CN110865409B - Seismic wave impedance inversion method based on wave impedance low-rank regularization - Google Patents

Seismic wave impedance inversion method based on wave impedance low-rank regularization Download PDF

Info

Publication number
CN110865409B
CN110865409B CN201911210757.2A CN201911210757A CN110865409B CN 110865409 B CN110865409 B CN 110865409B CN 201911210757 A CN201911210757 A CN 201911210757A CN 110865409 B CN110865409 B CN 110865409B
Authority
CN
China
Prior art keywords
wave impedance
rank
low
inversion
seismic
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN201911210757.2A
Other languages
Chinese (zh)
Other versions
CN110865409A (en
Inventor
陈雷平
李曙
贺达江
丁黎明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Linyi Gaoxin Tap Water Co ltd
Original Assignee
Huaihua University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huaihua University filed Critical Huaihua University
Priority to CN201911210757.2A priority Critical patent/CN110865409B/en
Publication of CN110865409A publication Critical patent/CN110865409A/en
Application granted granted Critical
Publication of CN110865409B publication Critical patent/CN110865409B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/30—Analysis
    • G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/62—Physical property of subsurface
    • G01V2210/622—Velocity, density or impedance
    • G01V2210/6226—Impedance

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Environmental & Geological Engineering (AREA)
  • Geology (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Geophysics (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

The invention discloses a seismic wave impedance inversion method based on wave impedance low-rank regularization. The method directly establishes a low-rank regularization item aiming at the seismic wave impedance, establishes a target function with low-rank wave impedance constraint on the basis, finally solves the problem by skillfully utilizing singular value decomposition, performs seismic inversion by utilizing the low-rank property of the wave impedance, fully utilizes the low-rank prior information of seismic wave impedance data and improves the accuracy of seismic wave impedance inversion.

