A method for automatic first arrival picking and correction of low signal-to-noise ratio seismic data

By improving the signal-to-noise ratio of seismic data and calculating time difference using multiple similarities, the initial automatic picking and correction of low signal-to-noise ratio seismic data is achieved, and the problems of low automatic picking accuracy and large workload in the prior art are solved. Improve the efficiency of seismic data processing and reliability of static correction.

CN113985478BActive Publication Date: 2024-11-05SINOPEC OILFIELD SERVICE CORPORATION +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111204249.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2021-07-28
Filing Date
2021-10-15
Publication Date
2024-11-05
Estimated Expiration
2041-10-15

AI Technical Summary

Technical Problem

The prior art is difficult to effectively automatically pick and correct the arrival of low signal-to-noise ratio seismic data, resulting in huge manual modification workload, affecting the efficiency of seismic data processing, and the automatic picking accuracy is low, reducing the effectiveness and reliability of the static correction algorithm.

Method used

The signal-to-noise ratio of seismic data is improved by cross-correlation method, the signal-to-noise ratio is calculated using power spectrum method, the initial arrival is automatically picked up, and the similarity coefficient and time difference are calculated based on multiple similarities, and the inaccurate initial arrival is corrected, and more than two software are used to The combination of automatic pickup and correction.

Benefits of technology

It realizes efficient automatic picking and correction of low signal-to-noise ratio seismic data, improves the accuracy and processing efficiency of the first arrival, reduces the need for manual modification, and improves the effectiveness and reliability of the static correction algorithm.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113985478B_ABST
    Figure CN113985478B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for automatically picking and correcting the first arrival of low signal-to-noise ratio seismic data, comprising the following steps: 1) using a cross-correlation method to improve the signal-to-noise ratio of seismic data; 2) automatically picking the first arrival on the data optimized in step 1); 3) using a power spectrum method to calculate the signal-to-noise ratio of each channel of the seismic data in step 1), and distinguishing the first arrival channels that cannot be picked up and the first arrival channels that can be picked up according to a standard; 4) for the pickable channels, calculating the first arrival absolute value difference of the same channels picked up by different methods; distinguishing the first arrival accurate channels and the first arrival channels that need to be modified according to a standard; 5) calculating the similarity coefficient of the first arrival channels that need to be modified and the accurate channels based on the similarity of multiple channels, and calculating the time difference when the two are most similar; 6) correcting the first arrival of the inaccurate channels according to the time difference calculated in step 5) and the first arrival time of the accurate channels. The present invention can be widely applied to the field of seismic exploration technology to automatically pick up and automatically correct the erroneous first arrival of low signal-to-noise ratio seismic data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of exploration technology, in particular to a method for automatically picking and correcting the first arrival of low signal-to-noise ratio seismic data. Background Art

[0002] Static correction methods based on first arrivals (including tomographic statics / refraction statics / first arrival residual statics, etc.) are the most effective methods to solve static correction problems in complex areas, and the first arrival is the key to ensure the successful application of such methods. Therefore, picking up the accurate first arrival becomes the most basic and important task in static correction processing.

[0003] Generally speaking, when the signal-to-noise ratio of seismic data is high, the existing methods or commercial software can basically quickly and automatically pick up accurate first arrivals. However, for low signal-to-noise ratio data, the existing automatic picking methods, especially mainstream commercial software, basically find it difficult to obtain satisfactory picking effects.

[0004] The traditional solution is to first use software to automatically pick the first arrival, and then manually modify the wrong first arrival. With the development of current high-density acquisition technology, the number of shots / receiving channels has increased significantly. The total number of shots is often hundreds of thousands or millions, and the number of receiving channels is basically tens of thousands. The workload of manual modification is very huge, which affects the efficiency of seismic data processing. In order to solve the problem of picking the first arrival of low signal-to-noise ratio seismic data, people have studied various automatic picking solutions, which are generally achieved through three ways: one is to try to improve the signal-to-noise ratio of seismic data, the second is to try to improve the noise resistance of the first arrival picking method, and the third is to automatically perform quality control on the picked first arrivals and then delete the wrong first arrivals.

