Seismic data noise attenuation method, device, equipment and medium

Abstract

The invention discloses a seismic data noise attenuation method, device and equipment and a medium. The seismic data noise attenuation method comprises the following steps: reading a group of two-dimensional seismic profile data; performing NSST multi-scale and multi-direction decomposition on the read-in two-dimensional seismic section data to obtain a low-pass coefficient and a series of high-pass coefficients in different scales and different directions; based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic section data, determining a denoising threshold of each direction sub-band under each scale, and using the denoising threshold to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients; and performing NSST inverse transformation on the low-pass coefficient and a series of denoised high-pass coefficients to reconstruct seismic data. According to the seismic data noise attenuation method, medium-high frequency noise in original seismic data can be effectively suppressed, the signal-to-noise ratio of acquired data is improved, and the method has good fidelity.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic data processing, and in particular to a seismic data noise attenuation method, device, computer equipment and computer-readable storage medium, which are used in the noise suppression link in seismic data processing, especially the noise suppression of seismic data collected by high-sensitivity detectors. Background Art

[0002] With the development of seismic acquisition technology, ultra-high sensitivity geophones with a sensitivity higher than 160V·m / s developed for surface environments with poor coupling conditions have moved from research and development to actual production. Through comparative tests between conventional geophones and ultra-high sensitivity geophones, it is found that ultra-high sensitivity geophones receive richer high-frequency weak reflection signals from earthquakes, but are also sensitive to high-frequency noise, and the difference between high-frequency noise and weak reflection signals is small, making it difficult to effectively distinguish them. Therefore, how to suppress the high-frequency noise of data collected by ultra-high sensitivity geophones while retaining weak reflection signals is a difficult problem in data processing for ultra-high sensitivity geophones. At present, there are few reports on noise suppression technology for ultra-high sensitivity geophones.

[0003] Based on this, there is still room for further improvement in the existing technology. Summary of the invention

[0004] The present invention is based on the Non-Subsampled Shearlet Transform (NSST) and combines the characteristics of ultra-high sensitivity detectors in collecting data to develop a noise attenuation method, device, equipment and medium for seismic data collected by ultra-high sensitivity detectors, which helps to give full play to the advantages of ultra-high sensitivity detectors and improve the resolution ability of seismic data for thin layers and micro-structures.

[0005] According to a first aspect of the present invention, a method for attenuating noise in seismic data is provided, which comprises the following steps:

[0006] Step 1: Read in a set of 2D seismic profile data;

[0007] Step 2: Perform NSST multi-scale and multi-directional decomposition on the read 2D seismic profile data to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions;

[0008] Step 3: Based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, determine the denoising threshold of each directional sub-band at each scale, and use the denoising threshold to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients;

[0009] Step 4: Perform NSST inverse transformation on the low-pass coefficients and the series of denoised high-pass coefficients to reconstruct seismic data.

[0010] According to one embodiment of the present invention, in step 3, a series of high-pass coefficients obtained by NSST decomposition are denoised using a Bayesian-based Stein threshold.

[0011] According to one embodiment of the present invention, the denoising threshold of each directional subband at each scale is calculated according to the following formula:

[0012]

[0013] Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

[0014] According to one embodiment of the present invention, the denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including:

[0015] The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold;

[0016] If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0;

[0017] If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

[0018] According to one embodiment of the present invention, in step 1, the read-in two-dimensional seismic profile data is stored as a matrix.

[0019] According to one embodiment of the present invention, the read-in two-dimensional seismic profile data is two-dimensional seismic profile data collected by an ultra-high sensitivity detector.

[0020] According to one embodiment of the present invention, in step 3, the low-pass coefficients obtained by NSST decomposition are not processed.

[0021] According to a second aspect of the present invention, a seismic data noise attenuation device is provided, comprising:

[0022] a data acquisition module configured to acquire two-dimensional seismic profile data;

[0023] A non-subsampled shear wave transform module is configured to perform NSST multi-scale and multi-directional decomposition on the two-dimensional seismic profile data acquired by the data acquisition module to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions;

[0024] A denoising module is configured to determine a denoising threshold for each directional sub-band at each scale based on a standard deviation of Gaussian noise and a size of the two-dimensional seismic profile data, and to denoise a series of high-pass coefficients obtained by decomposing the non-subsampled shearlet transform module NSST using the denoising threshold to obtain a series of denoised high-pass coefficients; and

[0025] The seismic data reconstruction module is configured to perform NSST inverse transformation on the low-pass coefficients obtained by the non-subsampled shearlet transform module NSST decomposition and the series of denoised high-pass coefficients obtained after denoising by the denoising module to reconstruct seismic data.

