Method and device for seismic data low-frequency recovery constrained by reflection structure

Abstract

A method for seismic data low-frequency recovery constrained by a reflection structure, which belongs to the field of seismic data processing. The method comprises: on the basis of original seismic data, determining a highest frequency to be recovered corresponding to low-frequency seismic data to be recovered; by using the original seismic data and an inversion algorithm, determining a reflection structure characterization operator; by using a reflection coefficient and the reflection structure characterization operator, constructing a regularization constraint term, and establishing a target functional for the reflection coefficient; reversely solving the target functional for the reflection coefficient to obtain the reflection coefficient; performing low-frequency filtering on the reflection coefficient to obtain the low-frequency seismic data; performing high-pass filtering on the original seismic data to obtain high-frequency seismic data; and combining the low-frequency seismic data and the high-frequency seismic data, so as to obtain seismic data after low-frequency recovery. The provided method can ensure the reliability of low-frequency seismic data, enhance the resolution of seismic data, and improve the accuracy of full-waveform velocity inversion. Further provided is a device for seismic data low-frequency recovery constrained by a reflection structure.

Description

Low-frequency recovery method and device for seismic data constrained by reflection structure Technical Field

[0001] The present invention relates to the technical field of seismic data processing, and in particular to a low-frequency recovery method for seismic data constrained by a reflection structure and a low-frequency recovery device for seismic data constrained by a reflection structure. Background Art

[0002] Due to factors such as excitation, reception, and surface-roll interference, low-frequency data in actual seismic recordings is suppressed, contaminated, and distorted. In some cases, low-frequency data in certain frequency bands is even missing. For example, low-frequency data excitation can be difficult during data acquisition, or the natural frequency of existing detectors can suppress low-frequency data during data reception. Another example is surface-roll interference, which contaminates low-frequency data and suppresses low-frequency data in actual seismic recordings. While various methods exist to suppress surface-roll interference, the process inevitably harms low-frequency data. Furthermore, field environmental noise surveys and analyses have shown that the frequency components of ambient noise are not ideally distributed white noise, and its low-frequency data often possesses higher energy. The loss of low-frequency data increases seismic wavelet sidelobes, exacerbates the interference between different strata, and reduces seismic data resolution. Low-frequency data also plays an indispensable role in impedance inversion.

[0003] To compensate for the impact of low-frequency loss on impedance inversion, existing techniques often use well logging data interpolation along horizons to construct low-frequency models. However, this approach limits the accuracy of well logging calibration and horizon interpretation, making it difficult to guarantee the reliability of low-frequency data. Low-frequency loss also severely reduces the accuracy of full-waveform velocity inversion.

[0004] Summary of the Invention

[0005] In response to the technical problems in the existing technology that low-frequency seismic data has low reliability, reduced the accuracy of full-waveform velocity inversion, and reduced the resolution of seismic data, the present invention provides a low-frequency recovery method and device for seismic data constrained by reflection structure. This method can ensure the reliability of low-frequency seismic data, enhance the resolution of seismic data, and improve the accuracy of full-waveform velocity inversion.

[0006] To achieve the above-mentioned purpose, the present invention provides a reflection structure constrained seismic data low-frequency recovery method comprising the following steps: obtaining original seismic data, determining the highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data; determining a reflection structure characterization operator using the original seismic data and an inversion algorithm; constructing a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establishing a target functional of the reflection coefficient through the regularization constraint term; inversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; performing a low-frequency filtering on the reflection coefficient with a high cutoff frequency of the highest frequency to be recovered to obtain the low-frequency seismic data; performing a high-pass filtering on the original seismic data with a low cutoff frequency of the highest frequency to be recovered to obtain high-frequency seismic data; and combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

[0007] Furthermore, determining the highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data includes: determining from the original seismic data that data with amplitude energy less than a set energy is the low-frequency seismic data; and determining the highest frequency of the low-frequency seismic data as the highest frequency to be recovered.

