Travel time data interpolation method and device based on data and model dual drive

Abstract

The invention provides a travel time data interpolation method and device based on data and model dual drive. The travel time data interpolation method based on data and model dual drive comprises the following steps: receiving coordinates of a point to be interpolated; and travel time data corresponding to the coordinates of the to-be-interpolated point is generated according to the coordinates of the to-be-interpolated point and a pre-constructed neural network model, and the neural network model is used for fitting a speed field and a travel time field of a work area where the coordinates of the to-be-interpolated point are located and is generated according to the speed field and the travel time field. From the perspective of physical guidance, the invention provides a travel time interpolation method and device based on data and model dual drive. The interpolation travel time data obtained according to the method provided by the embodiment of the invention has reliability and accuracy.

Description

Technical Field

[0001] The present application belongs to the field of oil and gas exploration technology, in particular to the field of seismic exploration data processing technology, and specifically relates to a travel time data interpolation method and device based on dual drive of data and model. Background Art

[0002] In the seismic exploration data processing process, due to the reasons of the observation system, there may be missing data, irregular shot points and receiver points, etc. Based on this, data interpolation plays an important role; for example, filling in missing seismic signals, regularizing to the user-specified grid points, etc.

[0003] As we all know, travel time data interpolation is widely used in ray superposition or migration. Existing travel time data interpolation algorithms are basically purely data-driven and do not involve model calculations. For example, weighted average of distance function or linear superposition of slowness are all purely data-driven.

[0004] The widespread application of neural networks has brought new ideas to interpolation algorithms. A common neural network algorithm uses data and labels as training sets, and then trains a neural network; then the neural network can be applied to other missing data. This is also a common process in the field of machine learning: that is, prepare data labels, train the network, and finally infer.

[0005] Both the conventional travel time data interpolation method and the neural network data interpolation method are purely data driven. The difference between the interpolated data and the real data depends on whether the data is complex. The simpler the data components, the more accurate the interpolation; conversely, the error will be larger. In the usual real data processing, due to the lack of reference data, it is impossible to judge the quality of the interpolated data itself; the accuracy of the interpolation can only be judged by the results of subsequent superposition or offset. If the interpolation is inaccurate, it is necessary to re-interpolate, and re-superimpose and offset. This slows down the data processing cycle to a certain extent and wastes a lot of computing resources. Summary of the invention

[0006] An object of the present invention is to provide a travel time data interpolation method based on data and model dual drive, the method obtains an accurate and reliable travel time data difference based on physical equations. The information of the eikonal equation in the error is fully absorbed into the neural network model, and the interpolation result is reliable and accurate.

[0007] Another object of the present invention is to provide a travel time data interpolation device based on dual drive of data and model. Another object of the present invention is to provide an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the travel time data interpolation method based on dual drive of data and model when executing the computer program. Another object of the present invention is to provide a readable medium, on which a computer program is stored, and when the computer program is executed by the processor, the steps of the travel time data interpolation method based on dual drive of data and model are implemented.

[0008] In order to solve the technical problems in the background technology of this application, the present invention provides the following technical solutions:

[0009] In a first aspect, the present invention provides a travel time data interpolation method based on data and model dual drive, comprising:

[0010] Receive the coordinates of the point to be interpolated;

[0011] The travel time data corresponding to the coordinates of the point to be interpolated are generated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and the travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field.

[0012] In one embodiment of the present invention, the step of constructing the neural network model includes:

[0013] Generate training data and verification data based on the single shot data set of the work area, and perform the following iterative operations:

[0014] Training an initial model or a training model of the pre-constructed neural network model according to the training data, wherein the training model is generated by training the initial model according to the training data;

[0015] Verifying the training model according to the verification data to generate a verification result;

[0016] When the verification result reaches a preset training cutoff condition, stop training to generate a training result;

[0017] The training results are constrained according to the velocity field and the travel time field to construct the neural network model.

[0018] In one embodiment of the present invention, training the pre-constructed initial model or training model of the neural network model according to the training data includes:

[0019] Update the weights of the input layer, convolution layer, activation function, pooling layer, and fully connected layer in the initial model or the training model according to the training data.

[0020] In one embodiment of the present invention, constraining the training results according to the velocity field and the travel time field to construct the neural network model includes:

[0021] Constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0022] The training results are constrained according to the augmented Lagrangian function to generate the neural network.

[0023] In one embodiment of the present invention, constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result includes:

[0024] Determine the data error and the eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0025] The augmented Lagrangian function is constructed according to the data error and the eikonal equation error.

