Surface layer Q field establishment method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a surface layer Q field establishment method and device, electronic equipment and a storage medium. The surface layer Q field establishment method comprises the following steps: acquiring first feature data of a plurality of surface layer Q value prediction points of a target work area; wherein the first feature data at least comprises an east coordinate, a north coordinate and a surface speed of a surface Q value prediction point; obtaining a plurality of groups of prediction input data according to the first feature data of the plurality of surface layer Q value prediction points; according to the multiple groups of prediction input data, utilizing a surface layer Q value neural network prediction model to obtain prediction values of surface layer Q values of the multiple surface layer Q value prediction points; and establishing a surface layer Q field of the target work area according to the predicted values of the surface layer Q values of the plurality of surface layer Q value prediction points and the surface layer Q values of the plurality of surface layer Q value survey points of the target work area.

Description

Technical Field

[0001] The present application relates to the field of petroleum geophysical exploration, and in particular to a surface Q field establishment method, device, electronic equipment and storage medium. Background Art

[0002] The surface Q value (hereinafter referred to as the surface Q value) of the formation quality factor is an important parameter for quantitatively describing the attenuation of seismic wave absorption. Compared with the deep medium, the surface absorbs and attenuates seismic waves more seriously. Accurately establishing the surface Q field and effectively compensating the seismic data are of great significance to improving the resolution of seismic data.

[0003] In the related art, there are many schemes for establishing the surface Q field, such as the surface Q value survey method, the fitting method, and the empirical formula method.

[0004] The surface Q value survey method mainly uses the double-well micro-logging and single-well micro-logging methods to conduct surface Q value surveys on the surface of the target work area to obtain the true surface Q value of the survey point. However, the cost of surface Q value surveys is high, the number of surface Q value survey points carried out in the actual target work area is small, and the control range is small, which cannot meet the requirements of finely establishing the surface Q field.

[0005] The surface Q field is usually established by fitting the relationship between the true surface Q value and the surface velocity of the surface Q value survey point, and the formula of the true surface Q value and the surface velocity is established. Then, the surface velocity of the surface Q value prediction point obtained by the first arrival of the ground seismic cannon or the previous small refraction and micro-logging data in the target work area is brought into the formula of the true surface Q value and the surface velocity to obtain the true surface Q value of the surface Q value prediction point. The process of establishing the formula of the surface true surface Q value and the surface velocity is subject to a large degree of human intervention, and it is usually difficult to establish an accurate formula of the surface true surface Q value and the surface velocity, resulting in a decrease in the accuracy of the established surface Q field.

[0006] The surface Q field is established by the empirical formula method. The surface velocity of the surface Q value survey point and the surface Q value prediction point obtained from the first arrival of the ground seismic cannon or the previous small refraction and micro-logging data of the target area is usually substituted into the empirical formula between the surface Q value and the surface velocity to obtain the surface relative surface Q value of the surface Q value survey point and the surface Q value prediction point. The relative surface Q value of the surface Q value survey point is calibrated by the real surface Q value of the surface Q value survey point to obtain the calibration coefficient of the surface relative surface Q value to the real surface Q value. The calibration coefficient is then used to convert the surface relative surface Q value of the surface Q value prediction point into the real surface Q value of the surface Q value prediction point, thereby establishing the surface Q field. The empirical formula method is complex to establish the surface Q field. It is usually difficult to obtain a calibration coefficient with strong applicability, and the calibration coefficient is greatly affected by human interference.

[0007] Therefore, how to accurately and conveniently establish the surface Q field is a technical problem that needs to be solved urgently. Summary of the invention

[0008] The embodiments of the present application provide a surface Q field establishment method, device, electronic device and storage medium for accurately and conveniently establishing a surface Q field.

[0009] One of the embodiments of the present application provides a method for establishing a surface Q field, the method comprising: obtaining first characteristic data of multiple surface Q-value prediction points of a target work area; wherein the first characteristic data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q-value prediction points; obtaining multiple groups of prediction input data based on the first characteristic data of the multiple surface Q-value prediction points; obtaining predicted values ​​of the surface Q values ​​of the multiple surface Q-value prediction points using a surface Q-value neural network prediction model based on the multiple groups of prediction input data; establishing the surface Q field of the target work area based on the predicted values ​​of the surface Q values ​​of the multiple surface Q-value prediction points and the surface Q values ​​of the multiple surface Q-value survey points of the target work area.

