Seismic wavelet and noise simultaneous estimation method and device

Abstract

The invention relates to a seismic wavelet and noise simultaneous estimation method and device, and the method comprises the steps: obtaining a first forward model of a reflection coefficient and seismic data, and a second forward model of seismic wavelets and seismic data, and building a probability graph model according to the relation between random variables in the first forward model and the second forward model; establishing a prior model based on the variables with the interdependency relationship being deterministic dependency in the probability graph model, and establishing a posterior model based on the variables with the interdependency relationship being probabilistic dependency in the probability graph model; logging data and seismological observation data are obtained, and simultaneous iteration calculation is performed on the seismic wavelets and the seismic noise through a Gibbs algorithm by using the logging data, the seismological observation data, the prior model and the posterior model to obtain a target seismic wavelet and noise sequence. According to the method, the coupling between the seismic wavelets and the seismic noise is considered, the seismic wavelets and the noise are iteratively solved at the same time through the Gibbs algorithm, and the seismic wavelet extraction precision is improved.

Description

Technical Field

[0001] The invention relates to a method and device for simultaneously estimating seismic wavelet and noise. Background Art

[0002] Seismic wavelets play an important role in seismic data processing, forward modeling and inversion, and are indispensable input information. At the same time, geophysical observation data usually contain strong noise and measurement errors, which makes the estimation of model parameters have great uncertainty. In the solution of inverse problems in the Bayesian theory system, the noise level has a significant impact on the stability and accuracy of the inversion results. When the noise level is estimated too high, the inversion result is close to the prior initial model, resulting in an overly smooth solution; when the noise level is estimated too low or approaches zero, the constraint of the prior information is reduced, and the inversion result is completely dependent on the seismic data. At this time, it corresponds to the solution of an ill-conditioned inverse problem, and the inversion result is unstable. In fact, the estimation of seismic wavelets and the estimation of seismic noise levels are mutually coupled, that is, the wavelets will affect the estimation of the seismic noise level, and conversely, the seismic noise level will also affect the estimation of the seismic wavelets, so it is necessary to consider both at the same time. Summary of the invention

[0003] In order to enrich the types of seismic wavelet and seismic noise estimation methods in the prior art, the present invention proposes a method and device for simultaneously estimating seismic wavelet and noise. The technical solution proposed by the present invention is as follows:

[0004] In a first aspect, the present invention provides a method for simultaneously estimating seismic wavelet and noise, comprising:

[0005] Obtain a first forward model of reflection coefficient and seismic data, and a second forward model of seismic wavelet and seismic data, and establish a probability graph model for seismic wavelet and seismic noise estimation based on the relationship between the random variables in the first forward model and the random variables in the second forward model; wherein the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms;

[0006] Establishing a priori model based on variables whose interdependence in the probabilistic graphical model is deterministic dependency, and establishing a posterior model based on variables whose interdependence in the probabilistic graphical model is probabilistic dependency;

[0007] The well logging data and the seismic observation data are acquired, and the seismic wavelet and the seismic noise are simultaneously iterated and obtained by the Gibbs algorithm using the well logging data, the seismic observation data, the prior model and the posterior model to obtain the target seismic wavelet and noise sequence.

[0008] In some embodiments, the logging measurement error term includes a reflection coefficient error term, the logging data includes a reflection coefficient sequence calculated from a logging curve, and the seismic observation data includes wellside seismic trace data; the method further includes: establishing a joint probability density function of all variables in the probability graph model according to a probability graph model calculation rule:

[0009] p(e r ,μ w ,w,Σ w ,Σ d ,r obs ,d obs )=p(e r )·p(μ w )·p(Σ w )·p(Σ d )·p(d obs |r obs ,e r ,w,Σ d )·p(w|μ w ,Σ w )

[0010] Among them, d obs =[d1,d2,…,d n ] T is the seismic trace data near the well, r obs =[r1,r2,…,r n ] T is the reflection coefficient sequence calculated from the logging curve, r is the reflection coefficient, w=[w1,w2,…,w m ] T is the seismic wavelet to be estimated, m and n are the seismic wavelet dimension and the number of seismic trace sampling points respectively, e r is the reflection coefficient error term, Σ d is the seismic noise e d The prior covariance matrix of r is the covariance matrix of the reflection coefficient error term, μ w is the wavelet expectation vector, Σ w is the wavelet prior covariance matrix.

[0011] In some embodiments, the seismic data is wellside seismic trace data, and the well logging data is a reflection coefficient sequence calculated from a well logging curve;

[0012] The method uses the well logging data, the seismic observation data, the prior model and the posterior model to iterate the seismic wavelet and the seismic noise simultaneously through the Gibbs algorithm to obtain the target seismic wavelet and noise sequence, including:

[0013] Randomly generate initial seismic wavelet w (0);

[0014] The reflection coefficient sequence r calculated according to the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance

[0015] According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance Determining a new reflection coefficient error term through the posterior model, and updating the reflection coefficient error term based on the new reflection coefficient error term to obtain an updated reflection coefficient error term;

[0016] determining a new reflection coefficient based on the updated reflection coefficient error term;

[0017] According to the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance

[0018] According to the conditional posterior expectation of the seismic wavelet and the conditional posterior covariance Determine a new seismic wavelet by using the a posteriori model;

[0019] Based on the new seismic wavelet, the seismic wavelet is updated to obtain an updated seismic wavelet, and the updated seismic wavelet and the new reflection coefficient are convolved to obtain new seismic noise;

[0020] Based on the new seismic noise, the seismic noise is updated to obtain updated seismic noise, and the covariance matrix Σ of the seismic noise is determined by the prior model according to the updated seismic noise. d ;

[0021] Based on the covariance matrix Σ of the updated seismic wavelet and the seismic noise d Repeat the above process of updating the reflection coefficient error term, seismic wavelet and seismic noise until the maximum number of iterations k is reached max , output the final seismic wavelet and seismic noise sequence The above target seismic wavelet and noise sequence are obtained.

