CN112698389A - Seismic data inversion imaging method and device - Google Patents
Seismic data inversion imaging method and device Download PDFInfo
- Publication number
- CN112698389A CN112698389A CN201911008369.6A CN201911008369A CN112698389A CN 112698389 A CN112698389 A CN 112698389A CN 201911008369 A CN201911008369 A CN 201911008369A CN 112698389 A CN112698389 A CN 112698389A
- Authority
- CN
- China
- Prior art keywords
- data
- imaging
- offset
- seismic
- result
- 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.)
- Granted
Links
- 238000003384 imaging method Methods 0.000 title claims abstract description 153
- 238000013508 migration Methods 0.000 claims abstract description 25
- 238000004088 simulation Methods 0.000 claims abstract description 25
- 230000006870 function Effects 0.000 claims description 60
- 238000000034 method Methods 0.000 claims description 44
- 230000008569 process Effects 0.000 claims description 25
- 230000005012 migration Effects 0.000 claims description 20
- 230000009466 transformation Effects 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 abstract description 8
- 238000004364 calculation method Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 5
- 230000006872 improvement Effects 0.000 description 5
- 238000009795 derivation Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000002939 conjugate gradient method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
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
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)
- Image Processing (AREA)
Abstract
The invention relates to a seismic data inversion imaging method and a device, comprising the following steps: acquiring seismic data, wherein the seismic data comprise an offset imaging result, an offset velocity field, a seismic source wavelet and observation data; calculating reverse offset simulation data according to the offset imaging result, the offset velocity field, the seismic source wavelet and the selected coding function; constructing a functional residual error in a logarithmic form according to the anti-migration simulation data and the super-cannon data combined by the observation data; and deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result according to the gradient until the residual error is smaller than a set low value to obtain a final imaging result. The invention greatly improves the tolerance and adaptability of the inversion imaging algorithm to the actual seismic data noise and improves the accuracy of the inversion imaging result.
Description
Technical Field
The invention relates to a seismic data inversion imaging method and device, and belongs to the technical field of oil-gas geophysical prospecting engineering.
Background
With the deep exploration and development of oil and gas, the requirement on the imaging precision of seismic data is higher and higher, and the conventional migration imaging method cannot meet the requirement of exploration and development. The seismic data inversion imaging method continuously approximates the inverse of a Hessian matrix in an iteration mode, adverse effects of an observation system, a wavelet band, a complex structure and the like are gradually eliminated from an imaging result, compared with a traditional migration imaging algorithm, the method has higher resolution, amplitude balance and fidelity, receives more and more attention in the industry, and is the next development direction of seismic imaging.
The seismic data inversion imaging method has many advantages, but the defect that the calculation amount is too large is irremediable. In order to improve the calculation efficiency of the seismic data inversion imaging algorithm, a multi-source strategy is often adopted to combine single guns into super guns, and the calculation time is greatly saved due to the reduction of the number of calculated guns. However, the multi-source strategy introduces more crosstalk noise, and the theory of the common suppression strategy (such as dynamic coding) is too simple and the effect is not obvious, so that the method still needs to be further improved. In addition, the adaptability of the inversion imaging method to actual data and the tolerance to seismic noise are directly determined by the quality of the residual of the functional of the seismic data inversion imaging, the current target functional is developed based on the L2 model of the residual, the theoretical derivation is established under a Bayes framework, the noise is required to accord with Gaussian distribution, namely only random noise can be processed, so that the tolerance to the seismic noise is low, the inversion imaging is inaccurate, the complex situation of the actual field cannot be adapted, and the popularization and the application of the inversion imaging method are greatly limited.
Disclosure of Invention
The invention aims to provide a seismic data inversion imaging method and a seismic data inversion imaging device, which are used for solving the problem that the inversion imaging result is inaccurate due to the fact that the tolerance of functional residual errors adopted by the existing inversion imaging method to seismic noise is low.