Description

Seismic wave impedance inversion method based on wave impedance low-rank regularization
Technical Field
The invention belongs to the field of seismic exploration, and particularly relates to a seismic wave impedance inversion method based on wave impedance low-rank regularization.
Background
Seismic wave impedance is an important earth medium parameter that is numerically equal to the product of the density of the subsurface rock and the velocity of the seismic waves as they propagate in the rock. It is closely related to the lithology, porosity, etc. of the rock. Therefore, how to effectively and accurately obtain the seismic wave impedance is always an important research content in seismic exploration. The seismic wave impedance inversion technology reversely deduces the wave impedance of the underground medium by using prior information according to data such as earthquake, well logging and the like obtained by observation, and is the most main way for obtaining the seismic wave impedance at present. With the development of seismic acquisition technology and signal processing technology, the quality of seismic data is higher and higher. The quality of the seismic inversion results is more and more obviously limited by the utilization of prior information. How to mine and effectively use prior information becomes a hot problem in the current seismic inversion field.
The low rank nature of the data is an important a priori information that has been widely focused and used in recent years. Such as: a large number of literature reports based on data low-rank prior exist in the fields of voice and image signal denoising, seismic data denoising/reconstruction and the like. In the field of seismic exploration, the low-rank characteristic is mainly used for the aspects of seismic data denoising, missing seismic channel completion, seismic data reconstruction and the like. In these studies, the low rank property of seismic data was utilized. Namely: the matrix formed by combining the seismic data of different sub-blocks with similar waveforms in the seismic data has the characteristic of low rank. When the low rank property is used, the denoising/reconstruction effect of the data is obviously improved compared with the traditional method. Thus, these methods experimentally demonstrate the advantage of low rank as a priori information for use in seismic data processing. It should be noted that these seismic data processing methods only utilize the low rank property of seismic data in the data domain. The output results after they are processed are still only seismic data.
Since seismic data has low rank properties between different sub-blocks, seismic data is closely related to the nature and structure of the subsurface rock formations. The former is a signalized representation of the latter. If there is similarity between different sub-blocks of the former, the corresponding position of the latter should also have such similarity. In this case, the matrix formed by the wave resistances of the sub-blocks, such as wave impedance, density, poisson ratio, and the like, has low rank. Low rank is a non-localized a priori information. In electromagnetic parametric inversion, the idea of using low rank approximation is mentioned in the article "A low-rank approximation for large-scale 3D controlled-source electromagnetic Gauss-Newton inversion" published in Geophysics journal by Manual Amaya et al (2016). However, their idea is to use its low rank approximation to replace the Hessian matrix in the inversion process to reduce the computational complexity, and not to directly utilize the low rank of the model parameters to be inverted. At present, no literature report is available on the utilization of the low rank property of wave impedance, and the wave impedance is obtained by using a seismic inversion method according to seismic data.
The existing seismic wave impedance inversion method mainly utilizes the seismic data or the localized prior information of the wave impedance. For example, the sparsity prior information utilized by sparse inversion is a localized prior information. It is a utilization of sparsity of local regions of seismic data or wave impedance under the L0, L1, or Lp (0< p <1) norm, lacking the utilization of a priori information between different sub-regions. In recent years, the seismic wave impedance inversion method based on total variation regularization which is concerned with uses gradient sparsity prior information of wave impedance, but the prior information is also localized, and non-localized prior information is not used.
In sum, the low rank characteristic has outstanding advantages in the fields of signal/data denoising, reconstruction and the like as important non-local prior information, so that the low rank characteristic plays an important role in seismic data processing. The seismic wave impedance should also have low rank. However, none of the existing seismic wave impedance inversion methods utilize this a priori information.
Disclosure of Invention
In order to solve the problems, the invention discloses a seismic wave impedance inversion method based on wave impedance low-rank regularization. The method directly establishes the low-rank regularization item aiming at the seismic wave impedance, establishes the target function with the low-rank constraint of the wave impedance on the basis, finally solves the problem by skillfully utilizing singular value decomposition, performs seismic inversion by utilizing the low-rank property of the wave impedance, fully utilizes the low-rank property of seismic wave impedance data and improves the accuracy of seismic wave impedance inversion.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a seismic wave impedance inversion method based on wave impedance low-rank regularization comprises the following steps:
step one, inputting seismic data and a wave impedance initial model, and constructing an initial low-rank approximate matrix of wave impedance: constructing an initial low-rank approximate matrix by adopting a similarity search-based method;
