Method for improving reconstruction quality of low-signal-to-noise-ratio data and related device

Abstract

The invention discloses a method for improving the reconstruction quality of low-signal-to-noise-ratio data and a related device. The method comprises the following steps: constructing a regular observation system based on an irregular observation system, and constructing a three-dimensional array; determining a first to-be-reconstructed channel or a second to-be-reconstructed channel in the regular observation system; according to a preset reconstruction method, obtaining corresponding first reconstruction data, and storing the first reconstruction data in a three-dimensional array; iterating the first reconstruction data in the three-dimensional array by adopting a preset iteration strategy until a preset termination condition is met; and for the first to-be-reconstructed trace, replacing the corresponding seismic trace data after iteration reaches a preset termination condition with the first reconstruction data corresponding to the first to-be-reconstructed trace. The method provided by the embodiment of the invention has good reliability and practicability, the quality of the reconstructed data basically meets the subsequent data processing requirement, and the exploration production requirement can be well met.

Description

Technical Field

[0001] The present invention relates to the technical field of petroleum geophysical exploration processing, and in particular, to a method and related device for improving the reconstruction quality of low signal-to-noise ratio data. Background Art

[0002] Original seismic data often has limitations. First, it is usually affected by no-go areas, obstacles, and ocean towed cable plume drift; second, restricted by construction costs, the data collected is generally irregular or sparse in the spatial direction; furthermore, the rejection of waste shots and waste traces during the processing stage is an important factor causing data irregularity. However, subsequent processing such as free surface multiple elimination and wave equation migration requires a complete data volume, so seismic data reconstruction has become an important link in seismic data processing.

[0003] In the prior art, seismic data reconstruction based on three-dimensional curvelet transform is an effective method. In application, usually, all time sampling points of the regular receiving points to be reconstructed are directly initialized to 0, and then the regular receiving points are reconstructed. Inevitably, this will generate noise interference and bring defects to the reconstruction quality of seismic data. Summary of the Invention

[0004] In view of the above problems, the purpose of the present invention is to provide a method and related device for improving the reconstruction quality of low signal-to-noise ratio data.

[0005] In a first aspect, an embodiment of the present invention provides a method for improving the reconstruction quality of low signal-to-noise ratio data, including:

[0006] Construct a regular observation system W1 based on the irregular observation system W0, determine the parameters of the regular observation system W1, and construct a three-dimensional array based on the parameters of the regular observation system W1, where the three-dimensional array is used to store all seismic trace data in the regular observation system W1;

[0007] Compare the seismic traces in the regular observation system W1 with the corresponding seismic traces in the irregular observation system W0, and determine whether each seismic trace in the regular observation system W1 is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison rule;

[0008] Reconstruct the seismic trace data corresponding to each first trace to be reconstructed and second trace to be reconstructed in the regular observation system W1 according to the seismic trace data of each seismic trace in the irregular observation system W0 and a preset reconstruction method to obtain corresponding first reconstruction data, and store the first reconstruction data of each first trace to be reconstructed and second trace to be reconstructed in the three-dimensional array;

[0009] Adopt a preset iteration strategy to iterate the first reconstruction data in the three-dimensional array until a preset termination condition is reached;

[0010] For the first trace to be reconstructed, the seismic trace data corresponding to the three-dimensional array after the iteration reaches a preset termination condition is replaced by the first reconstructed data corresponding to the first trace to be reconstructed.

[0011] In one embodiment, a preset iteration strategy is used to iterate the first reconstructed data in the three-dimensional array until a preset termination condition is reached, including:

[0012] Performing a three-dimensional curvelet transform on the three-dimensional array of the initial assignment to obtain preprocessed curvelet coefficients;

[0013] Performing a noise removal operation on the preprocessed curvelet coefficients; the noise is a curvelet coefficient whose curvelet coefficient is less than a preset noise threshold;

[0014] Performing a three-dimensional inverse curvelet transform on the de-noised curvelet coefficients to obtain a first iteration value of the seismic trace reconstruction data;

[0015] The first iteration value of the seismic trace reconstruction data is iterated again until the number of iterations reaches a preset total number of iterations or a threshold parameter is greater than a preset threshold parameter, and then the iteration is stopped.

[0016] In one embodiment, the preprocessed curvelet coefficients are obtained by the following method:

[0017] The curvelet coefficients obtained after performing three-dimensional curvelet transformation on the three-dimensional array with the initial assignment are sorted in order from small to large, and abnormal values ​​are eliminated by median filtering to obtain the pre-processed curvelet coefficients.

[0018] In one embodiment, performing a noise removal operation on the preprocessed curvelet coefficients includes:

[0019] Determine the average curve coefficient and the curve coefficient ratio formula according to the pre-processed curve coefficient;

[0020] According to the curvelet coefficient ratio of each sample point, a curvelet coefficient ratio curve is obtained, the curvelet coefficient ratio curve is normalized, and the sample point (curvelet coefficient) corresponding to the maximum value of the curvelet coefficient ratio is determined as the noise threshold in the first iteration;

[0021] Establishing a noise threshold function for each iteration according to the average curvelet coefficient and the noise threshold for the first iteration, wherein the noise threshold function is used to determine the noise threshold corresponding to each iteration;

[0022] According to the noise threshold function, the noise smaller than the noise threshold in each iteration process is set to zero.