[0022] In some embodiments, the reflection coefficient sequence r calculated according to the logging curve obs and the wellside seismic trace data d obs, the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance include:

[0023] Obtain the reflection coefficient measurement error data and seismic noise data in the historical work area information, and obtain the initial reflection coefficient error term and initial seismic noise;

[0024] According to the initial reflection coefficient error term and the initial seismic noise, the covariance matrix Σ of the corresponding reflection coefficient error term is determined by the prior model. r and the prior covariance matrix of seismic noise Σ d ;

[0025] Based on the covariance matrix Σ of the reflection coefficient error term r , the prior covariance matrix of the seismic noise Σ d , the reflection coefficient sequence r calculated from the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is determined by the posterior model and the conditional posterior covariance

[0026] In some embodiments, the covariance matrix Σ of the reflection coefficient error term is r Determined by the following formula:

[0027]

[0028] Among them, Σ r is the covariance matrix of the reflection coefficient error term, n r is the number of sampling points of the reflection coefficient error term, is the variance of the reflection coefficient error term, I is the unit matrix;

[0029] The prior covariance matrix of the seismic noise is Σ d Determined by the following formula:

[0030]

[0031] Among them, Σ 0d is the identity matrix, is a hyperparameter, Follows an inverse gamma distribution:

[0032]

[0033] Among them, γ d is the shape parameter, λ d is the scale factor.

[0034] In some embodiments, the conditional a posteriori expectation of the reflection coefficient error term is Determined by the following formula:

[0035]

[0036] Conditional a posteriori covariance of the reflection coefficient error term Determined by the following formula:

[0037]

[0038] Among them, W is the wavelet convolution matrix, and the wavelet convolution matrix W is:

[0039]

[0040] In some embodiments, the conditional a posteriori expectation based on the reflection coefficient error term and the conditional posterior covariance The new reflection coefficient error term is determined by the posterior model, including:

[0041] According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance Through the reflection coefficient error term e r The full conditional probability distribution of the reflection coefficient error term is determined; among them, the reflection coefficient error term e r The total conditional probability distribution of is:

[0042]

[0043] In some embodiments, the method according to the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance include:

[0044] According to the initial seismic wavelet w (0) , the wavelet expectation vector μ is determined by the prior model w and the wavelet prior covariance matrix Σ w ;

[0045] According to the wavelet expectation vector μ w , the wavelet prior covariance matrix Σ w , the prior covariance matrix of the seismic noise Σ d and wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is determined by the posterior model and the conditional posterior covariance

[0046] In some embodiments, the wavelet expectation vector μ w Determined by the following formula:

[0047]

[0048] in, and is the hyperparameter of the conjugate prior;

[0049] The wavelet prior covariance matrix Σ w Determined by the following formula:

[0050]

[0051] Among them, Σ 0w is the wavelet spatial correlation covariance matrix, is a hyperparameter, is a conjugate prior following an inverse gamma distribution:

[0052]

[0053] Among them, γ w is the shape parameter, λ w is the scale factor, n w is the number of sampling points of the seismic wavelet.

[0054] In some embodiments, the conditional posterior expectation μ of the seismic wavelet w|· Determined by the following formula:

[0055] μ w|· =μ w +Σ w R T (RΣ w R T +Σ d ) -1 (d obs -Rμ w )

[0056] Conditional posterior covariance Σ of seismic wavelet w|· Determined by the following formula:

[0057] Σ w|· =Σ w -Σ w R T (RΣ w R T +Σ d ) -1 RΣ w

[0058] Among them, R is the reflection coefficient convolution matrix, and the reflection coefficient convolution matrix R is:

[0059]

[0060] In some embodiments, the conditional a posteriori expectation of the seismic wavelet and the conditional posterior covariance Determining a new seismic wavelet by using the posterior model includes:

[0061] According to the conditional posterior expectation of the seismic wavelet and the conditional posterior covariance The new seismic wavelet is determined by the full conditional probability distribution of the seismic wavelet w; wherein the full conditional probability distribution of the seismic wavelet w is:

[0062]

[0063] In a second aspect, the present invention provides a device for simultaneously estimating seismic wavelets and noise, comprising:

[0064] A probability graph model building module, used to obtain a first forward model of reflection coefficient and seismic data, a second forward model of seismic wavelet and seismic data, and to build a probability graph model of seismic wavelet and seismic noise estimation according to the relationship between the random variables in the first forward model and the random variables in the second forward model; wherein the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms;

[0065] A model building module, used to build a priori model based on variables whose interdependence is deterministic dependence in the probabilistic graphical model, and to build a posterior model based on variables whose interdependence is probabilistic dependence in the probabilistic graphical model;

[0066] The estimation module is used to obtain well logging data and seismic observation data, and use the well logging data, the seismic observation data, the prior model and the posterior model to iterate the seismic wavelet and seismic noise simultaneously through the Gibbs algorithm to obtain the target seismic wavelet and noise sequence.

[0067] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for simultaneously estimating seismic wavelets and noise as described in the first aspect.

[0068] In a fourth aspect, the present invention provides an electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0069] Memory, used to store computer programs;

[0070] The processor is used to implement the method for simultaneously estimating seismic wavelets and noise as described in the first aspect when executing the program stored in the memory.

[0071] Based on the above technical solution, the beneficial effects of the present invention compared with the prior art are as follows:

[0072] The method for simultaneously estimating seismic wavelets and noise provided by the present invention obtains a first forward model of reflection coefficient and seismic data, a second forward model of seismic wavelets and seismic data, and establishes a probability graph model for estimating seismic wavelets and seismic noise according to the relationship between random variables in the first forward model and random variables in the second forward model; wherein, the random variables in the first forward model include seismic noise, and the random variables in the second forward model include logging measurement error terms; the quantitative relationship between seismic wavelets and seismic noise is described by introducing a probability graph model, and the dependency relationship between various parameters in seismic wavelet extraction under logging constraints is described by using a probability graph model, and a priori model and a posteriori model are established respectively according to the dependency relationship between various parameters, and the coupling between seismic wavelets and seismic noise is considered, and the seismic wavelets and noise are simultaneously iterated and obtained by the Gibbs algorithm, thereby improving the accuracy of seismic wavelet extraction. Moreover, in the simultaneous quantitative estimation of seismic wavelets and noise, the measurement error of logging data is considered, the accuracy of seismic wavelet and seismic noise estimation is improved, and a more accurate well-seismic calibration result can be obtained.