[0010] In some embodiments, the first feature data also includes a surface velocity fitting feature; wherein the surface velocity fitting feature is obtained by taking the square root of 2.2 of the surface velocity in the first feature data; and according to the first feature data of the plurality of surface Q value prediction points, a plurality of groups of prediction input data are obtained, including: for each of the first feature data of the surface Q value prediction point, performing the following processing: normalizing the first feature data; and using the normalized first feature data as a group of prediction input data.

[0011] In some embodiments, the surface Q value neural network prediction model is used according to the multiple groups of prediction input data to obtain the predicted values ​​of the surface Q value of the multiple surface Q value prediction points, including: for each surface Q value prediction point, performing the following processing: inputting the prediction input data of the surface Q value prediction point into the surface Q value neural network prediction model to obtain a first prediction value; performing denormalization processing on the first prediction value to obtain the predicted value of the surface Q value of the surface Q value prediction point.

[0012] In some embodiments, the surface Q field of the target work area is established based on the predicted values ​​of the surface Q values ​​of the multiple surface Q value prediction points and the surface Q values ​​of the multiple surface Q value investigation points of the target work area, including: obtaining the surface Q values ​​of the remaining positions of the target work area by using a spatial interpolation algorithm based on the predicted values ​​of the surface Q values ​​of the multiple surface Q value prediction points and the surface Q values ​​of the multiple surface Q value investigation points of the target work area; and establishing the surface Q field of the target work area by using the surface Q values ​​of all positions of the target work area.

[0013] In some embodiments, the surface Q-value neural network prediction model is a BP neural network model; the surface Q-value neural network prediction model is obtained by training in the following manner: obtaining a training data set; using the training data set to train an initial BP neural network model to obtain the surface Q-value neural network prediction model.

[0014] In some embodiments, the training data set includes multiple groups of training data and labels for each group of training data; obtaining the training data set includes: obtaining second feature data of multiple surface Q-value survey points in the target work area and surface Q-values ​​obtained using exploration data; wherein the second feature data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q-value survey point; for each of the surface Q-value survey points, performing the following operations: normalizing the second feature data of the surface Q-value survey point as a group of sample data; normalizing the surface Q-value of the surface Q-value survey point obtained using exploration data as a label for the sample data.

[0015] In some embodiments, the obtaining of second characteristic data of multiple surface Q-value survey points of the target work area includes: obtaining a set of basic data of all surface Q-value survey points of the target work area; wherein the basic data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q-value survey point; sorting the surface velocities in the set of basic data in descending order to form a velocity distribution interval, and removing the basic data corresponding to the surface velocities within a preset range at the beginning and the end of the velocity distribution interval from the set of basic data; and performing the following operations for the basic data of each surface Q-value survey point in the basic data set: performing a 2.2 power square root calculation on the surface velocity in the basic data to obtain the surface velocity fitting characteristics of the surface Q-value survey point; and obtaining the second characteristic data of multiple surface Q-value survey points of the target work area based on the basic data of all surface Q-value survey points in the basic data set and the surface velocity fitting characteristics.

[0016] In some embodiments, the use of the training data set to train the initial BP neural network model to obtain the surface Q-value neural network prediction model includes: using the initial BP neural network model to process sample data in the training data set; adjusting the parameters of the initial BP neural network model according to the difference between the output value of the initial BP neural network model and the label of the sample data until the convergence conditions of the model are met, thereby obtaining the surface Q-value neural network prediction model.

[0017] One of the embodiments of the present application provides a surface Q field establishment device, which includes: a first acquisition module, used to acquire first characteristic data of multiple surface Q value prediction points of a target work area; wherein the first characteristic data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q value prediction point; a second acquisition module, used to obtain multiple groups of prediction input data based on the first characteristic data of the multiple surface Q value prediction points; a third acquisition module, used to obtain predicted values ​​of surface Q values ​​of the multiple surface Q value prediction points based on the multiple groups of prediction input data using a surface Q value neural network prediction model; an establishment module, used to establish the surface Q field of the target work area based on the predicted values ​​of the surface Q values ​​of the multiple surface Q value prediction points and the surface Q values ​​of the multiple surface Q value survey points of the target work area.

[0018] An embodiment of the present application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the method described above when running the program.

[0019] An embodiment of the present application provides a storage medium for storing a computer-readable program. When the computer-readable program is executed, the method described above is executed.

