CN113866832A - Convolutional neural network elastic parameter prediction method based on virtual well training - Google Patents

Convolutional neural network elastic parameter prediction method based on virtual well training Download PDF

Info

Publication number
CN113866832A
CN113866832A CN202111065733.XA CN202111065733A CN113866832A CN 113866832 A CN113866832 A CN 113866832A CN 202111065733 A CN202111065733 A CN 202111065733A CN 113866832 A CN113866832 A CN 113866832A
Authority
CN
China
Prior art keywords
wave velocity
data
neural network
convolutional neural
residual
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.)
Pending
Application number
CN202111065733.XA
Other languages
Chinese (zh)
Inventor
杨继东
孙加星
李振春
黄建平
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.)
China University of Petroleum East China
Original Assignee
China University of Petroleum East China
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 China University of Petroleum East China filed Critical China University of Petroleum East China
Priority to CN202111065733.XA priority Critical patent/CN113866832A/en
Publication of CN113866832A publication Critical patent/CN113866832A/en
Pending legal-status Critical Current

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
    • G01V1/40—Seismology; Seismic or acoustic prospecting or detecting specially adapted for well-logging
    • G01V1/44—Seismology; Seismic or acoustic prospecting or detecting specially adapted for well-logging using generators and receivers in the same well
    • G01V1/48—Processing data
    • G01V1/50—Analysing data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/616—Data from specific type of measurement
    • G01V2210/6169—Data from specific type of measurement using well-logging

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Environmental & Geological Engineering (AREA)
  • Remote Sensing (AREA)
  • Geophysics (AREA)
  • Geology (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

The invention discloses a method for predicting elastic parameters of a convolutional neural network based on virtual well training, which belongs to the technical field of seismic exploration attribute analysis and fine interpretation. The method has the advantages that the parameter crosstalk artifact existing in the traditional inversion technology can be effectively avoided, the accuracy of the inversion result is ensured, and the inversion efficiency is improved.

