CN111239828A - Multiple suppression method based on optimal hyperbolic integral path superposition - Google Patents

Abstract

The invention belongs to the technical field of noise suppression in geophysical exploration, and relates to a multiple suppression method based on optimal hyperbolic integral path superposition. The invention optimizes and improves the conventional three-dimensional surface related multiple pressing method, utilizes the inner product of the cross survey line multiple contribution gather and the result obtained by parabolic Raon transformation of the cross survey line multiple contribution gather, determines the vertex coordinate of a hyperbola by picking up the local maximum value of the inner product result, performs statistics on the vertex coordinate, selects a few vertex coordinates with higher frequency as path superposition vertices, uses the vertex horizontal position coordinate as the offset selection result of the cross survey line multiple contribution gather to obtain a plurality of predicted multiple subsets, determines different intervals according to the time coordinate of the vertex position, selects and screens the multiple subsets according to the principle of maximum interval amplitude to obtain the final multiple prediction result, and finally realizes multiple pressing by matching and subtracting.

Description

Multiple Suppression Method Based on Optimal Hyperbolic Integral Path Superposition

technical field

The invention belongs to the technical field of noise suppression in geophysical exploration, in particular to a multiple-wave suppression method in marine seismic exploration, in particular to a multiple-wave suppression method based on the superposition of optimal hyperbolic integral paths.

Background technique

In marine seismic exploration, multiples are often regarded as a kind of coherent noise that needs to be removed because of their periodicity and kinematic morphological characteristics, which will reduce the signal-to-noise ratio of seismic data. At present, the more commonly used multiple suppression technology is 2D SRME technology. However, when the stratum is tilted, the real propagation path is a three-dimensional space function, and the 2D method cannot describe the change of the direction of the tie line. Therefore, the multiple wave prediction problem is transformed into 3D, and 3D SRME needs to be introduced. However, 3D prediction is limited by the sparse sampling of the observation system, and direct superposition will introduce prediction artifacts and cause waveform distortion, resulting in severe spatial aliasing in multiple-wave prediction.

The 2D SRME method was proposed by Berkhout and Verschuur et al. (Verschuur D J, Berkhout AJ, Wapenaar C P A. Adaptive surface-related multiple elimination [J]. GEOPHYSICS, 1992, 57(9): 1166-1177.) in 1992, The method removes all surface-related multiples by removing the effect of surface reflectivity in the data, designing multiple removal as an inversion process in which the source and surface reflectivity properties can be estimated, The inversion residual is the predicted multiple, which is data-driven and only suitable for simple terrain conditions. With the increasing difficulty of seismic exploration, the 3D SRME method was developed by E.J.van Dedem (Dedem E J V. 3D surface-related multiple elimination and interpolation[J].SEG Technical ProgramExpanded Abstracts,1999,17(1):1321.) in 1999 In 2008, it was proposed that on the basis of 2D SRME, the direction of the contact survey line was introduced to adapt to more complex terrain. However, the 3D prediction process is limited by the sparse sampling of the tie-line direction of the observation system, so a series of solutions have been proposed successively. The method of densely sampled data required for data reconstruction to obtain 3D SRME was introduced by Anatoly Baumstein et al. proposed in 2002; after that, the multiple suppression method based on sparse inversion was proposed in 2002 by Van Dedem and Verschuur (DedemE J V, Verschuur D J. 3D Surface-related Multiple Prediction Using SparseInversion: Experience With Field Data [J]. Seg Technical Program ExpandedAbstracts, 2002, 21(1):2094.) proposed that the sparse contact line multiple contribution gathers are reconstructed in the model space by sparse inversion, and then the predicted multiples are obtained by the correction and superposition of the amplitude and phase; Using curvelet transform to separate primary and multiple waves was proposed by Xiang Wu (Wu X, Hung B.Adaptive primary-multiple separation using 3D curvelet transform[J].ASEG Extended Abstracts,2015,2015(1):1155.) in 2015 ; 3D surface multiple suppression method based on data regularization and sparse inversion by Fang Yunfeng et al. (Fang Yunfeng, Nie Hongmei, Zhang Limei, et al. 3D surface multiple suppression method based on data regularization and sparse inversion[J] ] .Acta Geophysics, 2016, v.59(02):269-277.) proposed in 2016 that through data regularization and de-regularization techniques, the data distribution can meet the requirements of 3D SRME, and then sparse inversion is used instead of liaison Summation of line multiple contribution gathers. However, the sparse inversion methods and data reconstruction or curvelet transformation adopted by the above methods are less stable, and have a greater impact on different complex terrains.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a multiple suppression method based on the superposition of optimal hyperbolic integral paths, so as to overcome the problem of spatial aliasing generated by the conventional three-dimensional surface correlation multiple suppression method due to the sparse direction of the seismic data acquisition contact line. shortcoming.

