253 results about "Full waveform" patented technology

The invention discloses an underground anti-seepage structure leakage monitoring method and system based on discrete time, and relates to leakage detection. The method comprises the following steps: collecting transient electromagnetic response data generated by an excitation signal; according to the initial conductivity parameter and the electrode position parameter of the monitoring area, establishing a three-dimensional geophysical model of the monitoring area through a full waveform inversion algorithm; obtaining electromagnetic abnormal data by using a finite element forward modeling algorithm; according to the temperature data, a distributed temperature field analysis method is adopted to calculate the space gradient and the time change rate of the temperature, and temperature abnormal data are obtained; according to the pressure data and time sequence change characteristics of the pressure data, determining pressure abnormal data; fusing the electromagnetic abnormal data, the temperature abnormal data and the pressure abnormal data by using a Bayesian algorithm to obtain fused data; and inputting the fusion data into a pre-trained machine learning model to obtain a leakage detection result. Aiming at the low dynamic leakage monitoring precision of the underground anti-seepage structure, the dynamic monitoring precision is improved.

The invention relates to the technical field of atmospheric remote sensing, in particular to a satellite-borne laser radar-based multi-feature attention cloud top height estimation algorithm, which comprises the following steps of: full-waveform laser radar data acquisition and preprocessing: acquiring an original echo signal through a satellite-borne full-waveform laser radar system; performing time domain alignment, noise suppression and saturated signal elimination on the signals; a median filtering and wavelet threshold denoising algorithm is combined to eliminate high-frequency noise, and background radiation interference is eliminated through dynamic baseline correction; according to the method, a multi-stage noise reduction strategy combining sliding window median filtering and self-adaptive wavelet denoising is introduced in the waveform feature extraction stage, pulse interference and signal oscillation are effectively restrained, meanwhile, through a terrain matching error compensation mechanism, the detection accuracy is improved, and the detection accuracy is improved. The physical consistency of the elevation residual error under the real terrain background is enhanced, and the authenticity and the checkability of the point elevation return are improved.

The invention discloses a method for jointly detecting corrosion defects of a pressure pipeline by utilizing multiple guided wave modes, relates to an ultrasonic non-destructive testing technology, and is used for solving the problem that the accuracy is relatively low when a single guided wave mode is used for detecting. The method for detecting the corrosion defect of the pressure pipeline comprises the following steps: calculating ultrasonic guided wave dispersion data of the pressure pipeline, and determining the center frequency of a transducer according to the dispersion data; the method comprises the following steps: arranging ultrasonic transducer annular arrays on two sides of a to-be-detected pressure pipeline, and acquiring ultrasonic guided wave data detected by the pressure pipeline in an L (0, 1) mode and a T (0, 2) mode; preprocessing the array waveform data to obtain detection data which can be used for guided wave imaging; according to the L (0, 1) mode guided wave data, a full waveform inversion method is adopted to invert the wall thickness of the pipeline corrosion defect; judging whether the pipeline defect exceeds the detection depth range of the T (0, 2) mode, and if the pipeline defect exceeds the detection depth range, ending; and if not, inverting the wall thickness of the pipeline corrosion defect by adopting a full-waveform inversion method according to the T (0, 2) mode guided wave data.

The invention belongs to the technical field of seismic exploration, and discloses a robust full-waveform inversion method, system and device based on generative modeling, and a medium, and the method comprises the steps: obtaining seismic observation data; the method comprises the following steps: starting from random Gaussian noise, establishing a model space comprising a plurality of initial velocity models through an unconditional score model, and determining a global optimal initial velocity model of the model space by adopting a strategy search algorithm; de-noising is carried out on the global optimal initial velocity model based on back diffusion, seismological observation data are introduced as observation constraints, the velocity model is updated by calculating the mismatch gradient of forward modeling data of the current velocity model and the seismological observation data, and an implicit condition sample is obtained; performing forward diffusion processing on the implicit condition sample to obtain a speed model for a subsequent annealing time step; and when the annealing process reaches a preset condition, outputting a final speed model. According to the method, the robustness and accuracy of full-waveform inversion are improved, and the bottleneck problem of traditional full-waveform inversion is solved.

