METHOD AND APPARATUS FOR RECONSTRUCTING DIELECTRIC IMAGE USING ELECTROMAGNETIC WAVES

• US 20150042508A1
• Filed: 04/16/2014
• Published: 02/12/2015
• Est. Priority Date: 08/09/2013
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: KIM, Bo Ra, SIMONOV, Nikolai, SON, Seong-Ho, JEON, Soon Ik, KIM, Hyuk Je, LEE, Jong Moon, CHOI, Hyung Do
• Granted Patent: US 9,519,054 B2
• Time in Patent Office: Days
• US Class Current: 342/195
• CPC Class Codes: G01S 13/88 Radar or analogous systems specially adapted for specific applications electromagnetic prospecting or detecting of objects, e.g. near-field detection, G01V3/00
• CPC Class Codes: G01S 13/887 for detection of concealed objects, e.g. contraband or weapons
• CPC Class Codes: G01S 13/888 through wall detection
• CPC Class Codes: G01S 13/89 for mapping or imaging

Abstract

A method for reconstructing a dielectric image using electromagnetic waves, comprising: acquiring a measurement value of the electromagnetic waves; generating a matching system matrixes between the meshes; generating a smoothed dyadic Green'"'"'s function matrix; generating an electromagnetic wave calculation value; calculating a misfit error between the acquired electromagnetic wave measurement value and the generated electromagnetic wave calculation value, and checking whether a change of the calculated misfit error satisfies a predetermined optimization determination condition; updating dielectric parameters at the meshes; and outputting a reconstructed dielectric image in the image reconstruction region.

Claims

1. A method for reconstructing an image using electromagnetic waves, comprising:

acquiring a measurement value of the electromagnetic waves that are radiated by a transmitting antenna and propagate through an image reconstruction region;

setting a plurality of predetermined meshes in the image reconstruction region;

generating a matching system matrixes using basis functions related to the plurality of predetermined meshes;

generating a smoothed Green'"'"'s function matrix using the generated matching system matrix;

generating an electromagnetic wave calculation value through performing of a forward electromagnetic analysis with respect to the image reconstruction region on the basis of the generated smoothed dyadic Green'"'"'s function matrix;

calculating a misfit between the acquired electromagnetic wave measurement value and the generated electromagnetic wave calculation value, and checking whether a the calculated misfit error satisfies a predetermined optimization determination condition;

updating the dielectric parameters at the meshes if the predetermined optimization determination condition is not satisfied; and

outputting a reconstructed image in the image reconstruction region when the predetermined optimization determination condition is satisfied.

2. The method of claim 1, wherein the plurality of meshes include a fine mesh, a coarse mesh having a relatively larger spacing than the fine mesh, and an image reconstruction mesh having a relatively larger spacing than the coarse mesh.

3. The method of claim 2, wherein the plurality of meshes are in any one of rectangular grids, or hexagonal grids.

4. The method of claim 2, wherein the basis function corresponding to the fine mesh and the coarse mesh and the basis function corresponding to the coarse mesh and the image reconstruction mesh are in a sinc-type function, spline function type or wavelet type.

5. The method of claim 2, wherein the matrix B_c of matching system of the basis function, allocated at the fine mesh and centered at the coarse mesh nodes is generated as in an equation,

B_c=[b_c,1,b_c,2,b_c,3, . . . ,b_c,n, . . . ,b_c,Nc]

where, b_c,n denotes the vectors of basis function allocated to the fine mesh and centered at the n-th node of coarse mesh, and N_c denotes the total number of coarse mesh nodes.

6. The method of claim 5, wherein the matching system matrix satisfies an equation,

R_c≡

B_c·

{tilde over (B)}_c^T=[r_c,1,r_c,2,r_c,3, . . . ,r_c,n, . . . ,r_c,Nf]

where, R_c denotes a spatial resolution matrix, {tilde over (B)}_c^T;

denotes a transpose matrix, which is dual to B_c, r_c,n denotes a vector of a point spread function corresponding to the n-th fine mesh node, and N_f denotes the total number of fine mesh nodes.

7. The method of claim 6, wherein R_c has property of a spatial filter for smoothing an image reconstruction value.

8. The method of claim 6, wherein the matching system matrix simultaneously satisfies two conditions:

Condition 1;

The effective width of the basis function is equal to or close to the grid spacing of the coarse mesh.

Condition 2;

All rows and columns of the spatial resolution matrix correspond to functions having the same shape and the same peak values but centered at different nodes of the fine mesh.