In order to solve the technical problem, the invention provides a seismic data inversion imaging method, which comprises the following steps:
acquiring seismic data, wherein the seismic data comprise an offset imaging result, an offset velocity field, a seismic source wavelet and observation data;
calculating reverse offset simulation data according to the offset imaging result, the offset velocity field, the seismic source wavelet and the selected coding function;
constructing a functional residual error according to the anti-migration simulation data and super-cannon data combined by the observation data, wherein the functional residual error adopts a logarithmic form;
and deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result according to the gradient until the residual error is smaller than a set low value to obtain a final imaging result.
In order to solve the above technical problem, the present invention further provides a seismic data inversion imaging apparatus, including a processor and a memory, wherein the processor is configured to process instructions stored in the memory to implement the following method:
acquiring seismic data, wherein the seismic data comprise an offset imaging result, an offset velocity field, a seismic source wavelet and observation data;
calculating reverse offset simulation data according to the offset imaging result, the offset velocity field, the seismic source wavelet and the selected coding function;
constructing a functional residual error according to the anti-migration simulation data and super-cannon data combined by the observation data, wherein the functional residual error adopts a logarithmic form;
and deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result according to the gradient until the residual error is smaller than a set low value to obtain a final imaging result.
The invention has the beneficial effects that: by constructing the functional residual in a logarithmic form, the residual term appears in the numerator and denominator simultaneously in the process of obtaining the gradient after derivation, so that even if the data contains large noise, the gradient can still not be influenced, and the larger the noise is, the more obvious the suppression is. Therefore, the method greatly improves the tolerance and adaptability of the inversion imaging algorithm to the actual seismic data noise, and improves the accuracy of the inversion imaging result.
As a further improvement of the method and the apparatus, in order to construct a suitable functional residual to reduce the influence of noise on the imaging result, the functional residual is calculated by the following formula:
wherein j (m) is a functional residual, m is an imaging result, s is a data weighting parameter, σ is a weight adjustment factor, and L (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
As a further improvement of the method and apparatus, in order to obtain an accurate imaging result, the update formula corresponding to the offset imaging result is:
wherein, C-1For sparse inverse transform operators, mkAs a result of imaging in the k-th iteration, mk+1As a result of imaging for the (k + 1) th iteration, αkIs the update step size of the kth iteration, g is the gradient, LTFor the transpose of the forward modeling operator, abs () is an absolute valued function.
As a further improvement of the method and apparatus, in order to obtain an optimal coding function, the selected coding function is selected by: and respectively coding the observation data with the set proportion by adopting N coding functions to obtain super shot data, and carrying out imaging test on the super shot data to obtain a corresponding imaging signal-to-noise ratio, wherein the coding function corresponding to the maximum imaging signal-to-noise ratio is the selected coding function.
As a further improvement of the method and apparatus, in order to fully suppress the crosstalk noise introduced by the multi-source strategy, the sparse transformation operator C is a dictionary learning basis function.
Drawings
FIG. 1 is a schematic flow chart of a seismic data inversion imaging method of the present invention;
FIG. 2 is a schematic diagram of a dimple model velocity field in accordance with the present invention;
FIG. 3 is a schematic diagram of shot record data obtained by simulating forward calculation of field shot in the present invention;
FIG. 4 is a diagram illustrating a finally determined optimal encoding function according to the present invention;
FIG. 5 is a schematic representation of inversion imaging results in the present invention;
FIG. 6 is a diagram illustrating the results of prior art L2 mode inversion imaging;
FIG. 7 is a diagram illustrating the convergence curve of the functional residual according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Example 1:
the embodiment provides a seismic data inversion imaging method, which is mainly conceived as constructing a functional residual in a logarithmic form, so that in the process of deriving the functional residual to obtain a gradient, a residual item can appear in a numerator and a denominator simultaneously, and further offset noise introduced by abnormal noise in seismic data can be suppressed, so that the tolerance and the adaptability of an inversion imaging algorithm to actual seismic data noise are improved. In addition, in the process of iteratively updating the imaging result, the dictionary learning is utilized to carry out sparse expression on the gradient, and the crosstalk noise introduced by multiple seismic sources is further suppressed, so that the high-quality gradient updating direction is obtained, and the purpose of steady, rapid and high-quality imaging facing to actual data is achieved.
