CN103344989B - The analytical approach of impulse noise interference in vibroseis seismologic record - Google Patents

Abstract

The present invention relates to the analytical approach of impulse noise interference in a kind of vibroseis seismologic record.First according to the principle of work of vibroseis, direct-path signal determination reference amplitude value and the walkaway threshold value of vibroseis is utilized; Then according to detection threshold, the algorithm of impulsive noise in automatic detection record is designed; Finally, analyzing by detecting the impulsive noise obtained, evaluating the influence degree of vibroseis record by impulsive noise.Analytical approach disclosed by the invention extensively can be adapted to the sweep method of vibroseis linear processes, the noise analysis of Seismic Exploration with Vibrator data acquisition is fast and effeciently realized at earthquake data acquisition scene, the impulsive noise of real-time analysis vibroseis is on the impact of record, avoid because noise level is strong or comprise the not good situation of vibroseis relative recording quality that random high power pulse noise causes, thus reduce costs, effectively ensure that the acquisition quality of vibroseis.

Description

The analytical approach of impulse noise interference in vibroseis seismologic record

Technical field:

The present invention relates to the noise analytical technology in a kind of geophysical data, the analytical approach that especially in vibroseis seismologic record, impulsive noise is disturbed vibroseis seismologic record.

Background technology:

Impulsive noise to refer in seismologic record that the duration is short, the noise that amplitude is large.In actual seismic exploration, due to the enchancement factor in the external world, can produce a large amount of impulsive noise, and these impulsive noise interference to impact origins such as explosive sources being slight, is but serious to the interference of the relative recording of vibroseis.Because vibroseis is different from the principle of work of impact origin, when extracting the seismic signal of similar impact origin in vibroseis seismologic record, need to carry out coherent detection to the original seismic data of vibroseis, the impulsive noise be mixed in raw readings also have passed through coherent detection process, because related algorithm is a kind of nonlinear operation, so, the impulsive noise be mixed in raw readings there occurs nonlinearities change, it is made to be no longer local to the interference of signal, thus cause routinely noise level significantly to raise, serious reduction vibroseis seismologic record quality.

In actual Seismic Exploration with Vibrator, if directly record the geological data after being correlated with, the interference that the impulsive noise before relevant causes in relative recording will be difficult to remove, even if be also difficult to compacting in follow-up data processing; If in follow-up seismic data process process, when the noise in discovery vibroseis relative recording is too strong, field can only be got back to and re-start data acquisition, cause acquisition cost significantly to increase.

Summary of the invention:

The object of the invention is to for above-mentioned the deficiencies in the prior art, provide a kind of impulsive noise to the analytical approach of vibroseis relative earthquake record annoyance level.

Main thought of the present invention is: first according to the principle of work of vibroseis, utilizes direct-path signal determination reference amplitude value and the walkaway threshold value of vibroseis; Then according to detection threshold, the algorithm of impulsive noise in automatic detection record is designed; Finally, analyzing by detecting the impulsive noise obtained, evaluating the influence degree of vibroseis record by impulsive noise.

The present invention is achieved by the following technical solutions:

When vibroseis gathers, in background, impulsive noise comprises the following steps the analysis of annoyance level in relative earthquake record:

A, input the original vibroseis seismologic record of the irrelevant form of a big gun, according to irrelevant record, the initial time T of the primary wave signal of all valid traces in the effective recording channel of artificial pickup _b, T _b=(T _b(1), T _b(2), T _b(3) ..., T _b(i) ..., T _b(n _tr)), n _treffective number of channels of this big gun record, T _bi () is the initial time of the i-th valid trace, ask for the end time of corresponding i-th valid trace according to formula (1)

T _e(i)=T _b(i)+t _sweep………………………………（1）

In formula (1), T _ei () is the end time of the i-th valid trace, t _sweepfor the sweep time of vibroseis, i=1,2,3 ..., n _tr;

If T _ei () is greater than t writing time _recordmaximum moment t _max, then T _e(i)=t _max;