[0026] According to one embodiment of the present invention, the denoising module is configured to denoise a series of high-pass coefficients obtained by NSST decomposition using a Bayesian-based Stein threshold.

[0027] According to one embodiment of the present invention, the denoising threshold of each directional subband at each scale is calculated according to the following formula:

[0028]

[0029] Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

[0030] According to one embodiment of the present invention, the denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including:

[0031] The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold;

[0032] If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0;

[0033] If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

[0034] According to one embodiment of the present invention, the data acquisition module stores the acquired two-dimensional seismic profile data as a matrix.

[0035] According to an embodiment of the present invention, the two-dimensional seismic profile data acquired by the data acquisition module is two-dimensional seismic profile data collected by an ultra-high sensitivity detector.

[0036] According to a third aspect of the present invention, a computer device is provided, comprising:

[0037] At least one processor; and a memory, wherein the memory stores a computer program that can be run on the processor, wherein when the processor executes the computer program, a method for attenuating noise in seismic data is implemented, comprising:

[0038] Step 1: Read in a set of 2D seismic profile data;

[0039] Step 2: Perform NSST multi-scale and multi-directional decomposition on the read 2D seismic profile data to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions;

[0040] Step 3: Based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, determine the denoising threshold of each directional sub-band at each scale, and use the denoising threshold to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients;

[0041] Step 4: The low-pass coefficient and the series of denoised high-pass coefficients are

[0042] The inverse NSST transform is used to reconstruct seismic data.

[0043] According to one embodiment of the present invention, in step 3, a series of high-pass coefficients obtained by NSST decomposition are denoised using a Bayesian-based Stein threshold.

[0044] According to one embodiment of the present invention, the denoising threshold of each directional subband at each scale is calculated according to the following formula:

[0045]

[0046] Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

[0047] According to one embodiment of the present invention, the denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including:

[0048] The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold;

[0049] If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0;

[0050] If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

[0051] According to one embodiment of the present invention, in step 1, the read-in two-dimensional seismic profile data is stored as a matrix.

[0052] According to one embodiment of the present invention, in step 1, the two-dimensional seismic profile data read in is the two-dimensional seismic profile data collected by an ultra-high sensitivity detector.

[0053] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the seismic data noise attenuation method according to the first aspect of the present invention is implemented.

[0054] The present invention uses the non-subsampled shearlet transform (NSST) technology of adaptive threshold to decompose seismic signals in multiple scales and multiple directions. The high-frequency information obtained by different decomposition scales and directions is complementary, and mutually represents the edge and detail parts of the image, so that the different high-frequency information obtained by decomposition is helpful for analysis at different angles; then, the adaptive threshold method is used to calculate a relatively robust threshold, and the decomposed high-frequency signal is subjected to noise suppression, while protecting the high-frequency weak effective reflection signal, distinguishing the high-frequency important signal from the noise signal, and realizing accurate suppression of random noise. The method of the present invention can effectively suppress the medium and high-frequency noise in the original seismic data, improve the signal-to-noise ratio of the acquired data, and has good fidelity. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 Schematic diagram showing the NSST decomposition of synthetic seismic data;

[0057] Figure 2 shows the reconstruction error graphs of synthetic seismic data in the wavelet transform, curvelet transform and NSST transform domains;

[0058] Figure 3 A flow chart showing a method for attenuating noise in seismic data according to an embodiment of the present invention;

[0059] Figure 4The NSST domain seismic random noise attenuation analysis framework according to an embodiment of the present invention is shown;

[0060] Figure 5 Graphs showing data collected by an ultra-high sensitivity detector before and after being processed by the method of the present invention, wherein the left graph corresponds to the shot gather record before processing, the middle graph corresponds to the shot gather record after processing, and the right graph corresponds to the removed noise signal;

[0061] Figure 6 The figure shows the enlarged display of the stacked sections before and after the processing of the data collected by the ultra-high sensitivity detector using the method of the present invention, wherein the left figure corresponds to the enlarged display of the stacked sections before the processing, and the right figure corresponds to the enlarged display of the stacked sections after the processing;

[0062] Figure 7 A schematic diagram of a seismic data noise attenuation device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0063] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0064] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. The subsequent embodiments will not explain this one by one.

