Seismic data regularization method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide a seismic data regularization method and device, electronic equipment and storage medium; the method comprises: acquiring actual seismic data collected by an exploration data observation system, the data comprising actual shotpoint coordinates and travel time data in the exploration data observation system, acquiring coordinates of a target shotpoint input by a user, constructing a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and constructing a neural network equation guided based on an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network, inputting the actual shotpoint coordinates and the travel time data into the neural network equation guided based on the eikonal equation, and performing neural network training to obtain trained travel time and velocity networks; and inputting the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint. The method realizes regularization of seismic travel time data and ensures accuracy of data processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of seismic exploration, and in particular to a seismic data regularization method and device, an electronic device and a storage medium. BACKGROUND

[0002] In the field of seismic data, especially in the process of seismic exploration data processing in the industry, the integrity and regularity of data acquisition are crucial. Complete and regular data acquisition can ensure that the data coverage is relatively uniform, thereby improving the subsequent stacking and migration imaging.

[0003] However, in actual acquisition, especially in land seismic data acquisition, various defects in the observation system lead to problems such as data missing, irregular shot points and geophones in the acquired data. Currently, in view of these data defects, the acquired data is processed in the data domain, which is a data processing based on data characteristics, and the effect is poor.

[0004] In summary, the current seismic data processing method has the technical problem of poor data processing effect. SUMMARY

[0005] To alleviate the technical problem of poor data processing effect of the current seismic data processing method, the embodiments of the present application provide a seismic data regularization method, device, electronic device and storage medium.

[0006] In a first aspect, the embodiments of the present application provide a seismic data regularization method, comprising:

[0007] Obtaining actual seismic data acquired by an exploration data observation system, the actual seismic data comprising actual shot-receiver point coordinates and travel time data in the exploration data observation system;

[0008] Obtaining the coordinates of a target shot-receiver point input by a user;

[0009] According to the actual shot-receiver point coordinates and travel time data, constructing a travel time neural network and a velocity neural network, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network;

[0010] According to the neural network guided by the eikonal equation, inputting the actual shot-receiver point coordinates and travel time data, and performing neural network training to obtain trained travel time network and velocity network;

[0011] According to the trained travel time network, inputting the coordinates of the target shot-receiver point, obtaining the travel time data of the target shot-receiver point, and completing seismic travel time data regularization.

[0012] In some embodiments, the neural network guided by the eikonal equation inputs the actual shotpoint coordinates and travel time data, and is trained to obtain a trained travel time network and a velocity network, including:

[0013] According to the travel time data error of the travel time network and the eikonal equation of the target equation, a loss function is constructed.

[0014] According to the loss function and the actual seismic data, the neural network is trained to obtain a trained travel time network and a velocity network.

[0015] In some embodiments, the loss function includes an augmented Lagrangian function.

[0016] In some embodiments, the velocity network and the travel time network are physical information neural networks with the same network architecture.

[0017] In some embodiments, the velocity network and the travel time network are physical information neural networks with different network architectures.

[0018] In some embodiments, the velocity network contains 7 layers of hidden layers, each layer containing 20 neurons.

[0019] In some embodiments, the travel time network contains 9 layers of hidden layers, each layer containing 25 neurons.

[0020] In a second aspect, an embodiment of the present application provides a seismic data regularization device, including:

[0021] A first acquisition module is configured to acquire actual seismic data collected by an exploration data observation system, the seismic data including actual shotpoint coordinates and travel time data in the exploration data observation system.

[0022] A second acquisition module is configured to acquire coordinates of a target shotpoint input by a user.

[0023] A network construction module is configured to construct a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and to construct a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network.

[0024] A network training module is configured to input the actual shotpoint coordinates and the travel time data into the neural network guided by the eikonal equation, and to train the neural network to obtain a trained travel time network and a velocity network.

[0025] A regularization module is configured to input the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint, so as to complete seismic travel time data regularization.

[0026] In a third aspect, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, implements the method according to the first aspect.

[0027] In a fourth aspect, a computer readable storage medium is provided, wherein the computer readable storage medium stores a computer program, and the computer program, when executed by one or more processors, implements the method according to the first aspect.