step two, establishing a wave impedance low-rank regularization target function
Let the whole wave impedance profile be divided into K sub-data blocks with overlap, in which the ith sub-data block is denoted as diThen the wave impedance is adjustedThe optimal low rank approximation problem is expressed as:
Figure BDA0002297996710000021
Φiis a similarity measure and selection operator with diIs a reference block for selecting the sum d from the seismic wave impedance miQ sub-data blocks with the smallest euclidean distance of (d) are arranged into a matrix of size P × Q, P representing sub-data block diLength of (d); oiObtaining an optimal low-rank approximate matrix to be solved; | | non-woven hairFExpressing the Frobenius norm of the matrix, and rank () is the operation of solving the rank of the matrix; λ represents a low-rank constrained regularization parameter;
on the basis of the formula (1), constructing an objective function of a wave impedance low-rank regularization seismic wave impedance inversion problem:
Figure BDA0002297996710000022
g represents a forward operator during seismic wave impedance inversion, and S represents seismic data obtained through observation; eta is the regularization coefficient of the wave impedance low-rank regularization term used for balancing the data fidelity term
Figure BDA0002297996710000023
And the magnitude of the role played by the low-rank regularization term in inversion, wherein the larger the value is, the larger the role played by the low-rank regularization term is;
step three, solving the objective function
For the objective function of the formula (2), the solution is carried out in two steps, firstly, the optimal low rank approximation is solved, namely, the optimal value of the formula (1) is solved, and the optimal low rank approximation of the formula (1)
Figure BDA0002297996710000031
Writing into:
Figure BDA0002297996710000032
solving equation (3) using singular value decomposition and hard threshold shrinkage, let ΦiThe singular value of m is decomposed into:
Φim＝Αdiag(ω)ΒH (4)
a and B respectively represent two unitary matrixes generated after singular value decomposition; omega is a singular value vector obtained by singular value decomposition; diag (ω) represents a diagonal matrix spanned by singular value vectors ω obtained by singular value decomposition, and the element on the diagonal is ω; h represents a conjugate operator; optimal low rank approximation
Figure BDA0002297996710000033
The solution of (a) is written as:
Figure BDA0002297996710000034
where λ represents the threshold used in hard threshold contraction, HTλ() Represents a hard threshold function defined by the formula:
Figure BDA0002297996710000035
where Γ represents the input HTλ() Data in the function;
in finding OiAfter the optimal solution of (3), O is fixed in the objective function of equation (2)iThe update formula of the wave impedance inversion result is obtained as follows:
Figure BDA0002297996710000036
equation (7) is a least squares problem, and equation (7) is formulated with respect to
Figure BDA0002297996710000037
The equation of (c):
Figure BDA0002297996710000038
wherein G is*And phii *Are G and phi respectivelyiUsing conjugate gradient method to quickly solve to obtain inversion result
Figure BDA0002297996710000039
Step four, inversion results are obtained
Figure BDA00022979967100000310
And restoring the original wave impedance model into a 2-dimensional section or a 3-dimensional data volume according to a process opposite to the process of constructing the long column vector m, namely completing the whole inversion process.
In a further improvement, η is determined by a cross-validation method, namely: adjusting according to the difference between the inversion result and the true value, and selecting a value which enables the inversion error to be minimum, namely a value of eta; the seismic data are seismic signals recorded by a geophone.
In a further improvement, in the first step, the step of constructing an initial low-rank approximation matrix of wave impedance comprises the following steps: arranging an initial wave impedance model into a long column vector m according to columns; then determining a sub-data block d as a reference, and assuming that the length of the sub-data block d is P; sequentially selecting Q sub-data blocks with the minimum Euclidean distance from d from m; arranging the Q sub-data blocks into a matrix, and completing construction of an initial low-rank approximate matrix; the initial wave impedance is a 2-dimensional matrix or a 3-dimensional data volume.
Drawings
FIG. 1 is a flow chart of the construction of an initial low rank approximation matrix; in the figure, (a) represents a wave impedance profile, (b) represents similar blocks with similar euclidean distances, and (c) an initial low rank approximation matrix;
FIG. 2 is a flow chart of the present invention;
FIG. 3 is a plot of a portion of data taken from the Marmousi2 model as a true model of wave impedance;
FIG. 4 is an initial model of wave impedance;
FIG. 5 is seismic data for inversion;
FIG. 6 is a Rake wavelet with a dominant frequency of 40 Hz;
fig. 7 is a diagram of inversion results.
Detailed Description
The invention is further explained with reference to the drawings and the embodiments.
Example 1
In order to solve the problems that the existing seismic wave impedance inversion method does not utilize non-local prior information of wave impedance, cannot grasp the essential characteristics of the seismic wave impedance from the global angle and the like, the seismic wave impedance inversion method based on the low-rank regularization of the wave impedance is provided. The core of the method is as follows: the method comprises the steps of directly establishing a low-rank regularization term aiming at seismic wave impedance, establishing a target function with low-rank wave impedance constraint on the basis, and finally solving by skillfully utilizing singular value decomposition. At present, no research report for seismic inversion by using low rank property of wave impedance exists.