[0065] Non-subsampled shearlet transform (NSST) is an excellent multi-scale, multi-directional and optimal sparsity analysis method, which can not only accurately calculate the shearlet coefficients, but also provide an almost optimal approximation of the decomposed seismic effective signal. The present invention uses the non-subsampled shearlet transform technology with adaptive threshold to perform multi-scale and multi-directional decomposition of seismic signals. Figure 1As shown in the figure, a low-frequency coefficient image and a series of high-frequency coefficient images at different scales and directions are obtained through NSST decomposition. The high-frequency coefficients at different scales and directions clearly represent the edges and details of the image, and the edges and details of the image obtained by fusing the different high-frequency coefficients obtained by decomposition are better represented. Then, the Bayesian adaptive threshold method is used to calculate a relatively robust threshold, and the decomposed high-frequency signal is subjected to noise suppression. While protecting the high-frequency weak effective reflection signal, the high-frequency important signal and the noise signal are distinguished, so as to achieve accurate suppression of random noise. Since the non-subsampled shearlet transform is multi-directional, when the direction of the non-subsampled shearlet transform basis function is roughly the same as the direction of the seismic signal, the basis function coefficient is larger; and when the direction of the non-subsampled shearlet transform basis function is very different from the direction of the seismic signal, the basis function coefficient is smaller. Therefore, random noise is often distributed on small coefficients, so an appropriate threshold operator can be used to remove small coefficients to achieve the purpose of random noise attenuation.

[0066] Generally speaking, the effect of noise attenuation depends on the approximation of the decomposed effective signal. The greater the sparsity of the method, the better the denoising effect. Figure 2 The figure shows the reconstruction error curves of synthetic seismic data in the wavelet transform, curvelet transform and NSST transform domains. It can be clearly seen that the reconstruction error of NSST transform is basically 0 when only 5% of the coefficients are retained, that is, when the three transforms retain the same percentage of coefficients for reconstruction, NSST can obtain the least reconstruction error and has the best sparse property.

[0067] In the process of recovering the expected signal, it is assumed that the noisy seismic data to be processed is:

[0068] F(t,d)=X(t,d)+N(t,d), (1)

[0069] Among them, F(t,d) is the actual noisy seismic data collected, X(t,d) refers to the effective seismic signal, N(t,d) refers to random noise (Gaussian white noise); t and d refer to time and number of seismic channels respectively. The goal of NSST domain seismic data denoising is to recover the effective signal X(t,d) from the observed signal F(t,d).

[0070] According to a first aspect of the present invention, a method for attenuating noise in seismic data is provided. Figure 3 A flow chart of a method for attenuating noise in seismic data according to an embodiment of the present invention is shown. As shown in the figure, the method includes the following steps:

[0071] S1: Read in a set of 2D seismic profile data;

[0072] S2: Perform NSST multi-scale and multi-directional decomposition on the read 2D seismic profile data to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions;

[0073] S3: Based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, the denoising threshold of each directional sub-band at each scale is determined, and the denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients;

[0074] S4: Perform NSST inverse transform on the low-pass coefficients and a series of denoised high-pass coefficients to reconstruct the seismic data.

[0075] In some embodiments, step 2 includes: by NSST decomposition, the read seismic data is subjected to NSST decomposition according to a pre-set decomposition level J, thereby obtaining a low-frequency coefficient image (i.e., low-pass coefficient) and a series of high-frequency coefficient images at different scales and directions (i.e., high-pass coefficients). According to the set decomposition level J and decomposition direction number L, high-pass coefficients D at different scales and directions are obtained. jl (j=1,2,…,J;l=1,2,…,L). A series of high-pass coefficients generated by NSST coefficient decomposition can be obtained through formula (2). F(j,l,k) Indicates NSST decomposition of the data, including a low-frequency coefficient image and a series of high-frequency coefficient images D jl , and then the high-pass coefficient D jl Perform denoising.

[0076] NSSH F(j,l,k) =NSSH{F(t,d)}= <F(t,d),ψ j,l,k >. (2)

[0077] Among them, ψ j,l,k is the generating function, k is the location parameter, and NSSH{F(t,d)} refers to the NSST decomposition of the actual noisy seismic data.