[0008] Furthermore, the method of determining the reflection structure characterization operator using the original seismic data and the inversion algorithm includes: establishing a target functional of the reflection structure characterization operator using the seismic data and the reflection structure characterization operator; and using the conjugate gradient method to inversely solve the target functional of the reflection structure characterization operator to determine the reflection structure characterization operator.

[0009] Furthermore, the target functional of the reflection structure characterization operator is established in the following manner:

[0010] Wherein, e is the target functional of the reflection structure characterization operator, x is the surface position, t is the reflection time, i.e., different seismic traces, d(x, t) is the original seismic data, h(y, τ) is the reflection structure characterization operator, τ is the sample point of the reflection structure characterization operator in the time direction, y is the sample point of the operator in the direction of the seismic trace position, n x is the number of seismic traces, n t is the number of earthquake recording time samples, l x is the spatial length of the reflection structure characterization operator, l t The time length of the operator characterizing the reflection structure.

[0011] Furthermore, the reflection coefficient and the reflection structure characterization operator are used to construct a regularization constraint term, and a target functional of the reflection coefficient is established through the regularization constraint term, including: using a convolution model to establish an inversion target functional; using the reflection coefficient to establish a reflection coefficient sparse structure function as a first regularization constraint term of the inversion target functional; using the reflection structure characterization operator to establish a function of the reflection structure as a second regularization constraint term of the inversion target functional; and using the inversion target functional, the first regularization constraint term and the second regularization constraint term to establish the target functional of the reflection coefficient.

[0012] Furthermore, the inversion target functional is established in the following way:

[0013] Wherein, e(r) is the inversion target functional, w(τ) is the seismic wavelet, τ is the sample point of the seismic wavelet in the time direction, n x is the number of seismic traces, n t is the number of time samples of earthquake records, and r(x, t) is the reflection coefficient.

[0014] Furthermore, the seismic wavelet is obtained by: using a spectrum simulation method to fit the amplitude spectrum of the original seismic data to obtain a seismic wavelet amplitude spectrum; and performing Fourier transform on the seismic wavelet amplitude spectrum to obtain the seismic wavelet.

[0015] Furthermore, the reflection coefficient sparse structure function is established in the following way:

[0016] Wherein, φ1(r) is the reflection coefficient sparse structure function.

[0017] Furthermore, a function for reflecting the structure is established in the following way:

[0018] Among them, φ2(r) is a function of the reflection structure, H is a matrix composed of reflection structure characterization operators, F is a matrix composed of low-pass filter operators with a high cutoff frequency of f0, and r is a matrix composed of the reflection coefficients.

[0019] Furthermore, the target functional of the reflection coefficient is established in the following manner: α(r)=e(r)+λ1φ1(r)+λ2φ2(r);

[0020] Wherein, α(r) is the target functional of the reflection coefficient, λ1 is the regularization factor of the first regularization constraint term, and λ2 is the regularization factor of the second regularization constraint term.

[0021] Furthermore, the reversely solving the target functional of the reflection coefficient to obtain the reflection coefficient includes: using an iterative weighted least squares method to reversely solve the target functional of the reflection coefficient to obtain the reflection coefficient.

[0022] Furthermore, the low-frequency seismic data and the high-frequency seismic data are combined to obtain the seismic data after low-frequency recovery, including: weighted combination of the low-frequency seismic data and the high-frequency seismic data in the frequency domain to obtain a combination result; and transforming the combination result to the time domain to obtain the seismic data after low-frequency recovery.

[0023] According to a second aspect of the present invention, there is provided a low-frequency recovery device for seismic data constrained by a reflection structure, the low-frequency recovery device for seismic data constrained by a reflection structure comprising: a module for determining a maximum frequency to be recovered, for acquiring original seismic data, and determining the maximum frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data; a module for determining a reflection structure characterization operator, for determining a reflection structure characterization operator using the original seismic data and an inversion algorithm; a module for determining a target functional of a reflection coefficient, for constructing a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establishing a target functional of the reflection coefficient through the regularization constraint term; a module for determining a reflection coefficient, for inversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; a module for determining a low-frequency seismic data, for performing a low-frequency filtering on the reflection coefficient with a high cutoff frequency of the maximum frequency to be recovered to obtain the low-frequency seismic data; a module for determining a high-frequency seismic data, for performing a high-pass filtering on the original seismic data with a low cutoff frequency of the maximum frequency to be recovered to obtain the high-frequency seismic data; and a combining module, for combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

