阅读说明：本技术 一种地震波时频域初至自动拾取方法 (Automatic first arrival picking method for seismic wave time-frequency domain ) 是由 许卓 巩向博 韩立国 于 2020-06-05 设计创作，主要内容包括：本发明专利提供了一种地震波时频域初至自动拾取方法。该方法包括对炮集的初至波做时窗截取；通过S变换将地震道转换到时频域,对参考道及目标道的时频谱进行反褶积；在炮域内,对反褶积结果进行叠加生成虚拟时频谱；对虚拟时频谱和参考道时频谱进行褶积；在检波点域内,对褶积结果进行叠加得到超级虚拟时频谱；沿时间轴计算超级虚拟时频谱的能量；采用差分算子获得能量差分曲线；在差分曲线上找到局部最大值作为初至波到时；逐道完成时频域初至自动拾取。本发明抗噪能力强,拾取精度高,不受地震道波形不一致和近地表复杂条件的影响,大大减少人工修改初至波拾取的工作量,能很好地满足各种复杂条件下地震勘探对初至波自动拾取的高精度要求。(The invention provides a method for automatically picking up a first arrival of a seismic wave in a time-frequency domain. The method comprises the steps of carrying out time window interception on the first arrival waves of the shot set; converting the seismic channel into a time-frequency domain through S transformation, and performing deconvolution on time-frequency spectrums of the reference channel and the target channel; in the shot domain, the deconvolution results are superposed to generate a virtual time frequency spectrum; convolution is carried out on the virtual time frequency spectrum and the reference channel time frequency spectrum; in the wave detection point domain, overlapping convolution results to obtain a super virtual time frequency spectrum; calculating the energy of the super virtual time spectrum along a time axis; obtaining an energy difference curve by adopting a difference operator; finding a local maximum value on the difference curve as a first arrival time; and the time-frequency domain first arrival automatic pickup is completed channel by channel. The method has strong noise resistance and high picking precision, is not influenced by inconsistent seismic trace waveforms and near-surface complex conditions, greatly reduces the workload of manually modifying the first-arrival wave picking, and can well meet the high-precision requirement of seismic exploration on the automatic picking of the first-arrival waves under various complex conditions.)

1. A method for automatically picking up a first arrival of a seismic wave in a time-frequency domain, the method comprising:

(1) performing time window interception on the original shot gather data, and extracting a part containing a first arrival wave homophase axis;

(2) converting a reference track and a target track which are positioned at the position x and recorded by the detectors A and B respectively into a time-frequency domain through S transformation, wherein the reference track and the target track are respectively represented as S (A | x, tau, f) and S (B | x, tau, f);

(3) performing deconvolution calculation on the time frequency spectrums of the reference channel and the target channel in the time frequency domain;

(4) and (3) in the shot domain, performing step (3) on all shot gathers positioned at the stationary phase seismic source and performing superposition summation on all deconvolution calculation results to generate a virtual time frequency spectrum S (B | A, tau, f) between the detectors A and B^virt；

(5) Performing convolution calculation on the virtual time frequency spectrum and the time frequency spectrum of the reference channel in the time frequency domain;

(6) and in the wave detection point domain, performing step (5) on all reference channels positioned at a stationary phase wave detection point and performing superposition summation on all convolution calculation results to obtain a super virtual time frequency spectrum S (B | x, tau, f) of which the seismic source is positioned at x and recorded at a wave detection point B^super；

(7) Calculating the energy E (B | x, tau) of the super virtual time spectrum point by point along a time axis;

(8) obtaining an energy difference curve between adjacent time sampling points by adopting a difference operator for the energy calculated in the step (7);

(9) finding a local maximum value on the energy difference curve obtained in the step (8), wherein the time corresponding to the local maximum value is the arrival time of the first arrival wave of the target track;

(10) and (5) repeating the steps (2) to (9) one by one in the target shot set to finish the automatic time-frequency domain first arrival picking of the target shot set.

2. The method for automatically picking up the first arrivals in the time-frequency domain of seismic waves according to claim 1, wherein the length of the time window in step (1) is generally 2 to 3 wavelengths near the first arrivals.

3. The method for automatically picking up the first arrival of seismic waves in the time-frequency domain according to claim 1, wherein the automatic determination of the estimated position of the time window in the step (1) is realized by dividing the offset information of each trace in the shot gather by the approximate average velocity of the near-surface.