[0073] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0074] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0076] Figure 1 Schematic diagram of the process of simultaneous estimation of seismic wavelet and noise Figure 1 ;

[0077] Figure 2 Probabilistic graphical models for seismic wavelet and noise estimation;

[0078] Figure 3 The results of traditional wavelet calibration; (a) is the logarithmic domain wave impedance logging curve, (b) is the reflection coefficient curve, (c) is the Ricker wavelet synthetic seismic trace, and (d) is the actual wellside seismic trace;

[0079] Figure 4 The seismic wavelet estimation results, the black line represents the maximum a posteriori probability solution of the wavelet, and the gray line represents the random realization result of the wavelet;

[0080] Figure 5 is the seismic noise estimation result, the dotted line is the prior probability density function, and the histogram is the posterior probability density function;

[0081] Figure 6 The seismic wavelet calibration results extracted by the present invention; wherein, (a) is the logarithmic domain wave impedance logging curve, (b) is the reflection coefficient curve, (c) is the wavelet synthetic seismic trace obtained by the present invention, and (d) is the actual wellside seismic trace;

[0082] Figure 7 Schematic diagram of the process of the method for simultaneously estimating seismic wavelets and noise provided by the present invention Figure 2 ;

[0083] Figure 8 A schematic diagram of the structure of the device for simultaneously estimating seismic wavelets and noise provided by the present invention;

[0084] Fig. 9 This is a schematic structural diagram of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0085] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0086] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0087] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0088] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0089] According to whether well logging data is used, the commonly used seismic wavelet extraction methods can be divided into two categories: wavelet extraction without well constraints and wavelet extraction with well constraints. In wavelet extraction without well constraints, it is assumed that the reflection coefficient obeys a certain distribution and the wavelet is directly extracted from the seismic trace. Matsuoka et al. (Matsuoka, Toshifumi and Tad J. Ulrych. "Phaseestimation using the bispectrum." Proceedings of the IEEE 72 (1984): 1403-1411.) first used the bispectrum method to extract mixed-phase seismic wavelets; in order to improve the accuracy of wavelet extraction, Lazear (Lazear, Gregory D.. "Mixed-phase wavelet estimation using fourth-order cumulants." Geophysics 58 (1993): 1042-1051.) assumed that the reflection coefficient is a non-Gaussian distribution and proposed a parameterized wavelet estimation method based on the fourth-order cumulants; Ulrych et al. (Ulrych, TJ, Velis, DR, & Sacchi, MD (1995). Wavelet estimation Based on the Lazear method, the cepstrum stacking method is used to extract seismic wavelets without any assumptions about the distribution of reflection coefficients.

[0090] In well-constrained wavelet extraction, the reflection coefficient is usually calculated using the acoustic time difference and density logging curve, and then the seismic wavelet is inverted from the seismic records near the well according to the convolution model. Treitel and Robinson (Treitel, S., & Robinson, EA (1966). The Design of High-Resolution Digital Filters. IEEE Transactions on Geoscience and Remote Sensing, 4, 25-38.) first introduced the least squares Wiener filtering theory and constructed the reflection coefficient Toeplitz matrix to solve the seismic wavelet; Richard and Brac (Richard, V., & Brac, J. (1988). Wavelet Analysis Using Well Log Information. Seg Technical Program Expanded Abstracts.) used multi-channel coherence analysis to estimate the amplitude spectrum of the initial wavelet and seismic noise, and then estimated the wavelet amplitude spectrum and phase spectrum under the constraints of well logging data; Walden and White (Walden, AT, & Cristán, AC (1998). The phase–corrected undecimated discrete wavelet packet transform and its application to interpreting the timing of events. Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 454, 2243-2266.) proposed a method for fast wavelet extraction of noisy seismic data in the frequency domain. In recent years, the main progress of seismic wavelet extraction lies in the different improvements and applications of the above methods. For example, Yang Peijie (Yang Peijie. (2008). Blind extraction and nonlinear inversion of seismic wavelets. China University of Petroleum.) used a global optimization algorithm based on the high-order statistical quantum wave extraction method to study the blind extraction of seismic wavelets; Wang Yanghua (Wang Yanghua. (2020). Extraction method of space-varying generalized wavelets of seismic data. CN111427080A.) further considered the space-varying characteristics of seismic wavelets; Chen Xuehua et al. (Chen Xuehua, Zhang Jie, Yang Wei, Jiang Wei, Liu Yunfei, & Xu He. (2020). A model-based deep domain seismic wavelet extraction method. CN111708083A.) applied the ridge regression seismic wavelet extraction method to deep domain seismic data.

[0091] In fact, the estimation of seismic wavelets and the estimation of seismic noise levels are mutually coupled, that is, the wavelets will affect the estimation of the seismic noise level, and conversely, the seismic noise level will also affect the estimation of the seismic wavelets, so it is necessary to consider both at the same time. The above-mentioned seismic wavelet extraction methods cannot meet the expectations of the inventors. After further research and development, the inventors have made the present invention. The present invention introduces a probability graph model to describe the quantitative relationship between seismic wavelets and noise, and uses the Gibbs algorithm to iteratively obtain seismic wavelets and noise at the same time, thereby improving the accuracy of seismic wavelet extraction. From the published literature, in the field of seismic exploration, there is currently no effective method for seismic wavelet extraction that simultaneously considers seismic wavelets and noise. The present invention relates to the field of geophysical exploration technology, combines inversion theory with a probability graph model, and provides a method and device for simultaneous estimation of seismic wavelets and noise, which is used to improve the accuracy of seismic wavelet extraction.