Specifically, a flow chart corresponding to the seismic data inversion imaging method is shown in fig. 1, and specifically includes the following contents:
1) and acquiring seismic data such as a conventional offset imaging result, an offset velocity field, a seismic source wavelet and observation data acquired in the field, and presetting a control parameter-residual threshold (setting a low value).
Wherein the conventional offset imaging result can be obtained by a single-pass wave offset or reverse time offset imaging process. In practice, if the normal offset imaging process is not performed, the normal offset imaging result may be made 0. In this embodiment, the shot record data obtained by simulating the forward modeling of field shot is shown in fig. 3, the bar chart on the right side in the figure indicates the amplitude of the seismic data, the set low value is set to 1%, and the conventional migration imaging result is 0.
2) Selecting data with a set proportion (for example, about 5 percent) from the observation data in the step 1), and synthesizing the super shot data by utilizing five different coding functions. The five different coding functions are respectively polarity coding, frequency division coding, plane wave coding, amplitude coding and random time shift coding. And then, performing reverse time migration imaging test to obtain 5 migration profiles, respectively calculating the imaging signal-to-noise ratio of the migration profiles, and determining the coding function corresponding to the maximum signal-to-noise ratio as the optimal coding function E under the data.
In this embodiment, the dimpled model velocity field selected by the imaging test is shown in fig. 2, the size of the model grid is 260 × 188, the vertical and horizontal grid intervals are all 8m, the right-side bar graph in fig. 2 shows the velocity value size in m/s, and fig. 4 is the optimal coding function determined in this embodiment, that is, the polar coding function. The calculation formula of the imaging signal-to-noise ratio adopted by the embodiment is as follows:
wherein SNR is the imaging signal-to-noise ratio, mrefAs a reference model, mmigThe imaging result of the super shot data is obtained.
It should be noted that, when obtaining the optimal encoding function E, the type and data of the encoding function may be autonomously selected according to needs, and the method is not limited to the above-listed five encoding functions.
3) All observed data d are encoded by using the optimal encoding function EobsMake up into super big gun EdobsThen obtaining the inverse offset simulation data Ed according to the conventional offset imaging result, the offset velocity field, the seismic source wavelet and the optimal coding functioncal. Due to the composition of super cannon EdobsAnd acquiring the anti-offset analog data EdcalThe specific processes of (1) belong to the prior art, and are not described herein again.
4) Simulating data Ed according to the inverse offsetcalWith all observation data dobsCombined super gun EdobsConstructing a functional residual (namely a target functional), wherein a corresponding calculation formula is as follows:
wherein j (m) is a functional residual, m is an imaging result, s is a data weighting parameter, σ is a weight adjustment factor, and L (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
Unlike the target functional of the L2 model used in conventional seismic inversion imaging, the data fitting term of the target functional in equation (2) (i.e., the first term at the right end of equation 2) is in a weighted logarithmic form, and the data residual term after the derivation operation (i.e., L (m, ES) -Ed)obs) The method can simultaneously appear in the numerator and the denominator, so that even if the data contains large noise, the gradient can still not be influenced, and the noise is more suppressed and more obvious, so that the method greatly improves the adaptability of the algorithm to the actual data noise.
The embodiment utilizes the dictionary learning basis function based on the K-SVD as the sparse transformation operator C, has better sparsity compared with a fixed basis function (such as DCT, curvelet, Shearlet) and the like, and is more adaptive to each imaging section. And transforming the imaging result into a sparse domain through a sparse transformation operator C, and continuously constraining the sparsity in the inversion iteration process, so that crosstalk noise which cannot be well represented by the sparse transformation operator C is suppressed, and a high-quality clean imaging section is obtained.
In this embodiment, the dictionary learning is performed by using a K-SVD method, but is not limited to K-SVD, and as another embodiment, the dictionary learning may be performed by using a low rank graph preserving method, a matching pursuit method, a distributed dictionary learning method, or the like.
5) And (4) deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result based on the gradient and the sparsity of the gradient until the residual error value is smaller than a set low value to obtain a final imaging result.
