CN105700020A - Random noise suppression method and apparatus for seismic data - Google Patents

Abstract

The invention relates to the seismic exploration field, especially to a random noise suppression method and apparatus for seismic data. The method comprises: complementary ensemble empirical mode decomposition (CEEMD) is carried out on seismic data to obtain an intrinsic mode function (IMF) component sequence, and according to an interrelation between seismic data and IMF components, IMF components including random noises and IMF components not including random noises are determined; on the basis of differences of the random noises included by the IMF components including random noises, different threshold values are selected for the IMF components including random noises to carry out optimal curvelet iteration threshold de-noising processing, thereby obtaining the processed IMF components; and then seismic data after noise removing are obtained by reconstruction. According to the method provided by the embodiment of the application, on the basis of combination of the CEEMD and the optimal curvelet iteration threshold method, defects of effective signal loss due to an EMD method and de-noising method and single threshold selection because of a curvelet threshold de-noising method can be overcome. The effective signal can be kept well while random noises are suppressed, so that the signal to noise ratio of seismic data is improved.

Description

A kind of geological data stochastic noise suppression method and device

Technical field

The present invention relates to field of seismic exploration, particularly relate to a kind of geological data stochastic noise suppression method and device。

Background technology

By the feature that noise occurs on seismic profile, noise is divided into organized noise and random noise。Organized noise is primarily referred to as certain dominant frequency and the noise of certain apparent velocity, such as face ripple, alternating current disturbance, sound wave, shallow refraction etc.。Random noise and random noise, refer mainly to the ripple not having fixed frequency and the fixing direction of propagation, forms rambling background in the seismic data。It is very wide that random noise shows as frequency on earthquake record, without certain apparent velocity, thus is difficult with random noise and with the difference on frequency spectrum between significant wave or the difference on the direction of propagation, it is suppressed。Due to the complexity of surface conditions, often containing many random noises in seismic data, such as microseism, ambient interferences etc.。These noise profile are very wide, had a strong impact on the signal to noise ratio of seismic data。

In field of seismic exploration, the research of the random stochastic noise suppression method of geological data is constantly subjected to the extensive concern of related researcher。Current geological data stochastic noise suppression method has Fourier transform filter method, independent component analysis method and two-dimensional wavelet transformation method etc.。Fourier transform filter method needs putative signal to be stable, and is difficult to characterize the local features of signal, and when signal non-stationary, the method is difficult to reach desirable effect；Independent component analysis method requires that seismic signal and random noise are statistically separate, but the method tentation data road is the repeatedly observation to phase people having a common goal, and actual seismic track data be unsatisfactory for this hypothesis, indicating a kind of very rough being similar to, therefore it is unsatisfactory to the compacting ability of random noise。Two-dimensional wavelet transformation method is widely used in seismic data process, but owing to its basic function is isotropic, when processing geological data, the local maximum of conversion coefficient can only reflect that the position that this wavelet coefficient occurs is " mistake " edge, and the information at " edge " edge that is beyond expression, therefore image border data are processed unsatisfactory by two-dimensional wavelet transformation method。

Along with the development of technology, exploration targets shifts to depths and complex area, earth's surface gradually, and seismic data process requires also more and more higher, therefore, how effectively to suppress noise, thus the signal to noise ratio improving seismic data is the problem needing solution in field of seismic exploration badly。

Summary of the invention

The embodiment of the present application provides a kind of geological data stochastic noise suppression method and device, to keep useful signal, suppresses noise, thus improving the signal to noise ratio of seismic data。

For reaching above-mentioned purpose, on the one hand, the embodiment of the present application provides a kind of geological data stochastic noise suppression method, comprises the following steps:

The geological data obtained is carried out complementary experience state and decomposes CEEMD, it is thus achieved that intrinsic mode functions IMF vector sequence；

Determine the IMF component containing random noise in described IMF vector sequence and do not contain the IMF component of random noise；

To the described IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process；

Utilize the IMF component after described process and the described IMF component not containing random noise to be reconstructed and obtain the geological data removing random noise。

The geological data stochastic noise suppression method of the embodiment of the present application, described to the IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process, including:

Difference according to random noise contained in the described IMF component containing random noise, chooses different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process to the described IMF component containing random noise。

The geological data stochastic noise suppression method of the embodiment of the present application, described determines in IMF vector sequence containing noisy IMF component with without noisy IMF component, comprises the following steps:

Calculate the cross-correlation coefficient between described geological data and described IMF vector sequence, it is thus achieved that cross correlation Number Sequence；

Determine the minimum in described cross correlation Number Sequence；

Determine the IMF component containing random noise according to the IMF component positions that described minimum is corresponding and do not contain the IMF component of random noise。

The geological data stochastic noise suppression method of the embodiment of the present application, the cross-correlation coefficient between described calculating geological data and described IMF component, concrete formula is:

$R\left(j\right)=\frac{R_{data,{\mathrm{IMF}}_j}\left(0\right)}{R_{data}\left(0\right)R_{{\mathrm{IMF}}_j}\left(0\right)}$

Wherein, R (j) is cross-correlation coefficient, R_data(0) maximal peak point corresponding to zero moment after geological data auto-correlation,For maximal peak point corresponding to zero moment after the jth IMF auto-correlation separated,For the value corresponding to zero moment after initial data and jth IMF component cross-correlation。

The geological data stochastic noise suppression method of the embodiment of the present application, described utilization process after IMF component and the described IMF component contain random noise be reconstructed the computing formula of geological data obtaining removal random noise and be:

$X^{^}=\underset{m=1}{\overset{k}{\Σ}}{C}_m^{^}+\underset{n=k+1}{\overset{n}{\Σ}}{C}_n$

C^_mFor the IMF component after processing；C_nFor not containing the IMF component of random noise, wherein n is the CEEMD number decomposing the IMF component obtained；X^ is the geological data after removing noise；K is the IMF component positions that minimum is corresponding。

On the other hand, the embodiment of the present application also provides for a kind of geological data random noise pressure setting, including:

CEEMD resolving cell, decomposes CEEMD for the geological data obtained carries out complementary experience state, it is thus achieved that intrinsic mode functions IMF vector sequence；

IMF component analysis unit, for determining the IMF component containing random noise in described IMF vector sequence and not containing the IMF component of random noise；

Denoising unit, for the described IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process；

Reconfiguration unit, is reconstructed for the IMF component after utilizing described process and the described IMF component not containing random noise and obtains the geological data removing random noise。

The geological data random noise pressure setting of the embodiment of the present application, described to the IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process, including:

Difference according to random noise contained in the described IMF component containing random noise, chooses different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process to the described IMF component containing random noise。

The geological data random noise pressure setting of the embodiment of the present application, described IMF component analysis unit includes:

Cross-correlation coefficient computation subunit, for calculating the cross-correlation coefficient between described geological data and described IMF vector sequence, it is thus achieved that cross correlation Number Sequence；

Minimum determines subelement, for determining the minimum in described cross correlation Number Sequence；

Noise analysis subelement, for determining the IMF component containing random noise according to the IMF component positions that described minimum is corresponding and not containing the IMF component of random noise。

The geological data random noise pressure setting of the embodiment of the present application, the cross-correlation coefficient between described calculating geological data and described IMF component, concrete formula is:

$R\left(j\right)=\frac{R_{data,{\mathrm{IMF}}_j}\left(0\right)}{R_{data}\left(0\right)R_{{\mathrm{IMF}}_j}\left(0\right)}$

Wherein, R (j) is cross-correlation coefficient, R_data(0) maximal peak point corresponding to zero moment after geological data auto-correlation,For maximal peak point corresponding to zero moment after the jth IMF auto-correlation separated,For the value corresponding to zero moment after initial data and jth IMF component cross-correlation。

The geological data random noise pressure setting of the embodiment of the present application, described utilization process after IMF component and the described IMF component contain random noise be reconstructed the computing formula of geological data obtaining removal random noise and be:

$X^{^}=\underset{m=1}{\overset{k}{\Σ}}{C}_m^{^}+\underset{n=k+1}{\overset{n}{\Σ}}{C}_m$

