Irregular electromagnetic data processing method and device, electronic device, and storage medium

Abstract

The present disclosure provides a method and device for processing irregular electromagnetic data, an electronic device, and a storage medium, which belongs to the field of data processing technology. The method includes: processing multiple first acquisition data of each electromagnetic wave detection device based on the first configuration information of multiple electromagnetic wave detection devices to obtain multiple first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes the location information and parameter configuration information of the electromagnetic wave detection device; fusing the multiple first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map. The method and device for processing irregular electromagnetic data, the electronic device, and the storage medium provided by the present disclosure can improve the accuracy and efficiency of electromagnetic data processing.

Description

Technical Field

[0001] The present disclosure belongs to the field of data processing technology, and more specifically, relates to a method and device for processing irregular electromagnetic data, an electronic device, and a storage medium. Background Art

[0002] In geological exploration and engineering applications, electromagnetic data processing is a key link in understanding underground structures, identifying mineral resources, and assessing geological risks. However, traditional electromagnetic data processing methods are often based on idealized assumptions, that is, the detection devices are arranged regularly and the parameters are consistent, which is difficult to achieve in a complex and changeable natural environment. Especially in irregular terrains such as mountains and hills, the location of electromagnetic wave detection devices is difficult to plan uniformly, and due to environmental factors, the parameter configuration of each device also needs to be flexibly adjusted.

[0003] Most existing electromagnetic data processing technologies ignore these irregular factors, resulting in large position deviations and parameter errors in the processing results, making it difficult to accurately reflect the true distribution of underground geological waves. In addition, data collected by different detection devices often exist in isolation, lacking an effective fusion mechanism, and unable to form a comprehensive and continuous geological wave information map, limiting the depth and breadth of geological structure research. Summary of the invention

[0004] The purpose of the present disclosure is to provide a method and device for processing irregular electromagnetic data, an electronic device, and a storage medium to improve the accuracy and efficiency of electromagnetic data processing.

[0005] A first aspect of an embodiment of the present disclosure provides a method for processing irregular electromagnetic data, comprising:

[0006] Based on the first configuration information of the plurality of electromagnetic wave detection devices, multiple first collected data of each electromagnetic wave detection device are processed to obtain a plurality of first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes the position information and parameter configuration information of the electromagnetic wave detection device;

[0007] A plurality of first target data corresponding to each electromagnetic wave detection device are fused to generate a geological wave data map.

[0008] A second aspect of the embodiments of the present disclosure provides an irregular electromagnetic data processing device, comprising:

[0009] a data processing module, configured to process a plurality of first collected data of each electromagnetic wave detection device based on first configuration information of the plurality of electromagnetic wave detection devices to obtain a plurality of first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes position information and parameter configuration information of the electromagnetic wave detection device;

[0010] Data fusion module: used to fuse multiple first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map.

[0011] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned irregular electromagnetic data processing method when executing the computer program.

[0012] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned irregular electromagnetic data processing method are implemented.

[0013] The non-regular electromagnetic data processing method and device, electronic device, and storage medium provided by the embodiments of the present disclosure have the following beneficial effects:

[0014] First, the disclosed embodiment processes the collected data based on the first configuration information of multiple electromagnetic wave detection devices, which can fully take into account the location differences and parameter setting differences of different detection devices. The use of location information can better adapt to different terrains, landforms and geological environments, accurately evaluate the impact of environmental factors on data, and effectively eliminate errors caused by location. The inclusion of parameter configuration information improves the accuracy and reliability of the data. Secondly, the multiple first target data corresponding to each electromagnetic wave detection device are fused to generate a geological wave data map, which realizes the comprehensive integration of data, can present more complete and accurate geological wave information, and provide rich basic data for geological structure research. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 A schematic diagram of a flow chart of a method for processing irregular electromagnetic data provided by an embodiment of the present disclosure;

[0017] Figure 2 A flowchart of a method for processing irregular electromagnetic data provided by another embodiment of the present disclosure;

[0018] Figure 3 A schematic diagram of local spherical coordinates of a viewpoint provided in another embodiment of the present disclosure;

[0019] Figure 4A structural block diagram of an irregular electromagnetic data processing device provided by an embodiment of the present disclosure;

[0020] Figure 5 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0021] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0022] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0023] Please refer to Figure 1 , Figure 1 A flowchart of a method for processing irregular electromagnetic data provided by an embodiment of the present disclosure, the method comprising:

[0024] S101: Based on first configuration information of multiple electromagnetic wave detection devices, multiple first collected data of each electromagnetic wave detection device are processed to obtain multiple first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes position information and parameter configuration information of the electromagnetic wave detection device.