4. The method for automatically picking up the first arrival of seismic waves in the time-frequency domain according to claim 1, wherein the calculation formula of the virtual first arrival time spectrum in the step (4) is as follows:

wherein, represents complex conjugate, k is wave number, N_sThe number of stationary phase sources is the regularization parameter.

5. The method for automatically picking up the first arrival of seismic waves in the time-frequency domain according to claim 1, wherein the calculation formula for obtaining the super virtual first arrival time spectrum in the step (6) is as follows:

wherein N is_gIs the number of phase-stationary detectors.

6. The seismic wave time-frequency domain first arrival automatic picking method according to claim 1, wherein the energy calculated in the step (7) is defined as:

Technical Field

The invention relates to the technical field of geophysical exploration seismic data processing, in particular to a high-precision seismic wave first-motion automatic picking method.

Background

The first-arrival seismic wave picking method usually adopts a man-machine interaction mode, the traditional manual picking mode is high in relative precision, low in efficiency, time-consuming and labor-consuming, and difficult to meet the actual production requirement that the collected data volume is larger and larger, and the first-arrival automatic picking method is mainly adopted in the current production. In seismic data processing, it is very important to accurately pick up the first arrival waves, and the accuracy and quality of the result directly affect the precision of subsequent processing links such as static correction, borehole seismic exploration and tomography.

With the continuous deepening of the seismic exploration work, the seismic exploration target area gradually turns to an area with complex changes of the near-surface speed and the structure, and under the condition, the requirements on the precision and the quality of first-arrival picking are higher. However, under complex near-surface conditions, the first-arrival waveform state in the seismic data may be distorted, and the first-arrival waveform may be inconsistent between seismic traces due to the complex conditions, and furthermore, a large amount of interference noise often exists in the actual seismic data. The problems all cause that the first-arrival waves are difficult to identify under the condition of high background noise, so that the conventional first-arrival automatic pickup method is difficult to obtain a reliable result with high precision, and further the precision of subsequent related processing steps of the seismic data is influenced.

Therefore, there is a need for an effective first-arrival automatic picking method, which can improve the problem of low first-arrival picking precision of the conventional first-arrival automatic picking method under various complicated conditions.

Disclosure of Invention

The invention aims to provide a first arrival automatic picking method, which aims to improve the first arrival picking precision under various complex conditions including complex near-surface, inconsistent seismic gather waveforms, low signal-to-noise ratio and the like and provide reliable parameters for each data processing step for solving problems by utilizing the first arrival wave attributes.

In order to achieve the above object, the present invention comprises the steps of:

(1) performing time window interception on the original shot gather data, and extracting a part containing a first arrival wave homophase axis;

wherein the time window length is typically 2 to 3 wavelengths around the first arrival.

Further, the automatic determination of the estimated position of the time window can be realized by dividing the offset distance information of each track in the shot set by the approximate average speed of the near-surface.

(2) Converting a reference track and a target track which are positioned at the position x and recorded by the detectors A and B respectively into a time-frequency domain through S transformation, wherein the reference track and the target track are respectively represented as S (A | x, tau, f) and S (B | x, tau, f);

(3) performing deconvolution calculation on the time frequency spectrums of the reference channel and the target channel in the time frequency domain;

(4) and (3) in the shot domain, performing step (3) on all shot gathers positioned at the stationary phase seismic source and performing superposition summation on all deconvolution calculation results to generate a virtual time frequency spectrum S (B | A, tau, f) between the detectors A and B^virt；

Further, the calculation formula is:

wherein, represents complex conjugate, k is wave number, N_sThe number of stationary phase sources is the regularization parameter.

(5) Performing convolution calculation on the virtual time frequency spectrum and the time frequency spectrum of the reference channel in the time frequency domain;

(6) and in the wave detection point domain, performing step (5) on all reference channels positioned at a stationary phase wave detection point and performing superposition summation on all convolution calculation results to obtain a super virtual time frequency spectrum S (B | x, tau, f) of which the seismic source is positioned at x and recorded at a wave detection point B^super；

Further, the calculation formula is:

wherein N is_gIs the number of phase-stationary detectors.

(7) Calculating the energy E (B | x, tau) of the super virtual time spectrum point by point along a time axis;

further, the energy is defined as:

(8) obtaining an energy difference curve between adjacent time sampling points by adopting a difference operator for the energy calculated in the step (7);

(9) finding a local maximum value on the energy difference curve obtained in the step (8), wherein the time corresponding to the local maximum value is the arrival time of the first arrival wave of the target track;

(10) and (5) repeating the steps (2) to (9) one by one in the target shot set to finish the automatic time-frequency domain first arrival picking of the target shot set.