[0005] For example, Chinese invention patent publication CN202010667880.3 published a method for automatic first arrival picking based on multi-scale morphology, and Chinese invention patent publication CN201811282072.4 published a method for intelligent identification of abnormal channels based on first arrival information. However, these methods have three problems: 1) The development cost is high and it is difficult to play a role in actual production; 2) The abnormal first arrival can be automatically determined and then deleted, but automatic correction cannot be performed, which reduces the available first arrivals; 3) Some software actually used in production can automatically pick up the first arrival, but there is still a problem of low picking accuracy. A large number of erroneous first arrivals need to be manually modified or deleted, which reduces the effectiveness and reliability of the static correction algorithm based on the first arrival. Summary of the invention

[0006] The object of the present invention is to provide a method for automatically picking and correcting the first arrival of low signal-to-noise ratio seismic data to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for automatically picking and correcting the first arrival of low signal-to-noise ratio seismic data, comprising the following steps:

[0009] 1) Using cross-correlation method to improve the signal-to-noise ratio of seismic data;

[0010] 2) Automatically picking up the first arrival on the data optimized in step 1);

[0011] 3) Calculate the signal-to-noise ratio of each channel of the seismic data in step 1) using the power spectrum method, and distinguish the first arrival unpickable channels and pickable channels according to the standard;

[0012] 4) For the tracks that can be picked, calculate the absolute value difference of the first arrival of the same track picked by different methods; distinguish the first arrival accurate track and the first arrival track that needs to be modified according to the standard;

[0013] 5) Based on the similarity of multiple channels, calculate the similarity coefficient between the first arrival channel to be modified and the accurate channel, and calculate the time difference when the two are most similar;

[0014] 6) Correct the first arrival of the inaccurate track according to the time difference calculated in step 5) and the first arrival time of the accurate track.

[0015] As a further solution of the present invention: in step 1), the data that has undergone elevation static correction is subjected to linear dynamic correction using an appropriate speed V, and then the cross-correlation between a certain track i and the surrounding K tracks is calculated, and the track with a cross-correlation coefficient greater than 0.3 is superimposed on the track to form a new i-th track data.

[0016] As a further solution of the present invention: in step 2), two or more software / methods are used to automatically pick up first arrivals, and the picked up first arrivals are respectively recorded as Fb1 and Fb2.

[0017] As a further solution of the present invention: In step 3), the i-th seismic data is denoted as x i , the i+1th seismic data is x i+1 , then x i (f) and x i+1 (f) are their respective representations in the frequency domain, where X i (f) is X i The conjugate of (f) is given by the following formula:

[0018]

[0019] Where:

[0020]

[0021] As a further solution of the present invention: In step 3), a noise discrimination coefficient a is given 0 , when the signal-to-noise ratio of a channel is lower than a 0When the signal-to-noise ratio of a channel is greater than a 0 When , it is determined that the path can pick up the first arrival.

[0022] As a further solution of the present invention: the first arrivals of the channels marked as D in 3) in the first arrivals Fb1 and Fb2 are deleted, and the first arrivals of the channels Fb1a and Fb2a not marked as D are output to the next step.

[0023] As a further solution of the present invention: in step 5), each channel is identified according to the file number and channel number, and the absolute value difference FB_abs of the first arrival time of the channels with the same file number and channel number is calculated, FB_abs = Fb1a-Fb2a. If FB_abs is less than 5ms, it is determined that the first arrivals Fb1a and Fb2a of the channel are accurate, and the average value of Fb1a and Fb2a is calculated, that is, FB_mean = (Fb1a+Fb2a) / 2, and FB_mean is used as the new first arrival of the channel.

[0024] As a further solution of the present invention: in step 5), the similarity coefficient of each track within a certain time window is calculated by using a multi-track similarity formula, and the inclination time difference when the similarity coefficient takes the maximum value is the delay of the inaccurate track relative to the accurate track, and the first arrival time of the inaccurate track is calculated; for a track i with inaccurate first arrival, the first arrivals around it are accurate, and the similarity coefficient within a range of J tracks on a certain side of the track is calculated according to the formula; wherein j = {-2, -1, 1, 2}, ΔP is the inclination time difference, the maximum inclination time difference is P, then ΔP = -P, -P+st, -P+2*st, ..., P, where st is the sampling interval; the formula is:

[0025]

[0026] As a further solution of the present invention: in step 5), when the similarity coefficient reaches the maximum value, the inclination time difference is recorded as T, which is determined as the delay time when the current track is most similar to the track on its left or right.

[0027]

[0028] Set the first arrival time FB on the left side of the current track i-1 Add T to get the first arrival time of the current track,

[0029] FB i =FB i-1 +τ.

[0030] As a further solution of the present invention, the first arrival of the wrong track is calculated from the left while the first arrival of the wrong track is calculated from the right. When the two meet, the first arrival FB of the meeting track is calculated from the left. i左 and the first arrival FB obtained from the right side i右If the absolute value difference is less than 5ms, the sliding window is determined to be suitable. If the absolute value difference is greater than 5ms, the sliding window is determined to be unsuitable and ΔS needs to be reduced. Repeat the above steps until the absolute value difference is less than 5ms.