[0028] Compared with the prior art, one or more embodiments of the present application can bring at least the following beneficial effects:

[0029] Embodiments of the present application provide a seismic data regularization method and device, an electronic device and a storage medium; the method comprises: acquiring actual seismic data collected by an exploration data observation system, wherein the actual seismic data comprises actual shot-receiver coordinates and travel time data in the exploration data observation system; acquiring coordinates of a target shot-receiver input by a user; constructing a travel time neural network and a velocity neural network according to the actual shot-receiver coordinates and the travel time data, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network; inputting the actual shot-receiver coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training to obtain a trained travel time network and a velocity network; inputting the coordinates of the target shot-receiver into the trained travel time network to obtain travel time data of the target shot-receiver, so as to complete seismic travel time data regularization. In the method provided in the present solution, a neural network guided by physics is constructed, a trained travel time network and a velocity network are obtained by training, and then the coordinates of the target shot-receiver are input into the trained neural network to obtain travel time data corresponding to the target shot-receiver, so as to realize regularization of seismic travel time data. The method ensures the accuracy of data processing and improves the data processing effect. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0031] Figure 1 The first flowchart of the seismic data regularization method provided by the embodiments of the present application is shown in the figure.

[0032] Figure 2A schematic diagram of shot point distribution provided by an embodiment of the present application is shown in FIG. 1.

[0033] Figure 3 A schematic diagram of real travel time data of a test shot gather provided by an embodiment of the present application is shown in FIG. 2.

[0034] Figure 4 A schematic diagram of reconstructed travel time data obtained after neural network inversion provided by an embodiment of the present application is shown in FIG. 3.

[0035] Figure 5 A schematic diagram of error distribution of regularized data and original data provided by an embodiment of the present application is shown in FIG. 4.

[0036] Figure 6 A schematic diagram of the effect of drawing and displaying after converting the original data, the regularized data, and the error of the original data and the interpolated data into one-dimensional arrays provided by an embodiment of the present application is shown in FIG. 5.

[0037] Figure 7 A schematic diagram of one structure of a seismic data regularization device provided by an embodiment of the present application is shown in FIG. 6.

[0038] In the drawings, the same components are designated by the same reference numerals, and the drawings are not drawn to scale. DETAILED DESCRIPTION

[0039] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and embodiments, so that how the technical means of the present application is applied to solve the technical problems and achieve the corresponding technical effects can be fully understood and implemented. The embodiments of the present application and each feature in the embodiments can be combined with each other without conflict, and the formed technical solutions are all within the protection scope of the present application.

[0040] Meanwhile, in the following description, many specific details are set forth in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art, that the present application can be practiced without the specific details or in other specific ways.

[0041] Embodiment One

[0042] Figure 1 A first flowchart of a seismic data regularization method provided by an embodiment of the present application is shown in FIG. 1. Figure 1 The seismic data regularization method provided by the present embodiment includes:

[0043] Step S110: Obtain actual seismic data collected by an exploration data observation system.

[0044] In the present application, the actual seismic data includes actual shot-receiver point coordinates and travel time data in the exploration data observation system.

[0045] Step S120: Obtain the coordinates of the target shotpoint input by the user.

[0046] In this embodiment, the coordinates of the target shotpoint input by the user according to the data regularization requirement are obtained.

[0047] Step S130: According to the actual shotpoint coordinates and the travel time data, construct a travel time neural network and a velocity neural network, and construct a neural network equation guided by the eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network.

[0048] Specifically, the neural network (as shown in the following formula) is constructed, which respectively represents the travel time field network (corresponding to the travel time network in the foregoing) and the underground velocity field network (corresponding to the velocity network in the foregoing):

[0049]

[0050] wherein T is the travel time field, v is the underground velocity field, is the spatial coordinates of the target point / shotpoint, N t is the neural network outputting the travel time field, N v is the neural network outputting the underground velocity field, and these neural networks can be understood as a composite function.

[0051] The independent double network structure is adopted in this embodiment: the velocity field network and the travel time field network. The velocity field contains 7 layers of hidden layers, and each layer contains 20 neurons. The travel time field contains 9 layers of hidden layers, and each layer contains 25 neurons. The activation function of the neuron adopts the leaky-LU function, but at the output end, the sigmoid function is adopted in order to ensure that the neuron output result is a positive value (physically, the travel time and the velocity cannot be negative values).

