Semi-supervised time domain speed modeling method and device based on expert experience constraint

Abstract

The invention discloses a semi-supervised time domain speed modeling method and device based on expert experience constraint. The method comprises the following steps: acquiring an initial velocity function of an extracted seismic gather velocity spectrum constructed according to expert experience, and obtaining initial velocity functions of velocity spectrums of other seismic gathers in a research area through interpolation; based on an initial velocity function of a velocity spectrum, performing extrapolation according to a set rule to obtain a candidate region; for the velocity spectrum of each seismic trace set, identifying an energy cluster from the candidate region, and tracking the energy maximum position of the energy cluster to obtain a velocity function; and performing time interpolation on the velocity function, and obtaining a time domain velocity model of the research area according to the velocity function after interpolation of each seismic trace set. Through the semi-supervised time domain speed modeling method based on expert experience constraint, the influence of interference noise and multiple wave energy groups can be effectively avoided, the energy group pickup precision is ensured, the intellectualization and automation of time domain speed analysis are realized, the manpower is liberated, and the quality and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic data processing in the field of geophysical exploration, and in particular to a semi-supervised time domain velocity modeling method and device based on expert experience constraints. Background Art

[0002] Velocity analysis is the key to seismic data processing and is inseparable from the identification of velocity spectrum energy clusters in common center point gathers (CMP) and common reflection point gathers (CRP). Ensuring the accuracy of velocity picking is of great significance for subsequent processing. It requires experienced professional processing personnel to manually pick, which is time-consuming, labor-intensive and highly repetitive. Considering the labor cost, the manually picked velocity spectrum is usually low-density, and then the velocity of each CMP data is obtained by interpolation. At the same time, manual picking is often affected by subjective factors, and it is difficult to unify the picking standard. In order to improve the efficiency of velocity analysis, traditional automatic velocity picking methods mainly include iterative optimization methods, which are inversion methods that use optimization algorithms and maximum similarity strategies to invert the optimal solution of stacking velocity in the velocity spectrum. This type of method usually requires the assumption that the velocity model is linear, while noise and other interference signals make the nonlinear problems of this type of method prominent. At the same time, the iterative calculation amount is large, which reduces the picking efficiency. Therefore, this type of method is difficult to apply industrially. In recent years, computer performance has improved significantly, and machine learning has also been applied to velocity analysis, mainly including deep learning and unsupervised clustering methods. Deep learning can be regarded as a data-driven method. It establishes the optimal nonlinear mapping relationship between seismic data (such as common midpoint gathers, velocity spectra) and velocity labels through a neural network model of a certain depth and a back-propagation algorithm, and uses this relationship to predict velocity. Summary of the invention

[0003] The inventors found that supervised neural networks require processing personnel to manually pick up a large amount of rich label data for network training, which is time-consuming, labor-intensive and has poor generalization. When predicting seismic data with different characteristics, it is necessary to rebuild the label data and retrain the network. In addition, since the neural network is a huge composite function, it is difficult to be interpretable. For unsupervised clustering methods, seismic data features can be searched without establishing labels, data with the same characteristics can be grouped, and the approximate distribution of the data set can be obtained. The algorithm of this type of method is simple and easy to implement, with a small amount of calculation, high interpretability, and is applicable to seismic data of any characteristics. Therefore, it is more generalizable in industrial applications. However, unsupervised clustering algorithms often only consider the ability to track velocity spectrum energy groups, resulting in the velocity analysis being affected by multiple waves and other interference signal energy groups, reducing the accuracy of velocity analysis.

[0004] In order to at least partially solve the technical problems existing in the prior art, the inventors have made the present invention, and through a specific implementation method, provide a semi-supervised time domain velocity modeling method and device based on expert experience constraints, which can effectively avoid the influence of interference noise and multiple wave energy clusters, ensure the accuracy of energy cluster picking, realize the intelligence and automation of time domain velocity analysis, liberate manpower, and achieve quality improvement and efficiency enhancement.