[0023] In one embodiment, the threshold parameter is the ratio of the root mean square amplitude of all traces in the second set to be reconstructed after each iteration to the root mean square amplitude of all traces in the first set to be reconstructed.

[0024] In one embodiment, the parameters of the regular acquisition system include: receiver line spacing L, number of receiver lines S, regular trace spacing R, number of seismic traces M in each receiver line, and number of sampling points N in each seismic trace.

[0025] Constructing a three-dimensional array based on the parameters of the regular acquisition system includes:

[0026] Any element in the three-dimensional array is A[s][m][n], where s is the s-th receiver line, 1 ≤ s ≤ S, m is the m-th receiving point on the s-th receiver line, 1 ≤ m ≤ M, and n is the n-th sampling point of the seismic trace at the m-th receiving point on the s-th receiver line, 1 ≤ n ≤ N.

[0027] In one embodiment, determining whether each seismic trace in the regular acquisition system is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison method includes:

[0028] According to the actual position coordinates of the receiving points, determine the seismic trace in the irregular acquisition system W0 that is closest to each seismic trace in the regular acquisition system W1, calculate the distance between the two, and if the distance between the two is less than or equal to the preset distance value, mark the corresponding seismic trace in the regular acquisition system W1 as the first trace to be reconstructed, otherwise mark the corresponding seismic trace in the regular acquisition system W1 as the second trace to be reconstructed.

[0029] In one embodiment, the preset distance value r0 = ε·R, where ε is the boundary range coefficient and R is the regular trace distance.

[0030] In one embodiment, reconstructing the seismic trace data corresponding to each first trace to be reconstructed or second trace to be reconstructed in the regular acquisition system W1 according to the seismic trace data of the irregular acquisition system W0 and a preset reconstruction method to obtain the corresponding first reconstruction data includes:

[0031] Draw a circular region with the coordinate position of each seismic trace in the regular acquisition system W1 as the center and r as the preset radius, sort the seismic traces in the irregular acquisition system W0 included in the circular region in ascending order of the distance from the center to obtain the sorted seismic data S n (t), where n is the number of seismic traces in the irregular acquisition system W0 included in each different circular region, 1 ≤ n ≤ N0; N0 is the maximum number of seismic traces in the irregular acquisition system W0 included in each circular region.

[0032] Transform the seismic data S n (t) into the frequency domain FS n (t,f), and calculate the amplitude spectrum Amp n (t,f) and the phase spectrum phase n (t,f) of each seismic trace;

[0033] Take the logarithm of the amplitude spectrum Amp n (t,f) of each seismic trace to obtain the logarithmic amplitude spectrum LF n (t,f), and subtract the minimum value of the logarithmic amplitude spectrum LF n (t,f) of each seismic trace from all the spectral values of the corresponding trace to obtain the positive logarithmic amplitude spectrum LFP n (t,f);

[0034] Calculate the amplitude spectrum Amp0(t,f) and the positive logarithmic amplitude spectrum LFP0(t,f) of the trace to be reconstructed according to the amplitude spectrum corresponding to each frequency of each seismic trace included in the irregular observation system W0 within the circular domain;

[0035] Normalize the positive logarithmic amplitude spectrum LFP0(t,f) to obtain the normalized positive logarithmic amplitude spectrum LFPE0(t,f);

[0036] According to the phase spectrum phase1(t,f) when n = 1 and the positive logarithmic amplitude spectrum LFPE0(t,f), determine the reconstructed data LFPER0(t,f) of each seismic trace in the regular observation system W1 in the frequency domain FSn ( t,f);

[0037] Inverse-transform the reconstructed data LFPER0(t,f) to the time domain to obtain the first reconstructed data.

[0038] In one embodiment, the preset radius r is determined by the following method:

[0039] In the regular observation system, the preset radius r of each of the first traces to be reconstructed is ε·R;

[0040] In the regular observation system, the preset radius r of each of the second traces to be reconstructed satisfies being greater than ε·R and less than or equal to ε·βR, where β is a real number greater than 1 and is empirically valued; where ε is a boundary range coefficient and R is the regular trace distance.