[0026] In one embodiment of the present invention, generating travel time data corresponding to the coordinates of the to-be-interpolated points according to the coordinates of the to-be-interpolated points and a pre-built neural network model includes:

[0027] Generate a travel time field network of the work area according to the neural network model;

[0028] The travel time data corresponding to the coordinates of the point to be interpolated are generated according to the coordinates of the point to be interpolated and the travel time field network.

[0029] In one embodiment of the present invention, determining the data error and the eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result includes:

[0030] Generate a velocity field network according to the training result and the velocity field;

[0031] Generate a travel time field network according to the training results and the travel time field;

[0032] The data error and the eikonal equation error are determined according to the velocity field network and the travel time field network.

[0033] In a second aspect, the present invention provides a travel time data interpolation device based on dual drive of data and model, the device comprising:

[0034] A coordinate receiving module, used for receiving the coordinates of the points to be interpolated;

[0035] A travel time data generation module is used to generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-built neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0036] In one embodiment of the present invention, the travel time data interpolation device based on dual drive of data and model further includes:

[0037] A model building module is used to build the neural network model, and the model building module includes:

[0038] The data generation unit is used to generate training data and verification data according to the single shot data set of the work area, and perform the following iterative operations:

[0039] A model training unit, used to train an initial model or a training model of the pre-constructed neural network model according to the training data, wherein the training model is generated by training the initial model according to the training data;

[0040] A verification result generating unit, used to verify the training model according to the verification data to generate a verification result;

[0041] A training result generating unit, used to stop the training when the verification result reaches a preset training cutoff condition to generate a training result;

[0042] A model building unit is used to constrain the training results according to the velocity field and the travel time field to build the neural network model.

[0043] In one embodiment of the present invention, the model training unit includes:

[0044] The model training subunit is used to update the weights of the input layer, convolution layer, activation function, pooling layer, and fully connected layer in the initial model or the training model according to the training data.

[0045] In one embodiment of the present invention, the model building unit includes:

[0046] A Lagrangian function construction unit, used for constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0047] A training result constraint unit is used to constrain the training result according to the augmented Lagrangian function to generate the neural network.

[0048] In one embodiment of the present invention, the Lagrangian function construction unit includes:

[0049] An error determination unit, used for determining a data error and an eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0050] The Lagrangian function construction subunit is used to construct the augmented Lagrangian function according to the data error and the eikonal equation error.

[0051] In one embodiment of the present invention, the travel time data generating module includes:

[0052] A travel time field network generating unit, used for generating a travel time field network of the work area according to the neural network model;

[0053] The travel time data generating unit is used to generate the travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and the travel time field network.

[0054] In one embodiment of the present invention, the error determination unit includes:

[0055] A velocity field network generating unit, configured to generate a velocity field network according to the training result and the velocity field;

[0056] A travel time field network generating unit, used for generating a travel time field network according to the training result and the travel time field;

[0057] The error determination subunit is used to determine the data error and the eikonal equation error according to the velocity field network and the travel time field network.

[0058] In a third aspect, the present invention provides a computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the steps of a travel time data interpolation method based on dual drive of data and model.

[0059] In a fourth aspect, the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a travel time data interpolation method based on dual drive of data and model are implemented.

[0060] In a fifth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a travel time data interpolation method based on dual drive of data and model.

[0061] From the above description, it can be seen that an embodiment of the present invention provides a travel time data interpolation method and device based on dual drive of data and model, and the corresponding travel time data interpolation method based on dual drive of data and model includes: receiving the coordinates of the point to be interpolated; generating travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0062] The corresponding travel time data interpolation device based on dual drive of data and model includes: a coordinate receiving module, used to receive the coordinates of the point to be interpolated; a travel time data generating module, used to generate the travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-built neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field.

[0063] The travel time data interpolation method and device based on dual drive of data and model provided in the embodiment of the present invention propose a travel time interpolation method and device based on dual drive of data and model from the perspective of physical guidance. The interpolated travel time data obtained by the travel time data interpolation method and device based on dual drive of data and model provided in the embodiment of the present invention has reliability and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0065] Figure 1 A schematic flow chart of a travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0066] Figure 2 A second flow chart of a travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0067] Figure 3 It is a flowchart diagram of step 300 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0068] Figure 4 It is a flow chart of step 302 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0069] Figure 5It is a flow chart of step 305 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0070] Figure 6 It is a flow chart of step 3051 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0071] Figure 7 It is a flowchart diagram of step 200 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0072] Figure 8 It is a flow chart of step 30511 of the travel time data interpolation method based on data and model dual drive in an embodiment of the present invention;