The seismic wave time-frequency domain first arrival automatic picking method can effectively improve the picking precision of weak energy first arrival signals and low signal-to-noise ratio first arrival waves in seismic data, increase the number of first arrival gather capable of being picked, improve the picking error caused by the inconsistency of the waveforms of the first arrival waves and solve the problem of low first arrival picking precision under various complex conditions. Compared with the conventional first arrival picking method, the method has strong anti-noise capability and high picking precision, greatly reduces the workload of manually modifying first arrival picking, can meet the high requirement on the first arrival picking precision under the complex near-surface condition, and provides an effective basis for solving the subsequent related seismic data processing problem by using the first arrival.

Drawings

FIG. 1 is a complex near-surface geological model diagram of an embodiment of the automatic first arrival picking method for time-frequency domain seismic waves of the present invention.

FIG. 2 is a diagram of low SNR seismic shot gather data according to an embodiment of the present invention.

Fig. 3 is a single channel time-frequency domain automatic picking method according to an embodiment of the present invention.

FIG. 4 is a first arrival picking result of an embodiment of the method for automatically picking up a first arrival in time-frequency domain of seismic waves of the present invention.

Detailed Description

The invention is described in further detail below with reference to the figures and examples.

For a complex near-surface geological model shown in fig. 1, the recorded shot gather data with low signal-to-noise ratio and waveform inconsistency in fig. 2 comprises the following steps:

(1) and performing time window interception on the original shot gather data, and extracting a part containing a first arrival wave homophase axis.

The length of the time window is 2 to 3 wavelengths near the first arrival wave, and the automatic determination of the estimated position of the time window can be realized by dividing the offset information of each track in the shot set by the approximate average speed of the near-surface.

(2) The reference and target tracks, with the source located at x and recorded at detectors a and B respectively, are transformed into the time-frequency domain by an S-transform, denoted S (a | x, τ, f) and S (B | x, τ, f) respectively.

(3) And performing deconvolution calculation on the time frequency spectrums of the reference channel and the target channel in the time frequency domain.

(4) And (3) in the shot domain, performing step (3) on all shot gathers positioned at the stationary phase seismic source and performing superposition summation on all deconvolution calculation results to generate a virtual time frequency spectrum S (B | A, tau, f) between the detectors A and B^virtThe calculation formula is as follows:

wherein, represents complex conjugate, k is wave number, N_sThe number of stationary phase sources is the regularization parameter.

(5) And performing convolution calculation on the virtual time frequency spectrum and the time frequency spectrum of the reference channel in the time frequency domain.

(6) And in the wave detection point domain, performing step (5) on all reference channels positioned at a stationary phase wave detection point and performing superposition summation on all convolution calculation results to obtain a super virtual time frequency spectrum S (B | x, tau, f) of which the seismic source is positioned at x and recorded at a wave detection point B^superThe calculation formula is as follows:

wherein N is_gIs the number of phase-stationary detectors.

(7) Computing the energy E (B | x, tau) of the super virtual time spectrum point by point along the time axis, wherein the energy is defined as:

(8) and (5) adopting a difference operator to obtain an energy difference curve between adjacent time sampling points for the energy calculated in the step (7).

(9) And (4) finding a local maximum value on the energy difference curve obtained in the step (8), wherein the time corresponding to the maximum value is the arrival time of the first arrival wave of the target track.

(10) And (5) repeating the steps (2) to (9) one by one in the target shot set to finish the automatic time-frequency domain first arrival picking of the target shot set.

Fig. 3a and 3b are the display of the super virtual time-frequency spectrum and the time-frequency domain first arrival picking result on the corresponding single-track record, respectively. Therefore, for seismic data with low signal-to-noise ratio, the super virtual time-frequency spectrum obtained by the technology still has high quality, and the time-frequency domain first arrival automatic pickup is easy to carry out. Fig. 4 is a first arrival auto-picking result obtained using the technique of the present invention. Therefore, for single-shot seismic data acquired under the complex near-surface condition, the seismic wave time-frequency domain first-arrival automatic picking method has strong noise resistance, is not influenced by inconsistent waveforms and the complex near-surface condition, has high accuracy of the first-arrival automatic picking result, can well meet the high requirement of seismic exploration on the first-arrival picking accuracy under the complex condition, and improves the reliable basis for effectively solving the problems of static correction and the like of the seismic data by utilizing the first-arrival information subsequently.

The present invention is not intended to be limited to the embodiments shown herein, and any modifications, equivalents, improvements, etc. made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.