[0078] In some embodiments, in step 3, a series of high-pass coefficients D obtained by NSST decomposition are decomposed using a Bayesian Stein threshold jlDenoising. When the direction of the non-subsampled shearlet transform basis function is roughly the same as the direction of the seismic signal, the coefficient of the obtained basis function is large; when the two are very different, the coefficient of the obtained basis function is small. Therefore, the Stein threshold method, which is a compromise between the hard threshold (if the basis function coefficient is very small, it is judged as noise and set to zero; if the basis function coefficient is very large, it is judged to be reasonable) and the soft threshold (all coefficients are reduced to a certain extent, thereby reducing the estimated variance at the cost of higher deviation), is used for processing: first, the NSST is used to decompose the noise image into images of different scales and directions; then the NSST coefficient subbands are denoised using the Bayesian-based Stein threshold.

[0079] In some embodiments, the denoising threshold of each directional subband at each scale is calculated according to the following formula (3):

[0080]

[0081] Where j and l refer to the scale and direction of decomposition, respectively; T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband; S is the size of the 2D seismic profile data; δ j,l It refers to the standard deviation of Gaussian noise of the j-th scale and l-th direction subband, and its formula is:

[0082]

[0083] Where m×n is the number of sampling points in the seismic data matrix, x i is the value of the i-th sampling point, and μ is the mean of all sampling points.

[0084] For different two-dimensional seismic profile data, the size of S is the same; at the same time, the standard deviation of different seismic data can represent the discrete degree of the data, reflecting the distribution characteristics of seismic data in different blocks. Therefore, by processing different seismic data according to the above formula, the denoising threshold that conforms to their respective distribution characteristics can be obtained, highlighting its adaptability.

[0085] In some embodiments, a denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including: comparing the absolute value of the high-pass coefficient with the corresponding denoising threshold; if the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, setting the denoised high-pass coefficient to 0; if the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, setting the denoised high-pass coefficient to be equal to the original high-pass coefficient, that is, keeping it unchanged.

[0086] Specifically, the formula is as follows:

[0087]

[0088] In some embodiments, in step 1, the read-in two-dimensional seismic profile data is stored as a matrix.

[0089] In some embodiments, in step 1, the two-dimensional seismic profile data read in is two-dimensional seismic profile data collected by an ultra-high sensitivity detector.

[0090] In some embodiments, in step 3, the low-pass coefficients obtained by NSST decomposition are not processed.

[0091] Figure 4 The NSST domain seismic random noise attenuation analysis framework according to an embodiment of the present invention is shown. As shown in the figure, in this embodiment, in step S1, the noisy synthetic seismic data on the far left is obtained; in step S2, the noisy synthetic seismic data is decomposed by two-scale NSST to generate a low-pass transform coefficient image located in the middle left and high-pass transform coefficient images at two scales, the first-level scale setting decomposes 4 shear directions, and the second-level scale setting decomposes 2 shear directions; in step S3, a total of 6 high-frequency coefficient images obtained by NSST decomposition are denoised to obtain a denoised high-frequency coefficient image located in the middle right; finally, in step S4, the seismic data is reconstructed to obtain denoised synthetic seismic data. Note that no operation is performed on the low-pass transform coefficient image obtained by NSST decomposition. Specific embodiments

[0093] The method of the present invention is applied to the 150 area of ​​Liaohe Oilfield Town, and the newly collected data using an ultra-high sensitivity detector is subjected to noise suppression processing. Figure 5 The diagram shows the data collected by the ultra-high sensitivity detector before and after being processed by the method of the present invention, wherein the left diagram corresponds to the shot collection record before processing, the middle diagram corresponds to the shot collection record after processing, and the right diagram corresponds to the removed noise signal. Figure 5 It can be seen that the noise attenuation method of the seismic data collected by the ultra-high sensitivity detector of the present invention is effectively suppressed in the original seismic data, and the signal-to-noise ratio of the data collected by the ultra-high sensitivity detector is improved. At the same time, no effective signal is found from the removed noise signal, which shows that the technology has good fidelity. Figure 6 The figure shows the enlarged display of the stacked sections before and after the processing of the data collected by the ultra-high sensitivity detector using the method of the present invention, wherein the left figure corresponds to the enlarged display of the stacked sections before the processing, and the right figure corresponds to the enlarged display of the stacked sections after the processing. Figure 6 It can be found that after processing using the method of the present invention, high-frequency noise is better suppressed, while the high-frequency weak reflection signal phase axis is more continuous and clear, thereby further exerting the advantages of the ultra-high sensitivity detector and avoiding its shortcomings. As a supporting processing technology for ultra-high sensitivity detectors, it has very important practical significance for its promotion and application.