A method can include receiving seismic data of a seismic survey of a subsurface region; performing a full waveform inversion using the seismic data and a model of the subsurface region to determine one or more characteristics of the subsurface region, where the model of the subsurface region includes one or more angle-dependent model parameters; and outputting the one or more characteristics of the subsurface region.

The invention discloses a sound wave full waveform inversion method and system based on optimization effective boundary storage, and the method comprises the steps: constructing an initial geological velocity model, and obtaining seismic wave parameter information; performing forward modeling according to the initial geological velocity model and the seismic wave parameter information to obtain seismic simulation data and forward seismic source wave field data; acquiring actual seismological observation data of the geologic model to be measured, and constructing an objective function according to the actual seismological observation data and the seismic simulation data; re-sampling the wave field data of the target area according to the forward seismic source wave field data; restoring the resampled wave field data of the target area to an original data state by using an interpolation method, and then reconstructing a forward transmission seismic source wave field; zero-delay cross-correlation is carried out according to the reconstructed forward transmission seismic source wave field and the reconstructed backward transmission residual wave field to obtain a gradient; and carrying out iterative updating on the target function by adopting a gradient optimization algorithm until a final geological velocity model is output.

The invention relates to the technical field of tunnel engineering, and discloses a tunnel earthquake three-dimensional full-waveform inversion method, which comprises the following steps: S1, establishing a three-dimensional initial velocity model: establishing the three-dimensional initial velocity model based on a common reflection point migration method of ellipsoid expansion; s2, forward modeling three-dimensional full-space wave field characteristic analysis: performing forward modeling in the established typical unfavorable geologic body model, setting a seismic source point on a tunnel wall or a tunnel face, setting a detection point on the tunnel wall, and analyzing three-dimensional full-space wave field signal characteristics received at the positions of the tunnel wall and the tunnel face; s3, full-waveform inversion numerical simulation: realizing a full-waveform inversion numerical model experiment on the basis of the established initial velocity model to obtain simulation data; s5, performing three-dimensional full-waveform inversion imaging; and S6, result verification. The tunnel earthquake three-dimensional full-waveform inversion precision can be improved, and the tunnel earthquake three-dimensional advanced prediction method which is low in cost, safe, efficient and high in precision is achieved.

The invention provides a shale reservoir fracture prediction method and system based on three-dimensional full-waveform inversion and a medium, and relates to the technical field of data analysis, and the method comprises the steps: arranging a multi-channel seismic exploration instrument in a shale reservoir region, and collecting full-waveform seismic data, including longitudinal waves, transverse waves and reflected waves; performing three-dimensional full-waveform inversion based on the logging data, the geological structure map and the full-waveform seismic data to generate a shale reservoir attribute model; a multi-scale inversion strategy is constructed, fracture sensitivity analysis and fracture form dynamic prediction are carried out, and reservoir fracture distribution characteristics and reservoir fracture form characteristics are obtained; discrete fracture point connection and fracture network reconstruction are carried out, a reservoir three-dimensional fracture network model is constructed, and fracture prediction visualization display is carried out. The technical problems that in the prior art, only inference based on limited data can be provided, comprehensive capture of a complex fracture network is lacked, distribution and form of fractures cannot be comprehensively and accurately described, and reservoir analysis and mining decisions are affected are solved.