[0020] The above technical solution provided by the embodiment of the present application has at least the following advantages compared with the prior art:

[0021] In the embodiment provided by the present application, the first characteristic data of multiple surface Q value prediction points of the target work area are obtained; multiple groups of prediction input data are obtained according to the first characteristic data of multiple surface Q value prediction points; the surface Q value prediction values ​​of multiple surface Q value prediction points are obtained by using the surface Q value neural network prediction model according to the multiple groups of prediction input data; the surface Q field of the target work area is established according to the surface Q value prediction values ​​of multiple surface Q value prediction points and the surface Q values ​​of multiple surface Q value survey points of the target work area. Thus, the surface Q field can be established accurately and conveniently. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The present application will be further described in the form of exemplary embodiments, which will be described in detail by way of the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0023] Figure 1 is an exemplary flow chart of a surface Q field establishment method according to some embodiments of the present application;

[0024] Figure 2 is an exemplary flow chart of a training method for a surface Q-value neural network prediction model according to some embodiments of the present application;

[0025] Figure 3is an exemplary schematic diagram of the structure of the surface Q-value neural network prediction model according to some embodiments of the present application;

[0026] Figure 4A is an exemplary schematic diagram of a comparison between a surface Q value obtained using the method provided by the present application and a surface Q value obtained using exploration data according to some embodiments of the present application;

[0027] Figure 4B is another exemplary schematic diagram of a comparison between a surface Q value obtained using the method provided by the present application and a surface Q value obtained using exploration data according to some embodiments of the present application;

[0028] Figure 5A is an exemplary schematic diagram of the comparison between the surface Q value obtained according to the related art and the surface Q value obtained using the exploration data;

[0029] Figure 5B is another exemplary schematic diagram of a comparison between a surface Q value obtained according to a related art and a surface Q value obtained using exploration data;

[0030] Fig. 6A is an exemplary schematic diagram of a plane distribution diagram of a surface Q value of a control point according to some embodiments of the present application;

[0031] Figure 6B is another exemplary schematic diagram of a plane distribution diagram of a surface Q value of a control point according to some embodiments of the present application;

[0032] Fig. 7A is an exemplary schematic diagram of a Q field diagram obtained by using a spatial interpolation algorithm according to some embodiments of the present application;

[0033] Figure 7B is another exemplary schematic diagram of a Q field diagram obtained by using a spatial interpolation algorithm according to some embodiments of the present application;

[0034] Figure 8 It is a ground seismic profile before and after surface Q value compensation according to some embodiments of the present application;

[0035] Fig. 9A It is a comparison diagram of the spectrum of the ground seismic profile before and after the surface Q value compensation shown in some embodiments of the present application;

[0036] Fig. 9B It is another comparison diagram of the spectrum of the ground seismic profile before and after the surface Q value compensation shown in some embodiments of the present application;

[0037] Fig.10 is an exemplary schematic diagram of a surface Q field establishment device according to some embodiments of the present application;

[0038] Fig.11 It is a schematic diagram of an exemplary structure of an electronic device according to some embodiments of the present application. DETAILED DESCRIPTION

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0040] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0041] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprises" and "includes" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0042] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, the various steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0043] For ease of understanding, the technical solution of the present application is introduced below in conjunction with the accompanying drawings and embodiments.

[0044] Figure 1 It is an exemplary flow chart of a method for establishing a surface Q field according to some embodiments of the present application.

[0045] like Figure 1 As shown, the surface Q field establishment method includes the following steps:

[0046] Step S110, obtaining first characteristic data of a plurality of surface Q value prediction points of the target work area.

[0047] The first characteristic data is data related to the surface Q value, which at least includes the east coordinate, the north coordinate, and the surface velocity of the surface Q value prediction point.

[0048] In some embodiments, in order to obtain a more accurate prediction effect, the first characteristic data further includes a surface velocity fitting feature. The surface velocity fitting feature has a linear relationship with the surface Q value, and the surface velocity fitting feature can be obtained by performing a 2.2 power square root calculation on the surface velocity in the first characteristic data.

[0049] In the specific implementation process, the first feature data of the surface Q value prediction point corresponds to the second feature data input into the surface Q value neural network prediction model during training. For a detailed description of the surface Q value neural network prediction model, refer to the relevant content in step S130, which will not be repeated here. For a detailed description of the second feature data of the surface Q value neural network prediction model obtained by training, refer to Figure 2 The relevant content in will not be repeated here.

[0050] In the specific implementation process, multiple surface Q value prediction points can be selected in a variety of ways. For example, the test location points can be selected according to the distribution of the surface Q value survey points in the target work area so that the surface Q value prediction points and the surface Q value survey points are relatively evenly located in the target work area.