[0092] Embodiment 1

[0093] The embodiment of the present invention provides a method for simultaneously estimating seismic wavelet and noise, referring to Figure 1 As shown, including:

[0094] S101, obtaining a first forward model of reflection coefficient and seismic data, and a second forward model of seismic wavelet and seismic data, and establishing a probability graph model of seismic wavelet and seismic noise estimation according to the relationship between the random variables in the first forward model and the random variables in the second forward model; the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms;

[0095] According to the Robinson convolution model, earthquake observation data d obs It can be expressed as:

[0096] d obs (t) = w(t)*r(t)+e d (t) (1)

[0097] Among them, w(t) is the seismic wavelet, r(t) is the reflection coefficient, e d (t) is the seismic noise, and * represents the convolution operator. Rewriting the above formula into a matrix-vector form, we can get the forward model of reflection coefficient and seismic data, that is, the first forward model expression:

[0098] d obs =Wr+e d (2)

[0099] Or the forward model of seismic wavelet and seismic data, that is, the second forward model expression:

[0100] dobs =Rw+e d (3)

[0101] When the measurement error of logging data is taken into account, the reflection coefficient can be expressed as:

[0102] r=r obs +e r (4)

[0103] Among them, d obs =[d1,d2,…,d n ] T represents seismic observation data, which is the seismic trace data near the well, r obs =[r1,r2,…,r n ] T is the reflection coefficient sequence calculated from the logging curve, r is the actual underground reflection coefficient, w=[w1,w2,…,w m ] T is the seismic wavelet vector to be estimated, the subscripts m and n represent the seismic wavelet dimension and the number of seismic trace sampling points, respectively. d and e r are seismic noise and logging measurement error (i.e., reflection coefficient error), respectively. The wavelet convolution matrix W and the reflection coefficient convolution matrix R are

[0104]

[0105] S102, establishing a priori model based on variables whose interdependence in the probabilistic graphical model is deterministic dependence, and establishing a posterior model based on variables whose interdependence in the probabilistic graphical model is probabilistic dependence;

[0106] S103, acquiring well logging data and seismic observation data, and using the well logging data, the seismic observation data, the prior model and the posterior model, performing simultaneous iterative calculations on seismic wavelets and seismic noise through a Gibbs algorithm to obtain target seismic wavelet and noise sequences.

[0107] The method for simultaneously estimating seismic wavelets and noise provided by the present invention obtains a first forward model of reflection coefficient and seismic data, a second forward model of seismic wavelets and seismic data, and establishes a probability graph model for estimating seismic wavelets and seismic noise according to the relationship between random variables in the first forward model and random variables in the second forward model; wherein, the random variables in the first forward model include seismic noise, and the random variables in the second forward model include logging measurement error terms; the quantitative relationship between seismic wavelets and seismic noise is described by introducing a probability graph model, and the dependency relationship between various parameters in seismic wavelet extraction under logging constraints is described by using a probability graph model, and a priori model and a posteriori model are established respectively according to the dependency relationship between various parameters, and the coupling between seismic wavelets and seismic noise is considered, and the seismic wavelets and noise are simultaneously iterated and obtained by the Gibbs algorithm, thereby improving the accuracy of seismic wavelet extraction. Moreover, in the simultaneous quantitative estimation of seismic wavelets and noise, the measurement error of logging data is considered, the accuracy of seismic wavelet and seismic noise estimation is improved, and a more accurate well-seismic calibration result can be obtained.

[0108] The logging measurement error term in the above step S101 includes a reflection coefficient error term, the logging data in the above step S103 is a reflection coefficient sequence calculated by the logging curve, and the seismic observation data is well-behind seismic trace data. In an embodiment of the present invention, the simultaneous estimation of seismic wavelets and seismic noise is described based on the reflection coefficient error term, the reflection coefficient sequence calculated by the logging curve, and the well-behind seismic trace data.

[0109] The invention provides a method and device for simultaneously estimating seismic wavelets and noise, which can take into account the coupling between wavelets and noise, and the logging measurement error, and improve the accuracy of seismic wavelet extraction. The main invention contents include: a probability graph model, a priori model, a conditional a posteriori model, and a Gibbs sampling algorithm for seismic wavelet and noise estimation.

[0110] First, the probability graph model of seismic wavelet and noise estimation of the present invention is described:

[0111] The probability graph model can describe the "dependency" relationship between random variables in the probability distribution, thereby describing the joint probability distribution of a group of random variables. According to the relationship between the random variables in equations 2 and 3, the present invention constructs a reference Figure 2 The seismic wavelet and noise estimation probability graph model shown in Figure 1, where μ and Σ represent the expected vector and covariance matrix of random variables, thin lines represent probabilistic dependence between variables, and thick lines represent deterministic dependence between variables. When the relationship between two nodes is deterministic dependence, the connected nodes can be integrated into a single node. The known nodes in the DAG model are {r obs ,d obs}, the unknown node is {er ,μ w ,w,Σ w ,Σ d}, and the earthquake observation data d obs The parent node of the node is Pa(d obs )={r obs ,e r ,w,Σ d}, the parent node of the wavelet node w is Pa(w)={μ w ,Σ w According to the calculation rules of the DAG model, the joint probability density function expression of all nodes in this DAG model can be obtained:

[0112] p(e r ,μ w ,w,Σ w ,Σ d ,r obs ,d obs )=p(e r )·p(μ w )·p(Σ w )·p(Σ d )·p(d obs |r obs ,e r ,w,Σ d )·p(w|μ w ,Σ w ) (5) Secondly, the prior model for seismic wavelet and noise estimation provided by the present invention is described:

[0113] The prior model refers to the probability distribution of model parameters obtained based on past experience and cognition, which is represented by nodes without parent nodes (r obs 、e r , μ w ,Σ w ,Σ d ). In order to calculate the specific probability expression of the DAG model, it is necessary to specify the probability distribution of the prior model. obs As is known, the present invention estimates the DAG model of seismic wavelet and noise (refer to Figure 2 ) uses the following prior model for random variables whose interdependence is deterministic.