[0025] In this embodiment, the electromagnetic wave detection device can use the propagation characteristics of the physical phenomenon of seismic waves to conduct underground exploration and monitoring. By measuring the speed and direction of seismic waves propagating underground, the distribution, thickness, lithology and other information of underground rock layers can be revealed, providing basic data for geological structure research. The electromagnetic wave detection device can predict geological anomalies in front of the surrounding rock, such as faults, water-bearing fissures, and thinning areas of coal seams, providing safety guarantees for engineering design and construction. It can also be used to detect information such as the distribution, flow direction, flow velocity and water quality of groundwater, providing a scientific basis for the rational development and utilization of water resources.

[0026] Multiple electromagnetic wave detection devices perform data collection according to their own first configuration information. Each detection device is deployed at a specific geographical location. These locations can be determined based on pre-planned grids or specific exploration needs. Each detection device has different setting parameters, such as sensitivity, sampling rate, etc., which affect the quality and type of data.

[0027] Each electromagnetic wave detection device can perform multiple data collection (i.e., perform data collection for different viewing angles), which can increase the reliability of the data. After multiple data collections, the collected first collection data is processed according to the first configuration information, useful information is extracted, and multiple first target data corresponding to each electromagnetic wave detection device are obtained. The first target data more directly reflects the information of the geological structure or electromagnetic characteristics.

[0028] In one embodiment of the present disclosure, multiple first acquisition data of each electromagnetic wave detection device are processed based on the first configuration information of the multiple electromagnetic wave detection devices to obtain multiple first target data corresponding to each electromagnetic wave detection device, including:

[0029] Based on the position information and parameter configuration information of the electromagnetic wave detection device, error correction is performed on the multiple first collected data to obtain multiple second target data;

[0030] A plurality of first target data corresponding to each electromagnetic wave detection device is obtained based on the plurality of second target data.

[0031] In this embodiment, when multiple electromagnetic wave detection devices are collecting data, due to different location information and parameter configuration information, the first collected data may have errors. The location information determines the specific environment in which the detection device is located. Different locations may be affected by factors such as terrain, landform, and meteorology, resulting in data deviations. The parameter configuration information determines the working characteristics of the detection device, such as sensitivity, frequency response, etc. Different parameter settings may also cause differences in the data.

[0032] Therefore, in order to eliminate these errors, the error correction of multiple first collected data is performed according to the position information and parameter configuration information of the detection device to obtain multiple second target data. By analyzing the influence of position and parameters on the data, an error model can be established by dynamic error modeling. This model can accurately describe the errors caused by various factors to the data, and then perform adaptive error compensation, thereby effectively eliminating the detection errors caused by factors such as terrain, landform, and meteorology, and improving the accuracy and reliability of the data.

[0033] For example, in a large-scale urban rail transit construction project, it is necessary to conduct detailed surveys of the underground geological conditions along the subway line to ensure the safe construction of the subway tunnel and its subsequent stable operation.

[0034] According to the planning of the subway line and the needs of geological exploration, an electromagnetic wave detection device is deployed at a certain distance along the subway line. The locations of these devices are accurately recorded, including information such as longitude and latitude and altitude. Each detection device is configured with specific parameters according to the geological characteristics of its location and the detection task. For example, in areas close to rivers, the sensitivity of the detection device may be increased to better detect the impact of groundwater; while in areas with thicker rock layers, the sampling rate may be adjusted to obtain more detailed geological wave information. Data collection is carried out in different time periods, such as day and night, different seasons, etc., to capture the changes in geological conditions over time. At the same time, data collection is carried out from different angles, such as changing the transmission angle and receiving direction of the detection device to obtain more comprehensive underground information.

[0035] Since the detection devices at different locations are affected by factors such as terrain, landform, and weather, the first collected data may contain errors. For example, in low-lying areas, the interference of groundwater leads to data deviation; in areas with tall buildings, the influence of buildings changes the propagation path of geological waves.

[0036] According to the location information and parameter configuration information of the detection device, an error model is established by dynamic error modeling to analyze the influence of location and parameters on the data. For location factors, the influence of factors such as terrain undulation, geological structure, and groundwater level on the data can be considered. For parameter configuration factors, the error characteristics under different parameter settings can be determined through experiments or simulation analysis. According to the established error model, the first collected data is adaptively compensated for the error. For example, for data at a location that is greatly affected by the terrain, the data is corrected according to the terrain height and the error model; for data with unreasonable parameter settings, the parameters are adjusted or the data is compensated according to the relationship between the parameters and the errors. By continuously adjusting the error model and compensation method, the accuracy and reliability of the data can be improved.

