Weak signal compensation method based on seismic attribute analysis

Abstract

The invention discloses a weak signal compensation method based on seismic attribute analysis, which belongs to the field of seismic data processing, and comprises the following steps: processing post-stack data through a radial prediction filtering technology to obtain data with a high signal-to-noise ratio, gridding two-dimensional seismic data, and interpolating to form a three-dimensional data volume; then carrying out attribute analysis according to the signal-to-noise ratio and the continuity factor of the seismic data, scanning grid data of an inclination angle in a given range on the basis of the three-dimensional grid data, solving the signal-to-noise ratio and the continuity factor of the data, and selecting and extracting required superposition data according to attributes; and finally, correlation analysis is carried out through the CMP gather and extracted target superposition data, and weighted compensation of signals is carried out, so that effective seismic weak reflection signals are compensated, deep imaging quality is improved, and effective compensation of deep weak signals of seismic data is realized.

Description

Technical Field

[0001] The present invention belongs to the field of seismic data processing, and particularly relates to a weak signal compensation method based on seismic attribute analysis. Background Art

[0002] As the focus of marine oil and gas exploration and development gradually turns to the deep layer, due to the complexity of deep geological structures, the attenuation factors of seismic waves during propagation, and the influence of acquisition noise, deep seismic signals are often weak, resulting in low signal-to-noise ratio of seismic data and unclear formation continuity. Therefore, in order to effectively extract deep geological information, it is necessary to compensate and enhance weak signals, and at the same time improve the wave group characteristics of the formation to obtain an imaging profile that can meet the interpretation requirements of geologists.

[0003] Conventional two-dimensional seismic processing techniques for weak signal compensation mainly include: one is to restore and compensate weak signals through seismic denoising techniques, such as energy compensation methods, various Q compensation methods, etc.; the other is to effectively decompose complex signals based on compressive sensing and time-frequency analysis to achieve the purpose of extracting weak effective seismic signals. In the actual processing process, it is necessary to fully consider the characteristics of seismic data and the limitations of algorithms. These two types of seismic weak signal compensation methods have the following three problems: First, conventional energy compensation and Q compensation methods basically compensate the entire seismic data, and the compensation effect depends on the signal-to-noise ratio of the data, and it is easy to strengthen noise as well; second, methods using algorithms based on compressive sensing and time-frequency analysis are more dependent on the algorithms themselves, have poor adaptability to different types of seismic data, and are prone to aliasing during the decomposition process; third, the compensation process of two-dimensional seismic data is only limited to processing in a two-dimensional manner.

[0004] During the deep seismic exploration process, due to the influence of geometric spreading, formation absorption, and scattering effects during the propagation of seismic waves, deep seismic signals are severely attenuated. Coupled with the influence of acquisition equipment and environmental noise, the signal-to-noise ratio of deep seismic signals is further reduced, and the obtained seismic data has a low signal-to-noise ratio and unclear wave group characteristics. Conventional seismic denoising techniques are difficult to meet the compensation requirements for weak signals, which restricts the scientific research of scholars on deep sea formations. Summary of the Invention

[0005] In order to solve the problems that traditional seismic denoising techniques are difficult to meet the compensation requirements for deep weak signals, etc., the present invention fully considers the characteristics of data and proposes a seismic weak signal compensation method based on the control of signal-to-noise ratio and continuity factor to achieve effective compensation for deep weak signals.

[0006] The present invention is implemented by the following technical solutions: A weak signal compensation method based on seismic attribute analysis, comprising the following steps:

[0007] Step A: Process the stacked data through radial prediction filtering to improve its signal-to-noise ratio. On this basis, interpolate the processed stacked data based on three-dimensional gridding to form a three-dimensional data volume;

[0008] (1) First, on the stacked data, within a given range of traces, perform quadratic interpolation on the sampling interval of each trace to make it half of the input interval;

[0009] (2) Then, according to the combination of the set filter width and fan range, calculate the number of linear movements. For each movement, use the following formula to perform summation calculation within the specified time window to obtain the similarity value within the given trace range;

[0010] (3) For each time window, scan the similarities of all movements, find the maximum similarity and define it as the best movement of this time window. The corresponding adjacent traces have the greatest coherence; at the same time, define a linear weight function related to the number of traces of the stacked output trace and with the maximum value at the zero-offset position; apply the linear weight function to each trace, and then stack the main input traces together to generate one output trace;

[0011] (4) The filter moves one trace in a rolling manner, and repeat the above steps until the last trace is processed to complete the radial prediction filtering.