[0031] Compared with the prior art, the beneficial effects of the present invention are: the present invention can be widely applied to the field of seismic exploration technology, and automatically pick up and automatically correct the erroneous first arrival of low signal-to-noise ratio seismic data. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of the present invention.

[0033] Figure 2 This is an incorrect initial arrival schematic diagram in the present invention.

[0034] Figure 3 It is a schematic diagram of the original single gun in the present invention.

[0035] Figure 4 This is a schematic diagram of a single gun that has been optimized in the present invention.

[0036] Figure 5 This is a schematic diagram of the first arrival result automatically picked up by GeoEast software in the present invention.

[0037] Figure 6 This is a schematic diagram of the first arrival results automatically picked up by GMseis software in the present invention.

[0038] Figure 7 It is a schematic diagram of the result after automatic correction in the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] The present invention is achieved through the following technical solutions:

[0041] The method firstly applies the cross-correlation method to pre-process the data to improve the data signal-to-noise ratio; secondly, the first arrival is automatically picked up by two or more methods, and then the signal-to-noise ratio of each channel is calculated by the power spectrum method and a reasonable threshold value is set. The channels below the threshold value are determined as channels that cannot pick up the first arrival and are removed; then the absolute value difference of the first arrival of the same channel is calculated and a time difference limit (such as 5ms) is given. If the absolute value difference of the first arrival is less than the limit value, it is determined that the first arrivals picked up by the two methods of the channel are accurate, and the average value of the two first arrivals is calculated as the final first arrival. If the absolute value of the first arrival is greater than the limit value, it is determined that the first arrival of the channel is inaccurate and needs to be corrected. Based on the first arrival that has been determined to be accurate, the similarity coefficient of several accurate channels and inaccurate channels is calculated channel by channel, and the time difference between the inaccurate channel and the accurate channel is determined according to the maximum value of the similarity coefficient. Finally, the first arrival of the inaccurate channel is obtained by adding the time difference to the first arrival of the accurate channel.

[0042] like Figure 1 As shown, in an embodiment of the present invention, a method for automatically picking and correcting the first arrival of low signal-to-noise ratio seismic data includes the following steps:

[0043] S1. According to the characteristics of the work area data, various methods can be used to optimize the first arrival (Li Hong, Li Wandong, Zhu Zhe, et al. Discussion on issues that need to be paid attention to in the processing of controllable source seismic data [J]. Petroleum Geophysical Exploration, 2020, 59 (5) 758-767), such as minimum phase, denoising, filtering, gain processing or shaping processing. Through multiple experiments, the following method is used in the present invention to optimize the seismic data. Based on the surface consistency characteristics, the first arrivals in several channels have a certain similarity. Therefore, the seismic channels in several channels around a channel can be superimposed to form a new data channel for the channel. The specific method is that in the single shot data, the data that has been statically corrected for elevation is firstly linearly corrected using an appropriate velocity V, and then the correlation coefficient between the current channel i (i = 1, 2, ..., N, N is the total number of channels) and the surrounding K channels (which can be an odd number between 3-9) is calculated, and the channel with a correlation coefficient greater than 0.3 is superimposed with the i-th channel to form a new i-th channel data. Through the cycle, the new data of each channel is generated in turn. The signal-to-noise ratio of the processed seismic traces is greatly improved. The first arrival of the processed low-signal-to-noise ratio vibrator vibration record is relatively clear and crisp, which is conducive to automatic first arrival picking.

[0044] S2, we use commercial software Gmseis and GeoEast software (specific picking principle can refer to the random documents of the software) to automatically pick up the data respectively. When using the software to pick up the first arrival, it is necessary to select suitable picking parameters through experiments according to the data characteristics. Of course, the method or software for picking up the first arrival is not limited to these two. Any software that can have higher picking accuracy and efficiency can be used to pick up the first arrival. That is to say, the method described in the present invention is not important for specifically selecting which software to pick up. What is important is to use two or more methods to quickly pick up more accurate first arrivals. The first arrivals picked up are respectively recorded as Fb1 and Fb2.