[0114] ①Reflection coefficient error term e r The prior of follows a zero-mean white noise Gaussian distribution:

[0115]

[0116] in, is the covariance of the reflection coefficient error, is the variance of the reflection coefficient error term.

[0117] ②Wavelet expectation vector μ w The prior for follows a Gaussian distribution:

[0118]

[0119] in, and is the hyperparameter of the conjugate prior.

[0120] ③ Wavelet prior covariance matrix Σ w for:

[0121]

[0122] Among them, Σ 0w is the wavelet spatial correlation covariance matrix, and the hyperparameters is a conjugate prior following an inverse gamma distribution:

[0123]

[0124] Among them, γ w is the shape parameter, n w is the number of sampling points of the seismic wavelet, λ w is the scale factor,

[0125] ④ Prior covariance matrix of seismic noise Σ d for:

[0126]

[0127] Among them, the hyperparameters Follows an inverse gamma distribution:

[0128]

[0129] Among them, γ d is the shape parameter, n d is the number of sampling points of seismic noise; d is the scale factor, μ d is the noise expectation vector.

[0130] This section defines the error term e based on the reflection coefficient r (Equation 7), the wavelet expectation vector μ w (Equation 8), the wavelet prior covariance matrix Σ w (Equation 9), the seismic noise prior covariance matrix Σ d The prior model of (Equation 11).

[0131] The conditional posterior model of seismic wavelet and noise estimation is explained as follows:

[0132] Through the joint probability distribution of all variables in the DAG model (Formula 5), ​​the full conditional probability distribution of a single unknown variable under the constraints of other variables can be determined. Figure 2 The probabilistic graphical model shown uses the following posterior model for random variables whose interdependencies are probabilistically dependent:

[0133] ①Reflection coefficient measurement error term e r The total conditional probability distribution of is:

[0134] p(e r |·)=p(e r )p(d obs |r obs ,e r ,w,Σ d ) (12)

[0135] Under the condition of known seismic wavelet, the reflection coefficient error term e can be obtained by combining equations 2 and 4. r and earthquake observation data obs The joint conditional distribution of is:

[0136]

[0137] Then the reflection coefficient error term e r The total conditional probability distribution of can be further expressed as:

[0138]

[0139] Among them, the conditional posterior expectation of the reflection coefficient error term is and the conditional posterior covariance They are:

[0140]

[0141]

[0142] ②The total conditional probability distribution of the earthquake wavelet w is:

[0143] p(w|·)=p(d obs |r obs ,e r ,w,Σ d )p(w|μ w ,Σ w ) (17)

[0144] Under the condition of known reflection coefficient, according to the forward equation of formula 3, the seismic wavelet and seismic observation data dobs The joint conditional distribution of is:

[0145]

[0146] Then the full conditional probability distribution of the seismic wavelet w can be further written as:

[0147]

[0148] Among them, the conditional posterior expectation μ of the seismic wavelet is w|· and the conditional posterior covariance Σ w|· They are:

[0149] μ w|· =μ w +Σ w R T (RΣ w R T +Σ d ) -1 (d obs -Rμ w ) (20)

[0150] Σ w|· =Σ w -Σ w R T (RΣ w R T +Σ d ) -1 RΣ w (twenty one)

[0151] The present invention uses the Gibbs sampling algorithm to efficiently explore the posterior solution space. The principle of the Gibbs sampling algorithm is: simulate the joint distribution by sampling the conditional distribution, and then directly derive the conditional distribution through the simulated joint distribution, and iterate in this way. Each iteration samples a certain variable once, and the sampled variable value continues to participate in the calculation of the remaining variables as the current state until the conditional distribution of all variables has completed sampling. Then this iteration ends, and when the set number of iterations is reached, the results of multiple realizations can be output. In an embodiment of the present invention, a probability graph model is used to determine the interdependence between seismic observation data, well logging data and seismic wavelets, and establish the joint posterior probability distribution of each parameter, that is, the joint probability distribution on the left side of Formula 5. The fully conditional posterior probability distribution of each parameter is then obtained from the joint posterior probability distribution, that is, the conditional distribution (the right side of the equal sign in Formula 5) is based on reference Figure 2 The probability graph model shown is equivalent to simplifying the complex joint distribution into the product form of conditional distribution. Then the Gibbs algorithm is used for sampling to obtain relatively stable seismic wavelet and noise estimation results. According to the sampling principle of the Gibbs algorithm, the design reference Figure 6 and Figure 7 The seismic wavelet and seismic noise estimation method shown in the figure, the above step S103 specifically includes:

[0152] S200, reading input data, including reflection coefficient sequence calculated from logging curve and seismic trace data near the well;

[0153] S201, randomly generate initial seismic wavelet w (0) , set the maximum number of iterations k max And execute it in a loop;

[0154] S202, the reflection coefficient sequence r calculated according to the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance

[0155] S203: Conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance Determining a new reflection coefficient error term through the posterior model, and updating the reflection coefficient error term based on the new reflection coefficient error term to obtain an updated reflection coefficient error term;

[0156] S204, determining a new reflection coefficient based on the updated reflection coefficient error term;

[0157] S205, based on the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance

[0158] S206: Conditional posterior expectation of the seismic wavelet and the conditional posterior covariance Determine a new seismic wavelet by using the a posteriori model;

[0159] S207, updating the seismic wavelet based on the new seismic wavelet to obtain an updated seismic wavelet, and convolving the updated seismic wavelet with the new reflection coefficient to obtain new seismic noise;

[0160] S208, updating the seismic noise based on the new seismic noise to obtain updated seismic noise, and determining the covariance matrix Σ of the seismic noise through the prior model according to the updated seismic noise d ;

[0161] S209, based on the covariance matrix Σ of the updated seismic wavelet and the seismic noise d Repeat the above process of updating the reflection coefficient error term, seismic wavelet and seismic noise until the maximum number of iterations k is reached max , output the final seismic wavelet and seismic noise sequence The above target seismic wavelet and noise sequence are obtained.