[0037] In one embodiment of the present disclosure, multiple first collected data are error corrected using a first model formula to obtain multiple second target data;

[0038] The first model formula is:

[0039]

[0040] in,

[0041] Represented as the second target data, it is the data after error correction, which can more accurately reflect the information such as geological structure or electromagnetic characteristics;

[0042] It is represented as the first collected data;

[0043] The device error of the electromagnetic wave detection device is expressed as the device error. The device error will vary over time and in different spatial locations. For example, as the device is used for a longer time, its performance may change, resulting in different device errors at different time points; and at different spatial locations, due to the influence of environmental factors, the working state of the device is also different, resulting in different device errors;

[0044] Expressed as external factor error, it is the time, space position and external parameters Function, external parameters These factors may include terrain, wind speed, temperature, etc. These external factors may affect the propagation of geological waves and the operation of the detection device, resulting in data errors. For example, in areas with complex terrain, the propagation path of geological waves may change, affecting the accuracy of the data; factors such as wind speed and temperature may also affect the performance of the detection device, resulting in errors.

[0045] Expressed as an adjustment factor to take into account the noise variance The impact on the second target data, the noise variance reflects the uncertainty of the data, and the adjustment factor can adjust the strength of the error correction according to the size of the noise. When the noise is large, the adjustment factor will make the correction stronger to improve the overall reliability of the data; noise variance This means that the noise will be different at different times and spatial locations. For example, at certain times or locations, there may be more interference, resulting in an increase in noise variance.

[0046] , , , and Average time and spatial location functions that vary with time and spatial position.

[0047] In this embodiment, firstly, the data acquisition time and spatial position of the electromagnetic wave detection device are used as a dimension of the error model, that is, the characteristics of data changing over time and the spatial non-uniformity of data are considered. Secondly, by using the noise variance as the denominator of the first model formula, a stronger correction is performed on the data with large noise to improve the overall reliability of the data. Finally, this embodiment integrates external parameters into the error model to more comprehensively consider the impact of external factors on data errors.

[0048] In this embodiment, the device error of the electromagnetic wave detection device is Over time and device location In this embodiment, the device error can be modeled as a weighted sum of multiple time-related and position-related factors, and a nonlinear function is used to capture the relationship between the first acquisition data and the error. It is expressed as:

[0049]

[0050] in:

[0051] It is expressed as a time coefficient related to time. The time coefficient reflects the change of equipment error over time. For example, as the equipment is used for a longer time, it will age and perform poorly, causing the equipment error to change. The time coefficient can be obtained through time series analysis or machine learning models (such as LSTM). Time series analysis can analyze historical data to find out the trends, seasonality and other laws in the time series, thereby predicting equipment errors at different time points.

[0052] Expressed as a location coefficient related to location, the location coefficient reflects the difference in device error at different spatial locations. Different geographical locations have different geological conditions, environmental factors, etc. These factors will affect the working state of the equipment, resulting in different device errors. The location coefficient can be obtained through spatial interpolation or machine learning models (such as convolutional neural networks (CNNs) to process image data). Spatial interpolation can estimate the device error at an unknown location through interpolation methods based on the device error data at a known location. Machine learning models such as convolutional neural networks can learn features in spatial data and better capture the relationship between device error and spatial location.

[0053] It is expressed as a nonlinear function to capture the relationship between the first acquisition data and the error. Since the relationship between the first acquisition data and the device error is not linear, the introduction of a nonlinear function can more accurately describe this relationship. The nonlinear function can be a polynomial, an exponential function, etc. Different nonlinear functions can be selected according to actual conditions to better fit the relationship between the first acquisition data and the device error. For example, if the relationship between the first acquisition data and the device error presents the characteristics of a polynomial, a polynomial function can be selected as the nonlinear function.

[0054] In this embodiment, by modeling the device error as a weighted sum of the time coefficient and the position coefficient, the changes in the device error in time and space are fully considered. In this way, the device error can be evaluated and corrected more accurately, and the accuracy and reliability of the data can be improved. The nonlinear function can better capture the complex relationship between the first acquisition data and the device error.

[0055] In this embodiment, the external factor error is time, space position and external parameters (such as terrain, wind speed, temperature, etc.), this embodiment can represent the external parameters through Gaussian process regression (GPR) or Bayesian network modeling The nonlinear relationship between the error and external factors It is expressed as:

[0056]

[0057] in,

[0058] It is represented as a Gaussian process regression model, which is used to predict the external factor error based on the input time, spatial location and external parameters. It can naturally handle uncertainty and give confidence intervals for the prediction results. Expressed as hyperparameters of the Gaussian process regression model.