According to the functional residual error, the calculation process for obtaining the gradient g is as follows:
better gradient profiles can be obtained through the formula (3), the influence of noise in the seismic data can be effectively removed, and the crosstalk noise introduced by multiple seismic sources is still contained. In order to improve the calculation efficiency, a second item of processing is required, namely a sparse constraint item of the gradient, the display and solution of the item are complex, the item is fused into the conjugate gradient method updating process of the imaging profile, and the simplicity of the algorithm can be improved, and the specific process is as follows:
a general iterative threshold shrinkage algorithm is available for solving the following equation:
J(x)＝||Lx-y||+λ||x||1
the iterative threshold shrinkage solution is:
wherein x is a model variable, y is a data variable, L is a forward modeling operator, lambda is a regularization parameter, and g is a gradient.
When x is equal to Cm and is substituted into the above formula (4), the following are present:
Cmk+1＝T(C(mk-αkg),λ)
moving C to the right end of the formula, so as to obtain the final imaging result and update the formula as follows:
wherein, C-1For sparse inverse transform operators, mkAs a result of imaging in the k-th iteration, mk+1As a result of imaging for the (k + 1) th iteration, αkThe updating step length for the kth iteration can be obtained by a parabolic fitting or linear searching method, g is a gradient, and L isTFor the transpose of the forward modeling operator, abs () is an absolute valued function.
And the application process of the second gradient term is to extract a basis function from the imaging section by using K-SVD dictionary learning, and continuously restrict the sparsity of the imaging section by using an iterative shrinkage threshold method to suppress crosstalk noise. And finally, iteratively updating the imaging result by a conjugate gradient method until the residual threshold set in the step 1) is met, namely the residual threshold is reduced to be less than 1%, and terminating the updating of the imaging result. In this embodiment, the final inversion imaging result is shown in fig. 5.
Fig. 6 shows the inversion imaging result of the L2 mode (i.e. conventional least squares offset) commonly used in the prior art, and it can be seen from comparing fig. 5 and fig. 6 that the final imaging result (fig. 5) of the present invention is greatly improved compared with the inversion imaging result (fig. 6) in the prior art, and the specific improvements are as follows:
(1) not only the low-frequency noise is well suppressed, but also the shallow high-frequency noise introduced by filtering in the conventional offset is not contained;
(2) the longitudinal resolution is greatly improved, especially near a complex unmasked fault;
(3) the amplitude is more balanced, and the illumination compensation at the two sides of the section is better.
Fig. 7 is a convergence curve of functional residual, wherein a dotted line represents a convergence curve of functional residual obtained by using a model L2 commonly used in the prior art, which is denoted by old, and a solid line represents a convergence curve of functional residual obtained by using the inversion imaging method of the present invention, which is denoted by new. As can be seen from FIG. 7, the functional residuals of the present invention can converge to smaller values, indicating a better fit of the data and less influence by seismic noise.
Example 2:
the embodiment provides a seismic data inversion imaging method, which specifically comprises the following steps:
(1) and acquiring seismic data such as a conventional offset imaging result, an offset velocity field, a seismic source wavelet and observation data acquired in the field, and presetting a control parameter-residual threshold (setting a low value).
The seismic data acquired in step (1) is the same as the seismic data acquired in step 1) of embodiment 1, and will not be described herein again.
(2) And (3) selecting data with a set proportion (for example, about 5%) from the observation data in the step (1), and combining the super shot data by using four different coding functions. The four different coding functions are respectively polarity coding, frequency division coding, plane wave coding and amplitude coding. And then, performing an image test by adopting reverse time migration to obtain 4 migration profiles, respectively calculating the imaging signal-to-noise ratio of the migration profiles, and determining the coding function corresponding to the maximum signal-to-noise ratio as the optimal coding function E under the data.
The process of obtaining the optimal encoding function E in step (2) is the same as the process of obtaining the optimal encoding function E in step 2) of embodiment 1, and the difference is only that the number of the adopted encoding functions is different, and details are not described here.
(3) All observed data d are encoded by using the optimal encoding function EobsMake up into super big gun EdobsThen obtaining the inverse offset simulation data Ed according to the conventional offset imaging result, the offset velocity field, the seismic source wavelet and the optimal coding functioncal. Due to the composition of super cannon EdobsAnd acquiring the anti-offset analog data EdcalIs in the prior art, and is not described hereThe description is given.