[0162] After the above step S208 is completed, it is determined whether the maximum number of iterations k is reached. max If it is not reached, the updated seismic wavelet obtained in step S207 is used as the initial seismic wavelet of the next cycle, and the covariance matrix Σ of the seismic noise obtained in step S208 is calculated based on the updated seismic wavelet and the covariance matrix Σ of the seismic noise obtained in step S208. d , repeat the above steps S202-S208 until the maximum number of iterations k is reached max , output the seismic wavelet and seismic noise sequence obtained in the last iteration, and obtain the target seismic wavelet and noise sequence.

[0163] In a specific embodiment, the reflection coefficient sequence r calculated according to the logging curve in S202 is obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance include:

[0164] S2021. Obtain reflection coefficient measurement error data and seismic noise data in historical work area information to obtain initial reflection coefficient error terms and initial seismic noise;

[0165] S2022, according to the initial reflection coefficient error term and the initial seismic noise, determine the corresponding covariance matrix Σ of the reflection coefficient error term through the above formula 7 and formula 11 respectively r and the prior covariance matrix of seismic noise Σ d ;

[0166] S2023, based on the covariance matrix Σ of the reflection coefficient error term r , the prior covariance matrix of the seismic noise Σ d , the reflection coefficient sequence r calculated from the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is determined by the above equations 17 and 18 and the conditional posterior covariance

[0167] In a specific embodiment, the conditional posterior expectation of the reflection coefficient error term described in S203 above and the conditional posterior covariance The new reflection coefficient error term is determined by the posterior model, including:

[0168] According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance The reflection coefficient error term e is obtained by equation 16 above. r The full conditional probability distribution of is used to determine the new reflection coefficient error term.

[0169] According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance and the reflection coefficient error term e specified in Equation 16 r The full conditional probability distribution of can be used to infer the corresponding reflection coefficient error term to obtain a new reflection coefficient error term. For the specific process, those skilled in the art can refer to the description in the prior art, which will not be repeated here.

[0170] In some embodiments, the above step S205 is based on the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance include:

[0171] S2051, according to the initial seismic wavelet w (0) , the wavelet expected vector μ is determined by the above equations 8 and 9 w and the wavelet prior covariance matrix Σ w ;

[0172] S2052, according to the wavelet expected vector μ w , the wavelet prior covariance matrix Σ w , the prior covariance matrix of the seismic noise Σ d and wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is determined by equations 22 and 23 and the conditional posterior covariance

[0173] In a specific embodiment, the conditional posterior expectation of the seismic wavelet described in step S206 is and the conditional posterior covariance Determining a new seismic wavelet by using the posterior model includes:

[0174] According to the conditional posterior expectation of the seismic wavelet and the conditional posterior covariance The new seismic wavelet is determined by the full conditional probability distribution of the seismic wavelet w shown in Formula 21.

[0175] Conditional posterior expectation of earthquake wavelets and the conditional posterior covariance As well as the full conditional probability distribution of the specified seismic wavelet w, the corresponding seismic wavelet can be deduced to obtain a new seismic wavelet. For the specific process, those skilled in the art can refer to the description in the prior art, which will not be repeated here.

[0176] Embodiment 2

[0177] The present invention is applied to a certain exploration area, and the well logging data in the exploration area and the corresponding wellside seismic data are selected as input data. Figure 3 As shown in the figure. (a) is the logarithmic wave impedance curve obtained by acoustic wave and density logging curve, (b) is the reflection coefficient calculated from the logarithmic wave impedance curve, (c) is the synthetic seismic record result (Synthetic trace) obtained by convolution of 30Hz Ricker wavelet and reflection coefficient, which represents the calibration result of conventional method, and (d) is the actual borehole-side seismic trace (Borehole-side traces). It can be seen that the correspondence between the phase axis of the conventional wavelet synthetic seismic record result and the actual borehole-side seismic trace is poor. Figure 4 and Figure 5 They are the results of seismic wavelet and seismic noise estimation of the present invention, respectively. It can be seen that after the iterative calculation of the Gibbs algorithm, the distribution of the posterior probability density function (Post PDF, Posterior Pprobability Density Function) and the prior probability density function (Prior PDF, Prior Pprobability Density Function) of the seismic noise variance are quite different, indicating the effectiveness of Gibbs iteration for parameter updating. Figure 4 In the figure, the vertical axis Amplitude represents the amplitude, the black line (Estimatedwavelet) represents the maximum a posteriori probability solution of the wavelet, and the gray line (Simulations) represents the random realization result of the wavelet. Figure 6 The calibration results of the wavelet extracted by the present invention are shown in Figure 2. (a) is the logarithmic wave impedance logging curve, (b) is the reflection coefficient curve, (c) is the synthetic seismic record result obtained by convolution of the seismic wavelet obtained by the method of the present invention and the reflection coefficient curve, and (d) is the actual wellside seismic trace. Figure 3It can be seen that the correlation between the synthetic seismic recording results obtained by the method of the present invention and the well-side seismic trace is significantly improved, indicating the feasibility and effectiveness of the technology of the present invention. The present invention can consider the coupling relationship between seismic wavelets and noise, obtain more accurate seismic wavelet estimation results, and improve the well-seismic calibration effect.