[0059] In this embodiment, Gaussian process regression can give a confidence interval for the prediction result, which is very helpful for dealing with the uncertainty of external factors. Since external factors such as terrain, wind speed, temperature, etc. often have a certain degree of uncertainty, Gaussian process regression can take this uncertainty into account while predicting the error, thereby improving the accuracy and reliability of the prediction.

[0060] In this embodiment, the noise variance can be calculated by a heteroscedastic Gaussian process (HGP) .

[0061]

[0062] in, It is represented as a heteroscedastic Gaussian process model, which predicts the logarithm of the noise variance according to the temporal and spatial location of the input. Expressed as hyperparameters of the heteroskedastic Gaussian process model.

[0063] In this embodiment, the noise variance reflects the uncertainty of the data. In data processing, noise will affect the accuracy and reliability of the data, so it is necessary to estimate and correct the noise. Through the heteroscedastic Gaussian process model, the noise variance can be predicted according to the changes in time and spatial position, so as to better adapt to the noise conditions at different times and positions and improve the effect of data processing.

[0064] In one embodiment of the present disclosure, obtaining a plurality of first target data corresponding to each electromagnetic wave detection device based on a plurality of second target data includes:

[0065] The second target data is time-aligned and space-aligned to obtain a plurality of first target data corresponding to each electromagnetic wave detection device.

[0066] In this embodiment, after the error-corrected second target data is obtained, it is necessary to further perform time alignment and space alignment on it to obtain a plurality of first target data corresponding to each electromagnetic wave detection device.

[0067] Time alignment: Multiple electromagnetic wave detection devices may collect data at different times. In order to effectively integrate and analyze these data, their time bases need to be unified. According to the timestamp and Beidou positioning of the collected data, dynamic time regularization is used to process the data collected by different electromagnetic wave detection devices to the same time scale to ensure the consistency of the data in time.

[0068] Spatial alignment: The data collected by detection devices at different locations need to be spatially aligned and matched. First, the specific location of each detection device in space is determined by location information. Then, all data are mapped to a unified map at the same time through spatial geometric transformation. This makes the geological wave data continuous and consistent in space, providing a basis for subsequent data fusion and analysis.

[0069] In this embodiment, time alignment eliminates the timing differences caused by different data collection times, while space alignment ensures the consistency and continuity of the data of each device in the spatial coordinates. This not only improves the comparability of the data, but also enhances the accuracy of data fusion and analysis.

[0070] S102: Fusion the multiple first target data corresponding to the various electromagnetic wave detection devices to generate a geological wave data map.

[0071] In this embodiment, the multiple first target data obtained by each electromagnetic wave detection device have a certain degree of independence, but they all reflect the geological wave characteristics of different positions. In order to obtain more comprehensive and accurate geological wave information, these first target data need to be fused. The fusion process can comprehensively consider factors such as the positional relationship of each detection device and the reliability weight of the data. Through specific algorithms and data processing techniques, the first target data from different detection devices are integrated to generate a geological wave data map. This can more intuitively display the geological wave distribution of the detected area, providing an important reference for geological research, resource exploration and other fields.

[0072] From the above, it can be concluded that, firstly, this embodiment can fully take into account the location differences and parameter setting differences of different detection devices by processing the collected data based on the first configuration information of multiple electromagnetic wave detection devices. The use of location information can better adapt to different terrains, landforms and geological environments, accurately evaluate the impact of environmental factors on data, and effectively eliminate errors caused by location. The inclusion of parameter configuration information improves the accuracy and reliability of the data. Secondly, the multiple first target data corresponding to each electromagnetic wave detection device are fused to generate a geological wave data map, which realizes the comprehensive integration of data, can present more complete and accurate geological wave information, and provide rich basic data for geological structure research.

[0073] like Figure 2 As shown, in one embodiment of the present disclosure, a plurality of first target data corresponding to each electromagnetic wave detection device are fused to generate a geological wave data map, including:

[0074] S201: Acquire focus points of a plurality of first target data.

[0075] In this embodiment, the focus points are obtained from the multiple first target data corresponding to each electromagnetic wave detection device. These focus points can be locations with important features or potential values ​​in the data, such as areas with abnormal geological structures, areas with obvious changes in electromagnetic signals, etc.

[0076] S202: Rendering a local neighborhood of the focus point based on the coordinate information of the focus point to obtain multiple view patches.

[0077] In one embodiment of the present disclosure, a local neighborhood of a point of interest is rendered based on coordinate information of the point of interest to obtain multiple view patches, including:

[0078] Based on the SoftRas renderer, the local neighborhood of the focus point is rendered to obtain multiple view patches.