[0041] Second, the embodiment of the present invention provides a device for improving the reconstruction quality of low signal-to-noise ratio data, including:

[0042] A three-dimensional array construction module, configured to construct a regular acquisition system W1 based on an irregular acquisition system W0, determine parameters of the regular acquisition system W1, and construct a three-dimensional array based on the parameters of the regular acquisition system W1, where the three-dimensional array is used to store all seismic trace data in the regular acquisition system W1;

[0043] A to-be-reconstructed trace classification module, configured to compare seismic traces in the regular acquisition system W1 with corresponding seismic traces in the irregular acquisition system W0, and determine, according to a preset comparison rule, whether each seismic trace in the regular acquisition system W1 is a first to-be-reconstructed trace or a second to-be-reconstructed trace;

[0044] A first reconstruction module, configured to reconstruct seismic trace data corresponding to each first to-be-reconstructed trace and second to-be-reconstructed trace in the regular acquisition system W1 according to seismic trace data of each seismic trace in the irregular acquisition system W0 and a preset reconstruction method, to obtain corresponding first reconstruction data, and store the first reconstruction data of each first to-be-reconstructed trace and second to-be-reconstructed trace in the three-dimensional array;

[0045] An iteration module, configured to perform iteration on the first reconstruction data in the three-dimensional array by using a preset iteration strategy until a preset termination condition is met; for the first to-be-reconstructed trace, replace seismic trace data corresponding thereto in the three-dimensional array after the iteration reaches the preset termination condition with the first reconstruction data corresponding to the first to-be-reconstructed trace.

[0046] In a third aspect, an embodiment of the present invention provides a computing device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, where when the processor executes the program, the method for improving the reconstruction quality of low signal-to-noise ratio data described above is implemented.

[0047] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for improving the reconstruction quality of low signal-to-noise ratio data described above is implemented.

[0048] The beneficial effects of the above technical solutions provided by the embodiments of the present invention at least include:

[0049] The method and related device for improving the reconstruction quality of low signal-to-noise ratio data provided by the embodiments of the present invention construct a regular observation system according to an irregular observation system, and reconstruct the seismic trace data of all seismic traces (including the first trace to be reconstructed and the second trace to be reconstructed) in the regular observation system by using the seismic trace data of each seismic trace in the irregular observation system according to a preset reconstruction method and a preset iteration strategy. The classification of the first trace to be reconstructed and the second trace to be reconstructed can be divided, for example, according to the sparsity and discreteness of each seismic trace in the non-observed observation system. For the first trace to be reconstructed, the corresponding first reconstructed data is the final reconstructed data. For the second trace to be reconstructed, after obtaining the corresponding first reconstructed data, iteration is still required, and the reconstructed data until the preset termination condition is reached is used as the final reconstructed data. After multiple iterative reconstructions, the purpose of reducing the noise of the reconstructed data can be achieved. The method provided by the embodiments of the present invention has good reliability and practicability, and the quality of the reconstructed data basically meets the requirements of subsequent data processing and can well meet the requirements of exploration production.

[0050] Further, by using the irregular traces to assign values to each trace of the new regular observation system, the generation of random noise during the subsequent 3D curvelet transform for reconstructing traces can be reduced; a threshold function is constructed using the demarcation point between the noise and the curvelet coefficients of the effective wave for iterative calculation, which protects the weak effective signals to a certain extent; the preset termination condition is established: when the number of iterations reaches the preset total number of iterations or the threshold parameter is greater than the predetermined threshold parameter, the iteration is stopped, improving the quality of the reconstruction of low signal-to-noise ratio data. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0052] Figure 1 is a flowchart of the method for improving the reconstruction quality of low signal-to-noise ratio data in the embodiments of the present invention;

[0053] Figure 2 is a schematic diagram of the irregular observation system W0 constructed in the embodiments of the present invention;

[0054] Figure 3 is a schematic diagram of the regular observation system W1 constructed in the embodiments of the present invention;

[0055] Figure 4 is a schematic diagram of the result before the reconstruction of a certain shot gather data in the embodiments of the present invention;

[0056] Figure 5 is a schematic diagram of the result after the reconstruction of a certain shot gather data in the embodiments of the present invention;

[0057] Figure 6Schematic diagram of the device for improving the reconstruction quality of low signal-to-noise ratio data in the embodiments of the present invention. Detailed implementation manners

[0058] This embodiment provides a method and related device for improving the reconstruction quality of low signal-to-noise ratio data. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0059] The embodiments of the present invention provide a method for improving the reconstruction quality of low signal-to-noise ratio data. Referring to Figure 1 as shown, the method includes the following steps:

[0060] S11. Construct a regular observation system W1 based on the irregular observation system W0, determine the parameters of the regular observation system W1, and construct a three-dimensional array based on the parameters of the regular observation system W1. The three-dimensional array is used to store all seismic trace data in the regular observation system W1;

[0061] S12. Compare the seismic traces in the regular observation system W1 with the corresponding seismic traces in the irregular observation system W0, and determine whether each seismic trace in the regular observation system W1 is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison rule;

[0062] S13. Reconstruct the seismic trace data corresponding to each first trace to be reconstructed and second trace to be reconstructed in the regular observation system W1 according to the seismic trace data of each seismic trace in the irregular observation system W0 and a preset reconstruction method to obtain corresponding first reconstructed data, and store the first reconstructed data of each first trace to be reconstructed and second trace to be reconstructed in the three-dimensional array;

[0063] S14. Adopt a preset iterative strategy to iterate the first reconstructed data in the three-dimensional array until a preset termination condition is reached;

[0064] S15. For the first trace to be reconstructed, replace the corresponding seismic trace data in the three-dimensional array after the iteration reaches the preset termination condition with the first reconstructed data corresponding to the first trace to be reconstructed.