[0045] S3. Calculate the signal-to-noise ratio of seismic data by power spectrum method (Zhang Junhua, Zhou Zhenxiao, Zhong Lei, et al. Quantitative calculation and comparison of signal-to-noise ratio of seismic data. Oil and Gas Geophysics, 2008, 6(4): 9-14). If the i-th seismic data is x i , the i+1th seismic data is x i+1 , then x i (f) and x i+1 (f) are their respective representations in the frequency domain, where X i (f) is X i The conjugate of (f). The formula is as follows:

[0046]

[0047] Where:

[0048]

[0049] According to Formula 1, the signal-to-noise ratio of seismic data is calculated by selecting the appropriate number of channels / time window / frequency band range through experiments, and a noise discrimination coefficient a is given. 0 , when the signal-to-noise ratio of a channel is lower than a 0 When the signal-to-noise ratio of a channel is greater than a, the signal-to-noise ratio of the channel is determined to be very low and the first arrival cannot be picked up, which is represented by the symbol D. 0 , it is determined that the path can be picked up for the first arrival.

[0050] S4, the first arrivals of the roads marked as D in 3) are deleted from the first arrivals Fb1 and Fb2, and the first arrivals of the roads not marked as D, Fb1a and Fb2a, are output to the next step;

[0051] S5, based on the similarity of multiple tracks, calculate the similarity coefficient between the first track to be modified and the accurate track, and calculate the time difference when the two are most similar;

[0052] S5.1. Identify each channel according to the file number and channel number. Calculate the absolute difference FB_abs of the first arrival time of the channels with the same file number and channel number, that is, FB_abs = |Fb1a-Fb2a|. If FB_abs is less than 5ms, it is determined that the first arrival Fb1a and Fb2a of the channel are accurate. Calculate the average value of Fb1a and Fb2a, that is, FB_mean = (Fb1a+Fb2a) / 2, and use FB_mean as the new first arrival of the channel.

[0053] S5.2. Taking the first arrival accurate track as the benchmark, use the multi-track similarity formula to calculate the similarity coefficient of each track within a certain time window. The inclination time difference when the similarity coefficient takes the maximum value is the delay time of the inaccurate track relative to the accurate track. Then the first arrival time of the inaccurate track can be calculated. The specific method is as follows. For a track i with an inaccurate first arrival, if Figure 2 As shown, the first arrival of the third track is inaccurate, then the similarity coefficient within the range of J tracks (usually 1 or 2) on one side of the track can be calculated. In the formula, j = {-2, -1, 1, 2}, ΔP is the dip time difference, the maximum dip time difference is P, then ΔP = -P, -P+st, -P+2*st, ..., P, where st is the sampling interval.

[0054] The formula is

[0055]

[0056] S5.3. Then according to Formula 4, the inclination time difference when the similarity coefficient reaches the maximum value is recorded as T, that is, the delay time when the current track is determined to be most similar to the track on its left or right.

[0057]

[0058] Finally, the first arrival time FB on the left side of the current track i-1 Add T to get the first arrival time of the current track.

[0059] FB i =FB i-1 +τ Formula 5

[0060] S5.4. The maximum inclination angle P affects the accuracy of the final first arrival. Therefore, while calculating the first arrival of the wrong track from the left side using the method described in S5.2 and S5.3, calculate the first arrival of the wrong track from the right side. When the two meet, calculate the first arrival FB from the left side of the meeting track. i左 and the first arrival FB obtained from the right side i右The absolute value difference of the first arrival is calculated. If the absolute value difference is less than 5ms, the sliding time window is determined to be appropriate. If the absolute value difference is greater than 5ms, the sliding time window is determined to be inappropriate and needs to be reduced. Then reduce ΔS and repeat the above steps until the absolute value difference of the first arrival is less than 5ms. This method is called the "tunnel method", that is, the first arrival of each track is calculated from both sides. When they meet, the absolute value difference of the first arrival of the same track should be less than a certain value (the first arrival accuracy determination coefficient). If not, modify the time window and recalculate. However, for the tracks at both ends of a single shot, the first arrival can only be calculated from one side.

[0061] S6, correct the first arrival of the inaccurate track according to the time difference calculated in step S5 and the first arrival time of the accurate track; i Compare with the first arrival of the same track picked, that is, calculate FB respectively a =|FB i -Fb1a i | and FB b =|FB i -Fb2a i |If FB a or FB b If it is less than 5ms, then Fb1a can be determined i or Fb2a i is accurate, then calculate (FB i +FB gmsai ) / 2 or (FB i +FB geoai ) / 2 is the first arrival time of the current track.

[0062] The present invention is verified by actual data below. The method can improve the accuracy and efficiency of automatic first arrival picking to a great extent.

[0063] Figure 3 is the original single shot data, Figure 4 This is the single shot data after optimization processing. It can be seen in detail that after optimization processing, the signal-to-noise ratio of the single shot is higher, the first arrival is clearer, and it is conducive to first arrival picking.