(4) Simulating data Ed according to the inverse offsetcalWith all observation data dobsCombined super gun EdobsConstructing a functional residual (namely a target functional), wherein a corresponding calculation formula is as follows:
wherein j (m) is a functional residual, m is an imaging result, s is a data weighting parameter, σ is a weight adjustment factor, and L (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
In step (4), simulating data Ed according to the reverse offsetcalWith all observation data dobsCombined super gun EdobsFor a specific process of constructing a functional residual, refer to step 4 in embodiment 1), which is not described herein again.
(5) And (4) deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result based on the gradient and the sparsity of the gradient until the residual error value is smaller than a set low value to obtain a final imaging result.
In step (5), the process of acquiring the gradient and the process of acquiring the final imaging result according to the gradient and the sparsity thereof may refer to step 5 in embodiment 1), that is, the calculation formula of the gradient g is:
the imaging result updating formula is as follows:
example 3:
the embodiment provides a seismic data inversion imaging method, which specifically comprises the following steps:
(A) and acquiring seismic data such as a conventional offset imaging result, an offset velocity field, a seismic source wavelet and observation data acquired in the field, and presetting a control parameter-residual threshold (setting a low value).
The seismic data acquired in step (a) is the same as the seismic data acquired in step 1) of embodiment 1, and will not be described herein again.
(B) And (C) selecting a set proportion (for example, about 5 percent) of data from the observation data in the step (A), and combining the data into super shot data by utilizing five different coding functions. The five different coding functions are respectively polarity coding, frequency division coding, plane wave coding, amplitude coding and random time shift coding. And then, performing reverse time migration imaging test to obtain 5 migration profiles, respectively calculating the imaging signal-to-noise ratio of the migration profiles, and determining the coding function corresponding to the maximum signal-to-noise ratio as the optimal coding function E under the data.
The process of obtaining the optimal encoding function E in step (B) is the same as the process of obtaining the optimal encoding function E in step 2) of embodiment 1, and is not described herein again.
(C) All observed data d are encoded by using the optimal encoding function EobsMake up into super big gun EdobsThen obtaining the inverse offset simulation data Ed according to the conventional offset imaging result, the offset velocity field, the seismic source wavelet and the optimal coding functioncal. Due to the composition of super cannon EdobsAnd acquiring the anti-offset analog data EdcalThe specific processes of (1) belong to the prior art, and are not described herein again.
(D) Simulating data Ed according to the inverse offsetcalWith all observation data dobsCombined super gun EdobsConstructing a functional residual (namely a target functional), wherein a corresponding calculation formula is as follows:
wherein j (m) is a functional residual, m is an imaging result, s is a data weighting parameter, σ is a weight adjustment factor, and L (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
In the embodiment, a fixed basis function (such as DCT, curvelet, Shearlet) is used as a sparse transform operator C, the imaging result is transformed into a sparse domain through the sparse transform operator C, and sparsity is continuously constrained in an inversion iteration process, so that crosstalk noise which cannot be well characterized by the sparse transform operator C is suppressed, and a high-quality clean imaging profile is obtained.
(E) And (4) deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result based on the gradient and the sparsity of the gradient until the residual error value is smaller than a set low value to obtain a final imaging result.
In step (E), the process of acquiring the gradient and the process of acquiring the final imaging result according to the gradient and the sparsity thereof refer to step 5 in embodiment 1), that is, the calculation formula of the gradient g is:
the imaging result updating formula is as follows:
example 4:
the embodiment provides a seismic data inversion imaging device, which comprises a processor and a memory, wherein the processor is used for processing instructions stored in the memory so as to implement the seismic data inversion imaging method in the embodiment 1. For those skilled in the art, corresponding computer instructions may be generated according to the seismic data inversion imaging method in embodiment 1 to obtain the seismic data inversion imaging apparatus in this embodiment, which is not described herein again.