[0178] Embodiment 3

[0179] The embodiment of the present invention provides a device for simultaneously estimating seismic wavelet and noise, referring to Figure 8 As shown, including:

[0180] A probability graph model building module 301 is used to obtain a first forward model of reflection coefficient and seismic data, and a second forward model of seismic wavelet and seismic data, and to establish a probability graph model of seismic wavelet and seismic noise estimation according to the relationship between the random variables in the first forward model and the random variables in the second forward model; the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms;

[0181] A model building module 302 is used to build a priori models based on variables whose interdependencies in the probabilistic graphical model are deterministic dependencies, and to build a posteriori models based on variables whose interdependencies in the probabilistic graphical model are probabilistic dependencies;

[0182] The estimation module 303 is used to obtain well logging data and seismic observation data, and use the well logging data, the seismic observation data, the prior model and the posterior model to iterate the seismic wavelet and seismic noise simultaneously through the Gibbs algorithm to obtain the target seismic wavelet and noise sequence.

[0183] The implementation principle and technical effect of the device for simultaneous estimation of seismic wavelets and noise provided in the embodiment of the present invention are similar to those of any of the aforementioned method embodiments, and will not be described in detail here.

[0184] Embodiment 4

[0185] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for simultaneously estimating seismic wavelets and noise as described in any one of the aforementioned method embodiments is implemented.

[0186] The computer-readable storage medium may be included in the device / apparatus described in the above embodiment; or it may exist independently without being assembled into the device / apparatus. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0187] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0188] Embodiment 5

[0189] An embodiment of the present invention provides an electronic device, referring to Fig. 9 As shown, it includes a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112 and the memory 113 communicate with each other through the communication bus 114.

[0190] Memory 113, used for storing computer programs;

[0191] The processor 111 is used to implement the method for simultaneously estimating seismic wavelets and noise as described in any one of the aforementioned method embodiments when executing the program stored in the memory 113 .

[0192] The implementation principle and technical effect of the electronic device provided by the embodiment of the present invention are similar to those of any of the aforementioned method embodiments, and will not be repeated here.

[0193] The above-mentioned memory 113 can be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk or a ROM. The memory 113 has a storage space for program codes for executing any method steps in the above-mentioned method. For example, the storage space for program codes may include various program codes for implementing various steps in the above method respectively. These program codes can be read from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. Such computer program products are usually portable or fixed storage units. The storage unit may have a storage segment or storage space arranged similarly to the memory 113 in the above-mentioned electronic device. The program code can be compressed, for example, in an appropriate form. Generally, the storage unit includes a program for executing the method steps according to an embodiment of the present invention, that is, a code that can be read by, for example, a processor 111, and when these codes are run by an electronic device, the electronic device performs the various steps in the method described above.

[0194] In this document, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. The orientation or position relationship indicated by the terms "upper", "lower", etc. is based on the orientation or position relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0195] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict. The present invention is not limited to any single aspect, any single embodiment, or any combination and / or permutation of these aspects and / or embodiments. Each aspect and / or embodiment of the present invention can be used alone, or in combination with one or more other aspects and / or other embodiments.

[0196] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for simultaneously estimating seismic wavelets and noise, characterized in that: include: Obtain a first forward model of reflection coefficient and seismic data, and a second forward model of seismic wavelet and seismic data, and establish a probability graph model for seismic wavelet and seismic noise estimation based on the relationship between the random variables in the first forward model and the random variables in the second forward model; wherein the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms; Establishing a priori model based on variables whose interdependence in the probabilistic graphical model is deterministic dependency, and establishing a posterior model based on variables whose interdependence in the probabilistic graphical model is probabilistic dependency; The well logging data and the seismic observation data are acquired, and the seismic wavelet and the seismic noise are simultaneously iterated and obtained by the Gibbs algorithm using the well logging data, the seismic observation data, the prior model and the posterior model to obtain the target seismic wavelet and noise sequence.

2. The method for simultaneous estimation of seismic wavelet and noise according to claim 1, characterized in that: The logging measurement error term includes a reflection coefficient error term, the logging data includes a reflection coefficient sequence calculated from a logging curve, and the seismic observation data includes wellside seismic trace data; the method further includes: establishing a joint probability density function of all variables in the probability graph model according to a probability graph model calculation rule: p(e r ,m w ,w,S w ,S d ,r obs ,d obs )=p(e r )·p(μ w )·p(Σ w )·p(Σ d )·p(d obs |r obs ,e r ,w,S d )·p(w|μ w ,S w ) Among them, d obs =[d1,d2,…,d n ] T is the seismic trace data near the well, r obs =[r1,r2,…,r n ] T is the reflection coefficient sequence calculated from the logging curve, r is the reflection coefficient, w=[w1,w2,…,w m ] T is the seismic wavelet to be estimated, m and n are the seismic wavelet dimension and the number of seismic trace sampling points respectively, e r is the reflection coefficient error term, Σ d is the seismic noise e d The prior covariance matrix of r is the covariance matrix of the reflection coefficient error term, μ w is the wavelet expectation vector, Σ w is the wavelet prior covariance matrix.

3. The method for simultaneous estimation of seismic wavelet and noise according to claim 2, characterized in that: The seismic data is wellside seismic trace data, and the well logging data is a reflection coefficient sequence calculated from the well logging curve; The method uses the well logging data, the seismic observation data, the prior model and the posterior model to iterate the seismic wavelet and the seismic noise simultaneously through the Gibbs algorithm to obtain the target seismic wavelet and noise sequence, including: Randomly generate initial seismic wavelet w (0) ; The reflection coefficient sequence r calculated according to the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance Determining a new reflection coefficient error term through the posterior model, and updating the reflection coefficient error term based on the new reflection coefficient error term to obtain an updated reflection coefficient error term; determining a new reflection coefficient based on the updated reflection coefficient error term; According to the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance According to the conditional posterior expectation of the seismic wavelet and the conditional posterior covariance Determine a new seismic wavelet by using the a posteriori model; Based on the new seismic wavelet, the seismic wavelet is updated to obtain an updated seismic wavelet, and the updated seismic wavelet and the new reflection coefficient are convolved to obtain new seismic noise; Based on the new seismic noise, the seismic noise is updated to obtain updated seismic noise, and the covariance matrix Σ of the seismic noise is determined by the prior model according to the updated seismic noise. d ; Based on the covariance matrix Σ of the updated seismic wavelet and the seismic noise d Repeat the above process of updating the reflection coefficient error term, seismic wavelet and seismic noise until the maximum number of iterations k is reached max , output the final seismic wavelet and seismic noise sequence The above target seismic wavelet and noise sequence are obtained.