C^_mFor the IMF component after processing；C_nFor not containing the IMF component of random noise, wherein n is the CEEMD number decomposing the IMF component obtained；X^ is the geological data after removing noise；K is the IMF component positions that minimum is corresponding。

In the embodiment of the present application, decomposing acquisition IMF vector sequence by geological data being carried out CEEMD, then passing through the cross-correlation coefficient between geological data and IMF component and determine the IMF component containing random noise and do not contain the IMF component of random noise；Further according to the difference of random noise contained in the IMF component containing random noise, the described IMF component containing random noise is chosen different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process；The geological data removing noise is obtained finally by reconstruct。The method of the embodiment of the present application is decomposed and the combination of bent ripple optimum iteration method by CEEMD, overcome and choose single defect based on the loss of EMD method denoising method useful signal and curvelet threshold value denoising method threshold value, while Attenuating Random Noise, can better keep useful signal, thus improving the signal to noise ratio of geological data。

Accompanying drawing explanation

In order to be illustrated more clearly that the embodiment of the present application or technical scheme of the prior art, the accompanying drawing used required in embodiment or description of the prior art will be briefly described below, apparently, the accompanying drawing that the following describes is only some embodiments recorded in the application, for those of ordinary skill in the art, under the premise not paying creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings。

Fig. 1 is the geological data stochastic noise suppression method schematic diagram of the embodiment of the present application；

Fig. 2 is the geological data random noise pressure setting structural representation of the embodiment of the present application；

Fig. 3 (a)～3 (e) is the analog data denoising effect comparison diagram of the application one embodiment；

Fig. 4 (a)～4 (d) is the actual seismic data de-noising effect contrast figure of the application one embodiment。

Detailed description of the invention

In order to make those skilled in the art be more fully understood that the technical scheme in the application, below in conjunction with the accompanying drawing in the embodiment of the present application, technical scheme in the embodiment of the present application is clearly and completely described, obviously, described embodiment is only some embodiments of the present application, rather than whole embodiments。Based on the embodiment in the application, the every other embodiment that those of ordinary skill in the art obtain under not making creative work premise, all should belong to the scope of the application protection。

Below in conjunction with accompanying drawing, the detailed description of the invention of the embodiment of the present application is described in further detail。

With reference to Fig. 1, the geological data stochastic noise suppression method of the embodiment of the present application, comprise the following steps:

S1, the geological data obtained is carried out complementary experience state decompose CEEMD, it is thus achieved that intrinsic mode functions IMF vector sequence。

Empirical mode decomposition (empiricalmodedecomposition, EDM) technology is a kind of self adaptation signal processing method without any prior information, utilize EMD that signal decomposition can become a series of intrinsic mode functions (intrinsicmodefunction, IMF)。Due to the comprised frequency difference of each order component, it is possible to select different components to be reconstructed according to actual needs, to improve the signal to noise ratio of data；Therefore the screening process of IMF has multi-resolution characteristics。Although EMD has a lot of advantages, but causes modal overlap phenomenon owing to it decomposes unstable。Complementary set empirical mode decomposition (complementaryensembleempiricalmodedecomposition, CEEMD) it is that EMD algorithm has been improved, the aid in noise added adopts positive and negative paired form, this makes it possible to the remaining aid in noise eliminating in reconstruction signal well, and the noise set number of times added can be very low, computational efficiency is higher。

The geological data obtained is carried out CEEMD decomposition by the embodiment of the present application, it is thus achieved that intrinsic mode functions IMF vector sequence, specifically includes:

(1) adding n group auxiliary white noise in primary signal, aid in noise is to add in the way of positive and negative pair, thus generating two set IMF component set:

$\left[\begin{array}{c}{M}_1\\ M_2\end{array}\right]=\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]\left[\begin{array}{c}data\\ N\end{array}\right]$

Wherein, data is the geological data obtained, and N is aid in noise, M₁And M₂Respectively adding the signal after positive and negative paired noise, the number of the aggregate signal therefore obtained becomes 2n。