Example 5:
the embodiment provides a seismic data inversion imaging device, which comprises a processor and a memory, wherein the processor is used for processing instructions stored in the memory so as to implement the seismic data inversion imaging method in the embodiment 2. For those skilled in the art, corresponding computer instructions may be generated according to the seismic data inversion imaging method in embodiment 2 to obtain the seismic data inversion imaging apparatus in this embodiment, which is not described herein again.
Example 6:
the present embodiment provides a seismic data inversion imaging apparatus, which includes a processor and a memory, where the processor is configured to process instructions stored in the memory to implement the seismic data inversion imaging method in embodiment 3. For those skilled in the art, corresponding computer instructions may be generated according to the seismic data inversion imaging method in embodiment 3 to obtain the seismic data inversion imaging apparatus in this embodiment, which is not described herein again.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting the protection scope thereof, and although the present application is described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present application, various changes, modifications or equivalents of the embodiments of the present application can be made, and these changes, modifications or equivalents are within the protection scope of the claims of the present invention.
Claims (10)
1. A seismic data inversion imaging method is characterized by comprising the following steps:
acquiring seismic data, wherein the seismic data comprise an offset imaging result, an offset velocity field, a seismic source wavelet and observation data;
calculating reverse offset simulation data according to the offset imaging result, the offset velocity field, the seismic source wavelet and the selected coding function;
constructing a functional residual error according to the anti-migration simulation data and super-cannon data combined by the observation data, wherein the functional residual error adopts a logarithmic form;
and deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result according to the gradient until the residual error is smaller than a set low value to obtain a final imaging result.
2. The seismic data inversion imaging method of claim 1, wherein the functional residual is calculated by the formula:
wherein j (m) is a functional residual, m is an imaging result, s is a data weighting parameter, σ is a weight adjustment factor, and L (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
3. The seismic data inversion imaging method of claim 2, wherein the update formula corresponding to the migration imaging result is:
wherein, C-1For sparse inverse transform operators, mkAs a result of imaging in the k-th iteration, mk+1As a result of imaging for the (k + 1) th iteration, αkIs the update step size of the kth iteration, g is the gradient, LTFor the transpose of the forward modeling operator, abs () is an absolute valued function.
4. A method of seismic data inversion imaging according to any of claims 1 to 3, wherein the selected encoding function is selected by: and respectively coding the observation data with the set proportion by adopting N coding functions to obtain super shot data, and carrying out imaging test on the super shot data to obtain a corresponding imaging signal-to-noise ratio, wherein the coding function corresponding to the maximum imaging signal-to-noise ratio is the selected coding function.
5. The seismic data inversion imaging method of claim 2 or 3, wherein the sparse transformation operator C is a dictionary learning basis function.
6. A seismic data inversion imaging apparatus comprising a processor and a memory, said processor being configured to process instructions stored in the memory to implement the method of:
acquiring seismic data, wherein the seismic data comprise an offset imaging result, an offset velocity field, a seismic source wavelet and observation data;
calculating reverse offset simulation data according to the offset imaging result, the offset velocity field, the seismic source wavelet and the selected coding function;
constructing a functional residual error according to the anti-migration simulation data and super-cannon data combined by the observation data, wherein the functional residual error adopts a logarithmic form;
and deriving the functional residual error to obtain a gradient, and continuously and iteratively updating the offset imaging result according to the gradient until the residual error is smaller than a set low value to obtain a final imaging result.
7. The seismic data inversion imaging apparatus of claim 6, wherein the functional residual is calculated by the formula:
wherein J (m) is functional residual error, m is imaging result, s is data weighting parameter, and σ is weight adjustment factorL (m, ES) ═ EdcalFor the inverse migration simulation data, L is the forward simulation operator, λ is the regularization parameter, EdobsIs super-cannon data combined by observation data, C is a sparse transformation operator, ES is a super-cannon seismic source function, | Cm | luminance1L being a sparse coefficient1A norm constraint term.
8. The seismic data inversion imaging apparatus of claim 7, wherein the update formula corresponding to the migration imaging result is:
wherein, C-1For sparse inverse transform operators, mkAs a result of imaging in the k-th iteration, mk+1As a result of imaging for the (k + 1) th iteration, αkIs the update step size of the kth iteration, g is the gradient, LTFor the transpose of the forward modeling operator, abs () is an absolute valued function.