[0005] In a first aspect, an embodiment of the present invention provides a semi-supervised time domain speed modeling method based on expert experience constraints, comprising:

[0006] The initial velocity function is obtained based on the velocity picking result of the velocity spectrum of the extracted seismic gathers, and the initial velocity function of the velocity spectrum of other seismic gathers in the study area is obtained by interpolation;

[0007] Based on the initial velocity function of the velocity profile, the candidate area is obtained by extrapolation according to the set rules;

[0008] For the velocity spectrum of each seismic trace gather, energy clusters are identified from the candidate area, the maximum energy position of the energy cluster is tracked, and the velocity function is obtained;

[0009] The velocity function is interpolated in time, and the time domain velocity model of the study area is obtained from the interpolated velocity function of each seismic gather.

[0010] In a second aspect, an embodiment of the present invention provides a semi-supervised time-domain velocity modeling device based on expert experience constraints, comprising:

[0011] Initialization module, used to obtain the initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather, and to obtain the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation;

[0012] A candidate region determination module is used to obtain a candidate region by extrapolating the initial speed function of the speed spectrum according to a set rule;

[0013] Energy cluster tracking module, used to identify energy clusters in candidate regions based on the velocity spectrum of each seismic trace gather;

[0014] The velocity function building module is used to track the maximum energy position of the energy group and obtain the velocity function;

[0015] The time domain velocity model building module is used to perform time interpolation on the velocity function, and obtain the time domain velocity model of the study area from the velocity function after interpolation of each seismic gather.

[0016] In a third aspect, an embodiment of the present invention provides a computer storage medium, in which computer executable instructions are stored. When the computer executable instructions are executed by a processor, the above-mentioned semi-supervised time domain speed modeling method based on expert experience constraints is implemented.

[0017] In a fourth aspect, an embodiment of the present disclosure provides a server, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the semi-supervised time domain speed modeling method based on expert experience constraints is implemented.

[0018] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0019] (1) The semi-supervised time-domain velocity modeling method based on expert experience constraint provided by the embodiment of the present invention first introduces the geological background experience of the expert on the velocity field, obtains the initial velocity function constructed according to the expert experience of the velocity spectrum of the extracted seismic gather, obtains the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation, and obtains the candidate area based on the initial velocity function of the velocity spectrum according to the set rules; for the velocity spectrum of each seismic gather, the energy group is identified with the candidate area as a constraint, and the energy maximum position of the energy group is tracked to obtain the velocity function; the velocity function is interpolated in time, and the time-domain velocity model of the study area is obtained from the velocity function after interpolation of each seismic gather. This method realizes the intelligence and automation of time-domain velocity analysis, significantly improves the efficiency of time-domain velocity analysis, liberates manpower, avoids the burden brought by manual picking, and effectively realizes the improvement of quality and efficiency of velocity analysis; wherein, by manually picking a small number of velocity functions by experts and interpolating the initial velocity functions of all gathers, the noise resistance of the unsupervised clustering algorithm can be improved, the interference of noise and multiple wave energy groups can be avoided, and the evolution efficiency can be improved.

[0020] (2) The semi-supervised time-domain velocity modeling method based on expert experience constraints provided in the embodiment of the present invention introduces Gaussian kernel function and velocity spectrum energy value as constraints in the mean shift process, which can effectively improve the accuracy and efficiency of velocity analysis.

[0021] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0022] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0024] Figure 1 It is a flow chart of a semi-supervised time-domain speed modeling method based on expert experience constraints in Embodiment 1 of the present invention;

[0025] Figure 2 This is a specific implementation flow chart of the semi-supervised time-domain velocity modeling method based on expert experience constraints in the second embodiment of the present invention;

[0026] Figure 3 It is a velocity spectrum of a CMP gather of actual three-dimensional onshore data of the xx basin in the second embodiment of the present invention;

[0027] Figure 4 Yes Figure 3 The velocity spectrum constructs the segmentation candidate area of ​​the CV segmentation model according to expert experience;

[0028] Figure 5 Yes Figure 4 The candidate regions of are iteratively evolved using the CV segmentation model to obtain the effective energy cluster of the velocity spectrum;

[0029] Figure 6 Yes Figure 5 The energy peak of the effective energy cluster is tracked by mean shift unsupervised clustering to obtain the velocity function of the velocity spectrum of the track;

[0030] Figure 7 Yes Figure 6 The velocity function is interpolated at each sampling moment, and the velocity functions of all CMP gathers obtained through the same process are combined to form a three-dimensional stacked velocity body in the time domain;

[0031] Figure 8 Yes Figure 7 The stacked imaging volume is obtained by performing CMP gather dynamic correction stacking on the time domain three-dimensional stacked velocity volume;

[0032] Fig. 9 Schematic diagram of the structure of a semi-supervised time-domain velocity modeling device based on expert experience constraints in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0034] It should be understood that the terms described in the present invention are only for describing special embodiments and are not intended to limit the present invention. In addition, for the numerical range in the present invention, it should be understood that each intermediate value between the upper and lower limits of the scope is also specifically disclosed. Each smaller range between the intermediate value in any stated value or stated range and any other stated value or intermediate value in the described range is also included in the present invention. The upper and lower limits of these smaller ranges can be independently included or excluded in the scope.