(1) Constructing a low rank approximation matrix of wave impedance
And constructing a low-rank approximate matrix by adopting a similarity search-based method. The specific implementation method comprises the following steps: the initial wave impedance model is first arranged column by column into a long column vector m. Then, a sub-data block d is determined as a reference, assuming that it is of length P. And sequentially selecting Q sub-data blocks with the minimum Euclidean distance from d from m. Finally, the Q sub-data blocks are arranged into a matrix, and the obtained matrix is a low-rank matrix. The construction process is shown in figure 1. It should be noted that, in order to observe the similarity between sub-data blocks, the wave impedance in (a) in fig. 1 is not vectorized in advance.
(2) Establishing a wave impedance low-rank regularization target function
Suppose that the entire wave impedance profile can be divided into K sub-data blocks overlapping each other, where the ith sub-data block is denoted as diThen, the wave impedance optimal low rank approximation problem can be expressed as:
Figure BDA0002297996710000051
Φiis a similarity measure and selection operator which uses diIs a reference block for selecting the sum d from the seismic wave impedance miQ sub-data blocks having the smallest euclidean distance of (a) are arranged into a matrix of size P × Q. O isiAnd (4) obtaining an optimal low-rank approximate matrix to be solved. | | non-woven hairFThe Frobenius norm, rank () of the matrix is the operation to find the rank of the matrix. λ represents a low rank constrained regularization parameter. On the basis, an objective function of the wave impedance low-rank regularization seismic wave impedance inversion problem can be constructed:
Figure BDA0002297996710000052
wherein G represents a forward operator during seismic wave impedance inversion, and S represents seismic data obtained by observation. Eta is the regularization coefficient of the wave impedance low-rank regularization term used for balancing the data fidelity term
Figure BDA0002297996710000053
And the magnitude of the role played by the low-rank regularization term in the inversion, wherein the larger the value of the low-rank regularization term is, the larger the role played by the low-rank regularization term is.
(3) Solving of an objective function
For the objective function constructed in the previous step, the solution can be carried out in two steps. The first step is to solve for the optimal low rank approximation, i.e. to solve for the optimal value of:
Figure BDA0002297996710000054
the optimized expression for this equation can be written as:
Figure BDA0002297996710000055
the problem is solved using singular value decomposition and hard threshold shrinkage. Let phiiThe singular value of m is decomposed into:
Φim＝Αdiag(ω)ΒH
wherein, A and B respectively represent two unitary matrixes generated after singular value decomposition. And omega is a singular value vector obtained by singular value decomposition. diag (ω) denotes a diagonal matrix spanned by a singular value vector ω obtained by singular value decomposition, and the element on the diagonal is ω. In view of this, the optimal low rank can be approximated
Figure BDA0002297996710000056
The solution of (a) is written as:
Figure BDA0002297996710000057
where λ represents the threshold used in hard threshold contraction, HTλ() Represents a hard threshold function defined by the formula:
Figure BDA0002297996710000058
where Γ represents the input HTλ() Data in the function.
In finding OiAfter the optimal solution of (3), O is fixed in the objective functioniThe update formula of the wave impedance inversion result can be obtained as follows:
Figure BDA0002297996710000061
this is a least squares problem which can be collated as to
Figure BDA0002297996710000062
The equation of (c):
Figure BDA0002297996710000063
wherein G is*And phii *Are G and phi respectivelyiThe companion operator of (c). This problem can be solvedTo solve quickly using the conjugate gradient method.
Finally, inversion results are obtained
Figure BDA0002297996710000064
And restoring the original wave impedance model into a 2-dimensional section or a 3-dimensional data volume according to a process opposite to the process of constructing the long column vector m, namely completing the whole inversion process.
The specific calculation example of the invention is as follows: part of the data from the Marmousi2 model as a true model of wave impedance is shown in fig. 3, where there are 452 sample points in the depth direction and 512 samples in the distance direction. The initial model of wave impedance is a gaussian low-pass filter of the real model as shown in fig. 4; the seismic data used for inversion is obtained by obtaining a reflection coefficient by using a wave impedance real model, then performing convolution on the reflection coefficient and a Rake wavelet with a main frequency of 40Hz, and adding 20% of white Gaussian noise as shown in FIG. 5. The wavelet is a Rake wavelet with a dominant frequency of 40Hz, as shown in FIG. 6; the results are shown in FIG. 7.
It can be seen that the wave impedance inversion result obtained by the method is very similar to the real wave impedance model, and the boundary surface of the stratum can be clearly depicted in the inversion result.
While embodiments of the invention have been disclosed above, it is not limited to the applications set forth in the description and embodiments, which are fully applicable to various fields of endeavor for which the invention is intended, and further modifications may readily be effected therein by those skilled in the art, without departing from the general concept defined by the claims and their equivalents, which are to be limited not to the specific details shown and described herein.