(2) each signal in aggregate signal being carried out EMD decomposition, each signal obtains one group of IMF component, and wherein the jth IMF representation in components of i-th signal is IMF_ij。

(3) mode combined by multicomponent amount obtains the decomposition result of CEEMD:

${\mathrm{IMF}}_j=\frac{1}{2n}\underset{i=1}{\overset{2n}{\Σ}}{\mathrm{IMF}}_{ij}$

Wherein, IMF_jRepresent that CEEMD decomposes the jth IMF component finally given。

The embodiment of the present application utilizes CEEMD to decompose, non-stationary seismic signal in geological data is decomposed into the stationary signal of different frequency scope, avoid and cause modal overlap phenomenon owing to EMD decomposes unstable, overcome and lose based on EMD method denoising useful signal, thus laying a good foundation for the compacting of follow-up random noise。Meanwhile, in the embodiment of the present application geological data after CEEMD decomposes, it is thus achieved that IMF vector sequence arranged evenly from high frequency to low frequency, namely frequency content reduces successively, and the proportion that simultaneously noise is shared in each IMF component reduces, and useful signal strengthens gradually。

S2, determine in described IMF vector sequence the IMF component containing random noise and do not contain the IMF component of random noise。

In the embodiment of the present application, it is possible to determine the IMF component containing random noise according to the cross-correlation coefficient of described geological data and IMF vector sequence and do not contain the IMF component of random noise, specifically including following steps:

Calculate the cross-correlation coefficient between described geological data and described IMF vector sequence, it is thus achieved that cross correlation Number Sequence；

Determine the local minimum in described cross correlation Number Sequence；

Determine the IMF component containing random noise according to the IMF component positions that described local minimum is corresponding and do not contain the IMF component of random noise。

In the embodiment of the present application, the cross-correlation coefficient computing formula between geological data and IMF vector sequence is:

$R\left(j\right)=\frac{R_{data,{\mathrm{IMF}}_j}\left(0\right)}{R_{data}\left(0\right)R_{{\mathrm{IMF}}_j}\left(0\right)}$

Wherein, R (j) is cross-correlation coefficient, R_data(0) maximal peak point corresponding to zero moment after geological data auto-correlation,For maximal peak point corresponding to zero moment after the jth IMF auto-correlation separated,Corresponding to zero moment after initial data and jth IMF component cross-correlation value。

By asking for the cross-correlation coefficient between each IMF component and original earthquake data, it is hereby achieved that cross correlation Number Sequence。

The embodiment of the present application is decomposed, by CEEMD, the IMF vector sequence frequency content obtained and is reduced successively, noise proportion shared by IMF component is gradually lowered, useful signal strengthens gradually, therefore the feature being gradually lowered according to noise proportion in IMF vector sequence, it is possible to judge the IMF component containing random noise by the cross-correlation coefficient between geological data and IMF component。The embodiment of the present application judges to need the IMF component carrying out random noise compacting by the minimum point in cross correlation Number Sequence, first the cross-correlation coefficient of geological data and each IMF component is asked for, find the minimum in cross correlation Number Sequence, the IMF component positions k of correspondence is determined according to minimum, front k IMF component is the IMF component many containing random noise, it is the IMF component containing random noise, need to carry out denoising, k IMF component below in position is do not contain random noise or the little IMF component of random noise, it is not necessary to carry out denoising。The embodiment of the present application utilize noise proportion in IMF vector sequence be gradually lowered, the feature that useful signal strengthens gradually, the IMF component containing random noise is determined by the minimum of the cross-correlation coefficient between geological data and IMF component, thus only the IMF component containing random noise being carried out denoising, improve denoising efficiency。

S3, the described IMF component containing random noise is carried out curvelet threshold value optimum iterated denoising process, it is thus achieved that the IMF component after process。

To containing noisy IMF component IMF_iUseful signal s and the sum of noise n ' can be expressed as, it may be assumed that

IMF_i=s+n ' (i=1 ... k)

Theoretical according to sparse transformation denoising, the useful signal of signals and associated noises is as the sparse composition of this signal, the composition that namely coefficient is bigger, and noise is then that signals and associated noises is removed the remainder that sparse part obtains namely the part that coefficient is less。In the embodiment of the present application, useful signal s has relatively Daqu (massive raw stater for alcholic liquor) wave system number in warp wavelet territory, and the bent wave system number of noise is less。