9. The method of claim 5, wherein the matching system matrix B_r of the basis functions allocated to the fine mesh and centered at the image reconstruction mesh nodes is generated as in an equation,

B_r=[b_r,1,b_r,2,b_r,3, . . . ,b_r,n, . . . ,b_r,Nr]

where, b_r,n denotes the basis function allocated to the fine mesh and centered at the n-th node of image reconstruction mesh, and N_r denotes the total number of image reconstruction mesh nodes.

10. The method of claim 9, wherein the generating the matching system matrix comprises generating a spatial resolution matrix using multiplication of the matching system matrix and a transpose matrix, which is dual to the matching system matrix.

11. The method of claim 2, wherein the generating the smoothed Green'"'"'s function matrix comprises generating the smoothed Green'"'"'s function matrix by multiplying a Green'"'"'s function matrix defined at the fine mesh by the matching system matrix and a transpose matrix, which is dual to the matching system matrix.

12. The method of claim 11, wherein the generating the smoothed Green'"'"'s function matrixes comprises:

calculating the dyadic Green'"'"'s function matrix defined at all the fine mesh nodes in the image reconstruction; and

calculating the dyadic Green'"'"'s function matrix for calculating scattered waves incident on receiving antennas.

13. The method of claim 2, wherein the performing the forward electromagnetic analysis comprises analyzing electromagnetic wave scattering with respect to the coarse mesh and calculating the electromagnetic waves that can be acquired in positions of the receiving antennas.

14. The method of claim 1, wherein the checking comprises determining that the predetermined optimization condition is satisfied if a difference value between a current misfit error and a previously calculated misfit error is smaller than a predetermined allowable difference value or exceeds a predetermined number of calculations.

15. The method of claim 2, wherein the updating the dielectric parameters comprises:

generating a regularized Jacobian matrix using the generated matching system matrix with respect to the image reconstruction mesh, and the smoothed dyadic Green'"'"'s function matrix; and

updating the dielectric parameters at the coarse mesh using the generated regularized Jacobian matrix.

16. The method of claim 15, wherein the regularized Jacobian matrix J_r is generated as in an equation,

17. The method of claim 15, wherein the updating the dielectric parameters at the coarse mesh is expressed as in an equation,

∈

_k=∈

_k-1+α

_N−

∈

_b·

(J_r^T·

J_r)<sup>−

1</sup>J_r^T·

Δ

γ

where, α

_N·

∈

_b·

(J_r^T·

J_r)⁻

J_r^T. Δ

γ

denotes dielectric parameters update values, and in the update value term, α

_N denotes an increment coefficient, ∈

_b denotes a permittivity of a background material, J_r denotes a regularized Jacobian matrix, a superscript T denotes a transpose matrix, a superscript ‘

−

1’

denotes an inverse matrix, and Δ

γ

denotes an misfit between the electromagnetic wave measurement value and the electromagnetic wave calculation value.

18. The method of claim 1, wherein the outputting the reconstructed image comprises generating and outputting the reconstructed image as a visualization file that is visually identifiable or a data file.

19. An apparatus for reconstructing an image using electromagnetic waves, comprising:

a measurement value acquisition block acquiring a measurement value of the electromagnetic waves that are radiated by a transmitting antenna and propagate through an image reconstruction region;

a mesh setting block setting a plurality of predetermined meshes in the image reconstruction region;

a matching system matrix generation block generating a matching system matrix using basis functions related to the plurality of predetermined meshes;

a Green'"'"'s function matrix generation block generating a smoothed Green'"'"'s function matrix using the generated matching system matrix;

an electromagnetic wave calculation block generating an electromagnetic wave calculation value through performing of a forward electromagnetic analysis with respect to the image reconstruction region on the basis of the generated smoothed Green'"'"'s function matrix;

an optimization determination block calculating an misfit error between the acquired electromagnetic wave measurement value and the generated electromagnetic wave calculation value, and checking whether the calculated misfit error satisfies a predetermined optimization determination condition;

a parameter updating block generating a regularized Jacobian matrix using the matching system matrix, the image reconstruction mesh, and the smoothed Green'"'"'s function matrix, and updating dielectric parameters at the coarse mesh using the generated regularized Jacobian matrix if the predetermined optimization determination condition is not satisfied; and

an image reconstruction value generation block generating a reconstructed image in the image reconstruction region as a visualization file that is visually identifiable or a data file when the predetermined optimization determination condition is satisfied.

20. The apparatus of claim 19, wherein the mesh setting block sets a fine mesh, a coarse mesh having a relatively larger spacing than the fine mesh, and an image reconstruction mesh having a relatively larger spacing than the coarse mesh as the plurality of meshes.