[0035] Unless otherwise indicated, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which the invention belongs. Although the present invention describes only preferred methods and materials, any methods and materials similar or equivalent to those described herein may also be used in the implementation or testing of the present invention. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods and / or materials related to the documents. In the event of a conflict with any incorporated document, the content of this specification shall prevail.

[0036] Embodiment 1

[0037] Embodiment 1 of the present invention provides a semi-supervised time domain speed modeling method based on expert experience constraints, and its process is as follows Figure 1 As shown, the following steps are included:

[0038] Step S11: Obtain an initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather, and obtain the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation.

[0039] First, the velocity spectrum of each seismic trace in the study area is preprocessed as follows to ensure the accuracy of subsequent processing:

[0040] Median filtering is used to remove random salt and pepper noise and unify the speed scanning range of the speed spectrum.

[0041] Specifically, the velocity spectrum is a time domain stacking velocity spectrum or a root mean square velocity spectrum; the seismic gather is a common center point gather CMP, or a common reflection point gather CRP.

[0042] Experts’ geological background experience of velocity field is introduced, and a small number of velocity functions are manually picked up by experts and interpolated to all input CMP / CRP gather data.

[0043] Step S12: Based on the initial speed function of the speed map, the candidate area is obtained by extrapolation according to the set rules.

[0044] Step S13: for each seismic trace gather velocity spectrum, identify energy clusters from within the candidate region, track the energy maximum position of the energy cluster, and obtain the velocity function.

[0045] In some embodiments, the Chan-vese model segmentation method can be used to identify energy clusters from within the candidate regions.

[0046] In some embodiments, an unsupervised mean-shift clustering method may be used to track the maximum energy position of each energy cluster.

[0047] Furthermore, in the iterative process of unsupervised mean shift clustering, Gaussian kernel function and velocity spectrum energy value are introduced as constraints to track the maximum energy position of each energy cluster.

[0048] Specifically, the maximum energy position of each energy group is tracked by the following formula:

[0049]

[0050] Among them, x * is the maximum energy position of the energy cluster picked up by mean shift clustering in the current round, x is the maximum energy position of the energy cluster picked up in the previous round of mean shift clustering; G(•) is the Gaussian kernel function, and the value of Gaussian kernel function G is different for energy points at different distances h. The farther the distance, the smaller the G value; ω(x i ) is the energy group with position x i The energy value of the data point, the larger the energy value, the better the * The greater the contribution; i is the serial number of the data point in the energy group, i = 1, 2, ..., n, n is the total number of data points in the energy group, h i is the position x i The distance to position x.

[0051] By weighting the distance and the energy, the accuracy and time of the mean shift clustering search for the maximum energy value position of the velocity spectrum energy cluster can be improved.

[0052] The iterative termination condition of mean shift clustering can be that the number of iterations reaches a preset threshold; or the difference between the current maximum energy position and the maximum energy position obtained in the previous round meets the set error threshold; optionally, other termination conditions can also be used.

[0053] Step S14: Perform time interpolation on the velocity function, and obtain the time domain velocity model of the study area from the interpolated velocity function of each seismic gather.

[0054] The semi-supervised time-domain velocity modeling method based on expert experience constraint provided in the first embodiment of the present invention first introduces the geological background experience of the expert on the velocity field, obtains the initial velocity function constructed according to the expert experience of the velocity spectrum of the extracted seismic gather, obtains the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation, and obtains the candidate area by extrapolation based on the initial velocity function of the velocity spectrum according to the set rules; for the velocity spectrum of each seismic gather, the energy group is identified with the candidate area as a constraint, and the energy maximum position of the energy group is tracked to obtain the velocity function; the velocity function is interpolated in time, and the time-domain velocity model of the study area is obtained by the velocity function after interpolation of each seismic gather. The method realizes the intelligence and automation of time-domain velocity analysis, significantly improves the efficiency of time-domain velocity analysis, liberates manpower, avoids the burden brought by manual picking, and effectively realizes the quality and efficiency of velocity analysis; wherein, by manually picking a small number of velocity functions by experts and interpolating the initial velocity functions of all gathers, the noise resistance of the unsupervised clustering algorithm can be improved, the interference of noise and multiple wave energy groups can be avoided, and the evolution efficiency can be improved.