[0065] In step S11, according to the geological task, seismic data is excited and collected to construct an irregular observation system for the actual working condition. In the irregular observation system: the receiver line spacing is L, the number of receiver lines is S, and the number of sampling points of each seismic trace is N.

[0066] Based on the constructed non-regular observation system, a regular observation system is constructed. In the regular observation system: the receiver line spacing is L, the number of receiver lines is S, the regular trace spacing is R, there are M seismic traces on each receiver line, and the number of sampling points for each seismic trace is N.

[0067] According to the constructed regular observation system and related parameters, a three-dimensional array is constructed. Any element of this three-dimensional array is A[s][m][n], where s is the s-th receiver line, 1 ≤ s ≤ S, m is the m-th receiving point on the s-th receiver line, 1 ≤ m ≤ M, and n is the n-th sampling point of the seismic trace at the m-th receiving point on the s-th receiver line, 1 ≤ n ≤ N. This three-dimensional array is used to store the data of each seismic trace in the regular observation system. Since the data in all seismic traces in the regular observation system is empty at this time, the corresponding three-dimensional array does not contain any elements at this time.

[0068] In step S12, based on the actual position coordinates of the receiving points, the non-regular observation system and the regular observation system are superimposed, and the corresponding seismic trace in the non-regular observation system that is closest to each seismic trace in the regular observation system is determined. Calculate the distance between the two. If the distance between the two is less than or equal to the preset distance value r0(ε·R), then mark this seismic trace in the regular observation system as the first seismic trace to be reconstructed. Otherwise, mark this seismic trace in the regular observation system as the second seismic trace to be reconstructed. For the convenience of managing and analyzing the two types of seismic traces in the regular observation system later, all the first seismic traces to be reconstructed can be stored in the first set, and all the second seismic traces to be reconstructed can be stored in the second set.

[0069] When classifying the seismic traces in the regular observation system, the following method can be used. For example: taking a certain seismic trace in the regular observation system as a reference, calculate the distances between all seismic traces in the non-regular observation system and a certain seismic trace in the regular observation system in turn, and determine the seismic trace in the non-regular observation system that is the shortest distance from a certain seismic trace in the regular observation system as the closest seismic trace. If the distance between a certain seismic trace in the regular observation system and the closest seismic trace in the non-observation system is less than or equal to r0(ε·R), then it is determined as the first seismic trace to be reconstructed, otherwise it is determined as the second seismic trace to be reconstructed.

[0070] For the preset distance r0(ε·R), where ε is the boundary range coefficient and R is the regular trace distance, the smaller the value of the boundary range coefficient ε, the fewer the number of the first seismic traces to be reconstructed in the first set of the regular observation system. Its value can be taken according to experience. Preferably, ε = 0.5, which is not limited in the embodiments of the present invention.

[0071] It can be seen from the above that the basis for dividing the first seismic trace to be reconstructed or the second seismic trace to be reconstructed in the regular observation system is mainly determined by the distribution dispersion degree among the seismic traces in the non-regular observation system.

[0072] In step S13, according to all the seismic trace data in the irregular observation system, using a preset reconstruction method, all the seismic traces in the regular observation system are reconstructed, including the first reconstructed data corresponding to the first trace to be reconstructed and the second trace to be reconstructed, and all the first reconstructed data obtained are stored in a three-dimensional array. The first reconstructed data can be obtained by the following steps. For example:

[0073] (1) The reconstructed data of the seismic traces in the regular observation system are reconstructed from the seismic trace data in the irregular observation system according to a certain interpolation method. Therefore, it is necessary to determine the seismic trace data in the irregular observation system that each seismic trace in the regular observation system relies on during reconstruction respectively.

[0074] In the regular observation system, with the coordinates of each seismic trace as the center, circles corresponding to each seismic trace are drawn with different preset radii r corresponding to the first trace to be reconstructed and the second trace to be reconstructed, and all the seismic trace data S n (t) in the irregular observation system contained in each circular area are obtained. Obviously, n is the number of seismic traces in the irregular observation system contained in each different circular area, and 1 ≤ n ≤ N0; N0 is the maximum number of seismic traces in the irregular observation system contained in each circular area. For the first trace to be reconstructed, its preset radius r = ε·R. For the second trace to be reconstructed, its preset radius r satisfies being greater than ε·R and less than or equal to ε·βR, where β is a real number greater than 1 and is determined by experience. Preferably, β = 6. Where ε is the boundary range coefficient and R is the regular trace distance. The determined seismic trace data S n (t) are the seismic trace data that each seismic trace in the regular observation system relies on during interpolation reconstruction respectively.

[0075] (2) Transform the above-determined seismic trace data S n (t) to the frequency domain FS n (t,f) = FFT[S n (t)], and calculate the amplitude spectrum Amp n (t,f) and the phase spectrum phase n (t,f) of each seismic trace in each circular area.

[0076] The amplitude spectrum Amp n (t,f) and the phase spectrum phase n (t,f) are calculated respectively using the following formulas:

[0077]

[0078]

[0079] (3) For the amplitude spectrum Amp n(t, f) is logarithmically transformed to obtain the logarithmic amplitude spectrum LF n (t, f) = ln(Amp n (t, f)). The positive logarithmic amplitude spectrum LFP is obtained by subtracting the minimum value from all spectral values of each corresponding trace in the logarithmic amplitude spectrum of each trace n (t, f).