The embodiment of the invention provides a full waveform inversion method based on surface wave dispersion inversion driving, a control device and a storage medium, and belongs to the technical field of seismic inversion. The full-waveform inversion method comprises the following steps: acquiring a Rayleigh surface wave observation signal through a preset detection point according to a three-station interference method; according to one-dimensional surface wave frequency dispersion inversion, through the obtained Rayleigh surface wave observation signal, obtaining a corresponding one-dimensional shear wave velocity representing a central area of the detection point pair; constructing a corresponding three-dimensional velocity model based on the one-dimensional shear wave velocity body; and performing full-waveform inversion on the three-dimensional velocity model to obtain a full-waveform inversion result of the region. According to the embodiment of the invention, aiming at passive source seismic signals, based on a three-station interference method, the high-signal-to-noise-ratio Rayleigh surface wave signals among the one-dimensional linear stations are extracted and enhanced, the surface wave signals are used as the input of the whole inversion process, the inversion result is not influenced by background noise quality, and the precision is higher.

The invention relates to the technical field of seismic inversion, and particularly discloses a variational reasoning full-waveform inversion method and device based on geological prior information, and the method comprises the steps: constructing a plurality of filters; acting the plurality of filters obtained by construction on a seismic migration section to calculate a correlation coefficient matrix; according to a logging information fitting theory variation function, performing low-rank approximation on the calculated correlation coefficient matrix to obtain a correlation coefficient matrix after low-rank approximation; randomly generating a plurality of speed samples based on the correlation coefficient matrix after low-rank approximation; calculating the gradient of the objective function according to the plurality of speed samples; according to the gradient of the objective function, calculating the gradient of posteriori distribution by using a Bayesian formula; and based on the gradient of the full waveform inversion posteriori distribution, utilizing a Stein variational gradient descent algorithm to carry out multiple times of iterative updating on the plurality of speed samples, and outputting a target speed model obtained by inversion of the plurality of speed samples. According to the scheme, the multiplicity of inversion is reduced, and the inversion efficiency is improved.

The invention relates to the technical field of ground penetrating radar data processing, in particular to a ground penetrating radar fusion inversion fine imaging method, medium and equipment, and the method comprises the steps: scanning a certain detection region through a ground penetrating radar, and obtaining an original GPR data profile; analyzing wavefront continuity of the GPR data profile, and introducing a correction factor to carry out lossless separation on the GPR data profile to obtain a reflection data profile and a diffraction data profile; constructing an inversion initial model based on the hyperbolic characteristics of the diffraction data profile; constructing an inversion objective function based on the reflection data profile and the diffraction data profile; and taking the inversion initial model as an inversion initial solution, and solving an optimization problem of an inversion objective function by adopting an L-BFGS algorithm to obtain a diffracted wave-reflected wave fusion imaging result. According to the method, the imaging effect and precision of the ground penetrating radar processing result on the tiny diffraction anomalous body can be effectively improved, the inversion imaging performance is optimized, and the practicability of ground penetrating radar fine imaging and full-waveform inversion imaging is promoted.

The invention belongs to the technical field of seismic exploration and artificial intelligence, and relates to a physical perception collaborative fusion-based full waveform inversion neural network construction method, which comprises the steps of required model data construction, full waveform inversion neural network construction, loss function construction, network pre-training, and network fine tuning and application. Physical feature perception and collaborative fusion are carried out through a U-Net network and an adaptive residual learning module, so that the problem that traditional full-waveform inversion depends on synthetic data training and is not matched with an actually measured data field is solved. The designed pre-training strategy not only can get rid of dependence on a large-scale synthetic data set, but also can effectively relieve the domain offset problem through fine adjustment of measured data, thereby improving the generalization ability of the model. And the adaptive residual learning module enables the neural network to adaptively and dynamically optimize feature expression, and improves inversion precision through collaborative fusion of a physical perception mechanism. According to the method, the physical consistency and imaging reliability of speed model prediction can be remarkably improved while the labor cost is reduced.