[0012] (5) The three-dimensional gridding processing operation is as follows: First, copy the stacked data after improving the signal-to-noise ratio, and then define the size and resolution of the three-dimensional grid according to the data range. The grid defines the reference trace interval, and a three-dimensional data volume is formed through linear interpolation;

[0013] Step B: Based on the geological target requirements, perform attribute analysis based on the signal-to-noise ratio and continuity factor to obtain the optimized target stacked data, specifically including:

[0014] On the basis of the three-dimensional gridded data, scan the grid data of the dip angle within the given range, and calculate the signal-to-noise ratio and continuity factor of the data:

[0015] Signal-to-noise ratio SNR:

[0016]

[0017] Where P S is the signal power, P n is the noise power;

[0018] Continuity factor CF:

[0019]

[0020] Where S i,jIt represents the amplitude value of the i-th seismic record at the j-th sampling point, N is the number of seismic record traces, and M is the number of sampling points per seismic record trace;

[0021] Then, according to the requirements of the geological target for seismic data, weighted selection of parameters is carried out to obtain the optimized target stacked data.

[0022] Step C: Use correlation analysis to process the CMP gather and the optimized target stacked data, extract weak signals and target information from the CMP gather, and through weighted compensation with the original CMP gather, obtain the compensated gather, and after stacking processing, obtain the corresponding seismic section:

[0023] First, perform normal moveout correction on the CMP gather, calculate the signal-to-noise ratio and continuity factor of this gather on the CMP gather, and select the gather with multiple attributes at a certain interval;

[0024] Then, perform coherence analysis with the optimized target stacked data in a scanning manner, extract weak signals and target information from the CMP gather, and perform coherence analysis with the traces on the corresponding stacked section in a scanning manner at a certain interval. In the coherence analysis process, a time window containing at least 2 waveforms is selected within the gather to obtain the correlation coefficient of the corresponding signal-to-noise ratio and continuity factor between the optimized target stacked data and the CMP gather;

[0025] For different attributes, select the gather with a larger correlation to reconstruct into a new trace, and give a certain weight ratio according to the actual data needs and perform weighted averaging with the original CMP gather to obtain the compensated trace, thereby completing the processing of the entire gather, and realizing the compensation of weak signals after stacking.

[0026] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0027] In this solution, correlation analysis is carried out between the CMP gather and the extracted target stacked data, and weighted compensation of signals is performed to compensate for the effective weak seismic reflection signals and improve the deep imaging quality; the processing effect can be controlled by selecting attribute parameters, which has good applicability, and can also effectively improve the wave group characteristics of the formation, so the imaging effect is better; in addition, this solution is data-driven, and after two-dimensional data gridding, the attributes of the data are extracted for analysis, which has better accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow schematic diagram of the weak signal compensation method described in the embodiment of the present invention;

[0029] Figure 2 It is a schematic diagram of three-dimensional data gridding in the embodiment of the present invention;

[0030] Figure 3 Schematic diagram of the target profile extracted in the embodiment of the present invention;

[0031] Figure 4 CMP gather before and after compensation in the embodiment of the present invention, where (a) is before compensation and (b) is after compensation;

[0032] Figure 5 Schematic diagram of the stacked profile before compensation in the embodiment of the present invention;

[0033] Figure 6 Schematic diagram of the stacked profile after compensation in the embodiment of the present invention. Detailed implementation manners

[0034] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described below with reference to the accompanying drawings and embodiments. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0035] Embodiment, the present invention uses a weak signal compensation method based on seismic attribute analysis to effectively compensate for weak signals in deep seismic data. The main idea is as follows: First, the stacked data is processed by radial prediction filtering technology to obtain data with a high signal-to-noise ratio. After gridding the two-dimensional seismic data, it is interpolated to form a three-dimensional data volume; then, attribute analysis is performed according to the signal-to-noise ratio and continuity factor of the seismic data. Based on the three-dimensional gridded data, the grid data with dip angles within a given range is scanned to obtain the signal-to-noise ratio and continuity factor of the data, and the required stacked data is extracted according to the attributes; finally, correlation analysis is performed on the CMP gather and the extracted target stacked data, and weighted compensation of the signal is performed to compensate for the effective seismic weak reflection signal and improve the deep imaging quality.