The irrelevant record of b, vibroseis comprises auxiliary channels and effective recording channel two parts data, effective recording channel generally containing multiple tracks data, first takes out the first data of effective recording channel, is designated as u (t), t is sampling time sequence, t=1,2,3, ..., k, k are maximum sampling number, and this track data u (t) of automatic Picking is at T _b(1): T _e(1) all crest value p in the time and trough Value Data v;

p=(p(1),p(2),p(3),...,p(α),...,p(n _p))………………………………（2）

v=(v(1),v(2),v(3),...,v(β),...,v(n _v))………………………………（3）

In formula (2), n _pfor always counting of peak value p, p (α) is for u (t) is at T _b(1): T _e(1) α peak value in the time, α=1,2,3 ..., n _p; In formula (3), n _vfor always counting of peak value v, v (β) is for u (t) is at T _b(1): T _e(1) β valley in the time, β=1,2,3 ..., n _v; Ask for the absolute value of v

|v|=(|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(n _v)|)………………………………（4）

C, order

E=(p(1),p(2),p(3),...,p(α),...,p(n _p),|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(n _v)|)…………（5）

Then E comprises T _b(1): T _e(1) absolute value of all peak values and valley in the time, sorts by numerical values recited to E, obtains E _sort, E _sort=(e (1), e (2), e (3) ..., e (η) ... e (n _p+ n _v)), η is sequence number, η=1,2,3 ..., n _p+ n _v; At E _sortmiddle extracted valid data collection E _e, E _e=(e (η ₁), e (η ₁+ 1), e (η ₁+ 2) ..., e (n _p+ n _v-η ₁)), for rounding up, μ=0.05:0.4, by E _eask for the reference amplitude value of this track data

$\stackrel{\‾}{A}=\frac{1}{n_E}\underset{\mathrm{\ϵ}=1}{\overset{n_E}{\mathrm{\Σ}}}\left|E_e\left(\mathrm{\ϵ}\right)\right|\·\·\·\left(6\right)$

In formula (6), n _efor E _ealways count, ε is sequence number, ε=1,2,3 ..., n _e;

D, definition threshold value wherein M is weights, M>=1;

E, with this threshold value for thresholding, this track data u (t) is compared with thresholding A, according to formula (7), obtain impulsive noise n (t)

$n\left(t\right)=\left\{\begin{array}{cc}u\left(t\right)& \left|u\left(t\right)\right|\&GreaterEqual;A\\ 0& u\left(t\right)<A\end{array}\right.\&CenterDot;\&CenterDot;\&CenterDot;\left(7\right)$

F, ask for the averaged amplitude value W of n (t),

$W=\frac{1}{m}\underset{t=1}{\overset{k}{\mathrm{\&Sigma;}}}\left|n\left(t\right)\right|\&CenterDot;\&CenterDot;\&CenterDot;\left(8\right)$

In formula (8), m is always counting of the middle nonzero element of n (t), and k is the maximum sampling number of u (t);

The mean value W of g, definition impulsive noise and reference amplitude value ratio be pulse-response amplitude than a, the ratio of the sampling number of the number of impulsive noise and this trace record is impulse density q, and specific formula for calculation is as formula (9) and formula (10):

$a=\frac{W}{\stackrel{\&OverBar;}{A}}\&CenterDot;\&CenterDot;\&CenterDot;\left(9\right)$

q=m/k…………………………………（10）

H, by relevant to reference signal for this trace record u (t), obtain relative recording U (t) in this road;

I, n (t) is relevant to reference signal, obtain interference result N (t) of impulsive noise n (t) in vibroseis relative recording;

J, result according to h and i, make difference by relative recording and obtain S (t)=U (t)-N (t), determines useful signal according to S (t) and calculate the signal to noise ratio (S/N ratio) of the useful signal in U (t);

K, foundation U (t), N (t), S (t) image can qualitative analysis signal by noise situation, assess this road acquisition quality, then illustrate that this road acquisition quality is not good when the signal in S (t) is difficult to differentiate in U (t), otherwise, illustrate that acquisition quality is good;

L, can the degree that useful signal be disturbed in the domain of dependence of quantitative test impulsive noise according to the signal to noise ratio (S/N ratio) of pulse-response amplitude useful signal middle than a, impulse density q and U (t); Pulse-response amplitude is larger than a and impulse density q, illustrates that interference is stronger; Signal to noise ratio (S/N ratio) is less than 0, then think that useful signal is difficult to differentiate, acquisition quality is not good, otherwise, illustrate that acquisition quality is good;

All the other each effective recording channel data in m, successively input vibroseis record, according to the process of b to l, can obtain the acquisition quality evaluation of all valid traces and the characteristic parameter of impulse disturbances: pulse-response amplitude is than a and impulse density q;

N, acquisition quality evaluation result according to all valid traces, analyze the overall acquisition quality of these big gun data; The not good number of channels of acquisition quality exceedes effective number of channels then think this big gun defect of data.