Claims (3)

1. A seismic wave impedance inversion method based on wave impedance low-rank regularization is characterized by comprising the following steps:
step one, inputting seismic data and an initial wave impedance model, and constructing an initial low-rank approximate matrix of the wave impedance model: constructing an initial low-rank approximate matrix by adopting a similarity search-based method;
step two, establishing a wave impedance low-rank regularization target function
Let the whole wave impedance profile be divided into K sub-data blocks with overlap, in which the ith sub-data block is denoted as diThen, the wave impedance optimal low rank approximation problem is expressed as:
Figure FDA0003159529690000011
Φiis a similarity measure and selection operator to sub-data block diIs a reference block for selecting the sum d from the seismic wave impedance miQ sub-data blocks with the smallest euclidean distance of (d) are arranged into a matrix of size P × Q, P representing sub-data block diLength of (d); o isiObtaining an optimal low-rank approximate matrix to be solved; | | non-woven hairFExpressing the Frobenius norm of the matrix, and rank () is the operation of solving the rank of the matrix; λ represents a low-rank constrained regularization parameter;
on the basis of the formula (1), constructing an objective function of a wave impedance low-rank regularization seismic wave impedance inversion problem:
Figure FDA0003159529690000012
g represents a forward operator during seismic wave impedance inversion, and S represents seismic data obtained through observation; eta is the regularization coefficient of the wave impedance low-rank regularization term used for balancing the data fidelity term
Figure FDA0003159529690000013
And the magnitude of the role played by the low-rank regularization term in inversion, wherein the larger the value is, the larger the role played by the low-rank regularization term is;
step three, solving the objective function
For the objective function of the formula (2), the solution is carried out in two steps, firstly, the optimal low rank approximation is solved, namely, the optimal value of the formula (1) is solved, and the optimal low rank approximation of the formula (1)
Figure FDA0003159529690000014
Writing into:
Figure FDA0003159529690000015
solving equation (3) using singular value decomposition and hard threshold shrinkage, let ΦiThe singular value of m is decomposed into:
Φim＝Αdiag(ω)ΒH (4)
a and B respectively represent two unitary matrixes generated after singular value decomposition; omega is a singular value vector obtained by singular value decomposition; diag (ω) represents a diagonal matrix spanned by singular value vectors ω obtained by singular value decomposition; h represents a conjugate operator; optimal low rank approximation
Figure FDA0003159529690000021
The solution of (a) is written as:
Figure FDA0003159529690000022
wherein HTλ() Represents a hard threshold function defined by the formula:
Figure FDA0003159529690000023
where Γ represents the input HTλ() Data in the function;
in finding OiAfter the optimal solution of (3), O is fixed in the objective function of equation (2)iThe update formula of the wave impedance inversion result is obtained as follows:
Figure FDA0003159529690000024
equation (7) is a least squares problem, and equation (7) is formulated with respect to
Figure FDA0003159529690000025
The equation of (c):
Figure FDA0003159529690000026
wherein G is*And phii *Are G and phi respectivelyiUsing conjugate gradient method to quickly solve to obtain inversion result
Figure FDA0003159529690000027
Step four, inversion results are obtained
Figure FDA0003159529690000028
And restoring the original wave impedance model into a 2-dimensional section or a 3-dimensional data volume according to the process opposite to the process for forming the seismic wave impedance m, namely completing the whole inversion process.
2. The seismic wave impedance inversion method based on wave impedance low-rank regularization as claimed in claim 1, wherein η is determined by a cross-validation method, namely: and adjusting according to the difference between the inversion result and the true value, and selecting a value which enables the inversion error to be minimum, namely the value of eta.
3. The seismic wave impedance inversion method based on wave impedance low-rank regularization as claimed in claim 1, wherein in the first step, the step of constructing an initial low-rank approximation matrix of wave impedance comprises the following steps: arranging initial seismic wave impedance m into a long column vector according to columns; a sub-data block d is then determined as a referenceiSuppose a sub-data block diThe length is P; then sequentially selecting and d from miQ sub-data blocks with the smallest euclidean distance; and arranging the Q sub-data blocks into a matrix, and finishing the construction of the initial low-rank approximate matrix.
CN201911210757.2A 2019-12-02 2019-12-02 Seismic wave impedance inversion method based on wave impedance low-rank regularization Expired - Fee Related CN110865409B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911210757.2A CN110865409B (en) 2019-12-02 2019-12-02 Seismic wave impedance inversion method based on wave impedance low-rank regularization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911210757.2A CN110865409B (en) 2019-12-02 2019-12-02 Seismic wave impedance inversion method based on wave impedance low-rank regularization

Publications (2)

Publication Number Publication Date
CN110865409A CN110865409A (en) 2020-03-06
CN110865409B true CN110865409B (en) 2021-08-31

Family

ID=69658099

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911210757.2A Expired - Fee Related CN110865409B (en) 2019-12-02 2019-12-02 Seismic wave impedance inversion method based on wave impedance low-rank regularization

Country Status (1)

Country Link
CN (1) CN110865409B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113552631B (en) * 2021-08-16 2023-11-03 中煤科工集团西安研究院有限公司 Time-frequency double-domain regularized sparse deconvolution method and device for narrowband signals

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106204468A (en) * 2016-06-27 2016-12-07 深圳市未来媒体技术研究院 A kind of image de-noising method based on ReLU convolutional neural networks
CN107561576A (en) * 2017-08-31 2018-01-09 电子科技大学 Seismic signal method based on dictionary learning regularization rarefaction representation
CN108037531A (en) * 2017-11-24 2018-05-15 电子科技大学 A kind of seismic inversion method and system based on the full variational regularization of broad sense
CN108535775A (en) * 2018-03-30 2018-09-14 中国石油大学(北京) Non-stationary seismic data sound impedance inversion method and device
CN110208862A (en) * 2019-07-04 2019-09-06 电子科技大学 A kind of seismic inversion method based on the mixing sparse regularization of high-order fractional order ATpV

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4510128B1 (en) * 2009-05-29 2010-07-21 株式会社エイツー Earthquake judgment system and earthquake analysis method
US20150198729A1 (en) * 2014-01-13 2015-07-16 Cgg Services Sa Regularization of spatially aliased seismic data

