A method for processing pre-stack consistency based on shearlet domain continuous data

Abstract

The application provides a kind of based on Shearlet domain continuous data prestack consistency processing method, to the data of each block in continuous exploration area respectively carry out surface consistency amplitude compensation and surface consistency deconvolution and other consistency pretreatment, select a processing target model, the application selects a certain amount of good quality data as target model in the block of relatively good quality data in continuous exploration area, the other data in continuous exploration area is corrected as the standard of the target data, estimate the consistency correction factor of Shearlet domain, eliminate the statistical difference between target model data and data to be processed, realize the consistency processing of continuous exploration area. Process prestack data, can eliminate the difference of continuous data in amplitude, frequency and space, has good stability under noise environment.

Description

Technical Field

[0001] This invention relates to the field of seismic data processing for oil and gas exploration, and in particular to a method for pre-stack consistency processing of contiguous Shearlet domain data. Background Technology

[0002] Currently, the main methods for solving the problem of consistency in splicing contiguous data from different blocks include deconvolution parameter adjustment, matched filtering, and a combination of multiple techniques.

[0003] The deconvolution parameter adjustment method is mainly used to solve the problem of different resolutions of different blocks of data. The deconvolution parameter experiment is relatively complicated. When processing large contiguous areas of a dozen or more blocks, it is time-consuming and laborious to select reasonable parameters.

[0004] Matched filtering methods are divided into two categories: time-varying and time-invariant. Time-invariant matched filtering methods include wavelet processing and direct matched filtering, where wavelet processing is also called wavelet shaping or wavelet matching.

[0005] Wavelet processing assumes that the reflection coefficients are the same at the same location in contiguous exploration areas, and believes that differences in seismic records are caused by differences in wavelet characteristics. Therefore, the processing quality is severely affected by the quality of the spliced ​​data. Direct matched filtering has poor amplitude preservation and its physical basis is unclear.

[0006] Time-invariant matched filtering uses the same matching operator from shallow to deep, which is impractical. Time-varying matched filtering methods include wavelet transform-based wavelet processing and wavelet transform-based matched filtering, but these require stitching data with overlapping areas, which has significant limitations in practical processing.

[0007] The method of combining multiple technologies is to solve the consistency problem of contiguous data by combining multiple consistency processing methods such as amplitude, frequency, and phase.

[0008] In summary, the consistency correction factor of matched filtering is calculated from the overlapping area of ​​two datasets. Its applicability presupposes a certain degree of overlap between the two datasets, and the processing quality heavily depends on the magnitude of the quality difference in the overlapping data. Furthermore, it is generally only suitable for post-stack data splicing. Deconvolution parameter adjustment methods and combinations of various techniques are used to process pre-stack data, but parameter adjustments are cumbersome, and most lack time-varying processing capabilities. The processing results of existing technologies exhibit poor stability. Summary of the Invention

[0009] In view of the above problems, the present invention is proposed to provide a pre-stack consistency processing method for contiguous data based on Shearlet domains that overcomes or at least partially solves the above problems.

[0010] According to one aspect of the present invention, a method for pre-stack consistency processing of contiguous data based on Shearlet domains is provided, the processing method comprising:

[0011] Seismic data from various blocks within a contiguous exploration area are collected to obtain contiguous data.

[0012] Transform the contiguous data into a Shearlet field;

[0013] Calculate the spatiotemporal mean and root mean error of the contiguous data in the Shearlet domain, and establish the target model;

[0014] Calculate the statistical spectrum of the model data and the contiguous data to be processed in the frequency domain, respectively.

[0015] Using the statistical spectrum of the model data as the expectation, a frequency consistency correction operator is calculated, and statistical frequency consistency correction processing is performed on the contiguous data to be processed to obtain the correction processing result y;

[0016] The correction result y is transformed to the Shearlet domain to obtain the Shearlet coefficients S. y (j,s,m);

[0017] For the Shearlet coefficient S y (j,s,m) opens a rectangular window w of size a×b. a,b Alpha-trim mean filtering is performed, and the rectangular window w is slid. a,b The mean of the spatiotemporal variation of the correction result y is obtained. and mean square deviation

[0018] For the mean and the mean square deviation Two-dimensional Gaussian filtering is performed separately to obtain the Gaussian filtered mean. and Gaussian filter mean square error

[0019] Based on the model data and the Gaussian filter mean and the mean square error of the Gaussian filter Calculate the spatiotemporal consistency correction factor;

[0020] Based on the spatiotemporal consistency correction factor, the Shearlet coefficient of the correction result y is corrected to obtain the Shearlet domain result after the contiguous data consistency processing.