Beneficial effect: in conventional Seismic Exploration with Vibrator, if directly record the geological data after being correlated with, the interference that the impulsive noise before relevant causes in relative recording will be difficult to remove, even if be also difficult to compacting in follow-up data processing; If in follow-up seismic data process process, when the noise in discovery vibroseis relative recording is too strong, field can only be got back to and re-start data acquisition, cause acquisition cost significantly to increase.Empirically, in vibroseis seismologic record disclosed by the invention, the analytical technology of impulse noise interference extensively can be adapted to the sweep method of vibroseis linear processes, the noise analysis of Seismic Exploration with Vibrator data acquisition fast and effeciently can be realized at earthquake data acquisition scene, impulsive noise in the actual Seismic Exploration with Vibrator of real-time analysis is on the impact of physical record, avoid because noise level is strong or comprise the not good situation of vibroseis relative recording quality that random high power pulse noise causes, thus reduce costs, effectively ensure that the acquisition quality of vibroseis.

Accompanying drawing illustrates:

The vibroseis seismologic record of Fig. 1 correlation form

The interference that the impulsive noise that Fig. 2 detects causes in the domain of dependence

Fig. 3 removes the vibroseis record after noise

Embodiment:

Be described in further detail below in conjunction with drawings and Examples:

To the interference analysis of the impulsive noise in vibroseis record, the annoyance level of vibroseis record is realized with this impulsive noise of analysis mainly through the impulsive noise detected in vibroseis record.We adopt big gun vibroseis data to illustrate embodiment.If vibroseis record is containing 24 track datas, the 1st road is auxiliary channels, is also reference signal road, and 2-24 road is effective recording channel.Focus sweep time is 5s, sweep frequency 10-100hz, and record length is 6s, and sampling interval is 1ms.

A, input the original vibroseis seismologic record of the irrelevant form of a big gun, according to irrelevant record, the initial time T of the primary wave signal of all valid traces in the effective recording channel of artificial pickup _b, T _b=(0.05,0.1,0.15,0.2,0.25 ..., 0.95,1.0,1.05,1.10,1.15), unit is second (s), n _tr=23, the end time of corresponding i-th valid trace is asked for according to formula (1)

T _e(i)=T _b(i)+t _sweep………………………………（1）

In formula (1), t _sweep=5s, i=1,2,3 ..., 23;

T _e=(5.05,5.1,5.15,5.2,5.25 ..., 5.95,6,6,6,6), unit is second (s);

The irrelevant record of b, vibroseis comprises auxiliary channels and effective recording channel two parts data, effective recording channel is generally containing multiple tracks data, first take out the first data of effective recording channel, be designated as u (t), t is sampling time sequence, t=1,2,3 ..., 6000, all crest value p of this track data u (t) of automatic Picking within the 0.05:5.05s time and trough Value Data v;

p=(p(1),p(2),p(3),...,p(α),...,p(n _p))…………………………（2）

v=(v(1),v(2),v(3),...,v(β),...,v(n _v))……………………………（3）

In formula (2), n _p=456, p (α) is α the peak value of u (t) within the 0.05:5.05s time, α=1,2,3 ..., 456; In formula (3), n _v=454, v (β) is β the valley of u (t) within the 0.05:5.05s time, β=1,2,3 ..., 454; Ask for the absolute value of v

|v|=(|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(454)|)……………………（4）

C, order

E=(p(1),p(2),p(3),...,p(α),...,p(456),|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(454)|)…………（5）

Then E comprises the absolute value of all peak values and valley in the 0.05:5.05s time, sorts by numerical values recited to E, obtains E _sort, E _sort=(e (1), e (2), e (3) ..., e (η) ... e (910)), η is sequence number, η=1,2,3 ..., 910; At E _sortmiddle extracted valid data collection E _e, get μ=0.3, η ₁=273, then E _e=(e (273), e (274), e (275) ..., e (637)), by E _eask for the reference amplitude value of this track data

