[0113] In one possible implementation, an orange-red area appears near a depth of 2000 meters in the initial image, indicating a superior reservoir distribution supported by both vibration and gravity data.

[0114] It is understandable that calculating the gradient magnitude of the initial resource distribution characterization image can highlight the locations of abrupt changes in physical property parameters.

[0115] For example, after applying the gradient operator to each pixel of the image, it was found that the gradient magnitude at some locations exceeded 0.45, while the average value of the surrounding area was only 0.12. These high gradient bands indicate the interface where the internal properties of the reservoir change drastically, such as the transition zone where the porosity drops sharply from 18% to 6%, thereby identifying the key reservoir heterogeneity boundary feature region.

[0116] Preferably, performing morphological closing operations on the key reservoir heterogeneous boundary feature regions can eliminate minor noise and connect discontinuous boundaries.

[0117] In one embodiment, a 3×3 structuring element is used for closing operations, which fills the small gaps that were originally limited by data resolution, and the boundaries change from fragmented to continuous and smooth closed curves, forming a clear heterogeneous partition outline.

[0118] It should be noted that using Alpha blending technology to perform weighted overlay and fusion of the binarized labeled mask data and the initial resource distribution characterization image can highlight boundary features while preserving the original resource distribution information.

[0119] As one implementation method in this embodiment, the process of fusing and generating a spatiotemporally aligned joint observation dataset includes:

[0120] Acquire the initial wavefield data packets uploaded from each platform, which include wavefield response signals, spatial coordinates, and acquisition time labels;

[0121] The clock deviation is calculated based on the acquisition time tag and the additional transmission delay value, and the time synchronization observation sequence is obtained by correcting the clock deviation.

[0122] Based on the spatial coordinates in the time-synchronized observation sequence, grid mapping is performed to generate spatially aligned grid data;

[0123] Aggregate all spatially aligned grid data.

[0124] Step S106: Based on the structural distribution characteristics and the reservoir property judgment rules, if the vibration data indicates the existence of the reservoir but the electromagnetic signal shows an anomaly, then gravity change data is introduced for auxiliary verification to obtain the final reservoir property judgment result.

[0125] Vibration wave field records are constrained using structural distribution data to delineate vibration-indicating reservoir regions. If the electromagnetic response data corresponding to the vibration-indicating reservoir region is abnormal, the vibration-indicating reservoir region is marked as a signal conflict region. Microgravity variation data covering the signal conflict region is acquired, and the residual density distribution volume beneath the signal conflict region is obtained by inversion. The residual density distribution volume is spatially matched with the vibration-indicating reservoir region to generate a physical property verification set, and the final reservoir property determination result is output based on the physical property verification set.

[0126] Step S107: Based on the final reservoir property determination result, generate a resource distribution characterization image, mark key areas in the image, and output a comprehensive geological image suitable for resource exploration.

[0127] Obtain the final reservoir property determination result, and generate an initial resource distribution characterization image based on the final reservoir property determination result; calculate the gradient magnitude of the initial resource distribution characterization image, and determine the key reservoir heterogeneous boundary feature regions based on the gradient magnitude; perform morphological closing operation on the key reservoir heterogeneous boundary feature regions to generate binarized labeled mask data; use Alpha mixing technology to perform weighted superposition and fusion of the binarized labeled mask data and the initial resource distribution characterization image to output a comprehensive geological image suitable for resource exploration.