[0051] Step S120, obtaining multiple sets of prediction input data according to the first feature data of multiple surface Q-value prediction points.

[0052] In a specific implementation process, the first characteristic data of multiple surface Q-value prediction points may be preprocessed, and the preprocessed first characteristic data may be used as multiple groups of prediction input data.

[0053] In some embodiments, the following processing may be performed for the first feature data of each surface Q value prediction point:

[0054] The first feature data is normalized; and the normalized first feature data is used as a set of prediction input data.

[0055] As an example only, the detailed process of normalizing the first feature data is described below.

[0056] 1) The first characteristic data of the surface Q value prediction point obtained includes the east coordinate x′[i], the north coordinate y′[i], the surface elevation z′[i], the surface velocity v′[i], and the surface velocity fitting feature v′ 2.2 [i], surface thickness h′[i], and seismic wave propagation time t′[i] in the surface.

[0057] 2) Utilize Figure 2 The normalization parameters of the training data set features (for detailed description, see Figure 2First, the east coordinate of the surface Q value prediction point is normalized. The formula is as follows:

[0058]

[0059] Among them, x′[i] is the original value of the east coordinate of the i-th surface Q value prediction point, and x′[i]1 is the normalized feature of the east coordinate of the i-th surface Q value prediction point.

[0060] Similarly, the north coordinate y′[i], surface elevation z′[i], surface velocity v′[i], and surface velocity fitting feature v′ of the surface Q value prediction point are completed. 2.2 The normalized processing of surface thickness h′[i], surface propagation time t′[i] of seismic waves in the surface is performed to obtain the normalized features of the surface Q value prediction point y′[i]1, z′[i]1, v′[i]1, v′ 2.2 [i]1, h′[i]1, t′[i]1.

[0061] Step S130, based on multiple groups of prediction input data, using the surface Q value neural network prediction model, obtain the predicted values ​​of the surface Q values ​​of multiple surface Q value prediction points.

[0062] In the specific implementation process, a variety of machine learning models can be used to establish a surface Q value neural network prediction model, including but not limited to: convolutional neural network model, random forest model, etc. In some embodiments, the surface Q value neural network prediction model is a BP neural network model. The number of neurons in the input layer of the BP neural network model is consistent with the number of features of the predicted input data; the number of hidden layers n and the number of neurons in each hidden layer m are set as hyperparameters; the output layer has only one neuron.

[0063] Just as an example, Figure 3 As shown, the dimension of the first feature data is 6, so the number of neurons in the input layer of the surface Q-value neural network prediction model is also 6; the number of hidden layers is set to 2, of which the number of neurons in the first hidden layer is set to 9, and the number of neurons in the second hidden layer is set to 5.

[0064] The activation function used in each hidden layer is the sigmoid function, which is as follows:

[0065]

[0066] In some embodiments, the prediction input data is normalized in step S120, and the following processing can be performed for each surface Q-value prediction point: the prediction input data of the surface Q-value prediction point is input into the surface Q-value neural network prediction model, and the output layer of the surface Q-value neural network prediction model outputs a first prediction value; the first prediction value is denormalized to obtain the predicted value of the surface Q-value of the surface Q-value prediction point.

[0067] As an example only, the first feature data x′[i]1, y′[i]1, z′[i]1, v′[i]1, v′[i]1, v′[i]1 of the surface Q value prediction point normalized in step S120 may be 2.2 [i]1, h′[i]1, t′[i]1 are input into the trained surface Q value neural network prediction model to obtain the first prediction value q′[i] output by the surface Q value neural network prediction model. Then, the parameters for normalizing the labels of the training data set are used (see the detailed description for details). Figure 2 The data output by the neural network is denormalized. The denormalization formula is:

[0068] Q′[i]=q′[i]*(q max -q min )+q min (3)

[0069] Among them, q max and q min is the maximum and minimum value of the surface Q value in the training data set; the denormalized data Q′[i] is the predicted value of the surface Q value of the i-th surface Q value prediction point.

[0070] Step S140, establishing a surface Q field of the target work area according to the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points and the surface Q values ​​of the plurality of surface Q value investigation points of the target work area.