[0073] Fig. 9 It is a flow chart of a travel time data interpolation method based on data and model dual drive in a specific implementation manner of the present invention;

[0074] Fig.10 It is a schematic diagram of the original travel time data in a specific embodiment of the present invention;

[0075] Fig.11 It is a schematic diagram of extracting original travel time data in a specific embodiment of the present invention (extraction ratio 10%);

[0076] Fig.12 It is a schematic diagram of the travel time data interpolation result in a specific implementation manner of the present invention;

[0077] Fig.13 Schematic diagram of the error distribution between the interpolated data and the original data in a specific embodiment of the present invention Figure 1 ;

[0078] Fig.14 Schematic diagram of the error distribution between the interpolated data and the original data in a specific embodiment of the present invention Figure 2 ;

[0079] Fig.15 It is a block diagram of a travel time data interpolation device based on data and model dual drive in a specific embodiment of the present invention;

[0080] Fig.16 It is a schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0081] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0082] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0083] It should be noted that the terms "including" and "having" in the specification and claims of the present application and the above-mentioned drawings and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. In the absence of conflict, the embodiments in the present application and the features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0084] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of laws and regulations.

[0085] Embodiment 1:

[0086] The embodiment of the present invention provides a specific implementation of a travel time data interpolation method based on data and model dual drive, see Figure 1 , specifically including the following:

[0087] Step 100: receiving the coordinates of the point to be interpolated;

[0088] Step 200: Generate travel time data corresponding to the coordinates to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0089] From the above description, it can be seen that an embodiment of the present invention provides a travel time data interpolation method based on dual drive of data and model, and the corresponding travel time data interpolation method based on dual drive of data and model includes: receiving the coordinates of the point to be interpolated; generating travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-built neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field. This method analyzes the principles of the travel time field and the velocity field, so that the information of the eikonal equation in the error is completely absorbed into the neural network, and the interpolation result is not only reliable, but also accurate. In actual data processing, it can be used in various exploration areas.

[0090] With respect to step 100 , the coordinates of the points to be interpolated refer to the coordinates of the interpolation points specified by the user.

[0091] The neural network model in step 200 is a computational model consisting of a group of interconnected neurons that transmit information through weights. Neural networks can process and analyze large amounts of data through learning and optimization. The basic unit of a neural network is a neuron, also known as a node or perceptron. Each neuron receives input from other neurons and calculates output through an activation function. The connection weights between neurons determine the degree of influence of the input on the output. By adapting these weights, the neural network can learn patterns and relationships from the training data and make predictions on unseen data.

[0092] Neural networks usually consist of multiple layers, including input layers, hidden layers, and output layers. The input layer receives raw data, the hidden layer is used to process and extract features, and the output layer gives the final prediction results. Each layer contains multiple neurons, and the connections between them form a forward propagation network structure. The back propagation algorithm is used to train neural networks. The algorithm calculates the error between the predicted value and the true value and adjusts the weight of each connection according to the error to gradually optimize the model. By repeatedly iterating this process, the neural network can continuously improve its prediction ability.

[0093] In seismic exploration, travel time data refers to the time it takes for seismic waves to reach the seismometer from the source. Seismic waves are waves that propagate inside the earth after an earthquake event, and their speed and path are affected by underground structures. By analyzing the travel time data of seismic waves, we can infer underground structures and rock properties, and then understand underground mineral resources, groundwater, oil and gas, and other information.

[0094] In seismic exploration, a series of seismic detectors are usually arranged on the surface, and a certain number of seismic sources are buried underground. By measuring and analyzing the propagation path and travel time data of seismic waves, the physical properties and structure of the underground can be inferred. Travel time data is one of the most basic data types in seismic exploration, and it is an important data source for underground exploration and geological research. The methods of travel time data analysis include superposition, inversion, and migration. Among them, the superposition method is one of the most commonly used methods, which is to superimpose multiple seismic records together to enhance the strength and clarity of the signal. The inversion method is to infer the physical properties and structure of the underground by inverting the propagation path and travel time data of seismic waves. The migration method is to convert the propagation path and travel time data of seismic waves into an image of the underground physical structure by processing seismic records.

[0095] In addition, the interpolated travel time data can be used for overlay and ray-based migration applications.