[0079] In this embodiment, based on the coordinate information of the focus point, the SoftRas renderer framework can be used to render the local neighborhood of the focus point to obtain multiple view patches.

[0080] The viewpoint is used as a parameter that can be optimized during training, allowing the rendering signal to be back-propagated from rendered pixels to the viewpoint, thus achieving joint optimization of the rendering stage with other stages.

[0081] The specific steps of the SoftRas renderer framework to achieve joint optimization can involve the following key links:

[0082] 1. Data preprocessing

[0083] Input data collection: First, collect input data, which can include geometry, vertex appearance attributes, camera information, and lighting conditions.

[0084] Data cleaning and format conversion: Clean the collected data, remove noise and outliers, and convert it into a format that can be processed by the SoftRas renderer.

[0085] 2. Modeling

[0086] Rendering model construction: Use the SoftRas renderer framework to build a rendering model. SoftRas simulates the rasterization and Z-Buffering process of the standard graphics rendering pipeline through differentiable mathematical functions to ensure that the entire rendering process is differentiable.

[0087] Joint optimization model: According to the specific application scenario, a joint optimization model is designed. The joint optimization model can consider the output of the rendering stage and the other two stages at the same time to coordinate during the optimization process.

[0088] 3. Training and Joint Optimization

[0089] Initialization parameters: Set initial parameters for rendering models and other stages of the model.

[0090] Forward propagation: Forward propagation through the SoftRas renderer and other stages of the model to generate rendered images and other outputs.

[0091] Backpropagation and gradient calculation: Using the differentiable characteristics of the SoftRas renderer, the monitoring signal (such as the error of the rendered pixel) is backpropagated from the rendered pixel to the viewpoint and the gradient is calculated.

[0092] Parameter update: Update the parameters of the rendering model and other stage models according to the gradient information to achieve joint optimization.

[0093] Iterative training: Repeat the process of forward propagation, backpropagation, and parameter update until the convergence condition is met or the preset training rounds are reached.

[0094] 4. Verification and evaluation

[0095] Use validation data: During or after training, use validation data to evaluate the jointly optimized model.

[0096] Performance evaluation: Evaluate the model’s rendering quality, optimization effect, and generalization ability in different scenarios.

[0097] 5. Deployment and Application

[0098] Model deployment: Deploy the trained joint optimization model to the application environment.

[0099] Real-time monitoring and optimization: In actual applications, the real-time monitoring function of the SoftRas renderer is used to continuously collect feedback signals and fine-tune or re-optimize the model as needed.

[0100] In this embodiment, the plurality of first target data are multi-perspective viewpoints of each electromagnetic wave detection device. In this embodiment, the multi-perspective viewpoints can be represented by P.

[0101] In this embodiment, spherical coordinates are used to represent the regional multi-viewpoints, and each viewpoint is represented by two angles, a distance to a local origin, and a consistent upright direction. In this way, the local area of ​​the focus point can be observed and rendered from different angles and positions, providing multi-viewpoint data for subsequent feature extraction.

[0102] like Figure 3 As shown, the renderer settings are:

[0103] Using spherical coordinates Represents multi-viewpoints.

[0104] in, Expressed as the number of viewpoints, and Expressed as two angles from the viewpoint, Expressed as the distance from the viewpoint to the local origin, Represents a consistent upright orientation.

[0105] In this embodiment, a consistent upright direction refers to a fixed direction commonly recognized by observers at different perspectives during the rendering process, which is perpendicular to the radial distance of the spherical coordinates. However, it should be noted that the "upright direction" here is not geographically or physically upright in an absolute sense (i.e., perpendicular to the surface of the earth), but a certain logical or conventional upright direction relative to the rendered scene or object. In multi-perspective rendering, maintaining a consistent upright direction is critical to ensuring the consistency and accuracy of the rendering results. This means that no matter from which perspective the observer observes the scene, they should be able to recognize the same upright direction to understand the spatial relationships and object layouts in the scene.

[0106] To enforce the rotation invariance of each viewpoint feature, each rendered view patch is rotated in-plane by 90° intervals.

[0107] Rendering process: Each point is represented as a sphere, Represented as an arbitrary point on a sphere, the radius of which can be a fixed value or based on The average distance to its local neighbors is derived.

[0108] Use a specific viewpoint After perspective projection, each point Generate a probability plot , describing each output pixel being The probability of coverage is then rendered to the output view patch .

[0109] Each pixel in the view patch is calculated by the following formula:

[0110]

[0111] in,

[0112] Represented as a view patch No. pixels, Represented as a point Rendering properties (such as color or depth of field), Represents the default background value, Represented as a point The depth of Expressed as a weight function based on the probability of coverage and depth, Denoted as background weight, Expressed as The probability of covering pixels.