[0094] According to a second aspect of the present invention, a seismic data noise attenuation device is provided. Figure 7 A schematic diagram of a seismic data noise attenuation device according to an embodiment of the present invention is shown. As shown in the figure, the seismic data noise attenuation device 100 includes: a data acquisition module 110, configured to acquire two-dimensional seismic profile data; a non-subsampled shear wave transform module 120, configured to perform NSST multi-scale and multi-directional decomposition on the two-dimensional seismic profile data acquired by the data acquisition module 110, to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions; a denoising module 130, configured to determine the denoising threshold of each directional sub-band at each scale based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, and to denoise a series of high-pass coefficients obtained by NSST decomposition of the non-subsampled shear wave transform module 120 using the denoising threshold to obtain a series of denoised high-pass coefficients; and a seismic data reconstruction module 140, configured to perform NSST inverse transformation on the low-pass coefficients obtained by NSST decomposition of the non-subsampled shear wave transform module 120 and a series of denoised high-pass coefficients obtained by denoising the denoising module 130 to reconstruct seismic data.

[0095] In some embodiments, the denoising module 130 is configured to denoise a series of high-pass coefficients obtained from the NSST decomposition using a Bayesian-based Stein threshold.

[0096] In some embodiments, the denoising threshold of each directional subband at each scale is calculated according to the following formula:

[0097]

[0098] Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band, and S is the size of the two-dimensional seismic profile data.

[0099] In some embodiments, a denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including: comparing the absolute value of the high-pass coefficient with the corresponding denoising threshold; if the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, setting the denoised high-pass coefficient to 0; if the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, setting the denoised high-pass coefficient to be equal to the original high-pass coefficient.

[0100] In some embodiments, the data acquisition module 110 acquires two-dimensional seismic profile data from ultra-high sensitivity detectors and stores them in a matrix.

[0101] In some embodiments, the two-dimensional seismic profile data acquired by the data acquisition module 110 is two-dimensional seismic profile data collected by an ultra-high sensitivity detector.

[0102] According to a third aspect of the present invention, a computer device is provided, comprising: at least one processor; and a memory, wherein the memory stores a computer program executable on the processor, and when the processor executes the computer program, the seismic data noise attenuation method according to the first aspect of the present invention is implemented.

[0103] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the seismic data noise attenuation method according to the first aspect of the present invention is implemented.

[0104] In summary, the present invention uses the non-subsampled shearlet transform (NSST) technology with adaptive threshold to decompose seismic signals in multiple scales and directions, considers the correlation between multiple high frequencies decomposed, uses the adaptive threshold method to calculate a relatively robust threshold, suppresses the noise of the decomposed high-frequency signal, and distinguishes between high-frequency important signals and noise signals while protecting high-frequency weak effective reflection signals, thereby achieving accurate suppression of random noise. Based on the characteristics of ultra-high sensitivity detectors that can receive more abundant high-frequency weak signals, through the research and development and application of the present invention, the signal-to-noise ratio of high-frequency weak signals can be better improved than conventional noise suppression technology, which is of great significance to the promotion and application of high-sensitivity detectors.

[0105] It should be noted that a person skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0106] Furthermore, it should be appreciated that the computer-readable storage medium (eg, memory) herein may be either a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memory.

[0107] A person skilled in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the disclosure of the embodiments of the present invention (including the claims) is limited to these examples; under the concept of the embodiments of the present invention, the technical features in the above embodiments or different embodiments can also be combined, and there are many other changes in different aspects of the embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the protection scope of the embodiments of the present invention.

Claims

1. A method for attenuating noise in seismic data, characterized in that: The following steps are involved: Step 1: Read in a set of 2D seismic profile data; Step 2: Perform NSST multi-scale and multi-directional decomposition on the read 2D seismic profile data to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions; Step 3: Based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, determine the denoising threshold of each directional sub-band at each scale, and use the denoising threshold to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients; Step 4: Perform NSST inverse transformation on the low-pass coefficients and the series of denoised high-pass coefficients to reconstruct seismic data.

2. The method for attenuating seismic data noise according to claim 1, characterized in that: In step 3, the series of high-pass coefficients obtained by NSST decomposition are denoised using the Bayesian-based Stein threshold.

3. The method for attenuating seismic data noise according to claim 2, characterized in that: The denoising threshold of each directional subband at each scale is calculated according to the following formula: Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

4. The method for attenuating seismic data noise according to claim 3, characterized in that: The denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including: The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold; If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0; If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

5. The method for attenuating seismic data noise according to claim 1, characterized in that: In step 1, the read-in two-dimensional seismic profile data is stored as a matrix.