The structure of geological data has very strong geometry character and regularity, the features such as multi-direction, multiple dimensioned and anisotropy are possessed due to warp wavelet, so warp wavelet can automatically detect the position before seismic wave and direction, the optimum singularity characteristics expressing geological data curve；Meanwhile, geological data is very sparse in the expression of bent wave zone, just can capture the principal character of geological data with little Qu Bo, form big projection coefficient on these Qu Bo, and noise then forms little projection coefficient。The embodiment of the present application utilizes warp wavelet sparse representation, utilizes bent ripple optimum iteration threshold method the IMF component containing random noise can be carried out denoising, specifically include:

First it is make use of bent this openness of wave system number that noisy seismic signal is carried out sparse expression, for the given IMF component IMF containing random noise_i, and one group of sparse transformation base A, IMF_iCan by sparse representation, thus IMF component denoising is equivalent to solve IMF_iThe such indirect problem of coefficient vector x of bent wave zone in=Ax, according to solving the optimization function that coefficient vector x problem can build under following constraints:

$P_{\mathrm{\ϵ}}=\left\{\begin{array}{c}\begin{array}{ccc}\hat{x}={\min }_x{||x||}_1& s.t.& {||y-Ax||}_2\≤\mathrm{\ϵ}\end{array}\\ \hat{s}=A\hat{x}\end{array}\right.$

Wherein, P_εFor penalty, x is the coefficient vector of bent wave zone,For estimating effective curve wave system number vector,For the useful signal recovered, A be the song ripple inverse transformation factor, ε be can the maximum of allowable error, y is containing noisy IMF component。The above-mentioned optimization function having under constraints can be converted into the Regularization function solved under a unconfined condition by minimizing strategy:

Wherein, λ is regularization coefficient。

Then can using and based on the Landweber iteration threshold method declined, the Regularization function under above-mentioned unconfined condition be solved, its corresponding iterative formula is:

x^m+1=T_λ[x^m+A^T(y-Ax^m)]

T_λ=sgn (x) max (0, | x |-| t |)

Wherein, x is the coefficient vector of bent wave zone, x^m+1It is the m+1 time iteration result, T_λFor threshold function table, t is threshold value, A^TFor the bent ripple direct transform factor, y is containing noisy IMF component。

In the embodiment of the present application, the selection of threshold value is very big on denoising effect impact, and the excessive loss being easily caused useful signal of threshold value, threshold value is too small, can not effectively remove noise。In the embodiment of the present application, it is possible to the difference according to random noise contained in the described IMF component containing random noise, the described IMF component containing random noise is chosen different threshold value march ripple optimum iteration threshold and carries out denoising。For example, it is possible to the size according to contained random noise chooses corresponding threshold value, thus overcoming threshold value in curvelet threshold value denoising to choose single defect, reduce the loss of useful signal, thus reaching the denoising effect of the best。In the embodiment of the present application, first the size according to random noise contained in the described IMF component containing random noise chooses different threshold values, then threshold value is substituted in above-mentioned iterative formula and be iterated solving, leave big bent wave system number, namely retain the coefficient vector of the useful signal expressing seismic wave information；Remove little bent wave system number, namely remove the bent wave system number vector that those irregular noises are corresponding。Finally treated bent wave system number is carried out the seismic signal that contrary flexure wave conversion obtainsIt is the IMF component removing random noise, the IMF component after namely processing。

In another embodiment of the application, it is possible to use the Regularization function under above-mentioned unconfined condition is solved by base tracing algorithm (BP), matching pursuit algorithm (MP), orthogonal matching pursuit algorithm (OMP)。

In another embodiment of the application, it is possible to use the IMF component containing random noise is carried out denoising by fourier transform method, discrete cosine transform method (DiscreteCosine), Wavelet Transform。