[0080] The positive logarithmic amplitude spectrum LFP n (t, f) is calculated using the following formula:

[0081] LFP n (t, f) = LF n (t, f) - min(LF n (t, f))

[0082] (4) Calculate the amplitude spectrum Amp0(t, f) and the positive logarithmic amplitude spectrum LFP0(t, f) of the trace to be reconstructed according to the amplitude spectra corresponding to each frequency of each seismic trace in the irregular observation system within the circular domain in steps (2) and (3) for n = 1,..., N

[0083] The amplitude spectrum Amp0(t, f) and the positive logarithmic amplitude spectrum LFP0(t, f) of the trace to be reconstructed, including the first trace to be reconstructed and the second trace to be reconstructed, are calculated as follows respectively:

[0084]

[0085]

[0086] where: k is an adjustable parameter, generally taking the value of 2 (Euclidean distance) or 3 (Manhattan distance), and the optimal k value can also be determined by methods such as cross - validation. d i represents the distance between the trace in the irregular observation system within the circular domain and the trace to be reconstructed

[0087] (5) Normalize the positive logarithmic amplitude spectrum LFP0(t, f) in step (4) above to obtain the normalized positive logarithmic amplitude spectrum LFPE0(t, f). The normalization formula is as follows:

[0088]

[0089] (6) Determine the reconstructed data LFPE0(t, f) of each seismic trace in the regular observation system in the frequency domain FS n (t, f) according to the phase spectrum when n = 1 in step (2) and the normalized positive logarithmic amplitude spectrum LFPE0(t, f) in step (5).

[0090] The reconstructed data LFPER0(t, f) in the frequency domain is calculated using the following formula:

[0091]

[0092] (7) Inverse-transform the reconstructed data LFPER0(t,f) to the time domain to obtain the time-domain reconstructed data s0(t) = IFFT[LFPE0(t,f)], which is the first reconstructed data. Then, take the first reconstructed data obtained from each seismic trace in the regular acquisition system as an element and assign it to the constructed three-dimensional array.

[0093] At this time, in the three-dimensional array, the first reconstructed data stored in the corresponding second set must not be zero. In this way, in the subsequent three-dimensional curvelet transform reconstruction process, compared with directly setting the seismic data on the second trace to be reconstructed to zero in the prior art, the generation of random noise is reduced to a certain extent.

[0094] In step S14, the first reconstructed data stored in the three-dimensional array of the regular acquisition system can be iterated using the following method. For example:

[0095] (8) Perform a three-dimensional curvelet transform on each element A[s][m][n] in the three-dimensional array assigned with the first reconstructed data to obtain the curvelet coefficients at each scale and each angle after the first transform. Sort the obtained curvelet coefficients in ascending order, and use median filtering to remove outliers to obtain the preprocessed curvelet coefficients C j , where j is greater than or equal to 1 and less than or equal to Z, and Z is the total number of curvelet coefficients, and calculate the average curvelet coefficient as Preferably, median filtering generally uses five-point median filtering, which can ensure the filtering effect while improving the calculation efficiency.

[0096] (9) Establish a curvelet coefficient sample point ratio formula, and calculate the sample point ratio corresponding to each preprocessed curvelet coefficient C in step (8). j Calculate the sample point ratio corresponding to each curvelet coefficient.

[0097] The sample point ratio formula is as follows:

[0098]

[0099] Among them: R(r) represents the ratio of the rth sample point, w is the length of the sliding time window, r is the corresponding sample point number, α is a stability factor, and α is greater than zero and less than or equal to Preferably, take the value

[0100] (10) According to the ratio of curvelet coefficients corresponding to each sample point, a curvelet coefficient ratio curve is obtained. The curvelet coefficient ratio curve is normalized, and the curvelet coefficient at the same sample point corresponding to the maximum value of the normalized curvelet coefficient ratio is determined, which is the noise threshold at the first iteration and also the demarcation point between the noise and the effective wave curvelet coefficients, denoted as C div .

[0101] (11) According to the average curvelet coefficient and the noise threshold C at the first iteration div , a threshold function θ for each iteration is constructed.

[0102] The threshold function θ for each iteration is calculated according to the following formula:

[0103]

[0104] where: N is the total number of iterations, which is determined by experience, n is the nth iteration, C div If it is less than or equal to then take C div as μ·C max , where μ is greater than 0 and less than 1, and it should be ensured that μ·C max is greater than Preferably, μ is taken as 0.95.

[0105] (12) According to the established threshold function, in each iteration process, the curvelet coefficients less than the noise threshold are all set to 0, and then the inverse transform is performed on the curvelet coefficients after removing the noise.