Systems and methods for predicting seismic velocities for a subsurface formation include obtaining seismic data from a subsurface formation. The seismic data includes an amount of information about the seismic formation. A reduced seismic dataset is generated by applying an unsupervised machine learning model to cluster the seismic data. The reduced seismic dataset has less data than the seismic data, and the reduced seismic dataset includes the amount of information about the subsurface formation. Inversion models and forward modeling data are generated by performing a full waveform inversion for the reduced seismic dataset. The inversion models specify velocities in the subsurface formation and the forward modeling data specify response values of the inversion models. A supervised machine learning model is trained using the inversion models and the forward modeling data; and seismic velocities are predicted by providing the seismic data as input to the trained supervised machine learning model.

The invention discloses a stratigraphic water content layering inversion method based on a shallow seismic full waveform, and relates to the technical field of geophysical exploration data processing, and the method comprises the following steps: 1, under the same survey line condition, synchronously obtaining a seismic wave full waveform, complex impedance and an induced polarization frequency spectrum; 2, constructing a layered initial model containing a layer boundary line and an in-layer water content initial value; based on the layered initial model, executing dual-domain quantile alignment interface keeping iteration until convergence, and obtaining a final layer boundary line and an in-layer water content field; the double-domain quantile alignment interface maintaining iteration comprises the following steps: updating an in-layer moisture content field and a layer boundary line according to a seismic wave full waveform; and performing total variation regularization processing on the updated in-layer water content field to maintain the transition characteristic of the layer boundary line. According to the method, an efficient and reliable solution with actual guidance value is provided for mineral resource exploration, geological structure detection, groundwater survey and engineering geology evaluation.

A method is described for generating high-resolution images of transversely isotropic mechanical properties and attenuation properties of tissues associated with measured compressional waves. A pressure wave is transmitted to a region containing fibers that are oriented with transverse isotropy. Waves are received that propagate after the pressure wave has impinged upon the fibers. An initial velocity and attenuation model is created that takes into account the transverse isotropy of the fibers. Full Waveform Inversion (FWI) is used to iteratively minimize a difference between the received data and the initial velocity and attenuation model to create an image of the region that takes into account the transverse isotropy of the fibers. The image has a higher resolution than if created using the initial velocity and attenuation model.

The invention discloses a full-waveform inversion method for enhancing boundary perception. The full-waveform inversion method comprises the following steps: acquiring seismic data; constructing a seismic waveform inversion network model; the seismic waveform inversion network model comprises an encoder, a decoder and a jump connection; carrying out the size adjustment of the seismic data through an encoder, and then extracting the advanced features of the reserved low-frequency information, and obtaining a multi-scale coding feature map; inputting the multi-scale encoding feature maps extracted by the encoder into different levels of the decoder by utilizing jump connection; a decoder is used for obtaining distinguishing information of the spatial relation contained in the multi-scale coding feature map, and a prediction speed model is generated; multi-scale analysis is executed through a wavelet convolution layer in an encoder, and low-frequency features are enhanced; then, distinguishing information is extracted by utilizing the spatial relationship of the features through a spatial attention module in a decoder so as to improve the resolution; finally, a reflection coefficient is introduced into the loss function to emphasize punishment on the low-contrast boundary; therefore, the full-waveform inversion precision is improved.

In view of the problems and the like of the prior art, the purpose of the present invention is to provide a method for imaging body tissue rapidly and robustly using waveform inversion from ultrasound examination data.　The method for imaging body tissue rapidly and robustly using waveform inversion from ultrasound examination data according to the present invention comprises: (a) a step for acquiring ultrasound examination data; (b) a step for generating one wave field by means of a plurality of computers (CPUs); (c) a step for integrating squared errors (which constitute fc in expression (5)) from all CPUs, each squared error being the error between a waveform recorded at a reception point from the wave field and an observed waveform obtained in step (a); (d) a step for re-dividing a correction gradient and a pseudo Hessian matrix into regions with an equal number of CPUs in the x-direction; (e) a step for re-dividing from the uniform regions in the x-direction into regions with an equal number of CPUs in the y-direction; (f) a step for calculating a step length required for model updating in full-waveform inversion using a quasi-Newton method while maintaining the division of regions in the x-direction; and (g) a step for repeating the steps (b) to (f) to obtain imaging of a viscosity Q value in addition to a sound wave propagation velocity. Expression 5: Here, fc (scalar quantity) is a cost function representing the sum of squared residuals between the observed data (d_obs) and the synthetic data (d_syn) calculated from the initial model at all times of all reception points, and Ns, Nr, Nt are respectively the number of transmission points, the number of reception points, and the number of time samples.