9. The seismic data inversion imaging apparatus of any one of claims 6 to 8, wherein the selected encoding function is selected by: and respectively coding the observation data with the set proportion by adopting N coding functions to obtain super shot data, and carrying out imaging test on the super shot data to obtain a corresponding imaging signal-to-noise ratio, wherein the coding function corresponding to the maximum imaging signal-to-noise ratio is the selected coding function.
10. The seismic data inversion imaging apparatus of claim 7 or 8, wherein the sparse transform operator C is a dictionary learning basis function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911008369.6A CN112698389B (en) | 2019-10-22 | 2019-10-22 | Inversion imaging method and device for seismic data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911008369.6A CN112698389B (en) | 2019-10-22 | 2019-10-22 | Inversion imaging method and device for seismic data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112698389A true CN112698389A (en) | 2021-04-23 |
| CN112698389B CN112698389B (en) | 2024-02-20 |
Family
ID=75504985
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911008369.6A Active CN112698389B (en) | 2019-10-22 | 2019-10-22 | Inversion imaging method and device for seismic data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112698389B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120143506A1 (en) * | 2010-12-01 | 2012-06-07 | Routh Partha S | Simultaneous Source Inversion for Marine Streamer Data With Cross-Correlation Objective Function |
| US20130245954A1 (en) * | 2012-03-13 | 2013-09-19 | Seoul National University R&Db Foundation | Seismic imaging system using cosine transform in logarithmic axis |
| CN105388520A (en) * | 2015-10-22 | 2016-03-09 | 中国石油化工股份有限公司 | Seismic data pre-stack reverse time migration imaging method |
| US20170192118A1 (en) * | 2016-01-05 | 2017-07-06 | Schlumerger Technology Corporation | Amplitude Inversion on Partitioned Depth Image Gathers Using Point Spread Functions |
| CN107783190A (en) * | 2017-10-18 | 2018-03-09 | 中国石油大学(北京) | A kind of least square reverse-time migration gradient updating method |
| US20180164453A1 (en) * | 2015-05-29 | 2018-06-14 | Sub Salt Solutions Limited | Method for Improved Geophysical Investigation |
| CN108241173A (en) * | 2017-12-28 | 2018-07-03 | 中国石油大学(华东) | A kind of seismic data offset imaging method and system |
| CN108802813A (en) * | 2018-06-13 | 2018-11-13 | 中国石油大学(华东) | A kind of multi-component seismic data offset imaging method and system |
| CN109738952A (en) * | 2019-01-24 | 2019-05-10 | 吉林大学 | The direct offset imaging method in passive source based on full waveform inversion driving |
-
2019
- 2019-10-22 CN CN201911008369.6A patent/CN112698389B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120143506A1 (en) * | 2010-12-01 | 2012-06-07 | Routh Partha S | Simultaneous Source Inversion for Marine Streamer Data With Cross-Correlation Objective Function |
| US20130245954A1 (en) * | 2012-03-13 | 2013-09-19 | Seoul National University R&Db Foundation | Seismic imaging system using cosine transform in logarithmic axis |
| US20180164453A1 (en) * | 2015-05-29 | 2018-06-14 | Sub Salt Solutions Limited | Method for Improved Geophysical Investigation |
| CN105388520A (en) * | 2015-10-22 | 2016-03-09 | 中国石油化工股份有限公司 | Seismic data pre-stack reverse time migration imaging method |
| US20170192118A1 (en) * | 2016-01-05 | 2017-07-06 | Schlumerger Technology Corporation | Amplitude Inversion on Partitioned Depth Image Gathers Using Point Spread Functions |
| CN107783190A (en) * | 2017-10-18 | 2018-03-09 | 中国石油大学(北京) | A kind of least square reverse-time migration gradient updating method |
| CN108241173A (en) * | 2017-12-28 | 2018-07-03 | 中国石油大学(华东) | A kind of seismic data offset imaging method and system |
| CN108802813A (en) * | 2018-06-13 | 2018-11-13 | 中国石油大学(华东) | A kind of multi-component seismic data offset imaging method and system |