The bent ripple optimum iteration threshold denoising of the embodiment of the present application is big first with the bent wave system number of useful signal, the feature that the bent wave system number of noise is little, by warp wavelet method, the IMF component containing random noise is carried out sparse expression, then the size according to random noise contained in the IMF component containing random noise chooses different threshold values, the coefficient that namely big bent wave system number expresses the useful signal of seismic wave information is retained by iteration method, remove little bent wave system number, obtain the IMF component removing noise finally by contrary flexure wave conversion。The embodiment of the present application overcomes curvelet threshold value noise-removed threshold value and chooses single defect, simultaneously because select different threshold values according to the size containing random noise, while effective Attenuating Random Noise, it is also possible to better keep useful signal。

S4, utilize described process after IMF component and the described IMF component without random noise be reconstructed obtain remove noise geological data。

Application embodiment is reconstructed the geological data obtaining removing noise according to following formula:

$X^{^}=\underset{m=1}{\overset{k}{\Σ}}{C}_m^{^}+\underset{n=k+1}{\overset{n}{\Σ}}{C}_n$

C^_mFor the IMF component after processing；C_nFor without noisy IMF component, wherein n is the CEEMD number decomposing the IMF component obtained；X^ is the geological data after removing noise；K is the IMF component positions that local minimum is corresponding。

In the embodiment of the present application, decomposing acquisition IMF vector sequence by geological data being carried out CEEMD, then passing through the cross-correlation coefficient between geological data and IMF component and determine the IMF component containing random noise and do not contain the IMF component of random noise；Further according to the difference of random noise contained in the IMF component containing random noise, the described IMF component containing random noise is chosen different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process；The geological data removing noise is obtained finally by reconstruct。The method of the embodiment of the present application is decomposed and the combination of bent ripple optimum iteration method by CEEMD, overcome and choose single defect based on the loss of EMD method denoising method useful signal and curvelet threshold value denoising method threshold value, while Attenuating Random Noise, can better keep useful signal, thus improving the signal to noise ratio of geological data。

In order to the beneficial effect of the embodiment of the present application is clearly described, illustrate below in conjunction with accompanying drawing:

If Fig. 3 (a)～3 (e) is analog data denoising effect comparison diagram, this analog data is 50 roads laterally, are longitudinally 400 sampled points, and the sampling interval is 2ms。Fig. 3 (a) is noise free data, Fig. 3 (b) is the data after the stronger white Gaussian noise of addition, Fig. 3 (c) is the result after the data after adding noise are carried out denoising by the method based on EMD, Fig. 3 (d) is for Deconvolution to adding the result after the data after noise carry out denoising, and the data after adding noise are carried out the result after denoising by the method that Fig. 3 (e) is the embodiment of the present application。Table 1 is the contrast of signal to noise ratio after denoising, Y-PSNR, noise mean square deviation and noise removal capability index, and as can be seen from the table, after the method denoising of the embodiment of the present application, signal to noise ratio is the highest, and effective information loss is minimum, and noise removal capability is the highest；Contrast it can be seen that the denoising effect of the embodiment of the present application is best from Fig. 3 (a)～3 (e)。

The table of 1 three kinds of method denoising result evaluation index contrasts of table

	EMD denoising method	The Method of Deconvolution	The embodiment of the present application method
				Signal to noise ratio	0.55743	1.5095	3.0523
Y-PSNR/db	71.3761	75.7025	78.7605
				Noise energy mean square deviation	0.0047367	0.0017492	0.00086503
Noise removal capability	0.68684	1.1581	1.502

Wherein, every noise-removed technology index computing formula:

Signal to noise ratio:

Y-PSNR:

Noise energy mean square deviation: N_NR=(s₁-s)²/MN

Noise removal capability:

Wherein, s is for without making an uproar data, and y is the data after adding noise, s₁For the data after denoising, the columns of M and N respectively data and line number, S_NR0For original noisy data SNR。