The purpose of this invention is to realize through the following technical solutions:

First, the common shot gathers and the common receiver gathers are separated from the seismically collected data, and the multiple contribution gathers of the longitudinal survey line at different positions are obtained by convolving the common shot gathers and the common receiver gathers. The multiple contribution gathers of the survey line are summed along the direction of the longitudinal survey line to obtain the multiple contribution gathers of the connecting survey line. The inner product of the multiple contribution gathers of the connecting survey line and the result obtained by the parabolic Raon transform is used to pick the inner product. The local maximum value of the product result determines the hyperbolic vertex coordinates, and makes statistics on them. Select the few vertex coordinates with high frequency as the path superposition vertex, and use the vertex horizontal position coordinate as the offset of the multiple wave contribution gathers of the connecting survey line. From the selection results, several subsets of predicted multiples are obtained, different intervals are determined according to the time coordinates of the vertex positions, and the multiple subsets are selected and screened according to the principle of the maximum amplitude of the interval, and the final multiple prediction results are obtained. To achieve multiple suppression.

A multiple wave suppression method based on the superposition of optimal hyperbolic integral paths, comprising the following steps:

a. Complement the data collected by 3D marine seismic exploration through the reciprocity theorem to obtain the same seismic data of the shot track, and extract the gathers according to the shot number (fldr) in the track header information to obtain the common shot gather d (x _k , y _k , t, x _s , y _s ); extract the gathers according to the detection number (tracl) to obtain a common detection point gather r(x _r , y _r , t, x _k , y _k ), where (x _k , y _k ) is the position of any point (shot or detection point), (x _s , y _s ) is the coordinates of the shot point, (x _r , y _r ) is the coordinates of the detection point, and t is the time.

b. Perform Fourier positive on the common shot gather d(x _k ,y _k ,t,x _s ,y _s ) and the common receiver gather r(x _r ,y _r ,t,x _k ,y _k ) Transform to obtain the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain, where w is the angular frequency.

c. Multiply the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain to get Contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) of longitudinal line multiples corresponding to different positions (x _k , y _k ).

M _xy (x _r ,y _r ,x _s ,y _s ,x _k ,y _k ,w)=r ₀ R(x _r ,y _r ,w; x _k ,y _k )D(x _k ,y _k , w; x _s , y _s )

Among them, r ₀ represents the formation reflection coefficient, which is generally taken as -1.

d. The multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) of the longitudinal survey line at different positions are summed along the longitudinal survey line direction to obtain the multi-connected survey line The secondary wave contribution gathers M _y (x _r , y _r , x _s , y _s , y _k , w) are transformed back to the time domain by inverse Fourier transform to obtain the time domain multiple wave contribution traces of the tie line Set m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t).

e. Perform parabolic Radon transformation on the multiple contribution gathers of the tie-line survey line, and inner-product the result of the parabolic Radon transformation with the multiple wave contribution gathers of the tie-line survey line without parabolic Radon transformation.

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon(m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t))

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the result of Radon transformation, m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the inner product the result of.

f. Use findpeaks under the matlab platform to pick the local maximum value through the threshold limit and contrast amplitude value, and obtain the hyperbolic vertex coordinates.

[x _m ,t _m ]=findpeaks(m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t))

[x _m , t _m ] is a matrix composed of the lateral coordinates of the event axis vertices.

g. Count the coordinates of all event axis vertices, and select the ones with the highest frequency (usually 3-5, depending on the actual situation) as the vertices of the path superposition.