[0021] Optionally, transforming the contiguous data to the Shearlet field specifically includes:

[0022] Construction of the Shearlet transform:

[0023] set up Where a = 2 -j ,s∈R,A a Parabolic scale matrix, S s Shearing matrix;

[0024] Let ψ∈L 2 (R 2 And meet the following conditions:

[0025] in For the Fourier transform of ψ;

[0026] ψ1 is a continuous wavelet transform.

[0027] Specifically, when the interval is (-1, 1), And ||ψ2||=1;

[0028] Define Shearlet wave atom ψ j,s,m as follows:

[0029]

[0030] Where M AS =S s A a Given the parameters j (scale), s (direction), and m (position), the Shearlet transform of the function f is defined as follows:

[0031] S f (j,s,m)= <f,ψ j,s,m > (Equation 2).

[0032] Optionally, the step of calculating the spatiotemporal mean and root mean square error of the contiguous data in the Shearlet domain and establishing the target model specifically includes:

[0033] Obtain the gun set data and use it as model data to obtain the i-th model gun data x. i Where i = 1, 2, ..., N represents the gun number of the model gun data, and the Shearlet coefficients S after Shearlet transformation x (i,j,s,m);

[0034] The Shearlet coefficient S is determined using alpha-trim mean filtering. x The spatiotemporal mean and root mean of (i,j,s,m);

[0035] For the Shearlet coefficient S of the i-th model shot datax (i,j,s,m), open the rectangular window w of size a×b. a,b Where a and b are the length and width of the rectangular window, respectively. An alpha-trim mean filter is performed within the rectangular window to obtain the mean value at the center of the rectangular window.

[0036] Calculate the mean square error of the window data based on the mean, and then slide the rectangular window w. a,b Finally, the mean of the spatiotemporal variation of the i-th model gun data is obtained. and mean square deviation

[0037] The spatiotemporal mean and root mean deviation of multiple model gun data are statistically averaged to obtain the statistically derived spatiotemporal mean and root mean deviation:

[0038]

[0039]

[0040] in, and The values ​​represent the spatiotemporal mean and spatiotemporal root mean of the Shearlet coefficients of the target model data at scale j and direction s after multi-shot statistical averaging.

[0041] Optionally, the step of calculating a frequency consistency correction operator based on the statistical spectrum of the model data, and performing statistical frequency consistency correction processing on the contiguous data to be processed to obtain the correction processing result y specifically includes:

[0042] Calculate the model data x in the frequency domain respectively. i The statistical spectrum of the contiguous data z to be processed is denoted as follows: and

[0043] and

[0044] Where X represents the spectrum of each channel of the model data, and N x Z represents the total number of channels in the model data, Z represents the spectrum of each channel in the contiguous data to be processed, and N represents the total number of channels in the model data. z This indicates the total number of channels in the contiguous data to be processed, and abs() indicates modulo processing;

[0045] Using the statistical spectrum of the model data as the expectation, the frequency f consistency correction operator F(f) is calculated.

[0046]

[0047] in, This indicates the calculation for all values ​​of the frequency f. The maximum value of ε is the stability factor, which is a small constant.

[0048] The contiguous data z to be processed undergoes statistical frequency consistency correction. The result after processing is denoted as y, and its calculation formula is as follows:

[0049] y = IFFT(Z(f) * F(f))

[0050] Here, IFFT() represents the inverse Fourier transform.

[0051] Optionally, the mean and the mean square deviation The specific steps of performing two-dimensional Gaussian filtering include:

[0052] The formula for the two-dimensional Gaussian function is as follows:

[0053]

[0054] Where λ is the mean square error parameter, m c m d These are the filter parameters;

[0055] Using Equation 5 to evaluate the mean and the mean square deviation The formula for performing two-dimensional Gaussian filtering is as follows:

[0056]

[0057]

[0058] Among them, Gaussian filter mean and Gaussian filter mean square error r represents the radius of the filter template.