[0071] In the specific implementation process, the exploration data can be used to obtain the surface Q values ​​of multiple surface Q value survey points in the target work area. As an example only, the surface Q value survey of the target work surface can be carried out by using double-well micro-logging, single-well micro-logging, etc. to obtain the surface Q values ​​of multiple surface Q value survey points. Due to the high cost of the surface Q value survey, the number of surface Q values ​​of the surface Q value survey points obtained by the above method is small, and it is difficult to establish the surface Q field of the target work area based only on the surface Q values ​​of the surface Q value survey points.

[0072] In some embodiments, the predicted values ​​of the surface Q values ​​of multiple surface Q value prediction points and the surface Q values ​​of multiple surface Q value survey points can be used as the surface Q value control points of the target work area, and the surface Q values ​​of the remaining positions of the target work area can be obtained by using a spatial interpolation algorithm. The spatial interpolation algorithm can be an inverse distance interpolation formula, a Kriging interpolation formula, etc.

[0073] Finally, the surface Q values ​​of all positions in the target work area obtained by the above method can be used to establish the surface Q field of the target work area.

[0074] The technical effects of the embodiments provided in this application are described below.

[0075] Figure 4A and Figure 4B It is the prediction result of the test data set using the trained surface Q value neural network prediction model. In order to prevent the model from overfitting, the root mean square error of the prediction result of the test data set is 1.1117, and the change trend is basically consistent with the true value, which can meet the requirements for establishing the surface Q field.

[0076] Figure 5A and Figure 5B It is the prediction result of the test data set using the existing technology. The root mean square error of the prediction result of the test data set is 1.2513, which is larger than the surface Q value prediction error based on the neural network algorithm proposed in this application, and is basically inconsistent with the changing trend of the true value.

[0077] Fig. 6A and Figure 6B It is a plane distribution map of surface Q value control points composed of surface Q values ​​of surface Q value prediction points and surface Q values ​​of surface Q value investigation points calculated using the method provided in this application.

[0078] Fig. 7A and Figure 7B is based on Figure 5A and Figure 5B The surface Q value control points shown in FIG. 1 are used to obtain the surface Q field using the spatial interpolation algorithm. The surface Q compensation is performed on the surface seismic profile using the surface Q field. The surface seismic profiles before and after compensation are shown in FIG. Figure 8 It can be seen that after surface Q compensation, the event axis of the seismic profile is more continuous and the signal-to-noise ratio is higher (white arrow position). Fig. 9A and Fig. 9B It is the spectrum of the ground seismic profile before and after surface Q value compensation. It can be seen that after surface Q value compensation, the high-frequency end of the ground seismic profile is expanded from 48Hz to 66Hz, which proves the effectiveness of the surface Q field establishment method based on the neural network algorithm of the present application.

[0079] Figure 2 It is an exemplary flow chart of the training method of the surface Q-value neural network prediction model shown in some embodiments of the present application.

[0080] In this embodiment, the surface Q value neural network prediction model is constructed based on the BP neural network model, and the surface Q value neural network prediction model can be trained through the following steps:

[0081] Step S210, obtaining a training data set.

[0082] The training data set includes multiple sets of training data and labels for each set of training data.

[0083] In some embodiments, second characteristic data of multiple surface Q value survey points in the target work area and surface Q values ​​obtained using exploration data can be obtained; wherein the second characteristic data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q value survey points.

[0084] In some embodiments, a set of basic data of all surface Q value survey points of the target work area may be obtained. The basic data of the surface Q value survey points may include the east coordinate, north coordinate, surface elevation, surface velocity, surface thickness, propagation time of seismic waves in the surface, surface Q value obtained using exploration data, and other data of the surface Q value survey points.

[0085] The above-obtained set of basic data may contain inaccurate surface velocities (e.g., erroneous data obtained due to measurement errors or recording errors). In some embodiments, the surface velocities in the set of basic data may be sorted in descending order to form a velocity distribution interval, and the basic data corresponding to the surface velocities within a preset range at the beginning and end of the velocity distribution interval (e.g., 3% of the velocity distribution interval) are removed from the set of basic data, thereby removing abnormal data from the set of basic data.

[0086] Since there is a nonlinear relationship between the surface Q value and the surface velocity, in order to enable the surface Q value neural network prediction model to learn the nonlinear relationship and improve the accuracy of model prediction, in some embodiments, the training data may include surface velocity fitting features. In the specific implementation process, the following operation may be performed for the basic data of each surface Q value survey point in the basic data set: the surface velocity in the basic data is squared to the power of 2.2 to obtain the surface velocity fitting features of the surface Q value survey point.