$\stackrel{\&OverBar;}{A}=\frac{1}{n_E}\underset{\mathrm{\&epsiv;}=1}{\overset{n_E}{\mathrm{\&Sigma;}}}\left|E_e\left(\mathrm{\&epsiv;}\right)\right|\&CenterDot;\&CenterDot;\&CenterDot;\left(6\right)$

Calculate n _e=365, ε is sequence number, ε=1,2,3 ..., 365;

D, definition threshold value wherein M is weights, and in notebook data, M gets 1, then have A=0.99;

E, with this threshold value for thresholding, this track data u (t) is compared with thresholding A=0.99, according to formula (7), obtain impulsive noise n (t)

$n\left(t\right)=\left\{\begin{array}{cc}u\left(t\right)& \left|u\left(t\right)\right|\&GreaterEqual;0.99\\ 0& u\left(t\right)<0.99\end{array}\right.\&CenterDot;\&CenterDot;\&CenterDot;\left(7\right)$

F, ask for the averaged amplitude value W of n (t),

$W=\frac{1}{m}\underset{t=1}{\overset{k}{\mathrm{\&Sigma;}}}\left|n\left(t\right)\right|\&CenterDot;\&CenterDot;\&CenterDot;\left(8\right)$

Calculate W=30.04, m=527, k=6000;

G, by formula (9) and formula (10):

$a=\frac{W}{\stackrel{\&OverBar;}{A}}\&CenterDot;\&CenterDot;\&CenterDot;\left(9\right)$

q=m/k…………………………………（10）

Calculate a=30.24, q=0.0878;

H, by relevant to reference signal for this trace record u (t), obtain relative recording U (t) in this road, in Fig. 1, second is the relative recording in this road;

I, n (t) is relevant to reference signal, obtain interference result N (t) of impulsive noise n (t) in vibroseis relative recording, in Fig. 2, second is the interference that this road impulsive noise causes in the domain of dependence;

J, result according to h and i, relative recording is made difference and obtains S (t)=U (t)-N (t), in Fig. 3, second is S (t) result in this road, determines useful signal according to S (t) and calculates the signal to noise ratio (S/N ratio) of the useful signal in U (t);

K, according to U (t), N (t), S (t) image can qualitative analysis signal by noise situation, assess this road acquisition quality not good;

L, the signal to noise ratio (S/N ratio) calculating this road are-5.70, and think that useful signal is difficult to differentiate, acquisition quality is not good, and the characteristic parameter of this road impulse disturbances is a=30.24, q=0.0878;

All the other each effective recording channel data in m, successively input vibroseis record, according to the process of b to l, can obtain the acquisition quality evaluation of all valid traces and the characteristic parameter of impulse disturbances: pulse-response amplitude is than a and impulse density q; See attached list:

N, acquisition quality evaluation result according to all valid traces, analyze the overall acquisition quality of these big gun data; The not good number of channels of acquisition quality is 7 roads, exceedes effective number of channels think this big gun defect of data.

Claims (1)

1. the analytical approach of impulse noise interference in vibroseis seismologic record, is characterized in that, comprise the following steps:

A, input the original vibroseis seismologic record of the irrelevant form of a big gun, according to irrelevant record, the initial time T of the primary wave signal of all valid traces in the effective recording channel of artificial pickup _b, T _b=(T _b(1), T _b(2), T _b(3) ..., T _b(i) ..., T _b(n _tr)), n _treffective number of channels of this big gun record, T _bi () is the initial time of the i-th valid trace, ask for the end time of corresponding i-th valid trace according to formula (1)

T _e(i)＝T _b(i)+t _sweep………………………………(1)

In formula (1), T _ei () is the end time of the i-th valid trace, t _sweepfor the sweep time of vibroseis, i=1,2,3 ..., n _tr;

If T _ei () is greater than t writing time _recordmaximum moment t _max, then T _e(i)=t _max;

The irrelevant record of b, vibroseis comprises auxiliary channels and effective recording channel two parts data, effective recording channel generally containing multiple tracks data, first takes out the first data of effective recording channel, is designated as u (t), t is sampling time sequence, t=1,2,3, ..., k, k are maximum sampling number, and this track data u (t) of automatic Picking is at T _b(1) ~ T _e(1) all crest value p in the time and trough Value Data v;

p＝(p(1),p(2),p(3),...,p(α),...,p(n _p))…………………………(2)

v＝(v(1),v(2),v(3),...,v(β),...,v(n _v))……………………………(3)