[0036] As Figure 1 described, it includes the following steps;

[0037] Step A: Process the stacked data by the radial prediction filtering method to obtain stacked data with a high signal-to-noise ratio. On this basis, interpolate the data based on three-dimensional gridding to form a three-dimensional data volume;

[0038] Step B: Based on the geological target requirements, perform attribute analysis according to the signal-to-noise ratio and continuity factor to obtain optimized target stacked data;

[0039] Step C: Use correlation analysis to process the CMP gather and the optimized target stacked data, extract weak signals and target information on the CMP gather, and obtain the compensated gather through weighted compensation with the original gather. After stacking processing, the corresponding seismic profile is obtained.

[0040] Specifically:

[0041] The said step A is implemented in the following manner:

[0042] Step A1, radial prediction filtering improves the signal-to-noise ratio of the stacked data;

[0043] In order to better extract the effective signal, stacked data with a better signal-to-noise ratio is required (stacked data is a commonly used term in the industry, and the stacked data is the data after NMO correction and stacking processing of the CMP gather). In this embodiment, the radial prediction filtering technology is mainly adopted to process the stacked data. The main idea is to perform weighted summation on adjacent traces, and these adjacent traces are subjected to time-varying movement to align them at the maximum coherence angle, so as to enhance the most coherent signal within the specified dip range, while suppressing random noise and the coherent energy outside this range.

[0044] First, on the stacked data, by giving a certain number of traces (3 - 29) within a certain range, perform quadratic interpolation on the sampling interval of each trace to make it half of the input interval (to avoid spatial aliasing in the far-offset traces).

[0045] Then, according to the given combination of filter width and fan range, calculate the number of linear movements. At the far-offset trace, the movement increment is equal to the sampling interval of the input trace (usually 4 ms). If the number of movements required to cover the entire dip range exceeds 15 times, the movement increment at the far-offset will increase to 1.5 times the sampling interval. If the number of movements required to cover the dip range exceeds 45 times, the movement increment will further increase to two sampling intervals. For each movement, use formula 1 to perform summation calculation within the specified time window to obtain the similarity S within the given trace range C :

[0046]

[0047] where f ij is the amplitude value of the j-th sampling point of the i-th trace, M is the number of traces for sampling value summation, and N is the window length (number of samplings) centered on k.

[0048] For each time window, scan the similarities of all movements, find the maximum similarity, which is defined as the best movement of this time window, and the corresponding adjacent traces have the maximum coherence. At the same time, define a linear weight function related to the number of stacked output traces and with a maximum value at the zero-offset position, apply the linear weight function to each trace, and then stack the main input traces together to generate one output trace. The filter moves one trace in a rolling manner, and repeat the above steps until the last trace is processed. In the actual processing process, usually a certain percentage of the original input part is mixed to obtain a more natural signal-to-noise ratio improvement effect.

[0049] Step A2: 3D Meshing

[0050] The 3D grid data processing operations are as follows: First, copy the data after enhancing the signal-to-noise ratio in Step A1. To ensure the efficiency of the algorithm, generally, at least 10 copies are selected, and in this embodiment, 11 copies are preferably made. Then, define the size and resolution of the 3D grid according to the data range. The grid definition refers to the trace interval, and a 3D data volume is formed through linear interpolation.

[0051] In Step B, attribute analysis is performed based on the signal-to-noise ratio and the continuity factor to obtain the optimized target stacked data. The specific method is as follows:

[0052] Step B1: Selection of signal-to-noise ratio and continuity factor attributes

[0053] In this embodiment, attribute analysis is mainly performed based on the signal-to-noise ratio and the continuity factor of seismic data. On the basis of the 3D meshed data, scan the grid data of the dip angle within a given range, calculate the signal-to-noise ratio and the continuity factor of the data, and obtain the corresponding parameter setting basis.

[0054] Signal-to-noise ratio SNR:

[0055]

[0056] Among them, P S is the signal power, and P n is the noise power, which are respectively expressed as:

[0057]

[0058] Among them, s i,j and n i,j are respectively the signal and noise amplitude values corresponding to the jth sampling point of the ith seismic trace. N is the number of seismic traces, and M is the number of sampling points per seismic trace.