4. The method for simultaneous estimation of seismic wavelet and noise according to claim 3, characterized in that: The reflection coefficient sequence r calculated according to the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is calculated by the prior model and the posterior model and the conditional posterior covariance include: Obtain the reflection coefficient measurement error data and seismic noise data in the historical work area information, and obtain the initial reflection coefficient error term and initial seismic noise; According to the initial reflection coefficient error term and the initial seismic noise, the covariance matrix Σ of the corresponding reflection coefficient error term is determined by the prior model. r and the prior covariance matrix of seismic noise Σ d ; Based on the covariance matrix Σ of the reflection coefficient error term r , the prior covariance matrix of the seismic noise Σ d , the reflection coefficient sequence r calculated from the logging curve obs and the wellside seismic trace data d obs , the conditional posterior expectation of the reflection coefficient error term is determined by the posterior model and the conditional posterior covariance 5. The method for simultaneous estimation of seismic wavelet and noise according to claim 4, characterized in that: The covariance matrix of the reflection coefficient error term Σ r Determined by the following formula: Among them, Σ r is the covariance matrix of the reflection coefficient error term, n r is the number of sampling points of the reflection coefficient error term, is the variance of the reflection coefficient error term, I is the unit matrix; The prior covariance matrix of the seismic noise is Σ d Determined by the following formula: Among them, Σ 0d is the identity matrix, is a hyperparameter, Follows an inverse gamma distribution: Among them, γ d is the shape parameter, λ d is the scale factor.

6. The method for simultaneous estimation of seismic wavelet and noise according to claim 4, characterized in that: The conditional posterior expectation of the reflection coefficient error term is Determined by the following formula: Conditional a posteriori covariance of the reflection coefficient error term Determined by the following formula: Among them, W is the wavelet convolution matrix, and the wavelet convolution matrix W is:

7. The method for simultaneous estimation of seismic wavelet and noise according to claim 4, characterized in that: The conditional posterior expectation based on the reflection coefficient error term and the conditional posterior covariance The new reflection coefficient error term is determined by the posterior model, including: According to the conditional posterior expectation of the reflection coefficient error term and the conditional posterior covariance Through the reflection coefficient error term e r The full conditional probability distribution of the reflection coefficient error term is determined; among them, the reflection coefficient error term e r The total conditional probability distribution of is:

8. The method for simultaneous estimation of seismic wavelet and noise according to claim 3, characterized in that: The method according to the initial seismic wavelet and the wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is calculated by the prior model and the posterior model and the conditional posterior covariance include: According to the initial seismic wavelet w (0) , the wavelet expectation vector μ is determined by the prior model w and the wavelet prior covariance matrix Σ w ; According to the wavelet expectation vector μ w , the wavelet prior covariance matrix Σ w , the prior covariance matrix of the seismic noise Σ d and wellside seismic trace data d obs , the conditional posterior expectation of the seismic wavelet is determined by the posterior model and the conditional posterior covariance 9. The method for simultaneous estimation of seismic wavelet and noise according to claim 8, characterized in that: The wavelet expectation vector μ w Determined by the following formula: in, and is the hyperparameter of the conjugate prior; The wavelet prior covariance matrix Σ w Determined by the following formula: Among them, Σ 0w is the wavelet spatial correlation covariance matrix, is a hyperparameter, is a conjugate prior following an inverse gamma distribution: Among them, γ w is the shape parameter, λ w is the scale factor, n w is the number of sampling points of the seismic wavelet.

10. The method for simultaneous estimation of seismic wavelet and noise according to claim 8, characterized in that: The conditional posterior expectation μ of the seismic wavelet w|· Determined by the following formula: μ w|· =μ w +Σ w R T (RΣ w R T +Σ d ) -1 (d obs -Rμ w ) Conditional posterior covariance Σ of seismic wavelet w|· Determined by the following formula: S w|· =S w -S w R T (RΣ w R T +S d ) -1 RS w Among them, R is the reflection coefficient convolution matrix, and the reflection coefficient convolution matrix R is:

11. The method for simultaneous estimation of seismic wavelet and noise according to claim 8, characterized in that: The conditional posterior expectation of the seismic wavelet and the conditional posterior covariance Determining a new seismic wavelet by using the posterior model includes: According to the conditional posterior expectation of the seismic wavelet and the conditional posterior covariance The new seismic wavelet is determined by the full conditional probability distribution of the seismic wavelet w; wherein the full conditional probability distribution of the seismic wavelet w is:

12. A device for simultaneously estimating seismic wavelets and noise, characterized in that: include: A probability graph model building module, used to obtain a first forward model of reflection coefficient and seismic data, a second forward model of seismic wavelet and seismic data, and to build a probability graph model of seismic wavelet and seismic noise estimation according to the relationship between the random variables in the first forward model and the random variables in the second forward model; wherein the random variables in the first forward model include seismic noise, and the random variables in the second forward model include well logging measurement error terms; A model building module, used to build a priori model based on variables whose interdependence is deterministic dependence in the probabilistic graphical model, and to build a posterior model based on variables whose interdependence is probabilistic dependence in the probabilistic graphical model; The estimation module is used to obtain well logging data and seismic observation data, and use the well logging data, the seismic observation data, the prior model and the posterior model to iterate the seismic wavelet and seismic noise simultaneously through the Gibbs algorithm to obtain the target seismic wavelet and noise sequence.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for simultaneously estimating seismic wavelets and noise as described in any one of claims 1 to 11 is implemented.

14. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; The processor is used to implement the method for simultaneously estimating seismic wavelets and noise as described in any one of claims 1 to 11 when executing the program stored in the memory.

Images