[0052] The travel time field and the velocity field are related by the following eikonal equation:

[0053]

[0054] wherein is the partial differential operator, which represents the partial differential with respect to three directions (x, y, z), |. | represents the absolute value, and then:

[0055]

[0056]

[0057] is the coordinates of the shotpoint, represents that the travel time data is 0 at the shotpoint. v is the velocity field mentioned above, v -2 is the square of the velocity field.

[0058] Step S140: inputting the actual shotpoint coordinates and travel time data into the neural network guided according to the eikonal equation, and performing neural network training to obtain a trained travel time network and a velocity network.

[0059] In the present application, the neural network comprises a travel time network and a velocity network, wherein the velocity network is used to ensure the training accuracy of the travel time network; the trained neural network, i.e., the trained neural network, refers to the weight update on the neural network structure that has been built, so that the neural network reaches the optimal solution.

[0060] In some embodiments, the inputting the actual shotpoint coordinates and travel time data into the neural network guided according to the eikonal equation, and performing neural network training to obtain a trained travel time network and a velocity network comprises: constructing a loss function according to the travel time data error of the travel time network and the eikonal equation of the target equation; training the neural network according to the loss function and the actual seismic data to obtain a trained travel time network and a velocity network.

[0061] In some embodiments, the loss function comprises an augmented Lagrangian function.

[0062] Specifically, in order to fit the travel time data, the following augmented Lagrangian function is constructed:

[0063]

[0064] wherein i represents a 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 position, is the observed travel time, and N is the number of points of the underground velocity field. is a correction value given because the travel time at the source position is zero.

[0065] The augmented Lagrangian function can be used as a loss function to train a standardized neural network. After the training is completed, the velocity field network and the travel time field network can be obtained simultaneously.

[0066] In some embodiments, the velocity network and the travel time network are physical information neural networks with the same network architecture.

[0067] In some embodiments, the velocity network and the travel time network are physical information neural networks with different network architectures.

[0068] Specifically, in the embodiment, the velocity network and the travel time network can be the same physical information neural network, or can be two network physical information neural networks constructed independently.

[0069] In some embodiments, the velocity network comprises 7 layers of hidden layers, each layer having 20 neurons.

[0070] In some embodiments, the travel time network comprises 9 layers of hidden layers, each layer having 25 neurons.

[0071] Specifically, the activation function of the neuron adopts a leaky-LU function, but at the output end, a sigmoid function is adopted to ensure that the neuron output result is a positive value (physically, the travel time and the velocity cannot be negative values).

[0072] Step S150: inputting the coordinates of the target shot point into the trained travel time network to obtain travel time data of the target shot point, so as to complete the regularization of the seismic travel time data.

[0073] In some embodiments, according to the travel time network in the trained neural network, the coordinate information of the target shot point input by the user is input into the trained travel time field network to obtain travel time data corresponding to the target shot point.

[0074] Specifically, the trained neural network comprises a trained travel time field network; after obtaining the trained neural network, the coordinates of the target shot point input by the user are obtained, and the coordinates of the target shot point are input into the trained travel time field network to obtain travel time data corresponding to the target shot point, so as to complete the regularization of the seismic travel time data.

[0075] In summary, in the seismic data regularization method provided in the embodiment, the neural network based on physical guidance is constructed, the trained travel time network and velocity network are obtained by training, the coordinates of the target shot point are input into the trained neural network to obtain travel time data corresponding to the target shot point, and the regularization of the seismic travel time data is realized, which ensures the accuracy of data processing and improves the data processing effect.

[0076] Now, the seismic data is taken as the travel time data, and the data management is taken as the data regularization processing for description.

[0077] The travel time data regularization algorithm of the application comprises two steps:

[0078] First, data screening is performed on the observed travel time data to obtain shot points and inversion data that need to be regularized; then, a neural network based on physical guidance is established, including a travel time field network and an underground velocity field network, and inversion training based on underground velocity is performed using the inversion data; after network training is completed, a trained neural network is obtained. The neural network is determined by structure and weight, and the trained neural network refers to weight updating on the established neural network structure, so that the neural network reaches an optimal solution.