[0113] Traditional graphics rendering has the problem of non-differentiability, which limits its application in neural network training. This embodiment improves the SoftRas framework by considering the spherical representation of points and the probability map after perspective projection, making it more consistent with the actual graphics skeleton and differentiable.

[0114] S203: Perform feature extraction on the multiple view patches to obtain feature maps corresponding to the multiple view patches.

[0115] In one embodiment of the present disclosure, extracting features from a plurality of view patches to obtain feature maps corresponding to the plurality of view patches includes:

[0116] Feature extraction is performed on multiple view patches through a convolutional neural network to obtain feature maps corresponding to the multiple view patches.

[0117] In one embodiment of the present disclosure, feature extraction is performed on multiple view patches through a convolutional neural network to obtain feature maps corresponding to the multiple view patches, including:

[0118] In response to multiple view patches being fed into a convolutional neural network, feature information is obtained;

[0119] A feature map corresponding to a plurality of view patches is obtained based on the feature information.

[0120] In this embodiment, the obtained multiple view patches can be feature extracted by a convolutional neural network (CNN). Each view patch is fed into a convolutional neural network consisting of six stacked convolutional layers, each of which is followed by a normalization and ReLU layer. Through such a network structure, features can be extracted from each view patch. Rich feature information is extracted from the image, and feature maps corresponding to multiple view patches are obtained based on the feature information. These feature maps can be of size 8×8 with 128 channels, containing various features of different view patches such as color, texture, depth, etc.

[0121] In one embodiment of the present disclosure, in response to multiple view patches being fed to a convolutional neural network for feature extraction, feature information is obtained, including:

[0122] If the number of view patches is less than the first number, feeding the view patches to a convolutional neural network for feature extraction to obtain feature information;

[0123] If the number of view patches is greater than or equal to the first number, the multiple view patches are concatenated to obtain a target view patch, and the target view patch is fed to a convolutional neural network for feature extraction to obtain feature information.

[0124] In this embodiment, when the number of view patches is less than the first number, the view patches are directly fed to the convolutional neural network for feature extraction. Since the number of view patches is small, no additional splicing operation is required, and each view patch can be directly processed by the convolutional neural network. The convolutional neural network extracts features from the input view patches through its internal structures such as multiple convolutional layers, normalization layers, and ReLU layers. For example, the convolutional layer can extract features at different levels in the view patch, such as edges, textures, and other information; the normalization layer can normalize the data to improve the stability and convergence speed of the model; the ReLU layer can introduce nonlinear factors to enhance the expressiveness of the model. Through these operations, feature information is obtained.

[0125] When the number of view patches is greater than or equal to the first number, the multiple view patches are first spliced. In this way, the multiple view patches can be integrated into a larger target view patch to better utilize the processing power of the convolutional neural network. The splicing method can be designed according to specific needs, such as stacking, horizontally or vertically in a specific order. The spliced ​​target view patch is then fed to the convolutional neural network for feature extraction. The convolutional neural network processes the target view patch and extracts feature information. Since the target view patch contains information from multiple view patches, the convolutional neural network can extract features from a more macro perspective and capture the relationship between different view patches and the overall feature pattern.

[0126] In this embodiment, by flexibly processing different numbers of view patches, the capabilities of the convolutional neural network are effectively utilized, and the comprehensiveness and efficiency of feature extraction are improved, so that geological structure characteristics can be captured more accurately.

[0127] In an embodiment of the present disclosure, the method further includes: determining the first number based on the overlap of the plurality of view patches.

[0128] In this embodiment, the overlap of view patches refers to the degree of overlap of multiple view patches in space or features. A higher overlap means that there is more repeated information between multiple view patches, while a lower overlap means that the differences between the view patches are large. In this embodiment, the number of view patches to be processed can be reasonably adjusted according to the actual situation of the view patches to achieve the best feature extraction effect and calculation efficiency.

[0129] Specifically, first, the overlap between multiple view patches is calculated. The specific calculation method can be implemented by comparing the boundaries, pixel values ​​or features between patches. For example, the overlap between the features of the view patches can be measured by calculating the similarity between the feature vectors.

[0130] Then, the distribution of overlap is analyzed. For example, if the overlap between most patches is high, it means that they contain a lot of redundant information, and more patches can be spliced ​​together for processing. Based on the analysis results of the overlap, a threshold is set to determine the first number. This threshold can be set according to the requirements of the specific task, the limitations of computing resources, and the expected model performance. For example, an overlap threshold can be set, and when the overlap between two patches exceeds this threshold, they are considered to be spliced ​​together.