The invention relates to the technical field of seismic wave full-waveform inversion, in particular to a seismic wave full-waveform inversion method and system based on an RBF-FD fusion difference method, and the method comprises the steps: setting an artificial seismic source and a detection point, triggering a seismic source signal, and collecting actual measurement seismic waveform data; scanning a tunnel of a to-be-detected coal seam to generate point cloud data; constructing a tunnel three-dimensional geometric model according to the point cloud data, and dividing a regular surface area, an irregular surface area and a transition area; performing partition discrete solution to obtain a partition discrete solution result; presetting a medium parameter field of the to-be-measured coal seam; executing seismic wave forward modeling based on the seismic source signal, the medium parameter field of the current to-be-measured coal seam and the partition discrete solution result, generating a forward modeling wave field and extracting a forward modeling waveform; and performing residual comparison on the forward simulation waveform and the actually measured seismic waveform, calculating a residual norm, and updating the medium parameter field through an optimization algorithm until the precision requirement is met. According to the invention, high-fidelity seismic wave full-waveform inversion can be realized.

The invention discloses a robust full-waveform inversion method of a first-arrival constraint physical embedded recurrent neural network, which comprises the following steps of: solving a wave equation through forward propagation of the physical embedded recurrent neural network, generating a forward shot gather record, screening out a seismic trace with high matching degree with observation data by using first-arrival information, and participating in loss calculation to obtain a seismic trace with high matching degree with observation data; and along with the proceeding of the inversion process, the range of seismic traces participating in the inversion is gradually expanded, it is ensured that the inversion result gradually tends to a globally optimal solution, back propagation is carried out by using space-time field gradient information recorded by forward propagation of the recurrent neural network, the speed parameter is further optimized, and the inversion precision is improved. According to the method, the physical process of seismic wave propagation is embedded into the recurrent neural network, the first arrival constraint and the progressive inversion strategy are introduced, the dependence of the FWI on the initial model is remarkably reduced, and the underground velocity field structure can still be accurately reconstructed even if the initial model and a real model have remarkable deviation.

A method can include receiving seismic data of a subsurface region; performing an iterative full waveform inversion using at least a portion of the seismic data to generate a model of the subsurface region, where the performing includes, after one or more iterations of the full waveform inversion, automatically selecting one or more parts of the seismic data for inclusion in the at least a portion of the seismic data based on data classification using one or more quality attributes; and outputting the model of the subsurface region.

The invention relates to the field of exploration geophysics, and discloses a wave equation inversion method for constructing a middle-deep part background velocity field, which comprises the steps of S1, inputting observation gun record data and an initial velocity model, S2, performing forward modeling on the initial velocity field to calculate a forward propagation wave field, and performing back propagation on observation data to calculate a back propagation wave field; s3, calculating and simulating primary reflected wave data according to the migration imaging result and the initial velocity field back migration, solving a data residual error by using an optimal transmission distance, constructing an adjoint source, substituting the adjoint source into reflected wave inversion, calculating a gradient under the optimal transmission measure, and calculating a gradient under the optimal transmission measure; and S4, iteratively updating the background velocity field according to the calculated gradient, and performing full-waveform inversion by using the background velocity field information to obtain a final high-precision inversion result. Wave field residual errors are extracted more accurately based on an optimal transmission distance strategy, and background field gradient information more conforming to an underground structure is constructed.