[0059] Optionally, the step of using the model data and the Gaussian filter mean... and the mean square error of the Gaussian filter The calculation of the spatiotemporal consistency correction factor specifically includes:

[0060] The formula for the spatiotemporal consistency correction factor is:

[0061]

[0062]

[0063] in, and These represent the mean and root mean square error correction factor of the spatiotemporal variation, respectively.

[0064] Optionally, the step of correcting the Shearlet coefficient of the correction result y according to the spatiotemporal consistency correction factor specifically includes:

[0065] The correction formula is:

[0066]

[0067]

[0068] in, This represents the result of adjusting the time-varying mean of the shot gather data to be processed. This indicates the result after adjusting for the time-varying mean and standard deviation.

[0069] Optionally, the processing method further includes:

[0070] The Shearlet field result Shearlet coefficient The inverse transformation is applied to the spatiotemporal domain to obtain a consistent processing result.

[0071] This invention provides a pre-stack consistency processing method for contiguous data based on the Shearlet domain. After performing consistency preprocessing such as surface consistency amplitude compensation and surface consistency deconvolution on the data of each block in a contiguous exploration area, certain differences still exist in the temporal, frequency, and spatial dimensions of the contiguous data due to factors such as data quality and stratigraphic absorption differences. This invention selects a target model for processing. Within the blocks of relatively good data quality in the contiguous exploration area, a certain number of high-quality data are selected as the target model. Other data in the contiguous exploration area are corrected using this target data as the standard. The consistency correction factor of the Shearlet domain is estimated to eliminate the statistical differences between the target model data and the data to be processed, thus achieving consistency processing of the contiguous exploration area. Processing pre-stack data can effectively eliminate differences in amplitude, frequency, and space in contiguous data and exhibits good stability in noisy environments.

[0072] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0073] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 A flowchart of a pre-stack consistency processing method for contiguous data based on Shearlet domains provided in an embodiment of the present invention;

[0075] Figure 2 The image shows the shot gather data A of Example 1, with the source wavelet having a main frequency of 35Hz and a maximum value of 2.5 as the Rick wavelet.

[0076] Figure 3 This is a schematic diagram of the source wavelet B, which is the shot gather data of Example 1, with a main frequency of 30Hz and a maximum value of 1.

[0077] Figure 4 This is a comparison chart of the statistical spectrum of shot gather data A and B from Example 1;

[0078] Figure 5 This is a comparison of the amplitude statistical histograms of shot gather data A and B from Example 1;

[0079] Figure 6 The result of the shot collection data B from Example 1, with data A as the target, after consistency correction processing;

[0080] Figure 7 The statistical spectrum of shot gather data B after consistency correction in Example 1 is compared with the statistical spectrum of shot gather data A.

[0081] Figure 8 The amplitude statistical histogram of shot gather data B after consistency correction in Example 1 is compared with the amplitude statistical histogram of shot gather data A.

[0082] Figure 9 This is the shot collection data A for Example 2;

[0083] Figure 10 This is the shot collection data B from Example 2;

[0084] Figure 11 A comparison of the statistical spectra of shot gather data A and B in Example 2;

[0085] Figure 12 A comparison of amplitude statistical histograms for shot gather data A and B in Example 2;

[0086] Figure 13 This is the result of shot collection data B from Example 2, with data A as the target, after consistency correction processing;

[0087] Figure 14 The statistical spectrum of shot gather data B after consistency correction in Example 2 is compared with the statistical spectrum of shot gather data A.

[0088] Figure 15The amplitude statistical histogram of shot gather data B after consistency correction in Example 2 is compared with the amplitude statistical histogram of shot gather data A. Detailed Implementation

[0089] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0090] The terms "comprising" and "having," and any variations thereof, in the specification, embodiments, claims, and drawings of this invention are intended to cover non-exclusive inclusion, such as including a series of steps or units.

[0091] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0092] like Figure 1 As shown,

[0093] 1) Step 1: Perform Shearlet transformation on the contiguous data to transform it into the Shearlet field.

[0094] Consistency correction factor estimation and its correction processing require scale decomposition, which is performed in the transform domain. This invention chooses Shearlet transform, which is performed in the Shearlet domain. Shearlet transform has good adaptability to seismic data.