[0096] Embodiment 2:

[0097] In some embodiments of the present invention, see Figure 2 , based on the travel time data interpolation method driven by both data and model, also includes:

[0098] Step 300: Construct the neural network model, see Figure 3 , step 300 comprises:

[0099] Step 301: Generate training data and verification data according to the single shot data set of the work area, and perform the following iterative operations:

[0100] It should be pointed out that the Single Shot Data Set in seismic exploration refers to a seismic wave data set generated by a single source (seismic cannon) in a seismic exploration. In seismic exploration, in order to obtain information on the physical structure and properties of the underground, it is usually necessary to arrange multiple seismic detectors on the surface and bury a certain number of seismic sources underground. By measuring and analyzing the propagation path and travel time data of the seismic waves, the physical structure and properties of the underground can be inferred. A Single Shot Data Set refers to a seismic wave data set generated by a single source (seismic cannon) in this process. A Single Shot Data Set usually includes multiple seismic records, each of which corresponds to a seismic wave signal received by a seismic detector at different times.

[0101] First, the single-shot dataset is one of the most basic data types in seismic exploration. It contains the propagation path and travel time data of seismic waves and is an important data source for underground exploration and geological research. Second, the single-shot dataset can be used to evaluate the effect and quality of seismic exploration. By analyzing and processing the single-shot dataset, the accuracy and reliability of seismic exploration can be evaluated, providing more accurate data support for underground exploration and geological research.

[0102] Step 302: training an initial model or a training model of the pre-constructed neural network model according to the training data, wherein the training model is generated by training the initial model according to the training data;

[0103] The initial model of the neural network model is constructed by the following steps: First, determine the structure of the neural network, including the number of layers, the number of neurons in each layer, the activation function, etc. These parameters can be selected according to specific problems and needs. Next, initialize the weights. For the weights of each connection, you can use random numbers to initialize them. A common method is to randomly extract values ​​from a uniform distribution or a Gaussian distribution as the initial weights. This can be implemented using library functions or custom functions. Initialize the bias. For the bias of each neuron, you can also use random numbers to initialize it. You can also randomly extract values ​​from a uniform distribution or a Gaussian distribution as the initial bias. Select the activation function: Each neuron in a neural network usually has an activation function to introduce nonlinear characteristics. Common activation functions include sigmoid, ReLU, tanh, etc. Select a suitable activation function according to specific problems and needs.

[0104] It should be noted that the choice of initial model has a certain impact on the training and performance of the neural network. Generally speaking, a suitable initial model can accelerate the convergence of the model and improve the performance of the model. Different initialization methods may have different effects on the results, so experiments and adjustments are needed to find the best initialization method.

[0105] Step 303: verifying the training model according to the verification data to generate a verification result;

[0106] Step 304: When the verification result reaches a preset training cutoff condition, stop the training to generate a training result;

[0107] Training cutoff conditions include the following: Fixed number of rounds: Training can be set to stop after reaching a specified number of rounds. This is one of the simplest cutoff conditions, but it may not fully utilize the information of the dataset, resulting in underfitting or overfitting problems. Reach a specified accuracy or error: You can set a target accuracy or error value and stop training when the model reaches or approaches this target. This ensures that the model reaches a certain performance level. Early stopping: You can monitor the performance of the validation set and stop training when the performance no longer improves. You can set a tolerance (such as no performance improvement for multiple consecutive rounds) to determine when to stop training. Computational resource constraints: Training time may be limited by computing resources, such as time, memory, or computing power. You can set a time limit or resource limit and stop training when this limit is exceeded. Cross-validation: You can use cross-validation to evaluate the performance of the model and stop training when the cross-validation error no longer decreases.

[0108] It should be pointed out that it is necessary to choose the appropriate training cutoff condition according to the characteristics of the specific problem and data set. At the same time, techniques such as early stopping can be used to avoid overfitting and improve the generalization ability of the model. Early stopping means monitoring the performance of the validation set during training and stopping training when the performance on the validation set begins to decline to prevent the model from overfitting the training data.

[0109] Step 305: constraining the training results according to the velocity field and the travel time field to construct the neural network model.

[0110] The existing method for interpolating travel time data using a neural network model only considers the data aspect, that is, it is purely data-driven, and does not consider the physical constraints of the obtained results. That is, the travel time data obtained may not have practical significance, or may not match the corresponding velocity field. However, through step 305, the neural network model is physically constrained by the actual velocity field and travel time field of the work area, which can overcome the above problems.

[0111] In some embodiments of the present invention, see Figure 4 , step 302 comprises:

[0112] Step 3021: Update the weights of the input layer, convolution layer, activation function, pooling layer, and fully connected layer in the initial model or the training model according to the training data.