Description

Convolutional neural network elastic parameter prediction method based on virtual well training
Technical Field
The invention relates to the technical field of seismic exploration attribute analysis and fine interpretation, in particular to a convolutional neural network elastic parameter prediction method based on virtual well training.
Background
The deep learning technology is a data-driven learning algorithm based on a network, and the accuracy of a prediction result of the method has a direct relation with data of a training network. At present, large-scale training is carried out on a large number of pre-stack angle gather seismic data and corresponding label data to obtain a nonlinear mapping relation between the pre-stack angle gather seismic data and the label data when a convolutional neural network is adopted to predict elastic parameters, and then the relation is applied to the seismic data to estimate three elastic parameters of seismic longitudinal wave velocity, transverse wave velocity and density. The label data is real logging data, and the logging data is usually less than dozens of logs in the description of a typical geophysical reservoir, so that the accuracy of deep learning reservoir parameter prediction is severely limited.
Disclosure of Invention
The invention discloses a method for predicting elastic parameters of a convolutional neural network based on virtual well training, which aims to solve the technical problem that label data is lacked in intelligent learning elastic parameter prediction.
In order to achieve the purpose, the invention adopts the following technical scheme:
a convolutional neural network elastic parameter prediction method based on virtual well training comprises the following steps:
s1, acquiring input data;
s2: carrying out statistical analysis on the variation information of the logging curve to obtain the data variance and the spatial correlation length of the parameter to be inverted, and representing the spatial correlation of the parameter to be inverted by using a theoretical Gaussian variation function;
s3: constructing a covariance matrix by the variation function according to a certain arrangement mode;
s4: constructing a correlation matrix R of three elastic parameters of longitudinal wave velocity, transverse wave velocity and density;
s5: multiplying a matrix R representing the spatial correlation with a random non-correlation signal which is in accordance with standard normal distribution and has the length of 3 x n to obtain three-parameter residual curves of longitudinal wave velocity, transverse wave velocity and density; adding the obtained residual virtual well data to the input initial model to construct three elastic parameters of the virtual well;
s6: calculating a reflection coefficient by solving a Zoeppritz equation by using the constructed virtual well, and convoluteing the seismic wavelets w (t) to obtain a synthesized prestack angle gather and provide a required seismic gather and a label data set for training a convolutional neural network;
s7: and (3) building a convolutional neural network structure, training a network by using simulation data, and applying the trained neural network to predict three elastic parameters of longitudinal wave velocity, transverse wave velocity and density on the actual pre-stack angle seismic data.
As a further preferred aspect of the present invention, in step S1, the input data includes: well log x (t), initial three-parameter model m0(t), seismic angle domain gather data d (t), seismic wavelets w (t).
As a further preferred aspect of the present invention, in step S2, the theoretical gaussian variation function is: :
Figure BDA0003258303770000021
where h is the distance between the two parameters and a is the correlation length.
As a further preferred aspect of the present invention, in step S3, the covariance matrix is specifically formed as:
Figure BDA0003258303770000022
where γ is a variation function, dmaxThe maximum separation distance between data points of the well data curve.
As a further preferred aspect of the present invention, in step S4, simulating a plurality of elastic parameters, in addition to considering the spatial vertical correlation, also needs to consider the correlation among the plurality of parameters; when longitudinal wave velocity, transverse wave velocity and density parameters are inverted simultaneously, the method specifically comprises the following steps:
(1) a covariance matrix S is introduced to characterize the correlation between the parameters, expressed as,
Figure BDA0003258303770000023
wherein the content of the first and second substances,
Figure BDA0003258303770000024
Figure BDA0003258303770000025
and
Figure BDA0003258303770000026
respectively residual longitudinal wave, residual transverse wave and residual density variance, n is the number of sampling points of well data, p, s and den respectively represent the residual longitudinal wave velocity, residual transverse wave velocity and residual density, mu is the mean value of corresponding variables, sigma is the mean value of the corresponding variables, andp,s、σp,denand σden,sRespectively, covariance of the corresponding variables;
(2) the comprehensive covariance matrix K was calculated using the Kronecker product:
Figure BDA0003258303770000027
(3) the comprehensive covariance matrix is decomposed using Cholesky decomposition as:
K＝RRT (5)
wherein, the sizes of K and R are respectively 3 × n by 3 × n.
As a further preferred aspect of the present invention, step S5, parameter residual curve RtThe concrete form of (A) is as follows:
Rt＝Ru (6)；
the constructed virtual well may be represented as:
Figure BDA0003258303770000031
wherein m isp，msAnd mdenThe virtual well variation curve is in accordance with geological prior information.
The method has the advantages that firstly, the limited well data of the work area are subjected to statistical analysis to obtain a large number of virtual wells with physical significance, then the virtual wells are used for calculating the pre-stack angle gather seismic data based on an accurate forward equation, then the pre-stack angle gather seismic data and the corresponding virtual well data are calculated by adopting numerical simulation, the convolutional neural network is trained to obtain a complex nonlinear mapping relation between the pre-stack angle gather seismic data and the pre-stack angle gather seismic data, and finally three elastic parameters of longitudinal wave velocity, transverse wave velocity and density are predicted.
Compared with the traditional geophysical inversion technology, the technical method can effectively avoid the multi-parameter crosstalk phenomenon existing in the traditional inversion technology. In addition, when the initial model precision is not high, the elastic parameter resolution of the inversion is higher than the nonlinear inversion result of Cauchy constraint. The same phase axis in the inversion profile is continuous, the reservoir structure is clear, and the reservoir position inversion is accurate. The network prediction time only needs a few seconds, and the inversion efficiency is improved while the accuracy of the inversion result is ensured in the network elastic parameter prediction technology based on the virtual well training.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a longitudinal wave velocity model v according to an embodiment of the present inventionp(t) view;