[0128] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

[0129] Based on this, the present invention provides a multi-modal geophysical seismic exploration method. Through a pre-established multi-field information acquisition framework, it achieves synchronous high-precision acquisition of seismic vibration data, electromagnetic signals, and gravity change data, which are then standardized and noise-filtered to form a unified multi-source dataset. Subsequently, a feature mapping and information association mechanism based on physical constraints is constructed to cross-compare and mine consistency among the three types of data features. A multi-field weighted fusion algorithm is used to generate a comprehensive feature matrix. Furthermore, prior knowledge of deep subsurface exploration is introduced to construct a physical constraint model, intelligently correcting contradictory features in the matrix to obtain optimized feature descriptions that conform to geological laws. On this basis, a hierarchical analysis method is used to complete the subsurface structure division. Combined with reservoir property judgment rules, a three-field collaborative logic—primarily vibration data, secondarily electromagnetic anomalies, and verified by gravity changes—is used to finally determine the existence and properties of reservoirs, ultimately generating a comprehensive geological image of resource distribution in marked key areas. This invention solves the core problems of low prediction accuracy, high misjudgment rate, and difficulty in resolving contradictions in multi-source information under complex geological conditions by traditional single or loosely coupled exploration methods, significantly improving the reliability and exploration efficiency of deep reservoir identification.

[0130] Example 2

[0131] In this embodiment, a computer terminal device is provided, including:

[0132] One or more processors;

[0133] A memory, coupled to the processor, for storing one or more programs;

[0134] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the above-described polymorphic geophysical seismic exploration method.

[0135] In this embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described polymorphic geophysical seismic exploration method.

[0136] In this embodiment, an electronic device is also provided, including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the steps of the above-described polymorphic geophysical seismic exploration method.

[0137] In this embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the above-described polymorphic geophysical seismic exploration method.

[0138] The aforementioned program can run on a processor or be stored in memory (or a computer-readable medium). Computer-readable media includes both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random-access memory (SRAM), dynamic random-access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0139] These computer programs may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes can be implemented by different modules for different steps.

[0140] This embodiment provides such an apparatus or system. The system, referred to as a multi-modal geophysical seismic exploration system, includes:

[0141] The data acquisition module is used to acquire raw time-series data containing surface vibration and subsurface wave field response through a distributed sensor array;

[0142] The data processing module is used to perform time-frequency decomposition processing on the original time-series data to separate the effective reflection signal components and obtain a purified reflection wave field dataset.

[0143] The environmental detection module is used to detect whether the detection environment has undergone significant changes based on the purified reflected wave field dataset;

[0144] The decision optimization module is used to identify the dominant disturbance type and geological stratum adjustment needs using a convolutional neural network model when the significant change is detected, and to generate targeted detection parameter optimization instructions.

[0145] The collaborative control module is used to generate and send excitation source frequency and receiver array geometric synchronization adjustment instructions to each detection platform based on the targeted detection parameter optimization instructions;

[0146] The data fusion module is used to fuse the new round of wavefield response data collected by each detection platform after adjustment to generate a spatiotemporally aligned joint observation dataset.

[0147] The imaging reconstruction module is used to perform reverse time migration imaging processing on the spatiotemporally aligned joint observation dataset to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image.

[0148] The results output module is used to output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

[0149] As one implementation method in this embodiment, the data acquisition module includes:

[0150] The synchronous acquisition unit is used to synchronously acquire multi-channel analog signals using a clock synchronization protocol and generate multi-channel synchronous digital waveform data.

[0151] The signal interception unit is used to intercept the active signal range based on the amplitude change rate of the multi-channel synchronous digital waveform data.

[0152] The component division unit is used to divide the active signal interval into surface vibration components and subsurface wave field response components based on the frequency distribution characteristics of the active signal interval.

[0153] The signal labeling unit is used to perform spectral aliasing analysis and wavefront snapshot inversion on the divided components to generate a feature signal sequence with physical attribute labels.

[0154] As one implementation method in this embodiment, the data processing module includes:

[0155] The time-frequency mapping unit is used to map the original time-series data into a two-dimensional time-frequency distribution matrix using the generalized S-transform algorithm;

[0156] The feature separation unit is used to calculate the local spectral energy density based on the two-dimensional time-frequency distribution matrix and to distinguish the features of the effective reflection signal through singular value decomposition.