[0131] Finally, the first number is determined based on the set threshold and the result of the overlap calculation. This number represents the maximum number of view patches that can be processed individually before reaching a certain overlap threshold. If the number of patches in a certain area exceeds this number and the overlap between them is high enough, then these patches should be spliced ​​into a larger target patch for processing.

[0132] In this embodiment, by determining the first number according to the overlap, it is possible to avoid wasting computing resources when processing a large amount of view patches with repeated information, thereby improving the efficiency of feature extraction. At the same time, it is also possible to ensure that when the view patches are quite different, enough view patches are processed to fully extract feature information, thereby ensuring the effect of feature extraction.

[0133] S204: Obtaining a geological wave data map based on the characteristic map.

[0134] In one embodiment of the present disclosure, obtaining a geological wave data map based on a characteristic map includes:

[0135] The feature maps corresponding to multiple view patches are weightedly aggregated based on the weights of the feature maps to obtain a geological wave data map.

[0136] In this embodiment, a geological wave data map is obtained based on the extracted feature map. First, in order to fuse the features of all views, a soft view pooling module is designed.

[0137] The soft view pooling module is a technique for multi-view feature fusion, which achieves high-precision multi-view map fusion by adaptively aggregating all responses with the attention weights estimated by the sub-network. The following are the specific steps to implement this module:

[0138] Based on the sub-network, the feature maps corresponding to multiple view patches are evaluated to obtain the weight of each feature map;

[0139] The feature maps corresponding to multiple view patches are weightedly aggregated based on the weights of the feature maps to obtain a geological wave data map.

[0140] In this embodiment, the sub-network converts each feature map As input, the encoder-decoder design is followed to regress the corresponding weights, and after a series of convolutional layers and upconvolutional layers and the intermediate ReLU layer, the spatial size and channel depth are downsampled and upsampled to output the weight map. Then, for each position, the softmax function is applied to normalize the weights. Finally, the fused features are embedded into the multidimensional feature space through the fully connected layer and the normalization layer to generate the geological wave data map. The geological wave data map can integrate the information of multiple first target data and present the distribution of geological waves in a high-precision manner, providing strong support for geological research and engineering applications.

[0141] In this embodiment, by introducing the soft view pool module, using the sub-network to evaluate the weight of each view patch feature map, and performing weighted aggregation, a geological wave data map integrating multiple view information is generated. This not only improves the accuracy of multi-view data fusion, but also enhances the visualization effect of geological wave distribution, providing more comprehensive and accurate data support for geological exploration, engineering design and related research, and greatly promoting the interpretation and application of geological information.

[0142] Corresponding to the irregular electromagnetic data processing method of the above embodiment, Figure 4 This is a structural block diagram of an irregular electromagnetic data processing device provided by an embodiment of the present disclosure. For ease of description, only the parts related to the embodiment of the present disclosure are shown. Figure 2The irregular electromagnetic data processing device 20 includes: a data processing module 21 and a data fusion module 22 .

[0143] The data processing module 21 is used to process multiple first collected data of each electromagnetic wave detection device based on the first configuration information of the multiple electromagnetic wave detection devices to obtain multiple first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes the position information and parameter configuration information of the electromagnetic wave detection device;

[0144] The data fusion module 22 is used to fuse the multiple first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map.

[0145] In one embodiment of the present disclosure, the data processing module 21 is specifically used for:

[0146] Based on the position information and parameter configuration information of the electromagnetic wave detection device, error correction is performed on the multiple first collected data to obtain multiple second target data;

[0147] A plurality of first target data corresponding to each electromagnetic wave detection device is obtained based on the plurality of second target data.

[0148] In one embodiment of the present disclosure, the data processing module 21 is further configured to:

[0149] The second target data is time-aligned and space-aligned to obtain a plurality of first target data corresponding to each electromagnetic wave detection device.

[0150] In one embodiment of the present disclosure, the data fusion module 22 is specifically used to:

[0151] Acquire a plurality of focus points of first target data;

[0152] Rendering a local neighborhood of the focus point based on the coordinate information of the focus point to obtain multiple view patches;

[0153] Performing feature extraction on the multiple view patches to obtain feature maps corresponding to the multiple view patches;

[0154] A geological wave data map is obtained based on the characteristic map.

[0155] In one embodiment of the present disclosure, the data fusion module 22 is further configured to:

[0156] Based on the coordinate information of the point of interest, the local neighborhood of the point of interest is rendered to obtain multiple view patches, including:

[0157] Based on the SoftRas renderer, the local neighborhood of the focus point is rendered to obtain multiple view patches.