[0095] The Shearlet transform is constructed as follows:

[0096] set up Where: a = 2 -j ,s∈R. A a and S s These represent the parabolic scale matrix and the shear matrix, respectively.

[0097] Let ψ∈L 2 (R 2 And meet the following conditions:

[0098] ① in For the Fourier transform of ψ;

[0099] ②ψ1 is a continuous wavelet transform, supp

[0100] ③ supp Specifically, when the interval is (-1, 1), And ||ψ2||=1.

[0101] Then, define the Shearlet wave atom ψ j,s,m as follows:

[0102]

[0103] Where M AS =S s A a Let the parameters j, s, and m represent the scale, direction, and position (translation) respectively. Then, the Shearlet transform of the function f can be defined as follows:

[0104] S f (j,s,m)=|f,ψ j,s,m > (Equation 2).

[0105] 2) Step 2: Select data A as the target model data, and calculate the spatiotemporal mean and standard deviation after statistical analysis in the Shearlet domain.

[0106] Suppose that in work area A, shot gather data of N shots with relatively good quality are selected as model data, denoted as x. i Where i = 1, 2, ..., N represents the gun number of the model gun, and its Shearlet coefficients after Shearlet transformation are S. x (i,j,s,m).

[0107] The Shearlet coefficients S of the model data are determined using alpha-trim mean filtering. x The spatiotemporal mean and standard deviation of (i,j,s,m).

[0108] For the Shearlet coefficient S of the i-th model shot data x (i,j,s,m), open a rectangular window w of size a×b. a,b Here, a and b are the length and width of the rectangular window, respectively. Alpha-trim mean filtering is performed within the rectangular window (the alpha-trim mean filtering process is as follows: ① sort the coefficients from smallest to largest; ② discard the maximum and minimum values ​​of the proportion α in the total, where 0 ≤ α ≤ 0.5; ③ calculate the mean of the remaining portion), obtaining the mean value corresponding to the center position of the rectangular window. This mean value is then used to further calculate the root mean square error of the window data. The sliding rectangular window w... a,b Finally, the mean of the spatiotemporal variation of the i-th model gun data is obtained. and mean square deviation To improve stability, the spatiotemporal mean and root mean of multiple model gun data were statistically averaged to obtain the statistically derived spatiotemporal mean and root mean:

[0109]

[0110]

[0111] in, and The spatiotemporal mean and spatiotemporal mean variance of the Shearlet coefficients of the target model data at scale j and direction s after multi-shot statistical averaging.

[0112] 3) Step 3: Using the statistical spectrum of model data A as the expectation, perform statistical frequency consistency correction on the data to be processed, B.

[0113] Because the scale coefficients of the Shearlet transform correspond to a certain width of frequency domain support, its frequency consistency processing is not refined enough. Preliminary frequency domain consistency correction is needed before Shearlet domain consistency processing. The statistical spectra of the model data and the contiguous data to be processed are calculated separately in the frequency domain. Using the statistical spectrum of the model data as the expectation, a frequency consistency correction operator is calculated, and statistical frequency consistency correction processing is performed on the contiguous data to be processed. The result after processing is denoted as y. The specific steps are as follows:

[0114] ① Calculate the model data x in the frequency domain respectively. i The statistical spectra of the contiguous data z to be processed are respectively denoted as... and Calculated using the following formula:

[0115] and

[0116] Where X represents the spectrum of each channel of the model data, and N x Z represents the total number of channels in the model data, Z represents the spectrum of each channel in the data to be processed, and N represents the total number of channels in the model data. z This indicates the total number of channels of data to be processed, and abs() indicates modulo processing.

[0117] ② Using the statistical spectrum of the model data as the expectation, calculate the following frequency consistency correction operator F(f).

[0118]

[0119] in, This indicates that the calculation is performed for all possible values ​​of f. The maximum value of ε is the stability factor, which is a small constant.

[0120] ③ Perform statistical frequency consistency correction on the contiguous data z to be processed. The result after processing is denoted as y, and its calculation formula is:

[0121] y = IFFT(Z(f) * F(f))

[0122] Here, IFFT() represents the inverse Fourier transform.

[0123] 4) Step 4: Calculate the spatiotemporal mean and root mean square of data B after step 3 in the Shearlet domain.