[0106] Calculate the root mean square amplitude of all seismic traces in the second set after the dth reconstruction, denoted as J d , calculate the root mean square amplitude of all seismic traces in the first set after the first reconstruction, denoted as J0. Calculate the threshold parameter If ζ is greater than the predetermined threshold parameter, stop the calculation, otherwise continue the next curvelet transform until the preset termination condition (the number of iterations reaches the preset total number of iterations or ζ is greater than the given threshold parameter) is reached, then stop the iteration.

[0107] Finally, for the first trace to be reconstructed, replace the seismic trace data corresponding to it in the three-dimensional array after the iteration reaches the preset termination condition with the first reconstructed data corresponding to the first trace to be reconstructed, that is, the shot gather data (reconstructed seismic data) in the final regular observation system is obtained.

[0108] Take a specific example to illustrate the above method for improving the reconstruction quality of low signal-to-noise ratio data. Take a certain work area as an example, and the method for reconstructing seismic trace data is described in detail according to the above steps S11 - S13.

[0109] In the above step S11, seismic data is excited and collected according to geological tasks to construct an irregular observation system W0. Referring to Figure 2 as shown, where the receiver line spacing is 160 m, the number of receiver lines is 49, the trace interval varies in the range of 10 - 70 m, the shot gather data has a total of 12,872 seismic traces, the number of samples per trace is 1,750, the sampling interval is 4 ms, and the figure shows that there are 4,058 shots excited.

[0110] Based on the irregular observation system W0, a regular observation system W1 is constructed. Referring to Figure 3 as shown, where the receiver line spacing is 160 m, the number of receiver lines is 49, the regular trace interval is 20 m, each receiver line has 555 seismic traces, the shot gather data has a total of 27,195 seismic traces, the number of samples per trace is 1,750, the sampling interval is 4 ms, and the figure shows that there are 4,058 shots excited.

[0111] Therefore, based on the parameters of the regular observation system W1, for any element A[s][m][n] of the constructed three-dimensional array, where 1 ≤ s ≤ 49, 1 ≤ m ≤ 555, and 1 ≤ n ≤ 1,750.

[0112] In the above step S12, when classifying the seismic traces in the regular observation system W1, if the distance between any seismic trace in the regular observation system and the seismic trace closest to it is, for example, less than or equal to the preset distance value r0 = ε·R = 0.5×20 = 10 m, the corresponding seismic trace in the regular observation system is recorded as the first trace to be reconstructed; otherwise, the corresponding seismic trace in the regular observation system is recorded as the second trace to be reconstructed. In the embodiment of the present invention, the boundary range coefficient ε is taken as 0.5.

[0113] In the above step S13, when drawing a circular domain with each seismic trace in the regular observation system W1 as the reference, the preset radius of each first trace to be reconstructed is, for example, r = ε·R = 0.5×20 = 10 m. For the preset radius of each second trace to be reconstructed, according to experience, it is appropriate to take β as 6. In the embodiment of the present invention, r = ε·4R = 0.5×4×20 = 40 m.

[0114] After obtaining the first reconstructed data, the root mean square amplitude J0 of all seismic traces in the first set after the first reconstruction is calculated as 9,895.

[0115] In step S14, three-dimensional curvelet transform is performed on each element A[s][m][n] in the three-dimensional array assigned with the first reconstructed data. According to the preprocessed curvelet coefficient C j , the average curvelet coefficient is obtained as

[0116] According to the ratio of the curvelet coefficients corresponding to each sample point, the noise threshold C at the first iteration is determined. div= 19731.

[0117] In the embodiment of the present invention, when calculating the threshold function θ, μ is taken as 0.95, and the total number of iterations N is taken as 45.

[0118] In the embodiment of the present invention, the predetermined threshold parameter is taken as 0.95. When the number of iterations reaches the 42nd time, the root mean square amplitude J of all seismic traces in the second set after reconstruction d = 9403. At this time, the threshold parameter ζ = 0.95028, which is greater than the predetermined threshold parameter 0.95, and the iteration terminates.

[0119] In the embodiment of the present invention, referring to Figure 4 and Figure 5 shown, Figure 4 shows the 18th trace arrangement result of a certain shot gather data collected before the seismic data is reconstructed. There are 12,872 first traces to be reconstructed and 14,323 second traces to be reconstructed in this shot. Figure 5 shows the 18th trace arrangement result of this shot gather data after the seismic data is reconstructed. There are 27,195 seismic traces in this shot. Figure 4 Compared with Figure 3 the in-phase axis continuity of the single-shot profile after reconstruction is significantly improved, and the quality of the reconstructed data meets the requirements of subsequent data processing.

[0120] Based on the same inventive concept, the embodiment of the present invention also provides a device for improving the reconstruction quality of low signal-to-noise ratio data. Since the principle of the problem solved by this device is similar to that of the aforementioned method for improving the reconstruction quality of low signal-to-noise ratio data, the implementation of this device can refer to the implementation of the aforementioned method, and the repeated parts will not be elaborated.

