CN-116879963-B - Forward modeling physical equivalent method, device and equipment for ground penetrating radar and storage medium

CN116879963BCN 116879963 BCN116879963 BCN 116879963BCN-116879963-B

Abstract

The application discloses a physical equivalent method, device, equipment and storage medium for forward performance of a ground penetrating radar. The method comprises the steps of simulating a boundary domain where a PML layer is located, degrading a computational formula of the PML layer into a computational formula of a computational domain, vectorizing the computational formulas of the PML layer and the computational domain to obtain a vector computational formula, realizing degradation from the PML layer to the computational domain by controlling variable parameter values in the vector computational formula, reconstructing the vector computational formula into an RNN model, inputting a point source, an initial field and electrical parameters into the RNN model to obtain a field component with a specified time sequence, expanding the dimensions of the input parameters and the field component of the RNN model, inputting the expanded three-dimensional field component and parameter tensor into the RNN model, constructing a plurality of ground penetrating radar physical models, and simulating the ground penetrating radar to obtain a simulated B-Scan image. The application greatly improves the simulation efficiency, ensures the precision of the simulation result and improves the utilization rate of the computing equipment.

Inventors

CHENG XI
ZHAO YUNJIE
LI QIANG
ZHANG KUN
REN HONGYU

Assignees

新疆农业大学

Dates

Publication Date: 20260508
Application Date: 20230717

Claims (10)

1. A method of physical equivalents of forward performance of a ground penetrating radar, the method comprising: performing simulation calculation on a boundary domain where a PML layer is located, and degrading the calculation formula of the PML layer into the calculation formula of the calculation domain by cutting off the calculation domain in the boundary domain; Vectorizing the calculation formula of the calculation domain to obtain a vector calculation formula, and realizing the degradation from the PML layer to the calculation domain by controlling the variable parameter value in the vector calculation formula; Reconstructing the vector calculation formula into an RNN model, and inputting a point source, an initial field and electrical parameters into the RNN model to obtain a field component with a specified time sequence, wherein the field component comprises an electric field component and a magnetic field component; expanding the dimensions of the input parameters and the field components of the RNN model to obtain expanded three-dimensional field components and parameter tensors; And inputting the three-dimensional field component and the parameter tensor into the RNN model to construct a plurality of different ground penetrating radar physical models, and performing simulation calculation on the ground penetrating radar through each ground penetrating radar physical model to obtain a simulation B-Scan image.
2. The forward physical equivalent method of ground penetrating radar according to claim 1, wherein performing simulation calculation on a boundary domain where a PML layer is located, and by truncating a calculation domain in the boundary domain, degrading a calculation formula of the PML layer to a calculation formula of the calculation domain, comprises: performing simulation calculation on the boundary domain by using an FDTD method, and constructing a stepping formula of the PML layer by adopting the PML layer as a boundary condition in the boundary domain; and the calculation formula of the PML layer is degenerated into the calculation formula in the calculation domain by adjusting the correlation coefficient in the stepping formula.
3. The forward physical equivalent method of ground penetrating radar according to claim 1, wherein said vector computation comprises a first sub-vector matrix and a second sub-vector matrix, and vectorizing said computation of said computation domain to obtain a vector computation comprises: Vectorizing the calculation formula of the calculation domain to obtain two first sub-vector matrixes; And carrying out one-layer zero filling on the field component of the calculation domain according to the differential direction, and carrying out differential calculation on the two first sub-vector matrixes to obtain two second sub-vector matrixes.
4. The forward physical equivalent method of ground penetrating radar according to claim 1, wherein said variable parameter values include conductivity and magnetic loss, and wherein said realizing degradation of said PML layer to said computational domain by controlling variable parameter values in said vector computational formula comprises: Setting the position of the conductivity in the calculation domain and the position of the magnetic loss in the calculation domain to be the same position in the calculation domain; and setting the conductivity and the magnetic loss to be equal to the conductivity value in the physical model of the ground penetrating radar in the calculation domain so as to realize the unification of formulas of the PML layer and the calculation domain.
5. The forward physical equivalent method of ground penetrating radar according to claim 3, wherein said vector calculation formula further comprises four sub-vector calculation formulas, said vector calculation formulas are reconstructed into an RNN model, point sources, initial fields and electrical parameters are input into said RNN model, field components with specified time sequences are obtained, said field components comprise electric field components and magnetic field components, and said method comprises: Reconstructing the vector calculation formula into the RNN model, wherein the RNN model comprises a plurality of RNN layers, and inputting an electric field component and a magnetic field component of the previous time sequence into each RNN layer; and taking the sub-vector calculation formula as a parameter matrix of the RNN model, and inputting a point source, an initial field and electrical parameters into each RNN layer to obtain an electric field component and a magnetic field component of the appointed time sequence.
6. The forward physical equivalent method of ground penetrating radar according to claim 5, wherein expanding the dimensions of input parameters and field components of said RNN model to obtain expanded three-dimensional field components and parameter tensors comprises: expanding the dimensions of each sub-vector calculation formula, the electric field component and the magnetic field component to obtain a three-dimensional sub-vector calculation formula and a three-dimensional field component; and expanding the point source into a point source matrix, wherein the three-dimensional sub-vector calculation, the three-dimensional field component, the parameter tensor and the dimension of the point source matrix comprise the height and the width of the physical model of the ground penetrating radar and the quantity of the physical model of the ground penetrating radar to be constructed.
7. The method of claim 6, wherein constructing a plurality of different physical models of ground penetrating radar by inputting the three-dimensional field component and the parameter tensor into the RNN model, comprises: and inputting the three-dimensional field component, the parameter tensor and the point source matrix into corresponding RNN layers, and stacking each RNN layer for preset times to construct a plurality of physical models of the ground penetrating radar.
8. A ground penetrating radar forward physical equivalent apparatus, said apparatus comprising: The degradation module is used for carrying out simulation calculation on a boundary domain where the PML layer is located, and degrading the calculation formula of the PML layer into the calculation formula of the calculation domain by cutting off the calculation domain in the boundary domain; The simulation module is used for vectorizing the calculation formula of the calculation domain to obtain a vector calculation formula, and realizing the degradation from the PML layer to the calculation domain by controlling the variable parameter value in the vector calculation formula; the reconstruction module is used for reconstructing the vector calculation type into an RNN model, inputting a point source, an initial field and electrical parameters into the RNN model to obtain a field component with a specified time sequence, wherein the field component comprises an electric field component and a magnetic field component; The expansion module is used for expanding the dimensions of the input parameters and the field components of the RNN model to obtain expanded three-dimensional field components and parameter tensors; The construction module is used for inputting the three-dimensional field component and the parameter tensor into the RNN model to construct a plurality of different ground penetrating radar physical models, and carrying out simulation calculation on the ground penetrating radar through each ground penetrating radar physical model to obtain a simulation B-Scan image.
9. A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of the ground penetrating radar forward-looking physical equivalent method of any one of claims 1-7 when the computer program is executed.
10. A computer readable storage medium, characterized in that it stores a computer program which, when executed by a processor, implements the steps of the forward physical equivalent method of ground penetrating radar according to any one of claims 1 to 7.