[0124] Transform the data y to the Shearlet domain to obtain the Shearlet coefficients S. y (j,s,m), and open a rectangular window w of size a×b on it. a,b (Its size should be consistent with the window size in step 2), perform alpha-trim mean filtering in the same way as in step 2, and obtain the spatiotemporal mean of the data y to be processed by sliding the window. and mean square deviation

[0125] 5) Step 5: Perform Gaussian filtering on the result of step 4.

[0126] right and Two-dimensional Gaussian filtering is performed separately to eliminate local disturbances in the mean and root mean square error results of the spatiotemporal variation, thereby improving stability.

[0127] The formula for the two-dimensional Gaussian function is as follows:

[0128]

[0129] Where λ is the mean square error, the larger the value of λ, the flatter the Gaussian window function and the more obvious the filtering effect; the smaller the value of λ, the steeper the Gaussian window function and the lower the filtering effect. c m d These are the filter parameters.

[0130] Using Formula 5 and The formula for performing two-dimensional Gaussian filtering is as follows:

[0131]

[0132]

[0133] in, and They represent and The result after two-dimensional Gaussian filtering. r represents the radius of the filter template.

[0134] 6) Step 6: Calculate the consistency correction factor using the results of Step 2 and Step 5.

[0135] The consistency correction factor is calculated using the final spatiotemporal mean and root mean error of the model data and the data to be processed, y.

[0136] The formula for the consistency correction factor is:

[0137]

[0138]

[0139] in, and These represent the mean and root mean square error correction factor of the spatiotemporal variation, respectively.

[0140] 7) Step 7: Apply the consistency correction factor to the result of step 3 to obtain the consistency correction result of the Shearlet domain.

[0141] By applying a spatiotemporal consistency correction factor, the Shearlet coefficients of the data y to be processed are corrected to obtain the Shearlet domain result after contiguous data consistency processing.

[0142] The correction formula is:

[0143]

[0144]

[0145] in, This represents the result of adjusting the time-varying mean of the shot gather data to be processed. This indicates the result after adjusting for the time-varying mean and standard deviation.

[0146] 8) Step 8: Transform the result of step 7 into the spatiotemporal domain to obtain the final consistency processing result. Then, process the Shearlet coefficients... The inverse transformation to the spatiotemporal domain yields the final consistent processing result.

[0147] Example 1

[0148] A velocity model was designed, and different source wavelets were used to excite the model at different locations on the ground surface to perform forward modeling of shot gather records.

[0149] Figure 2 It is shot gather data A, with the source wavelet being a Ricker wavelet with a main frequency of 35 Hz and a maximum value of 2.5.

[0150] Figure 3 This is shot gather data B, with the source wavelet being a Ricker wavelet with a dominant frequency of 30Hz and a maximum value of 1. As can be seen from the figure, shot gather A has higher energy than shot gather B.

[0151] Figure 4The graph shows the statistical spectrum of gun sets A and B. As can be seen from the graph, compared with gun set B, gun set A has a slightly higher main frequency and a slightly wider bandwidth, indicating that there is a significant frequency difference between the two.

[0152] Figure 5 The comparison of statistical histograms of the amplitudes of gun sets A and B shows that the amplitudes of both gun sets A and B are mostly distributed around zero, but the amplitude distribution of gun set A is relatively more dispersed, with significantly more large and small values ​​than that of gun set B. Its statistical histogram is also wider, indicating that the root mean square error of gun set A is larger. Consistency correction is then performed on gun set B using gun set A as the target model.

[0153] Figure 6 It is the result of shot gather data B after consistency correction.

[0154] contrast Figure 2 , Figure 3 , Figure 6 After Shearlet domain consistency processing, the amplitude intensity of shot gather data B is basically consistent with the amplitude intensity of data A in the spatiotemporal domain. Furthermore, the characteristics of reflected and diffracted waves are consistent with the characteristics of the data before processing, indicating that the geometric and wave characteristics of the shot gather data before and after processing are well preserved.

[0155] Figure 7 The statistical spectrum of the shot collection data B after consistency correction is compared with the statistical spectrum of the shot collection data A. After processing by the method of this invention, the statistical spectrum of data B is basically consistent with the statistical spectrum of model data A.