[0024] The technical solution provided by the present invention has at least the following technical effects:

[0025] The method for low-frequency recovery of seismic data constrained by reflection structure of the present invention first obtains the original seismic data, determines the highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data, determines the reflection structure characterization operator using the original seismic data and the inversion algorithm, then constructs a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establishes a target functional of the reflection coefficient through the regularization constraint term, and inversely solves the target functional of the reflection coefficient to obtain the reflection coefficient. The reflection coefficient is subjected to low-frequency filtering with a high cutoff frequency of the highest frequency to be recovered to obtain low-frequency seismic data, and the original seismic data is subjected to high-pass filtering with a low cutoff frequency of the highest frequency to be recovered to obtain high-frequency seismic data. The low-frequency seismic data and the high-frequency seismic data are then combined to obtain seismic data after low-frequency recovery. The method for low-frequency recovery of seismic data constrained by reflection structure provided by the present invention can ensure the reliability of low-frequency seismic data, enhance the resolution of seismic data, and improve the accuracy of full waveform velocity inversion.

[0026] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0028] FIG1 is a flow chart of a method for low-frequency recovery of seismic data constrained by reflection structure according to an embodiment of the present invention;

[0029] FIG2 is a schematic diagram of seismic model data synthesized using a complex model in a low-frequency recovery method for seismic data constrained by reflection structure provided by an embodiment of the present invention;

[0030] 3 is a schematic diagram of low-frequency model data after high-cutoff low-pass filtering of seismic model data in the reflection structure-constrained seismic data low-frequency recovery method provided by an embodiment of the present invention;

[0031] 4 is a schematic diagram of high-frequency model data after low-cut high-pass filtering of seismic model data in the reflection structure-constrained seismic data low-frequency recovery method provided by an embodiment of the present invention;

[0032] 5 is a schematic diagram of low-frequency seismic data after seismic model data is recovered in the low-frequency seismic data recovery method with reflection structure constraints provided by an embodiment of the present invention;

[0033] 6 is a schematic diagram of seismic data after low-frequency recovery of seismic model data in the method for low-frequency recovery of seismic data with reflection structure constraints provided by an embodiment of the present invention;

[0034] FIG7 is a schematic diagram of actual original seismic data of a certain oil field in the low-frequency recovery method of seismic data constrained by reflection structure provided by an embodiment of the present invention;

[0035] FIG8 shows low-frequency seismic data to be recovered below 10 Hz of the actual original seismic data in the low-frequency seismic data recovery method with reflection structure constraints provided by an embodiment of the present invention;

[0036] FIG9 shows low-frequency seismic data recovered from actual original seismic data in the low-frequency seismic data recovery method with reflection structure constraints provided by an embodiment of the present invention;

[0037] FIG10 is seismic data obtained after low-frequency recovery of actual original seismic data in the seismic data low-frequency recovery method with reflection structure constraints provided by an embodiment of the present invention;

[0038] FIG11 is a schematic diagram of a low-frequency recovery device for seismic data constrained by reflection structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following describes the specific implementation of the embodiment of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present invention and is not used to limit the embodiment of the present invention.

[0040] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0041] In the present invention, unless otherwise specified, directional words such as "up, down, top, bottom" are usually used to describe the relative positions of components in the directions shown in the drawings or in the vertical, perpendicular or gravity directions.