为路径叠加的顶点，i＝1，2，3...。

Vertices for path stacking, i=1, 2, 3....

h. Select the x _max(i) of the multiple vertices superimposed on the path as the offset distance (offset) of the Radon transformation, perform the Radon transformation, and superimpose the results of the Radon transformation to obtain several predicted multiples m _(i) ( x _r , y _r , x _s , y _s , t), compare the size of t _max(i) to find different intervals in the time direction

m _y(i) (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon _(xmax(i)) (m _y (x _r ,y _r ,x _s ,y _s ,y _k , t))

中，选取上述若干个预测的多次波中振幅较大的部分组成最终的多次波预测结果。i, in different intervals

, the part with larger amplitude among the above-mentioned several predicted multiples is selected to form the final multiple prediction result.

j. Use time domain matching and subtraction to obtain seismic data without multiples.

Further, in step g, the number of vertices is generally 3-5, depending on the actual situation.

Compared with the prior art, the beneficial effects of the present invention:

The invention realizes the multiple wave suppression technology by selecting the optimal hyperbolic integral path superposition. Using the inner product of the multiple wave contribution gathers of the tie line and the result obtained by the parabolic Raon transform, the coordinates of the hyperbola vertexes are determined by picking the local maximum value of the inner product result, and a few vertices with higher occurrence frequency are selected according to statistics. The coordinates are used as the vertices of the path stacking, and the horizontal position coordinates of the vertices are used as the offset selection results of the multiple contribution gathers of the contact line, and several subsets of predicted multiples are obtained, and different intervals are determined according to the time coordinates of the vertex positions. According to the principle of maximum amplitude, multiple subsets are selected and screened, and the final multiple prediction result is obtained. Finally, multiple suppression is realized by matching and subtraction, which significantly improves the precision and accuracy of multiple suppression.

The present invention has the following advantages:

1. The multiple-wave suppression technology based on the superposition of the optimal hyperbolic integral paths is a data-driven method. It does not require the assumption and prior information of the underground medium, but only needs the information of the source and detection point, which can effectively predict and suppress multiple waves. purpose of the second wave.

2. This method effectively solves the problem that 3D SRME requires intensive data collection, and can obtain high-precision and high-stability multiple wave suppression effects for different geological models.

3. Compared with the multiple suppression method based on sparse inversion, this method has higher computational efficiency and reduces the cost of seismic data processing and interpretation.

Description of drawings

Figure 1. Flow chart of multiple suppression method based on optimal hyperbolic integral path superposition.

Fig. 2 Schematic diagram of the 3D marine seismic exploration data acquisition and observation system.

Figure 3 Schematic diagram of three-dimensional effect path propagation.

Fig. 4 Schematic diagram of the propagation path of surface multiples.

Fig. 5 Schematic diagram of the multiple contribution gathers of the longitudinal survey line.

Fig. 6 Schematic diagram of the multiple contribution gather of the tie line.

Fig. 7 Schematic diagram of the Radon transform result of the multiple contribution gather of the tie line.

Figure 8. Schematic diagram of the inner product of the multiple contribution gathers of the tie line and their Radon transform results

Figure 9 3D simulation data; Figure 9a simulates the geological model, Figure 9b Offset=-500 as a schematic diagram of multiples predicted by the path stacking vertex, Figure 9c Offset=-300 as a schematic diagram of multiples predicted by the path stack vertex, Figure 9dOffset=0 As a schematic diagram of multiples predicted by the path stacking vertex, Figure 9e is a schematic diagram of the multiple wave prediction result, Figure 9f is a schematic diagram of the original data, and Figure 9g is a schematic diagram of the multiple wave suppression result.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings and examples.

The present invention optimizes and improves the conventional three-dimensional surface-related multipleselimination (3D SRME) method, and utilizes the inner product of the multiple contribution gathers of the contact survey line and the result obtained by parabolic Raon transformation, and picks up The local maximum value of the inner product results determines the coordinates of the hyperbolic vertexes, and performs statistics on them. Select the result of the shift distance to obtain several subsets of predicted multiples, determine different intervals according to the time coordinates of the vertex positions, select and filter the multiple subsets based on the principle of the maximum amplitude of the interval, and obtain the final multiple prediction results, and finally pass the matching Subtract to achieve multiple suppression. The multiple suppression method based on the optimal hyperbolic integral path superposition is realized by Seismic Unix and MATLAB platform.