[0113] The input layer is used to accept raw data or features and pass them to the next layer for processing. The number of neurons in the input layer usually corresponds to the number of features in the input.

[0114] The convolutional layer is used to extract features of the input data. The convolutional layer generates a feature map by convolving the input data with a filter (also called a convolution kernel).

[0115] The activation function is a nonlinear function in a neural network, which is used to introduce nonlinear characteristics. The activation function performs nonlinear transformation on the output of the convolutional layer to increase the network's expressive power. Common activation functions include Sigmoid, ReLU, tanh, etc.

[0116] Pooling layers are used to reduce the spatial dimension of feature maps, thereby reducing the number of parameters and computation. Pooling operations usually take the maximum value (max pooling) or calculate the average value (average pooling) in each region. Pooling layers help extract the main features of the input data and reduce overfitting.

[0117] The fully connected layer connects the output of the convolutional layer or pooling layer to the output layer of the neural network. Each neuron in the fully connected layer is connected to all neurons in the previous layer, forming a fully connected structure. The fully connected layer is used to map features to the final output category or predicted value.

[0118] Specifically, a neural network is constructed to represent the travel time field and the velocity field respectively:

[0119]

[0120] Among them, T is the travel time field and v is the velocity field. is the spatial coordinate, N t is a neural network that outputs the travel time field, which can be understood as a composite function; N v is a neural network that outputs the velocity field. The travel time field and the velocity field are controlled or connected by the eikonal equation:

[0121] In some embodiments of the present invention, see Figure 5 , step 305 comprises:

[0122] Step 3051: constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0123] It can be understood that the augmented Lagrangian function in step 3051 is to fit the travel time data of the block.

[0124] Step 3052: constraining the training results according to the augmented Lagrangian function to generate the neural network.

[0125] The augmented Lagrangian function is a method for solving optimization problems with constraints. The original problem is transformed into an unconstrained optimization problem by introducing Lagrangian multipliers, and then gradually optimized in an iterative manner.

[0126] Specifically, based on formula (1), we have:

[0127]

[0128] in, is a partial differential operator, which represents the partial differential in three directions (x, y, z), and |.| represents the absolute value:

[0129]

[0130]

[0131] are the shot point coordinates, Indicates that the travel time at the shot point is 0. v is the velocity field mentioned above, v -2 It is the -2 power of the velocity field.

[0132] In some embodiments of the present invention, see Figure 6 , step 3051 includes:

[0133] Step 30511: determining the data error and the eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0134] Step 30512: Construct the augmented Lagrangian function according to the data error and the eikonal equation error.

[0135] Specifically:

[0136]

[0137] Among them, i represents the summation index, is a partial differential operator, N r is the number of receivers, is the value of the travel time field at the receiver location, is the observed travel time. N is the number of points in the discretized velocity field. The last term in equation 3 is because the travel time at the source location is zero.

[0138] In formula (3), is the data error, as well as is the error of the eikonal equation.

[0139] In some embodiments of the present invention, see Figure 7 , step 200 comprises:

[0140] Step 201: Generate a travel time field network of the work area according to the neural network model;

[0141] Step 202: Generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and the travel time field network.

[0142] In step 201 and step 202, formula (3) can be trained with a standardized neural network. After the training is completed, the velocity field network and the travel time field network can be obtained at the same time. Among them, the travel time field network is useful for data interpolation. For the points that need to be interpolated specified by the user, they are directly input into the travel time field network, and the travel time data of the target point can be obtained according to formula (1).

[0143] In some embodiments of the present invention, see Figure 8 , step 30511 includes:

[0144] Step 305111: Generate a velocity field network according to the training result and the velocity field;

[0145] Step 305112: Generate a travel time field network according to the training results and the travel time field;

[0146] Step 305113: Determine the data error and the eikonal equation error based on the velocity field network and the travel time field network.

[0147] From the above description, it can be seen that an embodiment of the present invention provides a travel time data interpolation method based on dual drive of data and model, and the corresponding travel time data interpolation method based on dual drive of data and model includes: receiving the coordinates of the point to be interpolated; generating travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-built neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field. This method analyzes the principles of the travel time field and the velocity field, so that the information of the eikonal equation in the error is completely absorbed into the neural network, and the interpolation result is not only reliable, but also accurate. In actual data processing, it can be used in various exploration areas.