[0157] The signal purification unit is used to construct a noise-reducing time-frequency mask and perform a Hadamard product operation with the two-dimensional time-frequency distribution matrix to retain effective reflected signal components.

[0158] The inverse transform unit is used to perform the inverse transform operation of the generalized S-transform on the retained effective reflected signal components.

[0159] As one implementation method in this embodiment, the environmental detection module includes:

[0160] The time-frequency decomposition unit is used to decompose the purified reflected wave field dataset into a time-frequency distribution matrix using the short-time Fourier transform algorithm;

[0161] The main frequency extraction unit is used to extract the instantaneous main frequency value based on the time-frequency distribution matrix and calculate the main frequency offset distribution map.

[0162] An attenuation calculation unit is used to calculate amplitude attenuation rate data based on the main frequency offset distribution map.

[0163] The determination unit is used to construct an environmental difference feature vector from the main frequency offset distribution map and the amplitude attenuation rate data, and to make a determination based on whether its modulus value exceeds a preset anomaly determination threshold.

[0164] As one implementation method in this embodiment, the collaborative control module includes:

[0165] The feature analysis unit is used to acquire detection environment data and process it to obtain spectral response features;

[0166] The parameter design unit is used to calculate the excitation source frequency value based on the spectral response characteristics and generate the geometric topology of the receiver array in combination with the platform spatial distribution data.

[0167] A timing correction unit is used to calculate transmission delay data based on the geometric topology to correct synchronization timing deviations;

[0168] The instruction encapsulation unit is used to fuse the corrected synchronization timing deviation and the excitation source frequency value to generate an adjustment instruction sequence, and encapsulate it into an adjustment instruction set.

[0169] As one implementation method in this embodiment, the data fusion module includes:

[0170] The data receiving unit is used to acquire the initial wavefield data packets uploaded by each platform, which include wavefield response signals, spatial coordinates, and acquisition time labels.

[0171] The time synchronization unit is used to calculate the clock deviation based on the acquisition time tag and the additional transmission delay value, and to use the clock deviation to correct and obtain the time-synchronized observation sequence.

[0172] A spatial alignment unit is used to perform grid mapping based on the spatial coordinates in the time-synchronized observation sequence to generate spatially aligned grid data.

[0173] The data aggregation unit is used to aggregate all spatially aligned grid data.

[0174] The system or apparatus is used to implement the functions of the methods in the above embodiments. Each module in the system or apparatus corresponds to each step in the method, as has been described in the method and will not be repeated here.

[0175] The above-described implementation method solves the problem of multi-modal geophysical seismic exploration in related technologies, thereby ensuring that the problems existing in the prior art are resolved.

[0176] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A multi-modal geophysical seismic exploration method, characterized in that, Includes the following steps: Raw time-series data containing surface vibration and subsurface wave field response were collected using a distributed sensor array. The original time-series data is subjected to time-frequency decomposition to separate the effective reflection signal components and obtain a purified reflection wavefield dataset. Based on the purified reflected wave field dataset, detect whether the detection environment has undergone significant changes; When the significant changes are detected, a convolutional neural network model is used to identify the dominant disturbance type and the geological stratum adjustment requirements, and to generate targeted detection parameter optimization instructions. Based on the targeted detection parameter optimization instructions, generate and issue instructions for adjusting the excitation source frequency and receiver array geometry synchronization to each detection platform; Based on the new round of wavefield response data collected by each detection platform after adjustment, a spatiotemporally aligned joint observation dataset is generated by fusion. The spatiotemporally aligned joint observation dataset is subjected to reverse time migration imaging to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image. Output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

2. The method according to claim 1, characterized in that, The process of acquiring raw time-series data using a distributed sensor array includes: A clock synchronization protocol is used to synchronously acquire multi-channel analog signals and generate multi-channel synchronous digital waveform data. The active signal range is extracted based on the amplitude change rate of the multi-channel synchronous digital waveform data; Based on the frequency distribution characteristics of the active signal range, it is divided into surface vibration components and subsurface wave field response components. The divided components are subjected to spectral aliasing analysis and wavefront snapshot inversion to generate a feature signal sequence with physical attribute annotations.