[0042] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0043] Please refer to Figure 1. An embodiment of the present invention provides a low-frequency recovery method for seismic data constrained by reflection structure, which includes the following steps: S101: acquiring original seismic data, and determining the highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data; S102: determining a reflection structure characterization operator using the original seismic data and an inversion algorithm; S103: constructing a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establishing a target functional of the reflection coefficient through the regularization constraint term; S104: inversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; S105: performing a low-frequency filtering on the reflection coefficient with a high cutoff frequency of the highest frequency to be recovered to obtain the low-frequency seismic data; S106: performing a high-pass filtering on the original seismic data with a low cutoff frequency of the highest frequency to be recovered to obtain high-frequency seismic data; S107: combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

[0044] Specifically, in an embodiment of the present invention, the original seismic data is first acquired. The highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered is determined based on the original seismic data. The reflection structure characterization operator is determined using the original seismic data and an inversion algorithm. The reflection coefficient and the reflection structure characterization operator are then used to construct a regularization constraint term. The target functional of the reflection coefficient is established through the regularization constraint term. The target functional of the reflection coefficient is then solved in reverse to obtain the reflection coefficient. The reflection coefficient is then subjected to a low-frequency filter with a high cutoff frequency of the highest frequency to be recovered to obtain low-frequency seismic data. The original seismic data is then subjected to a high-pass filter with a low cutoff frequency of the highest frequency to be recovered to obtain high-frequency seismic data. The low-frequency and high-frequency seismic data are then combined to obtain seismic data after low-frequency recovery.

[0045] The low-frequency recovery method for seismic data constrained by reflection structure provided by the present invention can ensure the reliability of low-frequency seismic data, enhance the resolution of seismic data, and improve the accuracy of full-waveform velocity inversion.

[0046] First, step S101 is performed: original seismic data is acquired, and the highest frequency to be restored corresponding to the low-frequency seismic data to be restored is determined according to the original seismic data.

[0047] Furthermore, determining the highest frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data includes: determining from the original seismic data that data with amplitude energy less than a set energy is the low-frequency seismic data; and determining the highest frequency of the low-frequency seismic data as the highest frequency to be recovered.

[0048] Specifically, in an embodiment of the present invention, raw seismic data d(x, t) is first acquired, where t is the reflection time and x is the surface position, i.e., different seismic traces. The raw seismic data d(x, t) is the entire profile of the post-stack two-dimensional data, missing low-frequency seismic data. Spectral analysis is performed on the raw seismic data d(x, t), and a segment of data with an amplitude energy (amplitude value) less than a set energy is identified as the low-frequency seismic data to be recovered. The highest frequency of the low-frequency seismic data is determined to be the highest frequency f0 to be recovered.

[0049] Then, step S102 is performed: using the original seismic data and the inversion algorithm, a reflection structure characterization operator is determined.

[0050] Furthermore, the method of determining the reflection structure characterization operator using the original seismic data and the inversion algorithm includes: establishing a target functional of the reflection structure characterization operator using the seismic data and the reflection structure characterization operator; and using the conjugate gradient method to inversely solve the target functional of the reflection structure characterization operator to determine the reflection structure characterization operator.

[0051] Specifically, in the implementation manner of the present invention, the effective data in the actual seismic data is spatially correlated, that is, the seismic data profile has a certain predictive relationship in both the time direction and the space direction. The reflection structure is the geometric characteristics and energy change characteristics of the seismic reflection axis in space, which represents the continuity and predictability of the seismic data in space. The reflection structure characterization operator h(x, t) is a space-time operator that describes the spatial relationship and energy relationship of the reflection structure. It can be considered that any sample point in the seismic profile can be obtained by convolving the surrounding sample points with the reflection structure operator, that is, the relationship between the sample points on the seismic data profile can be linearly described by the operator. Therefore, based on the spatial coherence of the seismic data, the reflection structure characterization operator h(x, t) can be calculated using the original seismic data d(x, t):

[0052] First, the target functional of the reflection structure characterization operator is established using the seismic data and the reflection structure characterization operator:

[0053] Where, e is the target functional of the reflection structure characterization operator, d(x, t) is the original seismic data, h(x, t) is the reflection structure characterization operator, and n x is the number of seismic traces, n t is the number of earthquake recording time samples, l x is the spatial length of the reflection structure characterization operator, l t is the time length of the reflection structure characterization operator. The meaning of this objective functional expression is that the best characterization operator obtained by solving can make any data point on the seismic data profile and its adjacent l x *l tThe error energy e is minimized between the data points and the data values ​​predicted after convolution of the representation operator.