[0079] Second, the coordinates of the target point set by the user or the shot point that needs to be regularized are input into the neural network to obtain the travel time data of the shot point after regularization.

[0080] Specifically, it includes:

[0081] The neural network is constructed to represent the travel time field network and the underground velocity field network:

[0082]

[0083] wherein T is the travel time field, v is the underground velocity field, is the spatial coordinates of the target point / shot point, N t is the neural network outputting the travel time field, N v is the neural network outputting the underground velocity field, and these neural networks can be understood as a composite function.

[0084] The independent double-network structure adopted in this embodiment is: the velocity field network and the travel time field network. The velocity field contains 7 hidden layers, each layer containing 20 neurons; the travel time field contains 9 hidden layers, each layer containing 25 neurons; the activation function of the neuron adopts a leaky-LU function, but at the output end, a sigmoid function is adopted to ensure that the neuron output result is positive (physically, travel time and velocity cannot be negative).

[0085] The travel time field and the velocity field are associated by the following function equation:

[0086]

[0087] wherein is a partial differential operator, representing the partial derivative in three directions (x, y, z), and |. | represents the absolute value, and further:

[0088]

[0089]

[0090] is the coordinates of the shot point, The travel time data is 0 at the shot point. v is the velocity field mentioned above, v -2 is the-2 power of the velocity field.

[0091] In order to fit the travel time data, the following augmented Lagrange function is constructed:

[0092]

[0093] Where i represents the summation index, is the partial differential operator, N r is the number of receivers, is the value of the travel time field at the receiver position, is the observed travel time, and N is the number of points of the underground velocity field, is the correction value given because the travel time is zero at the source position.

[0094] The augmented Lagrange function can be used as a loss function to train a standardized neural network. After training, the velocity field network and the travel time field network can be obtained. Among them, the travel time field network is useful for data regularization. For the target points or shot points specified by the user that need to be regularized, the target point travel time data can be obtained by directly inputting them into the travel time field network.

[0095] In summary, the present application comprises the following steps:

[0096] Step 1: Construct a neural network to fit the velocity field and the travel time field respectively. The velocity field network and the travel time field network can be the same or two independent networks;

[0097] Step 2: Construct an augmented Lagrange function to combine the travel time data error and the control travel time function equation error, and train the network to obtain the travel time field network;

[0098] Step 3: Input the coordinates of the points specified by the user that need to be regularized into the travel time field network to obtain the travel time data of the target points.

[0099] Based on the foregoing description, the present application will be described in detail in combination with specific scenarios.

[0100] To verify the seismic data regularization method of the embodiment of the present application, a batch of actual data is used for verification. A shot gather is randomly selected as a test shot gather; then the remaining shot gathers are used for neural network inversion. After the neural network training is completed, the original shot gather coordinates are input into the neural network to obtain a reconstructed travel time data, which can be compared with the original data.

[0101] As Figure 2As shown, the figure shows the overall distribution of the shot gather, and the overall distribution of the shot gather is relatively scattered, wherein the circle marks a randomly selected shot gather, which is used as a test shot gather and is not used in the inversion.

[0102] Figure 3 is a schematic diagram of real travel time data of a test shot gather provided by the embodiment of the application; in the training of the neural network, the test shot gather does not participate in the training, but is used for comparison with the regularized data obtained after the neural network is trained.

[0103] Figure 4 is a schematic diagram of reconstructed travel time data obtained after the neural network inversion provided by the embodiment of the application; in the embodiment, after the neural network inversion is completed, the coordinates of the test shot gather (i.e. the original data coordinates) are input to the neural network, and the reconstructed travel time data can be obtained, wherein the purpose of inputting the coordinates of the test shot gather is to compare with the original travel time data.

[0104] Figure 5 is a schematic diagram of error distribution of the regularized data and the original data provided by the embodiment of the application; from Figure 5 it can be seen that the error of the regularized data and the original data at all positions is within 0.08 seconds.