If Fig. 4 (a)～4 (d) is a certain area actual seismic data de-noising effect contrast figure, wherein Fig. 4 (a) is pending geological data section, Fig. 4 (b) is that the method based on EMD is to the result after seismic data noise attenuation, Fig. 4 (c) for Deconvolution, seismic data noise attenuation is processed after result, method seismic data noise attenuation that Fig. 4 (d) is the embodiment of the present application process after result。It can be seen that have lost high frequency useful signal based on the denoising result of EMD method, and some distortion of part area data；Deconvolution can protect useful signal, but can only press portion random noise, on section, noise residual is more；The method denoising result signal to noise ratio of the embodiment of the present application is the highest, and on section, noise residual is less, and lineups are more continuous, and denoising effect is best。

With reference to Fig. 2, corresponding with above-mentioned geological data stochastic noise suppression method, the geological data random noise pressure setting of the embodiment of the present application, including:

CEEMD resolving cell 21, decomposes CEEMD for the geological data obtained carries out complementary experience state, it is thus achieved that intrinsic mode functions IMF vector sequence；

IMF component analysis unit 22, for determining the IMF component containing random noise in described IMF vector sequence and not containing the IMF component of random noise；

Denoising unit 23, for the described IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process；

Reconfiguration unit 24, is reconstructed for the IMF component after utilizing described process and the described IMF component not containing random noise and obtains the geological data removing random noise。

Each ingredient of the device of the present embodiment is respectively used to realize each step of the method for previous embodiment, owing to, in embodiment of the method, each step being described in detail, not repeated them here。

In the embodiment of the present application, decomposing acquisition IMF vector sequence by geological data being carried out CEEMD, then passing through the cross-correlation coefficient between geological data and IMF component and determine the IMF component containing random noise and do not contain the IMF component of random noise；Further according to the difference of random noise contained in the IMF component containing random noise, the described IMF component containing random noise is chosen different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process；The geological data removing noise is obtained finally by reconstruct。The method of the embodiment of the present application is decomposed and the combination of bent ripple optimum iteration method by CEEMD, overcome and choose single defect based on the loss of EMD method denoising method useful signal and curvelet threshold value denoising method threshold value, while Attenuating Random Noise, can better keep useful signal, thus improving the signal to noise ratio of geological data。

In one or more exemplary designs, the above-mentioned functions described by the embodiment of the present application can realize in the combination in any of hardware, software, firmware or this three。If realized in software, these functions can store and on the medium of computer-readable, or be transmitted on the medium of computer-readable with one or more instructions or code form。Computer readable medium includes computer storage medium and is easy to so that allowing computer program transfer to the telecommunication media in other place from a place。Storage medium can be that any general or special computer can the useable medium of access。Such as, such computer readable media can include but not limited to RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage device, or other any may be used for carrying or storage with instruction or data structure and other can be read the medium of program code of form by general or special computer or general or special processor。

Particular embodiments described above; the purpose of the application, technical scheme and beneficial effect have been further described; it is it should be understood that; the foregoing is only the specific embodiment of the embodiment of the present application; it is not used to limit the protection domain of the application; all within spirit herein and principle, any amendment of making, equivalent replacement, improvement etc., should be included within the protection domain of the application。

Claims (10)

1. a geological data stochastic noise suppression method, it is characterised in that comprise the following steps:

The geological data obtained is carried out complementary experience state and decomposes CEEMD, it is thus achieved that intrinsic mode functions IMF vector sequence；

Determine the IMF component containing random noise in described IMF vector sequence and do not contain the IMF component of random noise；

To the described IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process；

Utilize the IMF component after described process and the described IMF component not containing random noise to be reconstructed and obtain the geological data removing random noise。

2. geological data stochastic noise suppression method as claimed in claim 1, it is characterised in that described to the IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process, including:

Difference according to random noise contained in the described IMF component containing random noise, chooses different threshold value march ripple optimum iteration threshold denoisings, it is thus achieved that the IMF component after process to the described IMF component containing random noise。

3. geological data stochastic noise suppression method as claimed in claim 1, it is characterised in that described determine in IMF vector sequence containing noisy IMF component with without noisy IMF component, comprises the following steps:

Calculate the cross-correlation coefficient between described geological data and described IMF vector sequence, it is thus achieved that cross correlation Number Sequence；

Determine the minimum in described cross correlation Number Sequence；

Determine the IMF component containing random noise according to the IMF component positions that described minimum is corresponding and do not contain the IMF component of random noise。