[0087] In some embodiments, the second characteristic data of multiple surface Q value survey points in the target work area can be obtained based on the basic data of all surface Q value survey points in the basic data set and the surface velocity fitting characteristics. The surface Q value obtained using the exploration data is used as a label for the corresponding second characteristic data.

[0088] In the specific implementation process, it is necessary to check and process the basic data collected above. Check to determine whether the basic data is missing. In the case where the basic data has missing values, the missing data can be supplemented in a variety of ways. As an example only, the data of multiple (for example, 5) surface Q-value survey points near the surface Q-value survey point corresponding to the missing value can be averaged, and the missing value can be replaced by the average value. For example, if the north coordinate feature of the 9th surface Q-value survey point is missing, the north coordinates of the 5 surface Q-value survey points closest to the 9th surface Q-value survey point can be averaged as the north coordinate feature of the 9th surface Q-value survey point.

[0089] In some embodiments, the following operations may be performed for each surface Q value survey point:

[0090] The second characteristic data of the surface Q value survey point are normalized and used as a set of sample data; the surface Q value of the surface Q value survey point obtained by using the exploration data is normalized and used as the label of the sample data.

[0091] In the specific implementation process, the data obtained from all surface Q value survey points can be normalized and divided into a training data set and a test data set of the surface Q value survey points in a ratio of 8:2. The training data set is used to train the surface Q value neural network prediction model. The test data set is used to test the trained surface Q value neural network prediction model to determine its prediction accuracy, and then the trained model that can accurately predict the surface Q value is used as the surface Q value neural network prediction model.

[0092] As an example only, a specific method for normalizing the data obtained from the surface Q value survey points is described in detail below.

[0093] 1) Normalize the data and labels in the training dataset.

[0094] First, normalize the east coordinate feature x[i] in the training data set of the surface Q value survey point. Find the minimum value x of the east coordinate feature x[i] in the training data set. min and the maximum value x max , x min 、x max It is called the normalization parameter of the east coordinate feature of the training data set. Then the normalized data x[i]1 of the east coordinate feature in the training data set is calculated using the following formula:

[0095]

[0096] Among them, x[i] is the original value of the east coordinate feature of the i-th surface Q value survey point in the training data set, x minis the minimum value of the east coordinate feature of the training data set, x max is the maximum value of the east coordinate feature of the training data set, and x[i]1 is the normalized feature of the east coordinate feature of the i-th surface Q value survey point in the training data set.

[0097] Similarly, find the north coordinate y[i], surface elevation z[i], surface velocity v[i], and surface velocity fitting feature v of the training data set 2.2 [i], layer thickness h[i], seismic wave propagation time t[i] in the surface layer, and the minimum and maximum values ​​in the training data set label q[i], and then the training data set north coordinate y[i], surface elevation z[i], surface velocity v[i], and surface velocity fitting feature v[i] are obtained. 2.2 Data y of the minimum and maximum values ​​of the surface thickness h[i], the propagation time of the seismic wave in the surface t[i] min ,y max 、z min 、z max 、v min 、v max 、v 2.2min 、v 2.2max 、h min 、h max ,t min ,t max , and get the minimum and maximum data q of the training data set label q[i] min ,q max .

[0098] According to formula (4), the features y[i], z[i], v[i], v[i] of all training data sets are completed in sequence. 2.2 Normalize h[i], t[i], and label q[i] to obtain the normalized features y[i]1, z[i]1, v[i]1, v 2.2 [i]1, h[i]1, t[i]1 and the normalized label q[i]1.

[0099] 2) Normalize the data and labels in the test dataset.

[0100] First, the parameter x is normalized using the east coordinate in the training dataset min 、x max , use the normalization formula shown below to normalize the east coordinate feature X[i] in the test data set. The normalization formula is:

[0101]

[0102] Among them, X[i] is the original value of the east coordinate feature of the i-th surface Q value survey point in the test data set, and X[i]1 is the normalized feature of the east coordinate feature of the i-th surface Q value survey point in the test data set.

[0103] Similarly, according to formula (5), the features Y[i], Z[i], V[i], V[i] of all test data sets are completed in sequence. 2.2 Normalize [i], H[i], T[i] and label Q[i] to obtain the normalized features Y[i]1, Z[i]1, V[i]1, V 2.2 [i]1, H[i]1, T[i]1 and the normalized label Q[i]1.

[0104] Step S220, using the training data set, train the initial BP neural network model to obtain a surface Q value neural network prediction model.