[0121] The embodiment of the present invention provides a device for improving the reconstruction quality of low signal-to-noise ratio data. Referring to Figure 6 shown, it includes:

[0122] A three-dimensional array construction module 61, configured to construct a regular acquisition system W1 based on the irregular acquisition system W0, determine the parameters of the regular acquisition system W1, and construct a three-dimensional array based on the parameters of the regular acquisition system W1. The three-dimensional array is used to store all seismic trace data in the regular acquisition system W1;

[0123] A trace to be reconstructed classification module 62, configured to compare the seismic traces in the regular acquisition system W1 with the corresponding seismic traces in the irregular acquisition system W0, and determine whether each seismic trace in the regular acquisition system W1 is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison rule;

[0124] The first reconstruction module 63 is configured to reconstruct the seismic trace data corresponding to each first trace to be reconstructed and each second trace to be reconstructed in the regular observation system W1 according to the seismic trace data of each seismic trace in the irregular observation system W0 and a preset reconstruction method, so as to obtain corresponding first reconstruction data, and store the first reconstruction data of each first trace to be reconstructed and each second trace to be reconstructed in the three-dimensional array;

[0125] The iteration module 64 is configured to iterate the first reconstruction data in the three-dimensional array by using a preset iteration strategy until a preset termination condition is reached; for the first trace to be reconstructed, replace the corresponding seismic trace data in the three-dimensional array after the iteration reaches the preset termination condition with the first reconstruction data corresponding to the first trace to be reconstructed.

[0126] An embodiment of the present invention provides a computing device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the method for predicting a favorable area for hydrocarbon accumulation as described above is implemented.

[0127] An embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for improving the reconstruction quality of low signal-to-noise ratio data as described above is implemented.

[0128] Obviously, those skilled in the art can make various changes to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes.

Claims

1. A method for improving the reconstruction quality of low signal-to-noise ratio data, characterized in that, include: Constructing a regular observation system W1 based on the irregular observation system W0, determining parameters of the regular observation system W1, and constructing a three-dimensional array based on the parameters of the regular observation system W1, wherein the three-dimensional array is used to store all seismic trace data in the regular observation system W1; Compare the seismic traces in the regular observation system W1 with the corresponding seismic traces in the irregular observation system W0, and determine whether each seismic trace in the regular observation system W1 is the first trace to be reconstructed or the second trace to be reconstructed according to a preset comparison rule; According to the seismic trace data in the irregular observation system W0 and a preset reconstruction method, the seismic trace data corresponding to the first trace to be reconstructed and the second trace to be reconstructed in the regular observation system W1 are reconstructed to obtain corresponding first reconstruction data, and the first reconstruction data of the first trace to be reconstructed and the second trace to be reconstructed are stored in the three-dimensional array; Adopting a preset iteration strategy to iterate the first reconstructed data in the three-dimensional array until a preset termination condition is reached; For the first trace to be reconstructed, the seismic trace data corresponding to the three-dimensional array after the iteration reaches a preset termination condition is replaced by the first reconstructed data corresponding to the first trace to be reconstructed.

2. The method according to claim 1, characterized in that, Adopting a preset iteration strategy to iterate the first reconstructed data in the three-dimensional array until a preset termination condition is reached, including: Performing a three-dimensional curvelet transform on the three-dimensional array of the initial assignment to obtain preprocessed curvelet coefficients; Performing a noise removal operation on the preprocessed curvelet coefficients; the noise is a curvelet coefficient whose curvelet coefficient is less than a preset noise threshold; Performing a three-dimensional inverse curvelet transform on the de-noised curvelet coefficients to obtain a first iteration value of the seismic trace reconstruction data; The first iteration value of the seismic trace reconstruction data is iterated again until the number of iterations reaches a preset total number of iterations or a threshold parameter is greater than a preset threshold parameter, and then the iteration is stopped.

3. The method according to claim 2, wherein The preprocessed curvelet coefficients are obtained by the following method: The curvelet coefficients obtained after performing three-dimensional curvelet transformation on the three-dimensional array with the initial assignment are sorted in order from small to large, and abnormal values ​​are eliminated by median filtering to obtain the pre-processed curvelet coefficients.

4. The method according to claim 2, characterized in that, The noise removal operation is performed on the preprocessed curvelet coefficients, including: Determine the average curve coefficient and the curve coefficient ratio formula according to the pre-processed curve coefficient; According to the curvelet coefficient ratio of each sample point, a curvelet coefficient ratio curve is obtained, the curvelet coefficient ratio curve is normalized, and the sample point (curvelet coefficient) corresponding to the maximum value of the curvelet coefficient ratio is determined as the noise threshold in the first iteration; Establishing a noise threshold function for each iteration according to the average curvelet coefficient and the noise threshold for the first iteration, wherein the noise threshold function is used to determine the noise threshold corresponding to each iteration; According to the noise threshold function, the noise smaller than the noise threshold in each iteration process is set to zero.

5. The method according to claim 2, wherein The threshold parameter is the ratio of the root mean square amplitude of all the channels in the second set to be reconstructed after each iteration to the root mean square amplitude of all the channels in the first set to be reconstructed.