[0055] Embodiment 2

[0056] Embodiment 2 of the present invention provides a specific application of a semi-supervised time domain velocity modeling method based on expert experience constraints, and its process is as follows: Figure 2 As shown, the following steps are included:

[0057] Step S21: Input the gather velocity spectrum obtained after conventional processing.

[0058] Step S22: pre-process the speed spectrum, use median filtering to remove random salt and pepper noise and unify the speed scanning range.

[0059] like Figure 3 As shown, it is the velocity spectrum of a single-channel CMP data set of a three-dimensional onshore actual data in the xx basin obtained after denoising and other processing methods, and the velocity range is unified to 1000m / s~3500m / s.

[0060] Step 23: Using the expert's geological background knowledge and experience on the velocity field, the candidate regions of the Chan-vese segmentation model are constructed.

[0061] A small number of seismic gathers are extracted for velocity spectra, and the initial velocity function is obtained based on the expert automatic velocity picking results. The initial velocity function of the velocity spectrum of other seismic gathers in the study area is obtained by interpolation. Based on the initial velocity function of the velocity spectrum, the candidate area is obtained by extrapolation according to the set rules.

[0062] like Figure 4 As shown, for Figure 3The velocity spectrum is used to construct the segmentation candidate area of ​​the CV segmentation model based on expert experience. The area surrounded by white lines is the candidate area, and the external area does not participate in the identification of energy clusters, which can improve the noise resistance of the unsupervised clustering algorithm and avoid the interference of multiple waves and other noise energy clusters.

[0063] Step 24: Use Chan-vese model to segment the velocity spectrum and identify energy clusters.

[0064] Figure 5 For Figure 4 The candidate regions are iteratively evolved using the CV segmentation model to obtain the effective energy cluster distribution of the velocity spectrum.

[0065] Step 25: Perform unsupervised mean shift clustering tracking on the effective energy cluster and pick up the energy peak position of the energy cluster.

[0066] Figure 6 For Figure 5 The energy peak of the effective energy cluster is tracked by mean shift unsupervised clustering to obtain the velocity function of the velocity spectrum.

[0067] Step 26: Interpolate the velocity function obtained from the analysis of all gather data to each moment and output the velocity field of the entire study area.

[0068] Figure 7 It is the time domain stacked velocity body obtained by interpolating and combining each time sampling point after obtaining the velocity function of all CMP gathers of the three-dimensional land data.

[0069] Figure 8 is to use Figure 7 The imaging result slices of the three-dimensional stacked velocity body after CMP gather dynamic correction and stacking show that the velocity modeling accuracy of this method can reach the accuracy of manual picking, and there are improvements in the local structure imaging details. The most critical thing is that this technology significantly improves the efficiency of time domain velocity analysis, and the efficiency of single-channel velocity spectrum analysis is increased by 30 times, which liberates manpower and avoids the burden of manual picking of massive data, effectively realizing the improvement of quality and efficiency of the velocity analysis process.

[0070] Based on the inventive concept of the present invention, the embodiment of the present invention also provides a semi-supervised time domain speed modeling device based on expert experience constraints, the structure of the device is as follows: Fig. 9 As shown, including:

[0071] Initialization module 91, used to obtain an initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather, and obtain the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation;

[0072] A candidate region determination module 92 is used to obtain a candidate region by extrapolation based on an initial speed function of the speed spectrum according to a set rule;

[0073] An energy cluster tracking module 93 is used to identify energy clusters in candidate regions based on the velocity spectrum of each seismic trace gather;

[0074] A velocity function building module 94 is used to track the maximum energy position of the energy group and obtain a velocity function;

[0075] The time domain velocity model building module 95 is used to perform time interpolation on the velocity function, and obtain the time domain velocity model of the study area from the interpolated velocity function of each seismic gather.