[0148] Embodiment three:

[0149] In a specific embodiment, the present invention also takes the structure of the neural network model as an independent dual network as an example, and provides a specific embodiment of the travel time data interpolation method based on data and model dual drive, see Fig. 9 , specifically including the following steps.

[0150] S1: Construct a neural network model, and fit the velocity field network and travel time field network respectively according to the neural network model.

[0151] For the observed travel time data of the work area, a physics-guided neural network model is established, and the single-shot data set is inverted. After the neural network model training is completed, a trained neural network model is obtained. Among them, the neural network model is determined by the structure and weight. The trained neural network model means that the weights are updated on the established neural network structure so that the neural network reaches the optimal solution.

[0152] It should be pointed out that the velocity field network and the travel time field network can be the same, or two networks can be constructed independently;

[0153] In the specific implementation of the present application, the neural network model (velocity field network and travel time field network) adopts an independent dual network structure. The velocity field network contains 7 hidden layers, each with 20 neurons; the travel time field network contains 9 hidden layers, each with 25 neurons; the activation function of the neuron adopts the leaky-LU function, but at the output end, in order to ensure that the output result of the neuron is positive (physically, travel time and speed cannot be negative), the sigmoid function is used.

[0154] S2: Construct the augmented Lagrangian function.

[0155] The data error and the eikonal equation error are fused together by augmenting the Lagrangian function, and the neural network model is trained to obtain the travel time field network.

[0156] S3: Input the interpolation point coordinates specified by the user into the travel time field network to obtain the travel time of the target point.

[0157] For the target point set by the user, the coordinates are input into the neural network (travel time field network) to obtain the interpolated travel time data.

[0158] Fig.10 What is shown is the original travel time data (single shot set). In order to simulate the real situation, the specific implementation method of this application extracts 10% of it as observation data, such as Fig.11 As shown. The specific implementation of this application uses this 10% data to simulate the real collected data. The neural network model is trained using Fig.11 After the neural network model is trained, the coordinates of the original data are input in the inference phase, and an intelligent interpolation result is obtained. Fig.12 is the interpolation result made by this application. In order to show the interpolation accuracy (compared with the original travel time data), Fig.13 The error between the original data and the interpolated travel time data is shown (the error is basically within 0.1). Fig.14 The original data, interpolated data, and error are displayed in lines. It can be seen that the maximum error is 0.1s, and most of the data errors are within 0.05s. Fig.14In the figure, the first row is the original data, the horizontal axis is the number of receivers, and the vertical axis is the travel time. The second row is the interpolated data. The third row is the error between the original data and the interpolated data. It can be seen that the maximum error is around 0.1s, and most errors are within 0.05s.

[0159] From the above description, it can be seen that an embodiment of the present invention provides a travel time data interpolation method based on dual drive of data and model, and the corresponding travel time data interpolation method based on dual drive of data and model includes: receiving the coordinates of the point to be interpolated; generating travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0160] The above method analyzes the travel time field and velocity field principle, so that the information of the eikonal equation in the error is completely absorbed into the neural network, and its interpolation result is not only reliable but also accurate. In actual data processing, it can be used in various exploration areas.

[0161] On the one hand, in view of the missing tracks and the non-uniform arrangement of receivers in the travel time data, the present invention proposes a travel time interpolation method based on dual drive of data and model. The purpose is to obtain an accurate and reliable travel time data difference based on the physical equation. On the other hand, in order to enhance flexibility, the travel time data interpolation method based on dual drive of data and model provided in the embodiment of the present invention only processes one common shot gather at a time.

[0162] It should be pointed out that the travel time data interpolation method based on dual drive of data and model provided in the embodiment of the present invention is not targeted at a certain region, but belongs to the research of basic methods.

[0163] Embodiment 4:

[0164] Based on the same inventive concept, the embodiment of the present application also provides a travel time data interpolation device based on data and model dual drive, which can be used to implement the method described in the above embodiment, such as the following embodiment. Since the principle of solving the problem by the travel time data interpolation device based on data and model dual drive is similar to the travel time data interpolation method based on data and model dual drive, the implementation of the travel time data interpolation device based on data and model dual drive can refer to the implementation of the travel time data interpolation method based on data and model dual drive, and the repetitions are not repeated. As used below, the term "unit" or "module" can implement a combination of software and / or hardware of a predetermined function. Although the system described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0165] The embodiment of the present invention provides a specific implementation of a travel time data interpolation device based on data and model dual drive, which can realize a travel time data interpolation method based on data and model dual drive, see Fig.15 The travel time data interpolation device based on data and model dual drive includes:

[0166] A coordinate receiving module 10 is used to receive the coordinates of the point to be interpolated;

[0167] The travel time data generation module 20 is used to generate the travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and the travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field.