| CN109738952A (en) * | 2019-01-24 | 2019-05-10 | 吉林大学 | The direct offset imaging method in passive source based on full waveform inversion driving |
Non-Patent Citations (7)
| Title |
|---|
| ALEKSANDR ARAVKIN等: "Robust full-waveform inversion using the Student’s t-distribution", SEG TECHNICAL PROGRAM EXPANDED ABSTRACTS, pages 2669 - 2673 * |
| QINGYANG LI等: "Cross-correlation least-squares reverse time migration in the pseudo-time domain", JOURNAL OF GEOPHYSICS AND ENGINEERING, vol. 14, no. 04, pages 841 - 851, XP020318524, DOI: 10.1088/1742-2140/aa6b33 * |
| WOODON JEONG等: "Full Waveform Inversion Using Student’st Distribution: a Numerical Study for Elastic Waveform Inversion and Simultaneous-Source Method", PURE AND APPLIED GEOPHYSICS, no. 172, pages 1491 - 1509 * |
| 任浩然;黄光辉;王华忠;陈生昌;: "地震反演成像中的Hessian算子研究", 地球物理学报, vol. 56, no. 07, pages 2429 - 2436 * |
| 李庆洋等: "基于Student‘s分布的不依赖子波最小二乘逆时偏移", 地球物理学报, vol. 60, no. 12, pages 4792 - 4798 * |
| 郭书娟;马方正;段心标;王丽;: "最小二乘逆时偏移成像方法的实现与应用研究", 石油物探, vol. 54, no. 03, pages 301 - 308 * |
| 黄金强等: "各向异性介质Low-rank有限差分法纯ｑＰ波叠前平面波最小二乘逆时偏移", 地球物理学报, vol. 62, no. 08, pages 3106 - 3129 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112698389B (en) | 2024-02-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7427748B2 (en) | Image processing systems and medical information processing systems | |
| CN105513026B (en) | One kind being based on the non local similar compressed sensing reconstructing method of image | |
| CN109671029B (en) | Image denoising method based on gamma norm minimization | |
| CN112578471B (en) | Clutter noise removing method for ground penetrating radar | |
| CN112819723A (en) | High-energy X-ray image blind restoration method and system | |
| CN111461146A (en) | Change detection method based on sparse cross reconstruction | |
| CN114418886B (en) | Robust denoising method based on depth convolution self-encoder | |
| CN109584330A (en) | One kind approaching L based on compressed sensing0The gradient projection image rebuilding method of norm | |
| CN115541693A (en) | Forward model constrained neural network magnetic particle imaging reconstruction method and system | |
| CN107301631B (en) | SAR image speckle reduction method based on non-convex weighted sparse constraint | |
| CN112435162A (en) | Terahertz image super-resolution reconstruction method based on complex field neural network | |
| CN116934999A (en) | NeRF three-dimensional reconstruction system and method based on limited view angle image | |
| CN106934398A (en) | Image de-noising method based on super-pixel cluster and rarefaction representation | |
| CN113866826A (en) | Mixed domain seismic migration hessian matrix estimation method | |
| CN112698389A (en) | Seismic data inversion imaging method and device | |
| CN112184567A (en) | Multi-channel blind identification adaptive optical image restoration method based on alternate minimization | |
| CN112598711A (en) | Hyperspectral target tracking method based on joint spectrum dimensionality reduction and feature fusion | |
| CN116626765A (en) | Multichannel seismic deconvolution method based on two-dimensional K-SVD and convolution sparse coding | |
| CN110926611A (en) | Noise suppression method applied to compressed sensing spectral imaging system | |
| CN110728728A (en) | Compressed sensing network image reconstruction method based on non-local regularization | |
| CN116011338A (en) | Full waveform inversion method based on self-encoder and deep neural network | |
| CN109655916B (en) | Method and system for separating effective waves and multiples in seismic data | |
| Yufeng et al. | Research on SAR image change detection algorithm based on hybrid genetic FCM and image registration | |
| CN106054247B (en) | Method for calculating high-precision reflection coefficient based on converted wave seismic data | |
| Novosadová et al. | Piecewise-polynomial signal segmentation using reweighted convex optimization |
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 |