6. The method for attenuating seismic data noise according to claim 1, characterized in that: In step 1, the two-dimensional seismic profile data read in are the two-dimensional seismic profile data collected by the ultra-high sensitivity detector.

7. The method for attenuating seismic data noise according to claim 1, characterized in that: In step 3, the low-pass coefficients obtained by NSST decomposition are not processed.

8. A seismic data noise attenuation device, characterized in that: include: a data acquisition module configured to acquire two-dimensional seismic profile data; A non-subsampled shear wave transform module is configured to perform NSST multi-scale and multi-directional decomposition on the two-dimensional seismic profile data acquired by the data acquisition module to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions; A denoising module is configured to determine a denoising threshold for each directional sub-band at each scale based on a standard deviation of Gaussian noise and a size of the two-dimensional seismic profile data, and to denoise a series of high-pass coefficients obtained by decomposing the non-subsampled shearlet transform module NSST using the denoising threshold to obtain a series of denoised high-pass coefficients; and The seismic data reconstruction module is configured to perform NSST inverse transformation on the low-pass coefficients obtained by the non-subsampled shearlet transform module NSST decomposition and the series of denoised high-pass coefficients obtained after denoising by the denoising module to reconstruct seismic data.

9. The seismic data noise attenuation device according to claim 8, characterized in that: The denoising module is configured to denoise a series of high-pass coefficients obtained by NSST decomposition using a Bayesian-based Stein threshold.

10. The seismic data noise attenuation device according to claim 9, characterized in that: The denoising threshold of each directional subband at each scale is calculated according to the following formula: Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

11. The seismic data noise attenuation device according to claim 10, characterized in that: The denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including: The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold; If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0; If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

12. The seismic data noise attenuation device according to claim 8, characterized in that: The data acquisition module stores the acquired two-dimensional seismic profile data as a matrix.

13. The seismic data noise attenuation device according to claim 8, characterized in that: The two-dimensional seismic profile data acquired by the data acquisition module is the two-dimensional seismic profile data collected by the ultra-high sensitivity detector.

14. A computer device comprising: at least one processor; as well as A memory storing a computer program executable on the processor, wherein the processor implements a method for attenuating noise in seismic data when executing the computer program, comprising: Step 1: Read in a set of 2D seismic profile data; Step 2: Perform NSST multi-scale and multi-directional decomposition on the read 2D seismic profile data to obtain low-pass coefficients and a series of high-pass coefficients at different scales and directions; Step 3: Based on the standard deviation of Gaussian noise and the size of the two-dimensional seismic profile data, determine the denoising threshold of each directional sub-band at each scale, and use the denoising threshold to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients; Step 4: Perform NSST inverse transformation on the low-pass coefficients and the series of denoised high-pass coefficients to reconstruct seismic data.

15. The computer device according to claim 14, characterized in that In step 3, the series of high-pass coefficients obtained by NSST decomposition are denoised using the Bayesian-based Stein threshold.

16. The computer device according to claim 15, characterized in that The denoising threshold of each directional subband at each scale is calculated according to the following formula: Where j and l refer to the scale and direction of decomposition, respectively. T(j,l) refers to the denoising threshold of the j-th scale and l-th direction subband. j,l It refers to the standard deviation of the Gaussian noise of the j-th scale and l-th direction sub-band. S is the size of the m×n matrix storing the two-dimensional seismic profile data. m and n are the number of sample points in the rows and columns of the two-dimensional seismic profile data matrix, respectively.

17. The computer device according to claim 14, characterized in that The denoising threshold is used to denoise a series of high-pass coefficients obtained by NSST decomposition to obtain a series of denoised high-pass coefficients, including: The absolute value of the high-pass coefficient is compared with the corresponding denoising threshold; If the absolute value of the high-pass coefficient is less than the corresponding denoising threshold, the denoised high-pass coefficient is set to 0; If the absolute value of the high-pass coefficient is not less than the corresponding denoising threshold, the denoised high-pass coefficient is set equal to the original high-pass coefficient.

18. The computer device according to claim 14, characterized in that In step 1, the read-in two-dimensional seismic profile data is stored as a matrix.

19. The computer device according to claim 14, characterized in that In step 1, the two-dimensional seismic profile data read in are the two-dimensional seismic profile data collected by the ultra-high sensitivity detector.

20. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the seismic data noise attenuation method according to any one of claims 1 to 7 is implemented.

Images