[0054] Then, the conjugate gradient method is used to reversely solve the target functional of the reflection structure characterization operator to determine the reflection structure characterization operator h(x,t).

[0055] Then, step S103 is performed: a regularization constraint term is constructed using the reflection coefficient and the reflection structure characterization operator, and a target functional of the reflection coefficient is established through the regularization constraint term.

[0056] Furthermore, the reflection coefficient and the reflection structure characterization operator are used to construct a regularization constraint term, and a target functional of the reflection coefficient is established through the regularization constraint term, including: using a convolution model to establish an inversion target functional; using the reflection coefficient to establish a reflection coefficient sparse structure function as a first regularization constraint term of the inversion target functional; using the reflection structure characterization operator to establish a function of the reflection structure as a second regularization constraint term of the inversion target functional; and using the inversion target functional, the first regularization constraint term and the second regularization constraint term to establish the target functional of the reflection coefficient.

[0057] Specifically, in the embodiment of the present invention, the convolution model d=w*r is used to establish the inversion target functional.

[0058] Among them, e(r) is the inversion target functional, w(τ) is the seismic wavelet, n x is the number of seismic traces, n t is the number of time samples of earthquake records, and r(x,t) is the reflection coefficient.

[0059] Furthermore, the seismic wavelet is obtained by: using a spectral simulation method to fit the amplitude spectrum of the original seismic data to obtain the seismic wavelet amplitude spectrum; performing Fourier transform on the seismic wavelet amplitude spectrum to obtain the seismic wavelet w(τ).

[0060] Furthermore, based on the probability distribution of the reflection coefficient and the assumption that the reflection coefficient is sparse, a sparse structure function of the reflection coefficient is established:

[0061] Among them, φ1(r) is the reflection coefficient sparse structure function.

[0062] Furthermore, based on the fact that the reflection coefficient has the same predictability as seismic records, a function of the reflection structure is established:

[0063] Where φ2(r) is a function of the reflection structure.

[0064] H is a matrix composed of reflection structure characterization operators:

[0065] Among them, h n (x, t) is the reflection structure characterization operator obtained for the nth seismic trace.

[0066] F is a matrix composed of low-pass filter operators with a high cutoff frequency of f0, representing the process of low-pass filtering of the reflection coefficient. This method uses the sinc function to construct a low-pass filter, then:

[0067] Where sinc(x) is the filter operator value used.

[0068] r is the matrix composed of reflection coefficients:

[0069] Among them, r n (x, t) is the reflection coefficient to be solved for the nth seismic trace.

[0070] Furthermore, the target functional of the reflection coefficient is established in the following manner: α(r)=e(r)+λ1φ1(r)+λ2φ2(r);

[0071] Where α(r) is the target functional of the reflection coefficient, λ1 is the regularization factor for the first regularization constraint, and λ2 is the regularization factor for the second regularization constraint. The regularization factor can be determined based on the actual situation to adjust the impact of the regularization constraint on the results. A larger λ1 results in a more pronounced temporal sparsity of the reflection coefficient, while a larger λ2 results in a stronger spatial continuity of the reflection structure.

[0072] Then, step S104 is performed: inversely solving the target functional of the reflection coefficient to obtain the reflection coefficient.

[0073] Furthermore, the target functional of the reflection coefficient is reversely solved to obtain the reflection coefficient, including: using an iterative reweighted least squares (IRLS) method to reversely solve the target functional α(r) of the reflection coefficient to obtain the reflection coefficient r(x, t).

[0074] Then, step S105 is executed: performing low-frequency filtering with a high cutoff frequency of the highest frequency to be recovered f0 on the reflection coefficient r(x, t) to obtain low-frequency seismic data s(x, t).

[0075] Then, step S106 is executed: the original seismic data r(x, t) is subjected to high-pass filtering with a low cutoff frequency of the highest frequency to be recovered f0 to obtain high-frequency seismic data.

[0076] Finally, step S107 is performed: combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

[0077] Furthermore, the low-frequency seismic data and the high-frequency seismic data are combined to obtain the seismic data after low-frequency recovery, including: weighted combination of the low-frequency seismic data and the high-frequency seismic data in the frequency domain to obtain a combination result; and transforming the combination result to the time domain to obtain the seismic data after low-frequency recovery.