[0156] Figure 8 The amplitude statistical histogram of shot gather data B after consistency correction is compared with the amplitude statistical histogram of shot gather data A. The distribution patterns of the two are basically similar. Except for the distribution characteristics near the zero value, the distribution of other amplitudes is basically the same, indicating that the amplitudes of the processed data and the model data are statistically consistent.

[0157] Example 2

[0158] To further verify the effectiveness and processing effect of the method of the present invention, consistency correction processing was performed on the shot gathering data of two blocks in the actual contiguous exploration area.

[0159] Figure 9 This is the shot set data for block A. Figure 10 This is the shot gather data from block B. As can be seen from the graph, the signal-to-noise ratio of the data from block A is better than that from block B.

[0160] Figure 11The statistical spectrum comparison between blocks A and B shows that the bandwidth of the two blocks is roughly the same, but the bandwidth of block B is slightly wider and the energy in the 40Hz to 60Hz range is significantly stronger than that of block A.

[0161] Figure 12 A comparison of the amplitude statistical histograms of blocks A and B reveals that the root mean square error of block B is slightly larger, while the number of amplitudes in block A falling within the range of -1000 to 1000 is significantly greater than that in block B, indicating a certain difference in amplitude distribution between the two. Consistency correction is then performed using block A as the target.

[0162] Figure 13 This is the final result of the data consistency correction process for Block B. It can be seen that after the consistency process, the energy distribution of the profile is more balanced and local noise is also suppressed.

[0163] Figure 14 The statistical spectrum of block B data after consistency correction is compared with that of block A data. It can be seen that the present invention can effectively correct the statistical frequency band difference between the two blocks of data. After processing, the main frequency, bandwidth and other parameters of the two are basically consistent.

[0164] Figure 15 The amplitude statistical histogram of block B data after consistency correction is compared with the amplitude statistical histogram of block A data. It can be seen that the method of the present invention greatly eliminates the difference in amplitude statistical distribution between the two blocks of data, and realizes the consistency of amplitude and frequency in time and space, thus verifying the effectiveness of the method of the present invention.

[0165] Beneficial effects: Reliability of the method's effectiveness. This invention utilizes the multi-scale and multi-directional nature of the Shearlet transform. Based on the optimized target data, it uses the alpha-trim filtering method to estimate the time-varying mean and root mean square error of the data to be processed and the target data at each scale and direction, extracting and eliminating the trend differences between the two. Compared with conventional methods, the results are more reasonable and have better stability in noisy environments.

[0166] The method is highly adaptable. This invention was developed for the consistency processing of contiguous data, and can be applied to both pre-stack and post-stack data, without requiring overlapping areas in the contiguous data.

[0167] The operation is simple and easy to implement. The steps of this invention are simple and convenient, and there is no need for tedious parameter adjustments and tests.

[0168] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for processing pre-stack consistency based on Shearlet domain patch data, characterized in that, The processing method comprises: Collecting seismic data of each block in a continuous exploration area to obtain continuous data; Transforming the continuous data to a Shearlet domain; Calculating a time-space variable mean and a mean square deviation of the continuous data in the Shearlet domain, and establishing a target model; Calculating statistical spectra of model data and to-be-processed continuous data respectively in a frequency domain; Taking the statistical spectrum of the model data as an expectation, calculating a frequency consistency correction operator, performing statistical frequency consistency correction processing on the to-be-processed continuous data to obtain a correction processing result y; transforming the correction processing result y to a Shearlet domain to obtain Shearlet coefficients S y (j, s, m); to the Shearlet coefficients S y (j,s,m) opens a rectangular window w of size a x b a,b , performs an alpha-trim mean filter processing and slides the rectangular window w a,b obtains a spatio-temporal varying mean of the correction processing result y and a mean square error For the mean and the mean square deviation Two-dimensional Gaussian filtering is performed separately to obtain the Gaussian filtered mean. and Gaussian filter mean square error calculating a spatio-temporal variation consistency correction factor based on the model data, the Gaussian filter mean and the Gaussian filter mean square deviation ; According to the time-space variable consistency correction factor, performing correction processing on Shearlet coefficients of the correction processing result y to obtain a Shearlet domain result of the consistency-processed continuous data.