[0168] In one embodiment of the present invention, the travel time data interpolation device based on dual drive of data and model further includes:

[0169] A model building module is used to build the neural network model, and the model building module includes:

[0170] The data generation unit is used to generate training data and verification data according to the single shot data set of the work area, and perform the following iterative operations:

[0171] A model training unit, used to train an initial model or a training model of the pre-constructed neural network model according to the training data, wherein the training model is generated by training the initial model according to the training data;

[0172] A verification result generating unit, used to verify the training model according to the verification data to generate a verification result;

[0173] A training result generating unit, used to stop the training when the verification result reaches a preset training cutoff condition to generate a training result;

[0174] A model building unit is used to constrain the training results according to the velocity field and the travel time field to build the neural network model.

[0175] In one embodiment of the present invention, the model training unit includes:

[0176] The model training subunit is used to update the weights of the input layer, convolution layer, activation function, pooling layer, and fully connected layer in the initial model or the training model according to the training data.

[0177] In one embodiment of the present invention, the model building unit includes:

[0178] A Lagrangian function construction unit, used for constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0179] A training result constraint unit is used to constrain the training result according to the augmented Lagrangian function to generate the neural network.

[0180] In one embodiment of the present invention, the Lagrangian function construction unit includes:

[0181] An error determination unit, used for determining a data error and an eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result;

[0182] The Lagrangian function construction subunit is used to construct the augmented Lagrangian function according to the data error and the eikonal equation error.

[0183] In one embodiment of the present invention, the travel time data generating module includes:

[0184] A travel time field network generating unit, used for generating a travel time field network of the work area according to the neural network model;

[0185] The travel time data generating unit is used to generate the travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and the travel time field network.

[0186] In one embodiment of the present invention, the error determination unit includes:

[0187] A velocity field network generating unit, configured to generate a velocity field network according to the training result and the velocity field;

[0188] A travel time field network generating unit, used for generating a travel time field network according to the training result and the travel time field;

[0189] The error determination subunit is used to determine the data error and the eikonal equation error according to the velocity field network and the travel time field network.

[0190] From the above description, it can be seen that an embodiment of the present invention provides a travel time data interpolation device based on dual drive of data and model, and the corresponding travel time data interpolation device based on dual drive of data and model includes: a coordinate receiving module, used to receive the coordinates of the point to be interpolated; a travel time data generation module, used to generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field.

[0191] The above device analyzes the travel time field and velocity field principle, so that the information of the eikonal equation in the error is completely absorbed into the neural network, and its interpolation result is not only reliable but also accurate. In actual data processing, it can be used in various exploration areas.

[0192] The travel time data interpolation device based on dual drive of data and model provided in the embodiment of the present invention proposes a travel time interpolation method and device based on dual drive of data and model from the perspective of physical guidance. The interpolated travel time data obtained by the travel time data interpolation method and device based on dual drive of data and model provided in the embodiment of the present invention has reliability and accuracy.

[0193] Embodiment five:

[0194] The embodiment of the present application also provides a specific implementation of an electronic device capable of implementing all steps in the travel time data interpolation method based on dual drive of data and model in the above embodiment, see Fig.16 , electronic equipment specifically includes the following:

[0195] Processor (processor) 1201, memory (memory) 1202, communication interface (CommunicationsInterface) 1203 and bus 1204;

[0196] The processor 1201, the memory 1202, and the communication interface 1203 communicate with each other through the bus 1204; the communication interface 1203 is used to realize information transmission between the server device and the client device and other related devices;

[0197] The processor 1201 is used to call the computer program in the memory 1202. When the processor executes the computer program, all steps in the travel time data interpolation method based on dual drive of data and model in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0198] Step 100: receiving the coordinates of the point to be interpolated;

[0199] Step 200: Generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0200] The electronic device analyzes the travel time field and velocity field principle, so that the information of the Eikonal equation in the error is completely absorbed into the neural network, and its interpolation result is not only reliable but also accurate. In actual data processing, it can be used in various exploration areas.

[0201] From the above description, it can be seen that the electronic device provided by the embodiment of the present invention, from the perspective of physical guidance, proposes a travel time interpolation method and device based on data and model dual drive. The interpolated travel time data obtained by the electronic device provided by the embodiment of the present invention is reliable and accurate.