[0105] In some embodiments, an initial BP neural network model can be used to process sample data in a training data set; based on the difference between the output value of the initial BP neural network model and the label of the sample data, the parameters of the initial BP neural network model are adjusted until the convergence conditions of the model are met, thereby obtaining a surface Q value neural network prediction model.

[0106] The convergence condition of the model can be that the loss function is reduced to below a set value, for example, less than 1e-6.

[0107] Fig.10 It is an exemplary schematic diagram of a surface Q field establishing device according to some embodiments of the present application.

[0108] like Fig.10 As shown, the surface Q field establishing device includes: a first acquisition module 1010 , a second acquisition module 1020 , a third acquisition module 1030 , and an establishing module 1040 .

[0109] The first acquisition module 1010 is used to acquire first characteristic data of multiple surface Q value prediction points of the target work area; wherein the first characteristic data at least includes the east coordinate, the north coordinate, and the surface velocity of the surface Q value prediction point.

[0110] The second acquisition module 1020 is used to obtain multiple groups of prediction input data according to the first feature data of the multiple surface Q-value prediction points.

[0111] The third acquisition module 1030 is used to obtain the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points according to the plurality of groups of prediction input data and using the surface Q value neural network prediction model.

[0112] The establishing module 1040 is used to establish the surface Q field of the target work area according to the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points and the surface Q values ​​of the plurality of surface Q value investigation points of the target work area.

[0113] In the above-mentioned embodiment of the surface Q field establishing device, the specific processing of each module and the technical effects brought about by it can be referred to the relevant description in the corresponding method embodiment, which will not be repeated here.

[0114] Fig.11 It is a schematic diagram of an exemplary structure of an electronic device according to some embodiments of the present application.

[0115] like Fig.11 As shown, the electronic device includes: at least one processor 1101, at least one communication interface 1102, at least one memory 1103 and at least one communication bus 1104; optionally, the communication interface 1102 can be an interface of a communication module, such as an interface of a GSM module; the processor 1101 may be a processor CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. The memory 1103 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Among them, the memory 1103 stores a program, and the processor 1101 calls the program stored in the memory 1103 to execute some or all of the above-mentioned method embodiments.

[0116] The present application relates to a storage medium for storing a computer-readable program. When the computer-readable program is executed, part or all of the above-mentioned method embodiments are executed.

[0117] Alternatively, the storage medium may be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

[0118] Based on the same inventive concept, an embodiment of the present application further provides a computer program product, including a computer program, which implements part or all of the above-mentioned method embodiments when executed by a processor.

[0119] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of the present application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and amendments to the present application. Such modifications, improvements and amendments are suggested in the present application, so such modifications, improvements and amendments still belong to the spirit and scope of the exemplary embodiments of the present application.

[0120] At the same time, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or multiple times in different positions in the present application does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present application can be appropriately combined.

[0121] In addition, unless explicitly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some invention embodiments that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this application. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0122] Similarly, it should be noted that in order to simplify the description of the disclosure of this application and thus help understand one or more embodiments of the invention, in the above description of the embodiments of this application, multiple features are sometimes combined into one embodiment, figure or description thereof. However, this disclosure method does not mean that the features required by the object of this application are more than the features mentioned in the claims. In fact, the features of the embodiments are less than all the features of the single embodiment disclosed above.

[0123] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining digits. Although the numerical domains and parameters used to confirm the breadth of their range in some embodiments of the present application are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0124] Each patent, patent application, patent application disclosure, and other materials, such as articles, books, instructions, publications, documents, etc., cited in this application are hereby incorporated by reference in their entirety. Except for application history documents that are inconsistent with or conflicting with the content of this application, documents that limit the broadest scope of the claims of this application (currently or later attached to this application) are also excluded. It should be noted that if the descriptions, definitions, and / or use of terms in the attached materials of this application are inconsistent or conflicting with the content described in this application, the descriptions, definitions, and / or use of terms in this application shall prevail.

[0125] Finally, it should be understood that the embodiments described in this application are only used to illustrate the principles of the embodiments of the present application. Other variations may also fall within the scope of the present application. Therefore, as an example and not a limitation, the alternative configurations of the embodiments of the present application may be considered to be consistent with the teachings of the present application. Accordingly, the embodiments of the present application are not limited to the embodiments explicitly introduced and described in the present application.