Description

Forward modeling physical equivalent method, device and equipment for ground penetrating radar and storage medium Technical Field The invention relates to the technical field of electromagnetics, in particular to a forward physical equivalent method, device and equipment of a ground penetrating radar and a storage medium. Background The ground penetrating radar is a non-invasive detection technology, which utilizes an antenna to emit electromagnetic pulse to a target in a medium below to detect and locate abnormal objects and structural characteristics in the ground or an object, and the technology is widely applied to the fields of civil engineering, geological disasters, military, archaeology and the like at present. Electromagnetic wave simulation is the most important topic in the technical research of ground penetrating radar, and is also the basis of applications such as abnormal body detection, target body inversion and the like. In electromagnetic wave forward modeling, common methods include an FDTD (FINITE DIFFERENCE TIME Domain, finite difference method in time Domain), a parallel method, a deep learning-based method and the like, but the methods can cause huge time consumption of the simulation, low calculation efficiency, complex custom operation, incapability of realizing parallel simulation of a plurality of physical models, incapability of considering generalization, high efficiency and accuracy, and difficulty in application to practical problems. Disclosure of Invention In view of the above, the present invention aims to overcome the defects in the prior art, and provide a method, a device and a storage medium for forward physical equivalence of a ground penetrating radar. The invention provides the following technical scheme: in a first aspect, in an embodiment of the present disclosure, there is provided a forward physical equivalent method of a ground penetrating radar, the method including: performing simulation calculation on a boundary domain where a PML layer is located, and degrading the calculation formula of the PML layer into the calculation formula of the calculation domain by cutting off the calculation domain in the boundary domain; Vectorizing a calculation formula of the PML layer and the calculation domain to obtain a vector calculation formula, and realizing degradation from the PML layer to the calculation domain by controlling variable parameter values in the vector calculation formula; Reconstructing the vector calculation formula into an RNN model, and inputting a point source, an initial field and electrical parameters into the RNN model to obtain a field component with a specified time sequence, wherein the field component comprises an electric field component and a magnetic field component; expanding the dimensions of the input parameters and the field components of the RNN model to obtain expanded three-dimensional field components and parameter tensors; And inputting the three-dimensional field component and the parameter tensor into the RNN model to construct a plurality of different ground penetrating radar physical models, and performing simulation calculation on the ground penetrating radar through each ground penetrating radar physical model to obtain a simulation B-Scan image. Further, performing simulation calculation on a boundary domain where a PML layer is located, and by cutting off a calculation domain in the boundary domain, degrading a calculation formula of the PML layer into a calculation formula of the calculation domain, including: performing simulation calculation on the boundary domain by using an FDTD method, and constructing a stepping formula of the PML layer by adopting the PML layer as a boundary condition in the boundary domain; and the calculation formula of the PML layer is degenerated into the calculation formula in the calculation domain by adjusting the correlation coefficient in the stepping formula. Further, the vector calculation formula includes a first sub-vector matrix and a second sub-vector matrix, and the vectorizing the calculation formula of the PML layer and the calculation domain to obtain a vector calculation formula includes: vectorizing the calculation formulas of the PML layer and the calculation domain to obtain two first sub-vector matrixes; And carrying out one-layer zero filling on the field component of the calculation domain according to the differential direction, and carrying out differential calculation on the two first sub-vector matrixes to obtain two second sub-vector matrixes. Further, the variable parameter values include conductivity and magnetic loss, and the degradation of the PML layer to the computational domain is achieved by controlling the variable parameter values in the vector computational formula, comprising: Setting the position of the conductivity in the calculation domain and the position of the magnetic loss in the calculation domain to be the same position in the calculation domain; an