3. The method according to claim 1, characterized in that, The process of performing time-frequency decomposition on the raw time-series data includes: The generalized S-transform algorithm is used to map the original time series data into a two-dimensional time-frequency distribution matrix; The local spectral energy density is calculated based on the two-dimensional time-frequency distribution matrix, and the characteristics of effective reflected signals are distinguished by singular value decomposition. A denoising time-frequency mask is constructed and subjected to a Hadamard product operation with the two-dimensional time-frequency distribution matrix to retain effective reflected signal components. Perform the inverse generalized S-transform operation on the retained effective reflected signal components.

4. The method according to claim 1, characterized in that, The process of detecting whether the detection environment has undergone significant changes based on a purified reflected wavefield dataset includes: The time-frequency distribution matrix of the purified reflected wave field dataset was obtained by decomposing it using the short-time Fourier transform algorithm. Extract the instantaneous main frequency value based on the time-frequency distribution matrix and calculate the main frequency offset distribution map; Calculate the amplitude attenuation rate data based on the main frequency offset distribution map; The main frequency offset distribution map and amplitude attenuation rate data are used to construct an environmental difference feature vector, and a judgment is made based on whether its modulus value exceeds a preset anomaly judgment threshold.

5. The method according to claim 1, characterized in that, The process of generating and distributing synchronization adjustment instructions to each detection platform includes: Acquire and process environmental data to obtain spectral response characteristics; The excitation source frequency value is calculated based on the spectral response characteristics, and the geometric topology of the receiver array is generated by combining the platform spatial distribution data. Based on the aforementioned geometric topology, transmission delay data is calculated to correct synchronization timing deviations; The corrected synchronization timing deviation and excitation source frequency value are integrated to generate an adjustment instruction sequence, which is then encapsulated into an adjustment instruction set.

6. The method according to claim 1, characterized in that, The process of fusing and generating a spatiotemporally aligned joint observation dataset includes: Acquire the initial wavefield data packets uploaded from each platform, which include wavefield response signals, spatial coordinates, and acquisition time labels; The clock deviation is calculated based on the acquisition time tag and the additional transmission delay value, and the time synchronization observation sequence is obtained by correcting the clock deviation. Based on the spatial coordinates in the time-synchronized observation sequence, grid mapping is performed to generate spatially aligned grid data; Aggregate all spatially aligned grid data.

7. A multi-modal geophysical seismic exploration system, characterized in that, The system for implementing the method of any one of claims 1-6 comprises: The data acquisition module is used to acquire raw time-series data containing surface vibration and subsurface wave field response through a distributed sensor array; The data processing module is used to perform time-frequency decomposition processing on the original time-series data to separate the effective reflection signal components and obtain a purified reflection wave field dataset. The environmental detection module is used to detect whether the detection environment has undergone significant changes based on the purified reflected wave field dataset; The decision optimization module is used to identify the dominant disturbance type and geological stratum adjustment needs using a convolutional neural network model when the significant change is detected, and to generate targeted detection parameter optimization instructions. The collaborative control module is used to generate and send excitation source frequency and receiver array geometric synchronization adjustment instructions to each detection platform based on the targeted detection parameter optimization instructions; The data fusion module is used to fuse the new round of wavefield response data collected by each detection platform after adjustment to generate a spatiotemporally aligned joint observation dataset. The imaging reconstruction module is used to perform reverse time migration imaging processing on the spatiotemporally aligned joint observation dataset to reconstruct the subsurface medium structure and obtain a high-resolution geological structure image. The results output module is used to output a three-dimensional description of underground structures to support resource assessment and engineering decisions.

8. A computer terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the steps of the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-6.

Images