In formula (2), n _pfor always counting of peak value p, p (α) is for u (t) is at T _b(1) ~ T _e(1) α peak value in the time, α=1,2,3 ..., n _p; In formula (3), n _vfor always counting of valley v, v (β) is for u (t) is at T _b(1) ~ T _e(1) β valley in the time, β=1,2,3 ..., n _v; Ask for the absolute value of v

|v|＝(|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(n _v)|)………………………………(4)

C, order:

E＝(p(1),p(2),p(3),...,p(α),...,p(n _p),|v(1)|,|v(2)|,|v(3)|,...,|v(β)|,...,|v(n _v)|)…………(5)

Then E comprises T _b(1) ~ T _e(1) absolute value of all peak values and valley in the time, sorts by numerical values recited to E, obtains E _sort, E _sort=(e (1), e (2), e (3) ..., e (η) ... e (n _p+ n _v)), η is sequence number, η=1,2,3 ..., n _p+ n _v; At E _smiddle extracted valid data collection E _e, for rounding up, μ=0.05 ~ 0.4, by E _eask for the reference amplitude value of this track data

$\stackrel{\&OverBar;}{A}=\frac{1}{n_E}\underset{\mathrm{\&epsiv;}=1}{\overset{n_E}{\mathrm{\&Sigma;}}}\left|E_e\left(\mathrm{\&epsiv;}\right)\right|...\left(6\right)$

In formula (6), n _efor E _ealways count, ε is sequence number, ε=1,2,3 ..., n _e;

D, definition threshold value wherein M is weights, M>=1;

E, with this threshold value for thresholding, this track data u (t) is compared with thresholding A, according to formula (7), obtain impulsive noise n (t)

$n\left(t\right)=\left\{\begin{array}{cc}u\left(t\right)& \left|u\left(t\right)\right|\&GreaterEqual;A\\ 0& u\left(t\right)<A\end{array}\right....\left(7\right)$

F, ask for the averaged amplitude value W of n (t),

$W=\frac{1}{m}\underset{t=1}{\overset{k}{\&Sigma;}}\left|n\left(t\right)\right|...\left(8\right)$

In formula (8), m is always counting of the middle nonzero element of n (t), and k is the maximum sampling number of u (t);

The averaged amplitude value W of g, definition impulsive noise and this track data reference amplitude value ratio be pulse-response amplitude be impulse density q than the ratio of a, m and k, specific formula for calculation is as formula (9) and formula (10):

$a=\frac{W}{\stackrel{\&OverBar;}{A}}...\left(9\right)$

q＝m/k…………………………………(10)

H, by relevant to reference signal for this track data u (t), obtain relative recording U (t) in this road;

I, n (t) is relevant to reference signal, obtain interference result N (t) of impulsive noise n (t) in vibroseis relative recording;

J, result according to h and i, make difference by relative recording and obtain S (t)=U (t)-N (t), determines useful signal and calculate the signal to noise ratio (S/N ratio) of the useful signal in U (t) according to S (t);

K, foundation U (t), N (t), S (t) image qualitative analysis signal is by noise situation, assess this road acquisition quality, then illustrate that this road acquisition quality is not good when the signal in S (t) is difficult to differentiate in U (t), otherwise, illustrate that acquisition quality is good;

The degree disturbing useful signal in the domain of dependence of the signal to noise ratio (S/N ratio) quantitative test impulsive noise of l, foundation pulse-response amplitude useful signal middle than a, impulse density q and U (t); Pulse-response amplitude is larger than a and impulse density q, illustrates that interference is stronger; Signal to noise ratio (S/N ratio) is less than 0, then think that useful signal is difficult to differentiate, acquisition quality is not good, otherwise, illustrate that acquisition quality is good;

All the other each effective recording channel data in m, successively input vibroseis record, according in the process of b to l to the disposal route of first data, process all the other each effective recording channel data, namely obtain the acquisition quality evaluation of all valid traces and the characteristic parameter of impulse disturbances: pulse-response amplitude is than a and impulse density q;

N, acquisition quality according to all valid traces, analyze the overall acquisition quality of these big gun data; The not good number of channels of acquisition quality exceedes effective number of channels then think this big gun defect of data.