The invention is suitable for the technical field of ground penetrating radar exploration, and provides a CUDA (Compute Unified Device Architecture) and PyTorch mixed compiled ground penetrating radar rapid full waveform inversion method, a PyTorch automatic differential function is self-defined by innovatively carrying out joint compilation on PyTorch and CUDA kernel functions, and a PyTorch platform is combined with CUDA mixed programming to realize automatic updating of two parameters of dielectric constant and conductivity. According to the method, GPU parallel calculation is used, the calculation efficiency of the ground penetrating radar data full-waveform inversion technology is remarkably improved, a two-parameter model can be quickly inverted, and the overall calculation efficiency is improved. The improvement of the efficiency is of great significance to the improvement of the overall effect of ground penetrating radar exploration and the practical application in the fields of urban underground space detection, road maintenance and the like.

The invention discloses a method for quantitatively evaluating the three-dimensional shape of a thinning defect in a plate under ultrasonic guided wave field undersampling scanning, belongs to the field of industrial nondestructive testing, and particularly relates to high-precision three-dimensional shape reconstruction and thickness loss quantification of a corrosion defect in a metal plate shell structure (such as aerospace skin, a pressure vessel wall, an oil pipeline and the like). According to the method, sparse sampling guided wave field reconstruction, full waveform inversion (FWI) and priori knowledge constraint are fused, Shannon sampling theorem limitation is broken through, efficient measurement of corrosion depth, contour and distribution under the conditions of a single excitation source and a limited view angle is achieved, and the method is suitable for rapid online detection in a complex industrial environment.

The invention relates to the technical field of coal seam inversion and imaging, in particular to a reflection channel wave imaging method and system considering the layering attenuation characteristic of a thick coal seam, and the method comprises the steps: constructing a coal seam layering geologic model, setting the initial wave velocity and initial quality factor of each layer, and forming an initial parameter field; setting an artificial seismic source and a plurality of detection points, exciting a seismic source signal at the artificial seismic source, recording seismic source information, and collecting a seismic wave response signal of each detection point; performing forward modeling by using a fractional order viscoelastic wave equation considering a spatial variable quality factor; carrying out full waveform inversion to obtain an inversion model; and on the basis of the inversion model, a level set function and a self-adaptive grid division strategy are adopted to express a barrier boundary in the coal seam to execute pre-stack wave field separation reverse time migration imaging, and a coal seam structure image is generated. The method can significantly improve the imaging resolution and calculation efficiency of the coal seam structure, and achieves the high-precision recognition of the boundary of the obstacle.

Seismic data is processed using a full-waveform inversion using a model with a pseudo-δ layer. The presence of the pseudo-δ layer in the model enables handing the difference at the water bottom between acoustically generated synthetic data and seismic data that corresponds to an elastic medium. The pseudo-δ layer may be less than 100 m thick and / or may be located directly underneath the water bottom. The pseudo-δ layer may have a negative value for S and a null value for ϵ (δ and ϵ being Thomsen's anisotropy parameters).

The invention relates to the technical field of data processing, in particular to a full-waveform radar data fitting method, device and equipment and a medium, and the method comprises the steps: controlling a transmitting wave transmitted when a full-waveform radar carries out distance detection, and obtaining each receiving wave detected by the full-waveform radar; sampling the transmitted wave and each received wave according to a set sampling period to obtain a first digital signal of the transmitted wave and a received wave digital signal of each received wave, and determining a second digital signal corresponding to the first digital signal from the received wave digital signal of each received wave; performing curve fitting on the first digital signal to obtain a first curve corresponding to the first digital signal; performing curve fitting on the second digital signal to obtain a second curve corresponding to the second digital signal; determining the detection distance of the full-waveform radar based on the first curve and the second curve; according to the invention, on the premise of ensuring the ranging accuracy, the full-waveform radar data processing speed is improved.