[0158] In one embodiment of the present disclosure, the data fusion module 22 is further configured to:

[0159] Feature extraction is performed on multiple view patches through a convolutional neural network to obtain feature maps corresponding to the multiple view patches.

[0160] In one embodiment of the present disclosure, the data fusion module 22 is further configured to:

[0161] The feature maps corresponding to multiple view patches are weightedly aggregated based on the weights of the feature maps to obtain a geological wave data map.

[0162] See also Figure 5 , Figure 5 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 3 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of the modules in the above-mentioned device embodiments, such as Figure 4 The functions of modules 21 to 22 are shown.

[0163] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0164] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0165] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0166] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of the irregular electromagnetic data processing method provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device described in the embodiments of the present disclosure, which will not be repeated here.

[0167] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0168] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0169] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0170] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0171] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0172] The units 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 on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0173] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0174] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A method for processing irregular electromagnetic data, characterized in that: include: Based on the first configuration information of the plurality of electromagnetic wave detection devices, multiple first collected data of each electromagnetic wave detection device are processed to obtain a plurality of first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes the position information and parameter configuration information of the electromagnetic wave detection device; Fusing a plurality of first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map; The step of fusing a plurality of first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map includes: Acquire a plurality of focus points of first target data; Rendering a local neighborhood of the focus point based on the coordinate information of the focus point to obtain a plurality of view patches; Performing feature extraction on the multiple view patches to obtain feature maps corresponding to the multiple view patches; Obtaining a geological wave data map based on the characteristic map; The method of processing a plurality of first collected data of each electromagnetic wave detection device based on the first configuration information of the plurality of electromagnetic wave detection devices to obtain a plurality of first target data corresponding to each electromagnetic wave detection device includes: Based on the position information and parameter configuration information of the electromagnetic wave detection device, error correction is performed on the multiple first collected data to obtain multiple second target data; Based on the plurality of second target data, a plurality of first target data corresponding to each electromagnetic wave detection device is obtained; The obtaining of a plurality of first target data corresponding to each electromagnetic wave detection device based on the plurality of second target data comprises: The second target data is time-aligned and space-aligned to obtain a plurality of first target data corresponding to each electromagnetic wave detection device.

2. The irregular electromagnetic data processing method according to claim 1, characterized in that: The rendering of the local neighborhood of the point of interest based on the coordinate information of the point of interest to obtain a plurality of view patches includes: The local neighborhood of the focus point is rendered based on the SoftRas renderer to obtain multiple view patches.

3. The irregular electromagnetic data processing method according to claim 1, characterized in that: The extracting features from the multiple view patches to obtain feature maps corresponding to the multiple view patches includes: Perform feature extraction on the multiple view patches through a convolutional neural network to obtain feature maps corresponding to the multiple view patches.

4. The irregular electromagnetic data processing method according to claim 1, characterized in that: The step of obtaining a geological wave data map based on the characteristic map includes: The feature maps corresponding to the multiple view patches are weightedly aggregated based on the weights of the feature maps to obtain a geological wave data map.

5. An irregular electromagnetic data processing device, characterized in that: include: a data processing module, configured to process a plurality of first collected data of each electromagnetic wave detection device based on first configuration information of the plurality of electromagnetic wave detection devices to obtain a plurality of first target data corresponding to each electromagnetic wave detection device, wherein the first configuration information includes position information and parameter configuration information of the electromagnetic wave detection device; Data fusion module: used for fusing a plurality of first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map; The step of fusing a plurality of first target data corresponding to each electromagnetic wave detection device to generate a geological wave data map includes: Acquire a plurality of focus points of first target data; Rendering a local neighborhood of the focus point based on the coordinate information of the focus point to obtain a plurality of view patches; Performing feature extraction on the multiple view patches to obtain feature maps corresponding to the multiple view patches; Obtaining a geological wave data map based on the characteristic map; The method of processing a plurality of first collected data of each electromagnetic wave detection device based on the first configuration information of the plurality of electromagnetic wave detection devices to obtain a plurality of first target data corresponding to each electromagnetic wave detection device includes: Based on the position information and parameter configuration information of the electromagnetic wave detection device, error correction is performed on the multiple first collected data to obtain multiple second target data; Based on the plurality of second target data, a plurality of first target data corresponding to each electromagnetic wave detection device is obtained; The obtaining of a plurality of first target data corresponding to each electromagnetic wave detection device based on the plurality of second target data comprises: The second target data is time-aligned and space-aligned to obtain a plurality of first target data corresponding to each electromagnetic wave detection device.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images