[0064] Figure 5 The first arrival results automatically picked by GeoEast software are shown in Figure 6 The first arrival results automatically picked by GMseis software are shown. Figure 7 The result after automatic correction by the method of the present invention is shown in FIG. 1 . It can be seen that all erroneous first arrivals have been corrected.

[0065] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0066] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for automatically picking up the first arrival of low signal-to-noise ratio seismic data and correcting the wrong first arrival, characterized in that: The following steps are involved: 1) Using cross-correlation method to improve the signal-to-noise ratio of seismic data; 2) Automatically picking up the first arrival on the data after the optimization process in step 1); 3) Calculate the signal-to-noise ratio of each channel of the seismic data in step 1) using the power spectrum method, and distinguish the first arrival unpickable channels and pickable channels according to the standard; 4) For the tracks that can be picked, calculate the absolute value difference of the first arrival of the same track picked by different methods; distinguish the first arrival accurate track and the first arrival track that needs to be modified according to the standard; 5) Based on the similarity of multiple channels, calculate the similarity coefficient between the first arrival channel that needs to be modified and the first arrival accurate channel, and calculate the time difference when the two are most similar; 6) Correct the first arrival of the track to be modified based on the time difference calculated in step 5) and the first arrival time of the accurate first arrival track; In step 5), the similarity coefficient of each track within a certain time window is calculated using the multi-track similarity formula based on the first arrival accurate track. The formula is: Where, j = {-2, -1, 1, 2}, ΔP is the dip time difference, the maximum dip time difference is P, then ΔP = -P, -P+st, -P+2*st, ..., P, where st is the sampling interval; i is the track with inaccurate first arrival, and the first arrivals around this track are accurate; J is the value range of the track selected on one side of the track with inaccurate first arrival; The inclination time difference when the similarity coefficient reaches the maximum value in step 5) is recorded as τ, and the formula for determining the delay time when the current track is most similar to the track on its left or right is: Add τ to the first arrival time FBi-1 on the left side of the current track to get the first arrival time of the current track, that is, FB i =FB i-1 +τ; While calculating the first arrival of the wrong track from the left side, calculate the first arrival of the wrong track from the right side. When the two meet, calculate the absolute value difference between the first arrival FBi left from the left side and the first arrival FBi right from the right side of the meeting track. If the absolute value difference is less than 5ms, it is determined that the sliding time window is appropriate. If the absolute value difference is greater than 5ms, it is determined that the sliding time window is inappropriate and ΔS needs to be reduced. Repeat the above steps until the absolute value difference of the first arrival is less than 5ms.

2. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 1, characterized in that: In step 1), the data that has been subjected to elevation static correction is subjected to linear dynamic correction using an appropriate speed V, and then the mutual correlation coefficients between the current i-th track and the surrounding K tracks are calculated respectively, and the tracks with mutual correlation coefficients greater than 0.3 are superimposed on the current i-th track to form new i-th track data.

3. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 2, characterized in that: In step 2), two or more software / methods are used to automatically pick up first arrivals, and the picked up first arrivals are respectively recorded as Fb1 and Fb2.

4. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 3, characterized in that: In step 3), the i-th seismic data is denoted as x i , the i+1th seismic data is x i+1 , then x i (f) and x i+1 (f) are their respective representations in the frequency domain, where For X i (f) is the conjugate; then the signal-to-noise ratio is calculated as follows: Where:

5. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 4, characterized in that: In step 3), a noise discrimination coefficient a0 is given. When the signal-to-noise ratio of a certain channel is lower than a0, it is determined that the first arrival cannot be picked up, which is represented by the symbol D. When the signal-to-noise ratio of a certain channel is greater than a0, it is determined that the first arrival of this channel can be picked up.

6. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 5, characterized in that: Delete the first arrivals of the tracks marked as D in step 3) from the first arrivals Fb1 and Fb2, and output the first arrivals of the tracks not marked as D to the next step.

7. The method for automatically picking up first arrivals of low signal-to-noise ratio seismic data and correcting wrong first arrivals according to claim 6, characterized in that: In step 5), each channel is identified according to the file number and channel number, and the absolute value difference FB_abs of the first arrival time of the channels with the same file number and channel number is calculated, FB_abs = |Fb1a-Fb2a|. If FB_abs is less than 5ms, it is determined that the first arrivals Fb1a and Fb2a of the channel are accurate, and the average value of Fb1a and Fb2a is calculated, that is, FB_mean = (Fb1a+Fb2a) / 2, and FB_mean is used as the new first arrival of the channel.