6. The method according to claim 1, wherein The parameters of the regular observation system include: the receiving line spacing L, the number of receiving lines S, the regular trace spacing R, the number of seismic traces M in each receiving line, and the number of sampling points N for each seismic trace; Constructing a three-dimensional array based on the parameters of the regular observation system includes: Any element in the three-dimensional array is A[s][m][n], where s is the s-th receiving line, 1 ≤ s ≤ S, m is the m-th receiving point on the s-th receiving line, 1 ≤ m ≤ M, and n is the n-th sampling point of the seismic trace at the m-th receiving point on the s-th receiving line, 1 ≤ n ≤ N.

7. The method according to claim 1, wherein Determining whether each seismic trace in the regular observation system is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison method includes: According to the actual position coordinates of the receiving points, determine the seismic trace in the irregular observation system W0 that is closest to each seismic trace in the regular observation system W1, calculate the distance between the two, and if the distance between the two is less than or equal to the preset distance value, record the corresponding seismic trace in the regular observation system W1 as the first trace to be reconstructed, otherwise record the corresponding seismic trace in the regular observation system W1 as the second trace to be reconstructed.

8. The method according to claim 7, wherein The preset distance value r0 = ε·R, where ε is the boundary range coefficient and R is the regular trace distance.

9. The method according to claim 1, characterized in that Reconstructing the seismic trace data corresponding to each first trace to be reconstructed or second trace to be reconstructed in the regular observation system W1 according to the seismic trace data of each seismic trace in the irregular observation system W0 and a preset reconstruction method to obtain the corresponding first reconstruction data includes: Taking the coordinate position of each seismic trace in the regular observation system W1 as the center, draw a circular region with a preset radius r, and sort the seismic traces in the irregular observation system W0 contained in the circular region in ascending order of the distance from the center to obtain the sorted seismic data S n (t), where n is the number of seismic traces in the irregular observation system W0 contained in each different circular region, 1 ≤ n ≤ N0; N0 is the maximum number of seismic traces in the irregular observation system W0 contained in each circular region; Transform the seismic data S n (t) into the frequency domain FS n (t,f), and calculate the amplitude spectrum Amp n (t,f) and the phase spectrum phase n (t,f); For the amplitude spectrum Amp n (t,f) of each seismic trace, take the logarithm to obtain the logarithmic amplitude spectrum LF n (t,f). Subtract the minimum value from all the spectral values of the logarithmic amplitude spectrum LF n (t,f) of each seismic trace respectively to obtain the positive logarithmic amplitude spectrum LFP n (t,f); Calculating the amplitude spectrum Amp0(t,f) and the positive logarithmic amplitude spectrum LFP0(t,f) of the trace to be reconstructed according to the amplitude spectrum corresponding to each frequency of each seismic trace in the irregular observation system W0 included in the circular domain; Normalizing the positive logarithmic amplitude spectrum LFP0(t,f) to obtain the normalized positive logarithmic amplitude spectrum LFPE0(t,f); Determine the reconstructed data LFPE0(t,f) of each seismic trace in the regular observation system W1 in the frequency domain FS according to the phase spectrum phase1(t,f) when n = 1 and the positive logarithmic amplitude spectrum LFPE0(t,f). n (t,f); Inverse-transforming the reconstructed data LFPE0(t,f) to the time domain to obtain the first reconstruction data.

10. The method according to claim 9, characterized in that The preset radius r is determined by the following method: In the regular observation system, the preset radius r of each of the first traces to be reconstructed is ε·R; In the regular observation system, the preset radius r of each of the second traces to be reconstructed satisfies being greater than ε·R and less than or equal to ε·βR, where β is a real number greater than 1 and is an empirical value; where ε is the boundary range coefficient and R is the regular trace distance.

11. An apparatus for improving the reconstruction quality of low signal-to-noise ratio data, characterized in that Includes: A three-dimensional array construction module, which is used to construct a regular observation system based on an irregular observation system, determine the parameters of the regular observation system, and construct a three-dimensional array based on the parameters of the regular observation system, and the three-dimensional array is used to store all seismic trace data in the regular observation system; A trace to be reconstructed classification module, which is used to compare the seismic traces in the regular observation system with the corresponding seismic traces in the irregular observation system, and determine whether each seismic trace in the regular observation system is a first trace to be reconstructed or a second trace to be reconstructed according to a preset comparison rule; A first reconstruction module, configured to reconstruct seismic trace data corresponding to each first trace to be reconstructed and each second trace to be reconstructed in the regular observation system according to the seismic trace data in the irregular observation system and a preset reconstruction method, so as to obtain corresponding first reconstruction data, and store the first reconstruction data of each first trace to be reconstructed and each second trace to be reconstructed in the three-dimensional array; An iteration module, configured to iterate the first reconstruction data in the three-dimensional array by using a preset iteration strategy until a preset termination condition is reached; For the first trace to be reconstructed, replace the corresponding seismic trace data in the three-dimensional array after the iteration reaches the preset termination condition with the first reconstruction data corresponding to the first trace to be reconstructed.

12. A computing device, characterized in that, Comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, the method for improving the reconstruction quality of low signal-to-noise ratio data according to any one of claims 1-10 is implemented.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the method for improving the reconstruction quality of low signal-to-noise ratio data according to any one of claims 1-10 is implemented.

Images