The present invention is based on the multiple wave suppression method superimposed on the optimal hyperbolic integral path, comprising the following steps:

a. Complement the data collected by the 3D marine seismic exploration and observation system (Fig. 2) through the reciprocity theorem to obtain the same seismic data of the gun track, and extract the gathers according to the gun number (fldr) in the track head information to obtain Common shot gather d(x _k , y _k , t, x _s , y _s ); extract gathers according to the detection signal (tracl) to obtain the common shot gather r(x _r , y _r , t, x _k , y _k ), where (x _k , y _k ) is the position of any point (shot point or receiver point), (x _s , y _s ) is the shot point coordinate, (x _r , y _r ) is the receiver point coordinate, t for time.

b. Perform Fourier positive on the common shot gather d(x _k ,y _k ,t,x _s ,y _s ) and the common receiver gather r(x _r ,y _r ,t,x _k ,y _k ) Transform to obtain the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain, where w is the angular frequency.

c. Considering the influence of three-dimensional effects (Fig. 3), according to the wave field motion characteristics of the multiple wave propagation path (Fig. 4), the common shot gather D(x _k , y _k , w, x _s , y _s ) and the common shot gather D(x k , y k , w, x s , y s ) The detection point gathers R(x _r , y _r , w, x _k , y _k ) are multiplied in the frequency domain to obtain the longitudinal survey line multiples contribution gathers M _xy (x _r ) corresponding to different positions (x _k , y _k ) , y _r , x _s , y _s , x _k , y _k , w) (Fig. 5).

M _xy (x _r ,y _r ,x _s ,y _s ,x _k ,y _k ,w)=r ₀ R(x _r ,y _r ,w; x _k ,y _k )D(x _k ,y _k , w; x _s , y _s )

Among them, r ₀ represents the formation reflection coefficient, which is generally taken as -1.

d. Sum the multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) along the longitudinal line direction to obtain the multiple contribution of the connecting line gather My (x _r , _{y r} , x _s , y _s , y _k , w), and transform it back to the time domain through inverse Fourier transform to obtain the time domain tie-line multiple contribution gather my _y (x _r , y _r , x _s , y _s , y _k , t) (Fig. 6).

e. Perform parabolic Radon transform on the multiple contribution gathers of the tie-line survey line, and inner-product the result of the parabolic Radon transform (Fig. 7) with the multiple contribution gathers of the tie-line survey line without parabolic Radon transform (Fig. 8).

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon(m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t))

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the result of Radon transformation, m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the inner product the result of.

f. Use findpeaks under the matlab platform to pick the local maximum value through the threshold limit and contrast amplitude value, and obtain the hyperbolic vertex coordinates.

[x _m ,t _m ]=findpeaks(m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t))

[x _m , t _m ] is a matrix composed of the lateral coordinates of the event axis vertices.

g. Count the coordinates of all the vertices of the event axis, and select the ones with the highest frequency (usually 3-5, depending on the actual situation) as the vertices of the path superposition.

为路径叠加的顶点，i＝1，2，3...。

Vertices for path stacking, i=1, 2, 3....

h. Select the x _max(i) of several vertices superimposed on the path as the offset distance (offset) of the Radon transformation, perform the Radon transformation, and superimpose the results of the Radon transformation to obtain several predicted multiples m _(i) ( x _r , y _r , x _s , y _s , t), compare the size of t _max(i) to find different intervals in the time direction

m _y(i) (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon _(xmax(i)) (m _y (x _r ,y _r ,x _s ,y _s ,y _k , t))

中，选取上述若干个预测的多次波中振幅较大的部分组成最终的多次波预测结果。i, in different intervals

, the part with larger amplitude among the above-mentioned several predicted multiples is selected to form the final multiple prediction result.

j. Match and subtract the predicted multiples from the original data in the time domain to obtain seismic data without multiples.