[0105] In order to more intuitively compare the data error, Figure 6 is a schematic diagram of the effect obtained by converting the original data, the regularized data, the error of the original data and the interpolated data in Figure 2 , Figure 3 , Figure 4 into one-dimensional arrays and drawing the graphs; as Figure 6 shown, the first row is the original data, the horizontal coordinate is the number of receivers, and the vertical coordinate is the travel time; the second row is the regularized data; the third row is the error of the original data and the interpolated data; from Figure 6 it can be seen that the maximum error is about 0.08 seconds, and most of the errors are within 0.03 seconds.

[0106] In summary, the method provided by the embodiment constructs a neural network based on physical guidance, trains to obtain a trained travel time network and a velocity network, and then inputs the coordinates of the target shot point based on the trained neural network to obtain the travel time data corresponding to the target shot point, so as to realize the regularization of the seismic travel time data, which ensures the accuracy of the data processing and improves the data processing effect.

[0107] Embodiment Two

[0108] Figure 7 is a structural schematic diagram of a seismic data regularization device provided by the embodiment of the application, please refer to Figure 7The embodiment provides a seismic data regularization device, which comprises:

[0109] The first acquisition module 710 is used for acquiring actual seismic data collected by an exploration data observation system, wherein the seismic data comprises actual shotpoint coordinates and travel time data in the exploration data observation system;

[0110] The second acquisition module 720 is used for acquiring coordinates of a target shotpoint input by a user;

[0111] The network construction module 730 is used for constructing a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used for connecting the travel time neural network and the velocity neural network;

[0112] The network training module 740 is used for inputting the actual shotpoint coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training, so as to obtain a trained travel time network and a velocity network;

[0113] The regularization module 750 is used for inputting the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint, so as to complete seismic travel time data regularization.

[0114] In some embodiments, the network training module 740 is further used for constructing a loss function according to travel time data errors of the travel time network and an eikonal equation of the target equation; and training the neural network according to the loss function and the actual seismic data, so as to obtain the trained travel time network and the velocity network.

[0115] The specific embodiments of the execution methods of the above modules have been described in detail in Embodiment One, and will not be described here again.

[0116] In summary, the seismic data regularization device provided by the embodiment realizes regularization of seismic travel time data by constructing a neural network guided by physics, training a trained travel time network and a velocity network, inputting coordinates of a target shotpoint into the trained neural network, and obtaining travel time data corresponding to the target shotpoint, which guarantees the accuracy of data processing and improves the data processing effect.

[0117] Embodiment Three

[0118] The embodiment provides an electronic device, which comprises a memory and a processor, and the memory stores a computer program, and the computer program is executed by the processor to realize the seismic data regularization method described in Embodiment One. It can be understood that the electronic device can further comprise an input / output (I / O) interface and a communication component.

[0119] The processor is configured to perform all or part of the steps of the seismic data regularization method as described in Embodiment One. The memory is configured to store various types of data, which can include, for example, instructions for any application programs or methods in the terminal device, and application-related data.

[0120] The processor can be an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic elements, configured to perform the seismic data regularization method as described in Embodiment One.

[0121] The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a Static Random Access Memory (SRAM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), an Erasable Programmable Read-Only Memory (EPROM), a Programmable Read-Only Memory (PROM), a Read-Only Memory (ROM), a magnetic storage, a flash memory, a magnetic disk, or an optical disk.

[0122] Embodiment Four

[0123] The present embodiment also provides a computer-readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read-Only Memory (ROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Programmable Read-Only Memory (PROM), a magnetic storage, a magnetic disk, an optical disk, a server, an App application store, etc., having a computer program stored thereon, wherein the computer program, when executed by a processor, can implement the following method steps:

[0124] obtaining actual seismic data collected by an exploration data observation system, the actual seismic data comprising actual shotpoint coordinates and travel time data in the exploration data observation system;

[0125] obtaining coordinates of a target shotpoint input by a user;

[0126] constructing a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network;

[0127] inputting the actual shotpoint coordinates and the travel time data into the neural network equation guided by the eikonal equation, and performing neural network training to obtain a trained travel time network and a trained velocity network;

[0128] inputting the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint, so as to complete seismic travel time data regularization.

[0129] The specific implementation process of the above method steps can be referred to the embodiment one, which will not be repeated here.