2. The method according to claim 1, wherein, The transformation of the continuous data to the Shearlet domain comprises: Construction of a Shearlet transform: Set where a = 2 -j , s e R, A a Parabolic scaling matrix, S s Shearing matrix; Let ψ∈L 2 (R 2 ), and satisfies the following conditions: wherein is the Fourier transform of ψ; ψ1is a continuous wavelet transform, wherein, when on the interval (-1, 1), and ||ψ2|| = 1; Definition of Shearlet wavelet atom ψ j,s,m As follows: where M AS = S s A a , parameters j scale, s direction and m position, the Shearlet transform of a function f is defined as follows: S f (j,s,m) = <f, ψ j,s,m > (Formula 2).

3. The method according to claim 1, wherein the method is characterized in that, The calculation of the time-space variable mean and the mean square deviation of the continuous data in the Shearlet domain, and the establishment of the target model comprise: Acquire shot gather data and take it as model data, obtain the i-th model shot data x i where i = 1, 2, …, N represents the shot number of the model shot data, wherein the Shearlet coefficients S x (i, j, s, m) after the Shearlet transform; determining the Shearlet coefficients S using alpha-trim mean filtering x spatio-temporal variable mean and mean square deviation of (i,j,s,m) Shearlet coefficients S of the i-th model gun data x (i,j,s,m), the rectangular window w with size a x b a,b wherein a and b are the length and width of the rectangular window respectively, and the alpha-trim mean filtering is performed in the rectangular window to obtain the corresponding mean value at the center position of the rectangular window. calculating a mean square error of the windowed data from the mean, sliding the rectangular window w a,b eventually obtaining a time-varying mean of the i-th model shot data and a mean square error Statistically averaging time-space variable means and mean square deviations of a plurality of model shot data to obtain statistical time-space variable means and mean square deviations: wherein and denote the spatio-temporal variable mean and the spatio-temporal variable mean square deviation of the Shearlet coefficients of the target model data after multi-shot statistical averaging in scale j and direction s.

4. The method according to claim 1, wherein, The taking of the statistical spectrum of the model data as the expectation, the calculation of the frequency consistency correction operator, and the statistical frequency consistency correction processing on the to-be-processed continuous data to obtain the correction processing result y comprise: calculating the model data x in the frequency domain i and the statistical spectrum of the to-be-processed continuous data z are respectively denoted as and and wherein X represents the frequency spectrum of each trace of the model data, N x represents the total number of traces of the model data, Z represents the frequency spectrum of each trace of the continuous data to be processed, N z represents the total number of traces of the continuous data to be processed, and abs() represents the modulus processing. Taking the statistical spectrum of the model data as an expectation, calculating a frequency f consistency correction operator F(f), wherein, represents the maximum value of the calculation of for all values of the frequency f, and ε is a stability factor, taken as a small constant; Performing statistical frequency consistency correction processing on to-be-processed continuous data z, and recording a processing result as y, and the calculation formula is, y = IFFT(Z(f) * F(f)) Where IFFT() represents an inverse Fourier transform.

5. The method according to claim 1, wherein, said mean value and said mean square deviation respectively performing two-dimensional Gaussian filtering processing specifically includes: A two-dimensional Gaussian function formula is as follows: where λ is a mean square error parameter, m c , m d is a filter parameter; The mean value is filtered by a two-dimensional Gaussian filter using formula 5 and the mean square error is filtered by a two-dimensional Gaussian filter using formula 6 The formula is as follows: wherein the Gaussian filter mean and the Gaussian filter mean square deviation r denotes the radius of the filter template.

6. The method according to claim 1, wherein, The spatial-temporal variation consistency correction factor is calculated according to the model data, the Gaussian filter mean value and the Gaussian filter mean square deviation The calculating the spatial-temporal variation consistency correction factor specifically comprises: A time-space variable consistency correction factor formula is as follows: where and denote the spatio-temporally varying mean and mean square error correction factors, respectively.

7. The method according to claim 1, wherein the method is characterized by, The correction processing on the Shearlet coefficients of the correction processing result y according to the time-space variable consistency correction factor comprises: A correction formula is as follows: wherein, represents the result of adjusting the time-varying mean of the data of the shot gather to be processed, represents the result of adjusting the time-varying mean and the mean square deviation.

8. The method according to claim 1, wherein the method is characterized in that, The processing method further comprises: The Shearlet domain result Shearlet coefficients Inverse transform to the space-time domain, get the consistency processing results.

Images