FIG. 3 is a transverse wave velocity model v according to an embodiment of the present inventions(t) view;
FIG. 4 is a graph of a density model den (t) in accordance with an embodiment of the present invention;
FIG. 5 is an initial longitudinal wave velocity model v according to an embodiment of the present inventionp0(t) view;
FIG. 6 is an initial shear velocity model v in an embodiment of the present inventions0(t) view;
FIG. 7 is a graph of an initial density model den in an embodiment of the present invention0(t) view;
FIG. 8 is a result of longitudinal wave velocity prediction using a conventional geophysical nonlinear inversion method;
FIG. 9 is a result of shear wave velocity prediction by a conventional geophysical nonlinear inversion method;
FIG. 10 is a graph of density results predicted by a conventional geophysical nonlinear inversion method;
FIG. 11 is a schematic diagram of a convolutional neural network compressional velocity prediction result based on virtual well training according to an embodiment of the present invention;
FIG. 12 is a schematic diagram of a velocity prediction result of a convolutional neural network based on virtual well training according to an embodiment of the present invention;
fig. 13 shows the result of predicting the density of the convolutional neural network based on the training of the virtual well according to the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A convolutional neural network elastic parameter prediction method based on virtual well training comprises the following steps:
(1) input data is acquired.
The input data includes: log data curve x (t), initial model m0(t), seismic data d (t), seismic wavelets w (t).
(2) And (3) carrying out statistical analysis on the well curve change information x (t), obtaining the data variance C and the spatial correlation length a of the parameter to be inverted, and representing the spatial correlation of the parameter to be inverted by replacing a correlation function with a theoretical Gaussian variation function gamma.
The specific expression of the gaussian variation function is as follows:
Figure BDA0003258303770000041
where h is the distance between the two parameters and a is the correlation length.
(3) The variance function constructs a covariance matrix C according to a certain arrangement mode: ,
Figure BDA0003258303770000042
where γ is a variation function, dmaxThe maximum separation distance between data points of the well data curve.
(4) And simultaneously constructing a plurality of elastic parameter models.
Simulating a plurality of elastic parameters, wherein the correlation among the plurality of parameters is also considered besides the spatial vertical correlation; when longitudinal wave velocity, shear wave velocity and density parameters are inverted simultaneously, the method specifically comprises the following steps:
4.1 introducing a covariance matrix S characterizing the correlation between the parameters, expressed as,
Figure BDA0003258303770000043
wherein the content of the first and second substances,
Figure BDA0003258303770000051
Figure BDA0003258303770000052
and
Figure BDA0003258303770000053
respectively residual longitudinal wave, residual transverse wave and residual density variance, n is the number of sampling points of well data, p, s and den respectively represent the residual longitudinal wave velocity, residual transverse wave velocity and residual density, mu is the mean value of corresponding variables, sigma is the mean value of the corresponding variables, andp,s、σp,denand σden,sRespectively, covariance of the corresponding variables;
4.2 calculate the comprehensive covariance matrix K using the Kronecker product: ,
Figure BDA0003258303770000054
4.3 decompose the covariance matrix K using Cholesky to:
K＝RRT (5)
wherein, the sizes of K and R are respectively 3 × n by 3 × n.
(5) Multiplying a random non-correlation signal u with the standard normal distribution length of 3 x n by a characterization space correlation and parameter correlation matrix R to obtain a parameter residual curve; and adding the obtained residual virtual well data to an initial model of the parameters to be inverted to construct a virtual well conforming to the geological significance.
Parameter residual curve RtThe concrete form of (A) is as follows:
Rt＝Ru (6)；
the constructed virtual well is represented as:
Figure BDA0003258303770000055
wherein m isp，msAnd mdenThe virtual well variation curve is in accordance with geological prior information.
(6) Calculating reflection coefficient by using accurate Zoeppritz equation, and performing convolution with seismic wavelets w (t) to calculate synthetic prestack angle gather seismic record dsynA large number of data sets are provided for training convolutional neural networks.
(7) And (3) building a convolutional neural network structure, training a network by using simulated virtual well data, and applying the trained network to predict three parameters of longitudinal wave velocity, transverse wave velocity and density on the actual pre-stack angle seismic data d (t).
The method disclosed by the method is applied to the SEAM model to invert the reservoir parameters, and ideal longitudinal wave velocity, transverse wave velocity and density inversion results are obtained. Longitudinal wave velocity model vp(t) (shown in FIG. 2), shear velocity model vs(t) (shown in FIG. 3), density model den (t) (shown in FIG. 4), and initial model of longitudinal wave velocity vp0(t) (shown in FIG. 5), initial model v of transverse wave velocitys0(t) (shown in FIG. 6), Density initialization model den0(t) (as shown in FIG. 7). The local SEAM model has 297 time sampling points longitudinally with 8ms time interval, 2200 CDPs transversely with 5 m distance between every two CDPs. Prestack seismic data were obtained using reflection coefficients calculated using the Zoeppritz equation and convolved with the raekwave. Using Cauchy constraints for contrasting inversion effectsThe inversion results obtained by the nonlinear inversion method are shown in fig. 8-10. Compared with the traditional geophysical inversion result, the virtual well training-based convolutional neural network elastic parameter prediction technology (shown in figures 11-13) is adopted, the parameter crosstalk phenomenon can be effectively avoided, the inversion result precision is high, particularly the density inversion result (shown in figure 13), the network prediction time is short, the inversion result precision is guaranteed, and the inversion efficiency is greatly improved.
The method comprises the steps of firstly carrying out statistical analysis on well data to simulate a large number of virtual wells with actual geological significance after acquiring input seismic data, seismic wavelets, well data and an initial model, then calculating a reflection coefficient by using a zoeppritz equation, convoluteing the seismic wavelets by the reflection coefficient to obtain pre-stack angle gather seismic data, using the pre-stack seismic data as input data of a network, using the virtual wells as output data of the network to train the network, and finally carrying out pre-measurement on the actual data by using the trained network to obtain a final inversion result.
It is to be understood that the above description is not intended to limit the present invention, and the present invention is not limited to the above examples, and those skilled in the art may make modifications, alterations, additions or substitutions within the spirit and scope of the present invention.