[0076] In some embodiments, the energy cluster tracking module 93, the energy cluster identified within the candidate region, is used to:

[0077] The Chan-vese model segmentation method is used to identify energy clusters from candidate regions.

[0078] In some embodiments, the velocity function building module 94, the tracking energy group's energy maximum position, is used to:

[0079] An unsupervised mean-shift clustering method is used to track the maximum energy position of each energy cluster.

[0080] In some embodiments, the velocity function building module 94, the tracking of the maximum energy position of each energy group, is used to:

[0081] Gaussian kernel function and velocity spectrum energy value are introduced as constraints to track the maximum energy position of each energy cluster.

[0082] In some embodiments, the speed function building module 94, the tracking of the maximum energy position of each energy group, includes:

[0083] The energy maximum position of each energy cluster is tracked by the following formula:

[0084]

[0085] Among them, x * is the maximum energy position of the energy cluster picked up by mean shift clustering in the current round, x is the maximum energy position of the energy cluster picked up in the previous round of mean shift clustering, G(·) is the Gaussian kernel function, ω(x i ) is the energy group with position x i The energy value of the data point, i is the serial number of the data point in the energy group, i = 1, 2, ..., n, n is the total number of data points in the energy group, h i is the position x i The distance to position x.

[0086] In some embodiments, the above device further includes a preprocessing module 96, which is used to perform the following preprocessing on the velocity spectrum of the seismic gather before the initialization module 91 obtains the initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather:

[0087] Median filtering is used to remove random salt and pepper noise and unify the speed scanning range of the speed spectrum.

[0088] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0089] Based on the inventive concept of the present invention, an embodiment of the present invention further provides a computer storage medium, in which computer executable instructions are stored. When the computer executable instructions are executed by a processor, the above-mentioned semi-supervised time domain speed modeling method based on expert experience constraints is implemented.

[0090] Based on the inventive concept of the present invention, an embodiment of the present invention also provides a server, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the semi-supervised time domain speed modeling method based on expert experience constraints is implemented.

[0091] Unless otherwise specifically stated, terms such as processing, computing, calculating, determining, displaying, etc. may refer to the actions and / or processes of one or more processing or computing systems, or similar devices, which operate and convert data represented as physical (e.g., electronic) quantities within registers or memories of a processing system into other data similarly represented as physical quantities within memories, registers, or other such information storage, transmission, or display devices of the processing system. Information and signals may be represented using any of a variety of different techniques and methods. For example, data, instructions, commands, information, signals, bits, symbols, and chips mentioned throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, light fields or particles, or any combination thereof.

[0092] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process can be rearranged without departing from the scope of protection of the present disclosure. The attached method claims present the elements of the various steps in an exemplary order and are not intended to be limited to the specific order or hierarchy described.

[0093] In the above detailed description, various features are grouped together in a single embodiment to simplify the disclosure. This method of disclosure should not be interpreted as reflecting an intention that an embodiment of the claimed subject matter requires more features than those set forth in each claim. On the contrary, as reflected in the appended claims, the invention is in a state of having less than all the features of the disclosed individual embodiments. Therefore, the appended claims are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate preferred embodiment of the invention.

[0094] Those skilled in the art will also appreciate that the various illustrative logic blocks, modules, circuits, and algorithmic steps described in conjunction with the embodiments herein can all be implemented as electronic hardware, computer software, or a combination thereof. In order to clearly illustrate the interchangeability between hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above around their functions. Whether such functions are implemented as hardware or software depends on specific applications and the design constraints imposed on the entire system. A skilled person can implement the described functions in an alternative manner for each specific application, but such implementation decisions should not be interpreted as departing from the scope of protection of the present disclosure.

[0095] The steps of the method or algorithm described in conjunction with the embodiments herein may be directly embodied as hardware, a software module executed by a processor, or a combination thereof. The software module may be located in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a mobile disk, a CD-ROM, or any other form of storage medium well known in the art. An exemplary storage medium is connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Of course, the storage medium may also be an integral part of the processor. The processor and the storage medium may be located in an ASIC. The ASIC may be located in a user terminal. Of course, the processor and the storage medium may also be present in a user terminal as discrete components.