[0059] Continuity factor CF:

[0060]

[0061] Among them, S i,j represents the amplitude value of the ith seismic trace at the jth sampling point. N is the number of seismic traces, and M is the number of sampling points per seismic trace.

[0062] Step B2: Obtain the optimized target stacked data;

[0063] The signal-to-noise ratio (SNR) and continuity factor (CF) of the post-stack data are obtained using Equations 2 and 4. These two attributes describe the data from the perspectives of effective reflection energy and characteristics, respectively, providing improved adaptability and accuracy during parameter selection and control. To facilitate attribute classification, both attributes are normalized. The optimization process primarily involves weighted parameter selection based on the geological requirements for seismic data to obtain optimized target stack data. In this invention, a weak signal is defined when the attribute factor ranges from 0.3 to 0.7. In areas with a good SNR, the continuity factor dominates; in areas with a good continuity factor, the SNR dominates.

[0064] The step C, performing the coherence analysis specifically adopts the following method: step C1, correlation analysis between the CMP gather and the target stacking data;

[0065] This embodiment mainly performs coherent analysis on the pre-stack CMP gathers and the optimized target stack data obtained in step B2 by scanning to obtain gathers that match the selected stack data. This is specifically achieved in the following manner:

[0066] First, a CMP gather is subjected to dynamic correction. The signal-to-noise ratio and continuity factor of the gather are calculated on the CMP gather. Gathers with multiple attributes are selected at a certain interval (5% of the difference between the maximum and minimum values). Then, coherence analysis is performed on the optimized target stack data by scanning to extract weak signals and target information on the CMP gather. By constraining the stack data, the attributes are further refined based on the pre-stack CMP gather, which is more conducive to the extraction of weak signals. During the coherence analysis, a time window containing at least two waveforms is selected in the gather to obtain the correlation coefficient of the corresponding signal-to-noise ratio and continuity factor between the optimized target stack data and the CMP gather. The calculation formula is as follows:

[0067]

[0068] Where r s,c (τ) is the cross-correlation function; s = s min ,s dif *0.05,s dif *0.1,…,s max ; where s dif =s max -s min , c=c min ,c dif *0.05,c dif *0.1,…,c max ; where c dif =c max -c min, j is the amplitude sampling serial number within the calculation time window; T1 and T2 represent the start and end times of the calculation time window; T2 - T1 is the time window length.

[0069] Step C2, weighted compensation of the target signal;

[0070] For different attributes, select the gather with greater correlation for reconstruction to form a new gather (generally taking more than 70%, where 1 is the maximum correlation). According to the actual data requirements, give a certain weight ratio and perform weighted averaging with the original CMP gather to obtain the compensated gather, and thus complete the processing of the entire gather. After stacking, the compensation for weak signals is realized, effectively improving the signal-to-noise ratio of seismic data and obtaining an improvement in the formation wave group characteristics.

[0071] Effect analysis

[0072] To verify the effectiveness and practicality of this invention, select the actual 2D seismic data in a certain sea area. To ensure a certain signal-to-noise ratio, the input original data is the data after conventional denoising. First, perform radial prediction filtering and gridding processing on the 2D data, and form a 3D data volume after interpolation ( Figure 2 as shown); then based on the feature analysis of the data, the main problems of the original data are the unclear formation continuity and wave group characteristics. Therefore, the continuity factor is selected as the main factor, and the weight mainly accounts for the range of 0.5 - 0.8, Figure 3 is the extracted target profile; finally, through the correlation analysis and weighted compensation with the CMP gather, the compensated CMP gather ( Figure 4 ) is obtained. By comparison, it can be seen that for the compensated CMP gather, the effective signals are highlighted, especially the weak reflection signals. Through comparison after stacking processing ( Figure 5 , Figure 6 ), it can be seen that the signal-to-noise ratio and effective reflection energy of the profile after weak signal compensation are enhanced, the wave group characteristics are also significantly improved, the formation structure becomes clearly visible, the problem of poor deep imaging in the sea area data is solved, and the imaging quality of the data can well meet the needs of geological interpreters.