Example 1

a. The simulated data model is a three-dimensional inclined stratigraphic model (Fig. 9a). The observation system is set up as 11 2000-meter-long survey lines that are distributed in near-parallel distances of 100 meters (-500m-500m), and evenly distributed detectors (0m-2000m) with an interval of 25 meters on each survey line. Seismic waves are excited at an interval of 25 meters at the horizontal position of the central cable, and the data is complemented by the reciprocity theorem to obtain the same seismic survey data on the central survey line. Extract the gather according to the shot number (fldr, 1-81) in the track header information, and obtain the common shot gather d(x _k , y _k , t, x _s , y _s ); according to the detection number (tracl, (1-81, 1-11)) extract the gathers to obtain the common detection point gathers r(x _r , y _r , t, x _k , y _k ), where (x _k , y _k ) is the position of any point ( shot point or detection point), (x _s , y _s ) are the coordinates of the shot point, (x _r , y _r ) are the coordinates of the detection point, and t is the time.

b. Perform Fourier positive on the common shot gather d(x _k ,y _k ,t,x _s ,y _s ) and the common receiver gather r(x _r ,y _r ,t,x _k ,y _k ) Transform to obtain the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain, where w is the angular frequency.

c. Multiply the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain to get Longitudinal line multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) corresponding to different positions (x _k , y _k ) (81*11 in total) .

M _xy (x _r ,y _r ,x _s ,y _s ,x _k ,y _k ,w)=r ₀ R(x _r ,y _r ,w; x _k ,y _k )D(x _k ,y _k , w; x _s , y _s )

Among them, r ₀ represents the formation reflection coefficient, which is generally taken as -1.

d. Sum the multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) along the longitudinal line direction to obtain the multiple contribution of the connecting line The gathers M _y (x _r , y _r , x _s , y _s , y _k , w) (81 in total), and transform them back to the time domain through inverse Fourier transform to obtain the time-domain contact line multiples Contribution gathers m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t).

e. Perform parabolic Radon transformation on the multiple wave contribution gathers of the tie line. In the parameter setting, the reference maximum offset distance is 2000m, the minimum time difference is 200, and the maximum time difference is 1000. The inner product of the parabolic Radon transform results and the multiple contribution gathers of the tie-line without parabolic Radon transform.

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon(m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t))

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the result of Radon transformation, m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t) is

The result of the inner product.

f. Use findpeaks under the matlab platform to pick the local maximum value through the threshold limit contrast amplitude value, and obtain the hyperbolic vertex coordinates (Table 1), where the amplitude threshold is set to 0.01.

[x _m ,t _m ]=findpeaks(m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t))

[x _m , t _m ] is a matrix composed of the lateral coordinates of the event axis vertices.

g. Count the coordinates of all the vertices of the event axis, and select the three with the highest frequency, which are x=-500, -300, and 0, respectively, as the vertices of the path superposition.

为路径叠加的顶点，i＝1，2，3...。

Vertices for path stacking, i=1, 2, 3....

h. Select x ₁ =-500, x ₂ =-300, and x ₃ =0 of the three vertices of the superimposed path respectively as the offset of Radon transformation, perform Radon transformation, and superimpose the results of Radon transformation to obtain 3 Predicted multiples (Fig. 9b, c, d), compare the size of t _max(i) to find different intervals in the time direction

m _y(i) (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon _(xmax(i)) (m _y (x _r ,y _r ,x _s ,y _s ,y _k , t))

中，选取上述3个预测的多次波中振幅较大的部分组成最终的多次波预测结果，可发现最终预测结果(图9e)相比于设置不同偏移距预测结果均好，预测到达时间准确，具有高精度高分辨率的优点。i, in different intervals

In the above three predicted multiples, the larger amplitude part is selected to form the final multiple prediction result. It can be found that the final prediction result (Fig. 9e) is better than the prediction results with different offsets, and the prediction reaches The time is accurate, and it has the advantages of high precision and high resolution.

j. Subtract the time domain matching from the original data (Fig. 9f) to obtain seismic data without multiples (Fig. 9g), and the multiples are effectively suppressed.