4. geological data stochastic noise suppression method as claimed in claim 3, it is characterised in that the cross-correlation coefficient between described calculating geological data and described IMF component, concrete formula is:

$R\left(j\right)=\frac{R_{data,{\mathrm{IMF}}_j}\left(0\right)}{R_{data}\left(0\right)R_{{\mathrm{IMF}}_j}\left(0\right)}$

Wherein, R (j) is cross-correlation coefficient, R_data(0) maximal peak point corresponding to zero moment after geological data auto-correlation,For maximal peak point corresponding to zero moment after the jth IMF auto-correlation separated,For the value corresponding to zero moment after initial data and jth IMF component cross-correlation。

5. geological data stochastic noise suppression method as claimed in claim 1, it is characterised in that described utilization process after IMF component and the described IMF component contain random noise be reconstructed the computing formula of the geological data obtaining removal random noise and be:

$X^{^}=\underset{m=1}{\overset{k}{\Σ}}{C}_m^{^}+\underset{n=k+1}{\overset{n}{\Σ}}{C}_n$

C^{^} _mFor the IMF component after processing；C_nFor not containing the IMF component of random noise, wherein n is the CEEMD number decomposing the IMF component obtained；X^{^}For the geological data after removal noise；K is the IMF component positions that minimum is corresponding。

6. a geological data random noise pressure setting, it is characterised in that including:

CEEMD resolving cell, decomposes CEEMD for the geological data obtained carries out complementary experience state, it is thus achieved that intrinsic mode functions IMF vector sequence；

IMF component analysis unit, for determining the IMF component containing random noise in described IMF vector sequence and not containing the IMF component of random noise；

Denoising unit, for the described IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process；

Reconfiguration unit, is reconstructed for the IMF component after utilizing described process and the described IMF component not containing random noise and obtains the geological data removing random noise。

7. geological data random noise pressure setting as claimed in claim 6, it is characterised in that described to the IMF component march ripple optimum iteration threshold denoising containing random noise, it is thus achieved that the IMF component after process, including:

Difference according to random noise contained in the described IMF component containing random noise, chooses different threshold value march ripple optimum iteration threshold and carries out denoising, it is thus achieved that the IMF component after process the described IMF component containing random noise。

8. geological data random noise pressure setting as claimed in claim 6, it is characterised in that described IMF component analysis unit includes:

Cross-correlation coefficient computation subunit, for calculating the cross-correlation coefficient between described geological data and described IMF vector sequence, it is thus achieved that cross correlation Number Sequence；

Minimum determines subelement, for determining the minimum in described cross correlation Number Sequence；

Noise analysis subelement, for determining the IMF component containing random noise according to the IMF component positions that described minimum is corresponding and not containing the IMF component of random noise。

9. geological data random noise pressure setting as claimed in claim 8, it is characterised in that the cross-correlation coefficient between described calculating geological data and described IMF component, concrete formula is:

$R\left(j\right)=\frac{R_{data,{\mathrm{IMF}}_j}\left(0\right)}{R_{data}\left(0\right)R_{{\mathrm{IMF}}_j}\left(0\right)}$

Wherein, R (j) is cross-correlation coefficient, R_data(0) maximal peak point corresponding to zero moment after geological data auto-correlation,For maximal peak point corresponding to zero moment after the jth IMF auto-correlation separated,For the value corresponding to zero moment after initial data and jth IMF component cross-correlation。

10. geological data random noise pressure setting as claimed in claim 6, it is characterised in that described utilization process after IMF component and the described IMF component contain random noise be reconstructed the computing formula of the geological data obtaining removal random noise and be:

$X^{^}=\underset{m=1}{\overset{k}{\Σ}}{C}_m^{^}+\underset{n=k+1}{\overset{n}{\Σ}}{C}_n$

C^{^} _mFor the IMF component after processing；C_nFor not containing the IMF component of random noise, wherein n is the CEEMD number decomposing the IMF component obtained；X^{^}For the geological data after removal noise；K is the IMF component positions that minimum is corresponding。