[0078] Specifically, in this embodiment of the present invention, the low-frequency seismic data s(x, t) and the high-frequency seismic data are transformed into the frequency domain, and the spectra of the two data are weighted and combined within the transition band to obtain a combined result. The combined result is then transformed into the time domain to form a seismic record after the low-frequency seismic data is recovered.

[0079] Example 1

[0080] To verify the effectiveness of this embodiment, the earthquake model data shown in Figure 2 was synthesized using a complex model. The earthquake model data of Figure 2 was low-pass filtered with a high cutoff frequency f0 = 10 Hz to obtain the low-frequency model data shown in Figure 3. The earthquake model data of Figure 2 was high-pass filtered with a low cutoff frequency f0 = 10 Hz to obtain the high-frequency model data shown in Figure 4. Figure 4 shows the high-frequency model data below 10 Hz that has been lost. Figure 5 shows the low-frequency earthquake data recovered from Figure 4 using the method of this embodiment. The low-frequency earthquake data is highly consistent with the low-frequency model data in Figure 3. Figure 6 shows the earthquake data after superimposing the low-frequency earthquake data recovered in Figure 5 with the high-frequency model data in Figure 4, effectively recovering the earthquake model data shown in Figure 2.

[0081] Example 2

[0082] The actual raw seismic data d(x, t) from an oilfield, as shown in Figure 7, was obtained. Spectral analysis of the raw seismic data d(x, t), as shown in Figure 8, determined that the data had weak energy below 10 Hz and required recovery of low-frequency seismic data below 10 Hz. The highest frequency f0 to be recovered for the low-frequency seismic data was 10 Hz. Using spectral simulation, the amplitude spectrum of the raw seismic data was fitted to obtain the seismic wavelet amplitude spectrum. A Fourier transform of the seismic wavelet amplitude spectrum was then performed to obtain the seismic wavelet w(t).

[0083] By solving a least-squares problem, the original seismic data and an inversion algorithm are used to determine the reflection structure characterization operator h(x,t). This operator represents the spatial continuity and predictability of the seismic data, minimizing the error energy between the original data point and the predicted data obtained by multiplying the characterization operator by multiple adjacent data points. A convolution model is used to establish a target functional for the reflection coefficient. Seismic data can be viewed as the convolution of a reflection coefficient sequence and a seismic wavelet. Determining the reflection coefficient sequence from the seismic data is the corresponding inverse problem, achieved by solving the target functional for the reflection coefficient. Inversion problems often have multiple solutions. To obtain more accurate results, the target functional is solved based on regularization theory. Regularization theory involves adding a constraint term to the inversion objective function to reduce singularities in the solution. Regularization constraints are then constructed separately: based on the probability distribution of the reflection coefficient and the assumption that the reflection coefficient is sparse, a reflection coefficient sparse structure function is established as the first regularization constraint, constraining the temporal position of the reflection events in the result; based on the fact that the reflection coefficient has the same predictability as the seismic record, a reflection structure function is established as the second regularization constraint, constraining the spatial continuity of the reflection events in the result. The two constraints are multiplied by their corresponding coefficients, which can be adjusted by the user. The larger the coefficient, the more significant the impact of the corresponding constraint on the result. In this embodiment, λ1 and λ2 are both 0.25, ensuring both the sparsity of the reflection coefficient and the lateral continuity of the reflection structure. The IRLS algorithm is then used to solve the target functional α(r) of the reflection coefficient with the regularization constraint added to obtain the reflection coefficient r(x, t). The low-frequency seismic data of the reflection coefficient is identical to the low-frequency component of the seismic data. A low-frequency filter with a high cutoff frequency of 10 Hz and a filter transition band of 7 Hz to 13 Hz is used to separate the low-frequency seismic data below 10 Hz, resulting in the low-frequency seismic data shown in Figure 9. Next, a high-pass filter with a low cutoff frequency of 10 Hz and a transition band of 7 Hz to 13 Hz is used to filter out low-frequency data below 10 Hz from the original seismic data in Figure 6, obtaining high-frequency seismic data. The low-frequency seismic data of the reflection coefficient and the high-frequency seismic data from which the low-frequency seismic data was filtered are then spectrally combined in the frequency domain. The spectra of the two data are weightedly combined within the transition band, i.e., the spectrum amplitude values ​​are weighted and added. For the low-frequency seismic data of the reflection coefficient, the weight is 1 at 7 Hz and 0 at 13 Hz, and the weight decreases linearly within the transition band. For the high-frequency seismic data of the original seismic data, the weight is 0 at 7 Hz and 1 at 13 Hz, and the weight increases linearly within the transition band. The combined spectrum is transformed into the time domain, and the low-frequency recovered seismic data shown in Figure 10 is finally obtained, completing the low-frequency recovery process constrained by the reflection structure.