[0096] For software implementation, the techniques described in this application can be implemented with modules (e.g., procedures, functions, etc.) that perform the functions described in this application. These software codes can be stored in a memory unit and executed by a processor. The memory unit can be implemented within the processor or outside the processor. In the latter case, it is coupled to the processor in a communication manner via various means, which are well known in the art.

[0097] The above description includes examples of one or more embodiments. Of course, it is impossible to describe all possible combinations of components or methods for the purpose of describing the above embodiments, but it should be recognized by those skilled in the art that the various embodiments may be further combined and arranged. Therefore, the embodiments described herein are intended to cover all such changes, modifications and variations that fall within the scope of protection of the appended claims. In addition, with respect to the term "comprising" used in the specification or claims, the word is covered in a manner similar to the term "including", just as "including," is explained as a transitional word in the claims. In addition, any term "or" used in the specification of the claims is intended to mean "non-exclusive or".

Claims

1. A semi-supervised time-domain velocity modeling method based on expert experience constraints, characterized in that: include: The initial velocity function is obtained based on the velocity picking result of the velocity spectrum of the extracted seismic gathers, and the initial velocity function of the velocity spectrum of other seismic gathers in the study area is obtained by interpolation; Based on the initial velocity function of the velocity profile, the candidate area is obtained by extrapolation according to the set rules; For the velocity spectrum of each seismic trace gather, energy clusters are identified from the candidate area, the maximum energy position of the energy cluster is tracked, and the velocity function is obtained; The velocity function is interpolated in time, and the time domain velocity model of the study area is obtained from the interpolated velocity function of each seismic gather.

2. The method according to claim 1, characterized in that The identifying of energy clusters from within the candidate regions comprises: The Chan-vese model segmentation method is used to identify energy clusters from candidate regions.

3. The method according to claim 1, characterized in that The tracking of the maximum energy position of the energy group includes: An unsupervised mean-shift clustering method is used to track the maximum energy position of each energy cluster.

4. The method according to claim 3, characterized in that The tracking of the maximum energy position of each energy group includes: Gaussian kernel function and velocity spectrum energy value are introduced as constraints to track the maximum energy position of each energy cluster.

5. The method according to claim 4, characterized in that The tracking of the maximum energy position of each energy group includes: The energy maximum position of each energy cluster is tracked by the following formula: Among them, x * is the maximum energy position of the energy cluster picked up by mean shift clustering in the current round, x is the maximum energy position of the energy cluster picked up in the previous round of mean shift clustering, G(•) is the Gaussian kernel function, ω(x i ) is the energy group with position x i The energy value of the data point, i is the serial number of the data point in the energy group, i = 1, 2, ..., n, n is the total number of data points in the energy group, h i is the position x i The distance to position x.

6. The method according to any one of claims 1 to 5, characterized in that: Before obtaining the initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather, the method further includes performing the following preprocessing on the velocity spectrum of the seismic gather: Median filtering is used to remove random salt and pepper noise and unify the speed scanning range of the speed spectrum.

7. The method according to any one of claims 1 to 5, characterized in that: The velocity spectrum is a time domain superposition velocity spectrum or a root mean square velocity spectrum; The seismic gathers are common center point gathers or common reflection point gathers.

8. A semi-supervised time domain velocity modeling device based on expert experience constraints, characterized in that: The device comprises: Initialization module, used to obtain the initial velocity function according to the velocity picking result of the velocity spectrum of the extracted seismic gather, and to obtain the initial velocity function of the velocity spectrum of other seismic gathers in the study area by interpolation; A candidate region determination module is used to obtain a candidate region by extrapolating the initial speed function of the speed spectrum according to a set rule; Energy cluster tracking module, used to identify energy clusters in candidate regions based on the velocity spectrum of each seismic trace gather; The velocity function building module is used to track the maximum energy position of the energy group and obtain the velocity function; The time domain velocity model building module is used to perform time interpolation on the velocity function, and obtain the time domain velocity model of the study area from the velocity function after interpolation of each seismic gather.

9. A computer storage medium, characterized in that: The computer storage medium stores computer executable instructions, and when the computer executable instructions are executed by a processor, the semi-supervised time-domain speed modeling method based on expert experience constraints described in any one of claims 1 to 7 is implemented.

10. A server, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the semi-supervised time-domain speed modeling method based on expert experience constraints as described in any one of claims 1 to 7 is implemented.

Images