[0130] Embodiment five

[0131] The application also provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device perform the seismic data regularization method in any of the above embodiments.

[0132] In summary, the embodiment of the present application provides a seismic data regularization method and device, an electronic device and a storage medium; the method comprises: obtaining actual seismic data collected by an exploration data observation system, wherein the actual seismic data comprises actual shot-receiver point coordinates and travel time data in the exploration data observation system; obtaining coordinates of a target shot-receiver point input by a user; constructing a travel time neural network and a velocity neural network according to the actual shot-receiver point coordinates and the travel time data, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network; inputting the actual shot-receiver point coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training to obtain a trained travel time network and a velocity network; inputting the coordinates of the target shot-receiver point into the trained travel time network to obtain travel time data of the target shot-receiver point, so as to complete seismic travel time data regularization. In the method provided in the present solution, a neural network guided by physics is constructed, a trained travel time network and a velocity network are obtained by training, and then the coordinates of the target shot-receiver point are input into the trained neural network to obtain travel time data corresponding to the target shot-receiver point, so that the seismic travel time data is regularized, the accuracy of data processing is ensured, and the data processing effect is improved.

[0133] The above sequence numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of implementation.

[0134] It should be noted that in this document, the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements not only includes those elements, but also includes other elements not explicitly listed, or inherent to such a process, method, article or device. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or device that includes the element.

[0135] Although the embodiments disclosed in the present application are as described above, the above description is only for the purpose of facilitating understanding of the embodiments of the present application, and is not intended to limit the present application. Any person skilled in the art without departing from the spirit and scope of the present application can make any modification and change in the form and details of implementation, but the patent protection scope of the present application shall be subject to the scope defined by the appended claims.

Claims

1. A method of seismic data regularization, characterized in that, The method comprises the following steps: acquiring actual seismic data collected by an exploration data observation system, wherein the actual seismic data comprises actual shotpoint coordinates and travel time data in the exploration data observation system; acquiring coordinates of a target shotpoint input by a user; constructing a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and constructing a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network; inputting the actual shotpoint coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training to obtain a trained travel time network and a trained velocity network; inputting the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint, so as to complete seismic travel time data regularization.

2. The seismic data regularization method of claim 1, wherein, The step of inputting the actual shotpoint coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training to obtain a trained travel time network and a trained velocity network comprises the following steps: constructing a loss function according to a travel time data error of the travel time network and the eikonal equation; training the neural network according to the loss function and the actual seismic data to obtain the trained travel time network and the trained velocity network.

3. The seismic data regularization method of claim 2, wherein, The loss function comprises an augmented Lagrangian function.

4. The seismic data regularization method of any of claims 1 to 3, wherein, The velocity network and the travel time network are physical information neural networks with the same network architecture.

5. The seismic data regularization method of any one of claims 1 to 3, wherein, The velocity network and the travel time network are physical information neural networks with different network architectures.

6. The seismic data regularization method of claim 5, wherein, The velocity network comprises 7 hidden layers, and each layer comprises 20 neurons.

7. The seismic data regularization method of claim 5, wherein, The travel time network comprises 9 hidden layers, and each layer comprises 25 neurons.

8. An apparatus for seismic data regularization, characterized by The method comprises the following steps: a first acquisition module is configured to acquire actual seismic data collected by an exploration data observation system, wherein the seismic data comprises actual shotpoint coordinates and travel time data in the exploration data observation system; a second acquisition module is configured to acquire coordinates of a target shotpoint input by a user; a network construction module is configured to construct a travel time neural network and a velocity neural network according to the actual shotpoint coordinates and the travel time data, and to construct a neural network equation guided by an eikonal equation, wherein the eikonal equation is used to link the travel time neural network and the velocity neural network; a network training module is configured to input the actual shotpoint coordinates and the travel time data into the neural network equation guided by the eikonal equation to perform neural network training to obtain a trained travel time network and a trained velocity network; a regularization module is configured to input the coordinates of the target shotpoint into the trained travel time network to obtain travel time data of the target shotpoint, so as to complete seismic travel time data regularization.

9. An electronic device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by one or more processors to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by one or more processors to implement the method according to any one of claims 1 to 7.

Images