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106204468A (en) * 2016-06-27 2016-12-07 深圳市未来媒体技术研究院 A kind of image de-noising method based on ReLU convolutional neural networks
CN107561576A (en) * 2017-08-31 2018-01-09 电子科技大学 Seismic signal method based on dictionary learning regularization rarefaction representation
CN108037531A (en) * 2017-11-24 2018-05-15 电子科技大学 A kind of seismic inversion method and system based on the full variational regularization of broad sense
CN108535775A (en) * 2018-03-30 2018-09-14 中国石油大学(北京) Non-stationary seismic data sound impedance inversion method and device
CN110208862A (en) * 2019-07-04 2019-09-06 电子科技大学 A kind of seismic inversion method based on the mixing sparse regularization of high-order fractional order ATpV

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Seismic acoustic impedance inversion with multi-parameter regularization;Shu Li et al.;《Journal of Geophysics and Engineering》;20170328;第520-532页 *
Seismic Data Denoising Based on Sparse and Low-Rank Regularization;Shu Li et al.;《energies》;20200113;第1-16页 *
WHEN SPARSITY MEETS LOW-RANKNESS: TRANSFORM LEARNING WITH NON-LOCAL LOW-RANK CONSTRAINT FOR IMAGE RESTORATION;Bihan Wen et al.;《2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS,SPEECH AND SIGNAL PEOCESSING》;20170619;第1-5页 *
基于ＡＴＶ 正则化与初始模型约束的波阻抗反演;李曙等;《吉首大学学报（ 自然科学版）》;20180930;第39卷(第5期);第33-37页 *
广义稀疏域地震反演方法研究及应用;李曙;《中国博士学位论文全文数据库 基础科学辑》;20200115;第23-26页 *

Also Published As

Publication number Publication date
CN110865409A (en) 2020-03-06

Similar Documents

Publication Publication Date Title
CN108037531B (en) A kind of seismic inversion method and system based on the full variational regularization of broad sense
CN111596366B (en) Wave impedance inversion method based on seismic signal optimization processing
Chen Automatic velocity analysis using high-resolution hyperbolic Radon transform
CN103954992B (en) Deconvolution method and device
CN106597545B (en) A kind of horizontal fracture earthquake prestack inversion method and apparatus
CN110687597B (en) Wave impedance inversion method based on joint dictionary
Liu et al. Smooth Deep Learning Magnetotelluric Inversion based on Physics-informed Swin Transformer and Multi-Window Savitzky-Golay Filter
Li et al. Magnetotelluric noise suppression via convolutional neural network
CN110865409B (en) Seismic wave impedance inversion method based on wave impedance low-rank regularization
CN104730572A (en) Diffracted wave imaging method and device based on L0 semi-norm
Yang et al. Wasserstein distance-based full-waveform inversion with a regularizer powered by learned gradient
Sun et al. Full-waveform inversion using a learned regularization
CN117787042A (en) Underground shallow closed space wave field split decoupling method
CN111273346B (en) Method, device, computer equipment and readable storage medium for removing deposition background
CN116068644B (en) Method for improving resolution and noise reduction of seismic data by using generation countermeasure network
CN109655891B (en) Method and system for overcoming full waveform inversion cycle skip
CN106291676A (en) A kind of geological data reconstructing method based on matching pursuit algorithm
Yuan et al. Horizon auto-picking with quantitative uncertainty evaluation by using a modified VQ-VAE framework
CN106291675A (en) A kind of geological data reconstructing method based on base tracer technique
CN110850473B (en) Seismic wave impedance inversion method based on sparse transform learning
Yu et al. Seismic detection method for small-scale discontinuities based on dictionary learning and sparse representation
ZHANG et al. Estimation of Regular Parameters for Impedance Inversion
Zhao et al. Smoothness: The key factor in well-log information-assisted PINNtomo
Yang et al. Seismic random noise attenuation using DnCNN with stratigraphic dip constraint
Luo et al. Low-frequency prediction based on multiscale and cross-scale deep networks in full-waveform inversion

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right

Effective date of registration: 20230104

Address after: 276000 Building 001, Torch Building, Keji Avenue, Linyi City, Shandong Province

Patentee after: Linyi Gaoxin tap water Co.,Ltd.

Address before: 418000 Huai Dong Road, Hecheng District, Huaihua, Hunan Province, No. 180

Patentee before: HUAIHUA University

TR01 Transfer of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210831

CF01 Termination of patent right due to non-payment of annual fee