[0073] The above are only the preferred embodiments of the present invention, and are not limitations to the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A weak signal compensation method based on seismic attribute analysis, characterized in that, It includes the following steps: Step A: Process the stacked data through radial prediction filtering to improve its signal-to-noise ratio. On this basis, interpolate the processed stacked data based on three-dimensional gridding to form a three-dimensional data volume. Step B: According to the geological target requirements, perform attribute analysis based on the signal-to-noise ratio and continuity factor to obtain the optimized target stacked data. Step C: Use correlation analysis to process the CMP gather and the optimized target stacked data. Extract weak signals and target information from the CMP gather, and through weighted compensation with the original CMP gather, obtain the compensated gather. After stacking processing, obtain the corresponding seismic profile.

2. The weak signal compensation method based on seismic attribute analysis according to claim 1, wherein: In the said Step B, based on the three-dimensional gridded data, scan the grid data of the dip angle within a given range, and calculate the signal-to-noise ratio and continuity factor of the data: Signal-to-noise ratio SNR: Among them, P S signal power, P n is the noise power; Continuity factor CF: Among them, S i,j represents the amplitude value of the i-th seismic record at the j-th sampling point, N is the number of traces of the seismic record, and M is the number of sampling points of each seismic record; Then, perform weighted selection of parameters according to the requirements of the geological target for seismic data to obtain the optimized target stacked data.

3. The weak signal compensation method based on seismic attribute analysis according to claim 1, characterized in that: In the said Step C, perform coherence analysis on the pre-stack CMP gather and the optimized target stacked data obtained in Step B by scanning to obtain the gather matching the selected target stacked data: First, perform NMO correction on the CMP gather, calculate the signal-to-noise ratio and continuity factor of the gather on the CMP gather, and select the gather with multiple attributes at a certain interval. Then, perform coherence analysis with the optimized target stacked data by scanning, extract weak signals and target information on the CMP gather, perform coherence analysis with the traces on the corresponding stacked section by scanning at a certain interval. In the coherence analysis process, select a time window containing at least 2 waveforms within the gather to obtain the correlation coefficient of the corresponding signal-to-noise ratio and continuity factor between the optimized target stacked data and the CMP gather. For different attributes, select the gather with large correlation for reconstruction to form a new trace. Given a certain weight ratio according to the actual data needs, perform weighted averaging with the original CMP gather to obtain the compensated trace, thereby completing the processing of the entire gather. After stacking, realize the compensation of weak signals.

4. The weak signal compensation method based on seismic attribute analysis according to claim 3, characterized in that: In the said Step C, the calculation formula of the said correlation coefficient is as follows: where r s,c (τ) is the cross-correlation function; s = s min , s dif *0.05, s dif *0.1, …, s max ; where s dif = s max - s min , c = c min , c dif *0.05, c dif *0.1, …, c max ; where c dif = c max - c min , and j is the amplitude sampling sequence number within the calculation time window; T1 and T2 represent the start and end times of the calculation time window; T2 - T1 is the time window length.

5. The weak signal compensation method based on seismic attribute analysis according to claim 1, characterized in that: In the said Step A, improve the signal-to-noise ratio of the stacked data through radial prediction filtering. The specific implementation method is as follows: (1) First, on the stacked data, given the number of traces within a certain range, perform quadratic interpolation on the sampling interval of each trace to make it half of the input interval. (2) Then, according to the combined filter width and fan-shaped range set, calculate the number of linear movements. For each movement, use the following formula to perform a summation calculation within the specified time window to obtain the similarity value S within the given trace range. C : where f ij is the amplitude value of the j-th sampling point in the i-th trace, M is the number of traces for summing the sampling values, and N is the window length centered on k; (3) For each time window, scan all moving similarities, find the maximum similarity and define it as the best movement of the time window. The corresponding adjacent traces have the maximum coherence; at the same time, define a linear weight function related to the number of traces of the stacked output trace and with the maximum value at the zero offset position; apply the linear weight function to each trace, and then stack the main input traces together to generate one output trace. (4) The filter moves one trace in a rolling manner, and repeats the above steps until the last trace is processed to complete the radial prediction filtering.

6. The weak signal compensation method based on seismic attribute analysis according to claim 1, wherein: In the step A, the three-dimensional grid processing is performed as follows: First, the stacked data after improving the signal-to-noise ratio is copied. Then, the size and resolution of the three-dimensional grid are defined according to the data range. The grid definition refers to the trace interval, and a three-dimensional data volume is formed after linear interpolation.

Images