[0084] It can be seen that this embodiment effectively recovers the low-frequency seismic data of the actual original seismic data. The seismic data after the low-frequency seismic data recovery can more accurately reflect the seismic response of the actual underground structure and oil and gas reservoir.

[0085] 11 , a second aspect of the present invention provides a low-frequency recovery device for seismic data constrained by reflection structure, the device comprising: a module for determining a maximum frequency to be recovered, for acquiring original seismic data and determining the maximum frequency to be recovered corresponding to the low-frequency seismic data to be recovered based on the original seismic data; a module for determining a reflection structure characterization operator using the original seismic data and an inversion algorithm; a module for determining a target functional of a reflection coefficient, for constructing a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establishing a target functional of the reflection coefficient through the regularization constraint term; a module for determining a reflection coefficient, for inversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; a module for determining a low-frequency seismic data, for performing a low-pass filtering on the reflection coefficient with a high cutoff frequency of the maximum frequency to be recovered to obtain the low-frequency seismic data; a module for determining a high-frequency seismic data, for performing a high-pass filtering on the original seismic data with a low cutoff frequency of the maximum frequency to be recovered to obtain the high-frequency seismic data; and a combining module for combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

[0086] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various simple modifications can be made to the technical solution of the present invention, and these simple modifications all fall within the scope of protection of the present invention.

[0087] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

[0088] In addition, the various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A low-frequency recovery method for seismic data constrained by reflection structure, characterized in that: The reflection structure constrained seismic data low-frequency recovery method comprises: Acquire original seismic data, and determine the highest frequency to be restored corresponding to the low-frequency seismic data to be restored according to the original seismic data; Determining a reflection structure characterization operator using the original seismic data and an inversion algorithm; A regularization constraint term is constructed using the reflection coefficient and the reflection structure characterization operator, and a target functional of the reflection coefficient is established through the regularization constraint term; Reversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; Performing low-frequency filtering with a high cutoff frequency of the highest frequency to be recovered on the reflection coefficient to obtain the low-frequency seismic data; Performing high-pass filtering with a low cutoff frequency of the highest frequency to be recovered on the original seismic data to obtain high-frequency seismic data; The low-frequency seismic data and the high-frequency seismic data are combined to obtain seismic data after low-frequency recovery.

2. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 1, characterized in that: The step of determining the highest frequency to be restored corresponding to the low-frequency seismic data to be restored according to the original seismic data comprises: Determining from the original seismic data that the data with amplitude energy less than the set energy is the low-frequency seismic data; The highest frequency of the low-frequency seismic data is determined as the highest frequency to be restored.

3. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 1, characterized in that: The method of determining the reflection structure characterization operator by using the original seismic data and the inversion algorithm comprises: Using the seismic data and the reflection structure characterization operator, a target functional of the reflection structure characterization operator is established; The conjugate gradient method is used to reversely solve the target functional of the reflection structure characterization operator to determine the reflection structure characterization operator.

4. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 3, characterized in that: The target functional of the reflection structure characterization operator is established in the following way: Wherein, e is the target functional of the reflection structure characterization operator, x is the surface position, t is the reflection time, i.e., different seismic traces, d(x, t) is the original seismic data, h(x, t) is the reflection structure characterization operator, τ is the sample point of the reflection structure characterization operator in the time direction, y is the sample point of the operator in the direction of the seismic trace position, n x is the number of seismic traces, n t is the number of earthquake recording time samples, l x is the spatial length of the reflection structure characterization operator, l t The time length of the operator characterizing the reflection structure.

5. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 1, characterized in that: The method of constructing a regularization constraint term by using the reflection coefficient and the reflection structure characterization operator, and establishing a target functional of the reflection coefficient by using the regularization constraint term, includes: The inversion target functional is established using the convolution model; Using the reflection coefficient, a reflection coefficient sparse structure function is established as the first regularization constraint term of the inversion target functional; Using the reflection structure characterization operator to establish a reflection structure function as a second regularization constraint term of the inversion target functional; The target functional of the reflection coefficient is established by using the inversion target functional, the first regularization constraint term and the second regularization constraint term.

6. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 5, characterized in that: The inversion target functional is established in the following way: Wherein, e(r) is the inversion target functional, w(τ) is the seismic wavelet, τ is the sample point of the seismic wavelet in the time direction, n x is the number of seismic traces, n t is the number of time samples of earthquake records, r(x,t) is the inverse Emission coefficient.

7. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 6, characterized in that: The seismic wavelet is obtained by: Using a spectrum simulation method, a seismic wavelet amplitude spectrum is obtained by fitting the amplitude spectrum of the original seismic data; The seismic wavelet is obtained by performing Fourier transform on the seismic wavelet amplitude spectrum.

8. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 7, characterized in that: The sparse structure function of the reflection coefficient is established in the following way: Wherein, φ1(r) is the reflection coefficient sparse structure function.

9. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 8, characterized in that: The function that builds the reflection structure is done in the following way: Among them, φ2(r) is a function of the reflection structure, H is a matrix composed of reflection structure characterization operators, F is a matrix composed of low-pass filter operators with a high cutoff frequency of f0, and r is a matrix composed of the reflection coefficients.

10. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 9, characterized in that: The target functional of the reflection coefficient is established in the following way: α(r)=e(r)+λ1φ1(r)+λ2φ2(r); Wherein, α(r) is the target functional of the reflection coefficient, λ1 is the regularization factor of the first regularization constraint term, and λ2 is the regularization factor of the second regularization constraint term.

11. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 1, characterized in that: The reversely solving the target functional of the reflection coefficient to obtain the reflection coefficient includes: The iterative weighted least squares method is used to reversely solve the target functional of the reflection coefficient to obtain the reflection coefficient.

12. The method for low-frequency recovery of seismic data constrained by reflection structure according to claim 1, characterized in that: The combining the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery includes: Performing weighted combination of the low-frequency seismic data and the high-frequency seismic data in the frequency domain to obtain a combination result; The combined result is transformed into the time domain to obtain seismic data after low-frequency recovery.

13. A low-frequency recovery device for seismic data constrained by reflection structure, characterized in that: The reflection structure constrained seismic data low-frequency recovery device comprises: A module for determining the highest frequency to be restored, used to obtain original seismic data, and determine the highest frequency to be restored corresponding to the low-frequency seismic data to be restored according to the original seismic data; A reflection structure characterization operator determination module, used to determine a reflection structure characterization operator using the original seismic data and an inversion algorithm; A target functional determination module for the reflection coefficient, used to construct a regularization constraint term using the reflection coefficient and the reflection structure characterization operator, and establish a target functional of the reflection coefficient through the regularization constraint term; A reflection coefficient determination module, used for reversely solving the target functional of the reflection coefficient to obtain the reflection coefficient; A low-frequency seismic data determination module, used for performing low-frequency filtering with a high cutoff frequency of the highest frequency to be restored on the reflection coefficient to obtain the low-frequency seismic data; A high-frequency seismic data determination module, configured to perform high-pass filtering on the original seismic data with a low cutoff frequency being the highest frequency to be recovered, to obtain high-frequency seismic data; The combination module is used to combine the low-frequency seismic data and the high-frequency seismic data to obtain seismic data after low-frequency recovery.

Images