The invention discloses an intelligent exploration data analysis system and method, and relates to the technical field of geological exploration, and the system comprises a test module, an analysis module and a prediction module. According to the method, geological electromagnetic and stratum original data are collected in a multi-dimensional mode, forward modeling and inversion are carried out on electromagnetic data, the relation between underground conductivity and an electromagnetic field can be accurately simulated, data errors are effectively reduced, the electromagnetic data processing accuracy is improved, a geological unit calculates a seismic wave field based on wave equation numerical simulation, and the electromagnetic data processing accuracy is improved. The prediction wave field and the observation wave field are substituted into an objective function, the stratum velocity and density are accurately inverted, the geological data processing precision is guaranteed, the joint unit organically combines an electromagnetic inversion objective function, an earthquake full-waveform inversion objective function and a parameter correlation constraint term through a joint optimization formula, multi-source data depth fusion is achieved, and the geological data processing precision is guaranteed. The limitation of single data processing is avoided, and the underground geological condition is reflected more comprehensively.

The invention discloses a plate structure defect Lamb wave full waveform inversion method based on U-Net neural network regularization, and belongs to the field of nondestructive testing. Firstly, a two-dimensional sound wave equation is solved through a frequency domain finite difference method, and analog signals of all receiving points are obtained; the method comprises the following steps: constructing a target function to represent a relative deviation between an analog signal and a sensor actual measurement signal, and converting an inversion problem of defects in a metal plate into an optimization problem of the target function; performing derivation on the target function by using an adjoint state method to obtain a gradient, updating the velocity field along the negative direction of the gradient, and performing regularization processing on the velocity field through a specific coding-decoding structure of the U-Net neural network to suppress noise and artifacts; a frequency serial inversion strategy is adopted, and inversion is carried out step by step from low frequency to high frequency. According to the method, artifacts in the inversion process are suppressed by using the U-Net neural network, dependence of inversion on initial model precision is reduced by using a frequency domain serial inversion strategy, and high-precision inversion of plate structure defects is realized.

The invention provides a cross-correlation correction full-waveform inversion ultrasonic tomography method based on source coding. The method comprises the following steps of: 1, acquiring an actually measured sound pressure signal, setting an initial sound velocity model and the number of iterations of each stage, and initializing the number of iterations; step 2, calculating travel time difference between a prediction signal dm (xn, t) and a measured signal gm (xn, t) in the first iteration of each stage based on the initial sound velocity model in the step 1 and the number of iterations of each stage, generating an intermediate signal of the stage based on travel time, and not updating the travel time difference in the stage except the first iteration; 3, coding the intermediate signal and the prediction signal based on the intermediate signal and the prediction signal generated in the step 2; 4, calculating an adjoint wave field and gradient based on the intermediate signal and the prediction signal coded in the step 3, determining the step length of sound velocity model iteration, and completing model updating; and 5, obtaining a final model. According to the method, the stability of a traditional FWI-based ultrasonic tomography algorithm is improved while the calculated amount is not remarkably increased.

The invention relates to the technical field of elastic waves, and discloses an elastic wave full-waveform inversion method, which comprises the steps of inputting observed seismic data into a preset neural network encoder, and extracting to obtain high-dimensional feature representation; inputting the high-dimensional feature representation into a preset neural network decoder to obtain a gradient update value; obtaining an elastic parameter model based on the gradient update value and the initial elastic parameter model; obtaining Lame parameters based on the longitudinal wave velocity, the transverse wave velocity and the medium density; based on a Lame parameter and an elastic wave equation, obtaining synthetic seismic data; obtaining an objective function based on the observed seismic data and the synthesized seismic data; using the target function to iteratively update the preset neural network weight until the loss function reaches a first preset threshold value or the number of iterations reaches a second preset threshold value, and obtaining a target gradient update value; and based on the target gradient update value and the target elastic parameter model, obtaining a final target elastic parameter model so as to relieve discomfort caused by simultaneous inversion of multiple elastic parameters.