[0202] Embodiment six:

[0203] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps in the travel time data interpolation method based on dual drive of data and model in the above embodiment. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps of the travel time data interpolation method based on dual drive of data and model in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0204] Step 100: receiving the coordinates of the point to be interpolated;

[0205] Step 200: Generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

[0206] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present invention proposes a travel time interpolation method and device based on data and model dual drive from the perspective of physical guidance. The interpolated travel time data obtained according to the computer-readable storage medium provided in the embodiment of the present invention has reliability and accuracy.

[0207] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the hardware + program embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0208] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0209] Although the present application provides method operation steps such as embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the actual device or client product is executed, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0210] For the convenience of description, the above devices are described in various modules according to their functions. Of course, when implementing the embodiments of this specification, the functions of each module can be implemented in the same or more software and / or hardware, or the module implementing the same function can be implemented by a combination of multiple sub-modules or sub-units. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0211] Those skilled in the art also know that, in addition to implementing the controller in a purely computer-readable program code, the controller can be made to implement the same function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered as a hardware component, and the devices for implementing various functions included therein can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules for implementing the method and structures within the hardware component.

[0212] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0213] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0214] Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. In the description of this specification, the description of the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of this specification. In this specification, the schematic representation of the above terms does not necessarily target the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, in the absence of contradiction, a person skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0215] The above is only an example of the embodiment of the present specification and is not intended to limit the embodiment of the present specification. For those skilled in the art, the embodiment of the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiment of the present specification shall be included in the scope of the claims of the embodiment of the present specification.

Claims

1. A travel time data interpolation method based on data and model dual drive, characterized in that: include: Receive the coordinates of the point to be interpolated; The travel time data corresponding to the coordinates of the point to be interpolated are generated according to the coordinates of the point to be interpolated and a pre-constructed neural network model, wherein the neural network model is used to fit the velocity field and the travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated according to the velocity field and the travel time field.

2. The travel time data interpolation method according to claim 1, characterized in that: The steps of constructing the neural network model include: Generate training data and verification data based on the single shot data set of the work area, and perform the following iterative operations: Training an initial model or a training model of the pre-constructed neural network model according to the training data, wherein the training model is generated by training the initial model according to the training data; Verifying the training model according to the verification data to generate a verification result; When the verification result reaches a preset training cutoff condition, stop training to generate a training result; The training results are constrained according to the velocity field and the travel time field to construct the neural network model.

3. The travel time data interpolation method according to claim 2, characterized in that: Training the pre-built initial model or training model of the neural network model according to the training data includes: Update the weights of the input layer, convolution layer, activation function, pooling layer, and fully connected layer in the initial model or the training model according to the training data.

4. The travel time data interpolation method according to claim 2, characterized in that: Constraining the training results according to the velocity field and the travel time field to construct the neural network model includes: Constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result; The training results are constrained according to the augmented Lagrangian function to generate the neural network.

5. The travel time data interpolation method according to claim 4, characterized in that: Constructing an augmented Lagrangian function corresponding to the travel time data according to the velocity field, the travel time field and the training result, including: Determine the data error and the eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result; The augmented Lagrangian function is constructed according to the data error and the eikonal equation error.

6. The travel time data interpolation method according to claim 5, characterized in that: Generating travel time data corresponding to the coordinates of the to-be-interpolated points according to the coordinates of the to-be-interpolated points and a pre-built neural network model, including: Generate a travel time field network of the work area according to the neural network model; The travel time data corresponding to the coordinates of the point to be interpolated are generated according to the coordinates of the point to be interpolated and the travel time field network.

7. The travel time data interpolation method according to claim 5, characterized in that: Determining a data error and an eikonal equation error corresponding to the travel time data according to the velocity field, the travel time field and the training result includes: Generate a velocity field network according to the training result and the velocity field; Generate a travel time field network according to the training results and the travel time field; The data error and the eikonal equation error are determined according to the velocity field network and the travel time field network.

8. A travel time data interpolation device based on data and model dual drive, characterized in that: include: A coordinate receiving module, used for receiving the coordinates of the points to be interpolated; A travel time data generation module is used to generate travel time data corresponding to the coordinates of the point to be interpolated according to the coordinates of the point to be interpolated and a pre-built neural network model, wherein the neural network model is used to fit the velocity field and travel time field of the work area where the coordinates of the point to be interpolated are located, and is generated based on the velocity field and the travel time field.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the travel time data interpolation method based on dual drive of data and model as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the travel time data interpolation method based on dual drive of data and model as described in any one of claims 1 to 7 are implemented.

Images