Claims (6)

1. A convolutional neural network elastic parameter prediction method based on virtual well training is characterized by comprising the following steps:
s1, acquiring input data;
s2: carrying out statistical analysis on the variation information of the logging curve to obtain the data variance and the spatial correlation length of the parameter to be inverted, and representing the spatial correlation of the parameter to be inverted by using a theoretical Gaussian variation function;
s3: constructing a covariance matrix by the variation function according to a certain arrangement mode;
s4: constructing a correlation matrix R of three elastic parameters of longitudinal wave velocity, transverse wave velocity and density;
s5: multiplying a matrix R representing spatial correlation by a random non-correlation signal which is in accordance with standard normal distribution and has the length of 3 x n to obtain three parameter residual curves of longitudinal wave velocity, transverse wave velocity and density; adding the obtained residual virtual well data to the input initial model to construct three elastic parameters of the virtual well;
s6: calculating a reflection coefficient by solving a Zoeppritz equation by using the constructed virtual well, and convoluteing the seismic wavelets w (t) to obtain a synthesized prestack angle gather and provide a required seismic gather and a label data set for training a convolutional neural network;
s7: and (3) building a convolutional neural network structure, training a network by using simulation data, and applying the trained neural network to predict three elastic parameters of longitudinal wave velocity, transverse wave velocity and density on the actual pre-stack angle seismic data.
2. The method for predicting elastic parameters of a convolutional neural network based on virtual well training as claimed in claim 1, wherein in step S1, the input data comprises: logging data curve x (t), initial parameter model m0(t), seismic angle domain gather data d (t), seismic wavelets w (t).
3. The method for predicting elastic parameters of convolutional neural network based on virtual well training as claimed in claim 1, wherein in step S2, the gaussian variation function can be expressed as:
Figure FDA0003258303760000011
where h is the distance between the two parameters and a is the correlation length.
4. The method for predicting elastic parameters of a convolutional neural network based on virtual well training as claimed in claim 1, wherein in step S3, the covariance matrix is specifically formed as:
Figure FDA0003258303760000012
where γ is a variation function, dmaxThe maximum distance between well data curve data points.
5. The method of claim 1, wherein in step S4, simulating a plurality of elastic parameters, in addition to the spatial vertical correlation, requires considering the correlation between the plurality of parameters; when longitudinal wave velocity, shear wave velocity and density parameters are inverted simultaneously, the method specifically comprises the following steps:
(1) a covariance matrix S is introduced to characterize the correlation between the parameters, expressed as,
Figure FDA0003258303760000021
wherein the content of the first and second substances,
Figure FDA0003258303760000022
and
Figure FDA0003258303760000023
respectively the variance of the residual longitudinal wave velocity, the residual transverse wave velocity and the residual density, n is the number of sampling points of well data, subscripts p, s and den respectively represent the residual longitudinal wave velocity, the residual transverse wave velocity and the residual density, mu is the mean value of corresponding variables, sigma is the average value of the corresponding variablesp,s、σp,denAnd σden,sCovariance of the corresponding variables, respectively;
(2) using the covariance matrices S and C, a Kronecker (Kronecker) product is calculated, obtaining a three-parameter integrated covariance matrix K:
Figure FDA0003258303760000024
(3) applying Cholesky decomposition to the matrix K to obtain a matrix R,
K＝RRT
where the matrices K and R are 3n by 3n, respectively.
6. The method for predicting elastic parameters of convolutional neural network based on virtual well training as claimed in claim 1, wherein in step S5, the parameter residual curve RtThe concrete form of (A) is as follows:
Rt＝Ru；
the constructed virtual well is represented as:
Figure FDA0003258303760000025
wherein m isp，msAnd mdenThe virtual well variation curve is in accordance with geological prior information.
CN202111065733.XA 2021-09-13 2021-09-13 Convolutional neural network elastic parameter prediction method based on virtual well training Pending CN113866832A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111065733.XA CN113866832A (en) 2021-09-13 2021-09-13 Convolutional neural network elastic parameter prediction method based on virtual well training