Claims

1. A surface Q field establishment method, characterized in that: The method comprises: Acquire first characteristic data of a plurality of surface Q value prediction points in a target work area; wherein the first characteristic data at least includes the east coordinate, the north coordinate, and the surface velocity of the surface Q value prediction points; Obtaining multiple sets of prediction input data according to the first feature data of the multiple surface Q value prediction points; According to the plurality of groups of prediction input data, using a surface Q value neural network prediction model, obtaining predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points; A surface Q field of the target work area is established according to the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points and the surface Q values ​​of the plurality of surface Q value investigation points of the target work area.

2. The method according to claim 1, characterized in that The first characteristic data also includes a surface velocity fitting feature; wherein the surface velocity fitting feature is obtained by performing a 2.2 power square root calculation on the surface velocity in the first characteristic data; The step of obtaining multiple sets of prediction input data based on the first feature data of the multiple surface Q value prediction points includes: For each of the first feature data of the surface Q value prediction point, the following processing is performed: performing normalization processing on the first feature data; The normalized first feature data is used as a set of prediction input data.

3. The method according to claim 2, characterized in that The method of obtaining the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points by using the surface Q value neural network prediction model according to the plurality of sets of prediction input data includes: For each surface Q value prediction point, perform the following processing: Inputting the prediction input data of the surface Q value prediction point into the surface Q value neural network prediction model to obtain a first prediction value; The first prediction value is subjected to inverse normalization processing to obtain a predicted value of the surface Q value of the surface Q value prediction point.

4. The method according to claim 1, characterized in that: The step of establishing the surface Q field of the target work area according to the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points and the surface Q values ​​of the plurality of surface Q value investigation points of the target work area comprises: According to the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points and the surface Q values ​​of the plurality of surface Q value survey points of the target work area, the surface Q values ​​of the remaining positions of the target work area are obtained by using a spatial interpolation algorithm; The surface Q field of the target work area is established by using the surface Q values ​​of all positions of the target work area.

5. The method according to claim 1, characterized in that The surface Q value neural network prediction model is a BP neural network model; The surface Q value neural network prediction model is obtained by training in the following way: Get the training dataset; The initial BP neural network model is trained using the training data set to obtain the surface Q value neural network prediction model.

6. The method according to claim 5, characterized in that The training data set includes multiple groups of training data and labels for each group of training data; The step of obtaining a training data set includes: Acquire second characteristic data of a plurality of surface Q value survey points of the target work area and surface Q values ​​obtained using exploration data; For each of the surface Q value survey points, perform the following operations: The second characteristic data of the surface Q value survey point are normalized to obtain a set of sample data; The surface Q value of the surface Q value survey point obtained by using the exploration data is normalized and used as the label of the sample data.

7. The method according to claim 6, characterized in that The step of obtaining second characteristic data of a plurality of surface Q value survey points of the target work area includes: Acquire a set of basic data of all surface Q value survey points of the target work area; wherein the basic data at least includes the east coordinate, north coordinate, and surface velocity of the surface Q value survey points; Sorting the surface velocities in the basic data set in descending order to form a velocity distribution interval, and removing the basic data corresponding to the surface velocities within a preset range at both ends of the velocity distribution interval from the basic data set; and For the basic data of each surface Q value survey point in the basic data set, perform the following operations: Perform 2.2 power square root calculation on the surface velocity in the basic data to obtain the surface velocity fitting characteristics of the surface Q value survey point; According to the basic data of all the surface Q value survey points in the basic data set and the surface velocity fitting characteristics, the second characteristic data of a plurality of surface Q value survey points in the target work area are obtained.

8. A surface Q field establishment device, characterized in that: The device comprises: A first acquisition module is used to acquire first characteristic data of a plurality of surface Q value prediction points in a target work area; wherein the first characteristic data at least includes the east coordinate, the north coordinate, and the surface velocity of the surface Q value prediction point; A second acquisition module, configured to obtain multiple sets of prediction input data according to the first feature data of the multiple surface Q value prediction points; A third acquisition module is used to obtain the predicted values ​​of the surface Q values ​​of the plurality of surface Q value prediction points according to the plurality of groups of prediction input data and using the surface Q value neural network prediction model; An establishing module is used to establish the surface Q field of the target work area according to the predicted values ​​of the surface Q values ​​of the multiple surface Q value prediction points and the surface Q values ​​of the multiple surface Q value investigation points of the target work area.

9. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the method according to any one of claims 1 to 7 when running the program.

10. A storage medium for storing a computer-readable program, wherein when the computer-readable program is executed, the method according to any one of claims 1 to 7 is executed.

Images