Table 1. Local maximum value to obtain hyperbolic vertex coordinate table

serial number x(m) t(s) 1 -300 3.05 2 0 3.70 3 -500 3.80 4 0 3.81 5 -300 4.19 6 0 4.37 7 -300 4.83 8 -500 5.30 9 -100 5.33 10 -300 5.63 11 -300 5.77

Claims (2)

1. A multiple wave suppression method based on optimal hyperbolic integral path superposition, which is characterized in that it does not require assumptions and prior information of underground media, which solves the problem that 3D SRME requires intensive data collection, and can be used for different geological models. The multiple suppression effect with high precision and stability is obtained, and compared with the multiple suppression method based on sparse inversion, it has higher computational efficiency and reduces the cost of seismic data processing and interpretation, including the following steps: a. Complement the data collected by 3D marine seismic exploration through the reciprocity theorem to obtain the same seismic data of the shot track, and extract the gathers according to the shot numbers in the track header information to obtain the common shot gather d(x _k , y _k , t, x _s , y _s ); extract the gathers according to the detection number to obtain a common detection point gather r(x _r , y _r , t, x _k , y _k ), where (x _k , y _k ) is the position of any point (shot point or detection point), (x _s , y _s ) is the coordinates of the shot point, (x _r , y _r ) is the coordinates of the detection point, and t is the time; b. Perform Fourier positive on the common shot gather d(x _k ,y _k ,t,x _s ,y _s ) and the common receiver gather r(x _r ,y _r ,t,x _k ,y _k ) Transform to obtain the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain, where, w is the angular frequency; c. Multiply the common shot gather D(x _k ,y _k ,w,x _s ,y _s ) and the common receiver gather R(x _r ,y _r ,w,x _k ,y _k ) in the frequency domain to get Longitudinal line multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) corresponding to different positions (x _k , y _k ); M _xy (x _r ,y _r ,x _s ,y _s ,x _k ,y _k ,w)=r ₀ R(x _r ,y _r ,w; x _k ,y _k )D(x _k ,y _k , w; x _s , y _s ) Among them, r ₀ represents the formation reflection coefficient, which is -1; d. The multiple contribution gathers M _xy (x _r , y _r , x _s , y _s , x _k , y _k , w) of the longitudinal survey line at different positions are summed along the longitudinal survey line direction to obtain the multi-connected survey line The secondary wave contribution gathers M _y (x _r , y _r , x _s , y _s , y _k , w) are transformed back to the time domain by inverse Fourier transform to obtain the time domain multiple wave contribution traces of the tie line set m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t),

e. Perform parabolic Radon transformation on the multiple contribution gathers of the tie-line survey line, and take the inner product of the result of the parabolic Radon transformation and the multiple wave contribution gathers of the tie-line survey line without parabolic Radon transformation, m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t)=Radon(m _y (x _r ,y _r ,x _s ,y _s ,y _k ,t))

m _yradon (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the result of Radon transformation, m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t) is the inner product the result of; f. Pick the local maxima through the threshold limit and contrast the amplitude value, and obtain the hyperbolic vertex coordinates; [x _m ,t _m ]=findpeaks(m _d (x _r ,y _r ,x _s ,y _s ,y _k ,t)) [x _m ,t _m ] is a matrix composed of the lateral coordinates of the event axis vertices; g. Count the coordinates of all the vertices of the event axis, and select the vertices with the highest frequency as the vertices of the path superposition,

For the vertices of the path stack, i=1, 2, 3...; h. Select the x _max(i) of the multiple vertices superimposed on the path as the offset of the Radon transformation, perform the Radon transformation, and superimpose the results of the Radon transformation to obtain several predicted multiples m _(i) (x _r , y _r , x _s , y _s , t), compare the size of t _max(i) to find different intervals in the time direction

i, in different intervals

, select the part with larger amplitude among the above-mentioned several predicted multiples to form the final multiple prediction result; j. Use time domain matching and subtraction to obtain seismic data without multiples. 2 . The multiple-wave suppression method based on the superposition of optimal hyperbolic integral paths according to claim 1 , wherein in step g, the number of vertices is 3-5. 3 .

Images