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111065733.XA CN113866832A (en) 2021-09-13 2021-09-13 Convolutional neural network elastic parameter prediction method based on virtual well training

Publications (1)

Publication Number Publication Date
CN113866832A true CN113866832A (en) 2021-12-31

Family

ID=78995524

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111065733.XA Pending CN113866832A (en) 2021-09-13 2021-09-13 Convolutional neural network elastic parameter prediction method based on virtual well training

Country Status (1)

Country Link
CN (1) CN113866832A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115616665A (en) * 2022-09-30 2023-01-17 中国科学院地质与地球物理研究所 Convolutional neural network processing method and device and electronic equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111507048A (en) * 2020-04-23 2020-08-07 中国石油大学(北京) Method, device, equipment and system for predicting gas content of tight sandstone reservoir
CN112100906A (en) * 2020-08-21 2020-12-18 中海石油(中国)有限公司 Data-driven large-scale density modeling method, computing device and storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111507048A (en) * 2020-04-23 2020-08-07 中国石油大学(北京) Method, device, equipment and system for predicting gas content of tight sandstone reservoir
CN112100906A (en) * 2020-08-21 2020-12-18 中海石油(中国)有限公司 Data-driven large-scale density modeling method, computing device and storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
洪忠 等: "基于岩石物理的致密碎屑岩气藏岩性及流体概率预测", 石油物探, vol. 54, no. 6, pages 735 - 744 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115616665A (en) * 2022-09-30 2023-01-17 中国科学院地质与地球物理研究所 Convolutional neural network processing method and device and electronic equipment

Similar Documents

Publication Publication Date Title
CN109709603B (en) Seismic horizon identification and tracking method and system
Genton Separable approximations of space‐time covariance matrices
Bouttier et al. Data assimilation concepts and methods March 1999
CN105954804B (en) Shale gas reservoir fragility earthquake prediction method and device
CN105353412A (en) Calculating method and system of well-to-seismic integration average speed field
CN104597490A (en) Multi-wave AVO reservoir elastic parameter inversion method based on precise Zoeppritz equation
CN111522063B (en) Pre-stack high-resolution fluid factor inversion method based on frequency division joint inversion
CN114662414B (en) Oil reservoir production prediction method based on graph wavelet neural network model
CN113687433B (en) Bi-LSTM-based magnetotelluric signal denoising method and system
CN113947198A (en) Logging curve reconstruction method based on nonlinear autoregressive neural network model
CN109100791B (en) Velocity inversion method based on longitudinal and transverse space constraints
CN111948712A (en) Pre-stack linear inversion method based on depth domain seismic record
CN113866832A (en) Convolutional neural network elastic parameter prediction method based on virtual well training
CN116047583A (en) Adaptive wave impedance inversion method and system based on depth convolution neural network
CN111058840A (en) Organic carbon content (TOC) evaluation method based on high-order neural network
CN104991272A (en) Earthquake speed disturbance modeling method for well-free earthquake reversion
AU608496B2 (en) Method for identifying anomalous noise amplitudes in seismic data
CN115862324A (en) Space-time synchronization graph convolution neural network for intelligent traffic and traffic prediction method
US6847921B2 (en) Method for analyzing spatially-varying noise in seismic data using Markov chains
CN114114414A (en) Artificial intelligence prediction method for 'dessert' information of shale reservoir
CN111368680B (en) Wave atom transformation-based deep learning anti-aliasing seismic data regularization method
CN112147677A (en) Oil and gas reservoir parameter tag data generation method and device
CN115598714B (en) Time-space coupling neural network-based ground penetrating radar electromagnetic wave impedance inversion method
CN112100906A (en) Data-driven large-scale density modeling method, computing device and storage medium
CN113253350B (en) Porosity inversion method based on joint dictionary

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