Precise infrastructure mapping using full-waveform inversion of ground penetrating radar signals

• US 10,234,552 B1
• Filed: 06/27/2018
• Issued: 03/19/2019
• Est. Priority Date: 06/27/2018
• Status: Active Grant

• Current Assignee: University of South Florida (State University System of Florida)
• Original Assignee: University of South Florida (State University System of Florida)
• Inventors: Jazayeri, Sajad, Kruse, Sarah
• Primary Examiner(s): Brainard, Timothy A
• Application Number:
  US16/020,665
• Time in Patent Office: 265 Days
• Field of Search: 342 22
• CPC Class Codes: G01S 13/885 for ground probing prospecting or detecting using electromagnetic waves G01V3/12
• CPC Class Codes: G01S 13/89 for mapping or imaging
• CPC Class Codes: G01S 7/295 Means for transforming co-ordinates or for evaluating data, e.g. using computers
• CPC Class Codes: G01S 7/414 Discriminating targets with respect to background clutter

Abstract

A new method for FWI of common-offset GPR data, particularly targeting the dimensions and infilling material of buried pipes. The method is useful in situation where clear isolated diffraction hyperbolas indicate the presence of a pipe, but pipe dimensions and filling may be unknown. The invention is a method of taking GPR data and applying advanced numerical methods to get the depth and size of an underground pipe in a very accurate manner. An embodiment of the method consists of five main steps: GPR data processing, ray-based analysis to set a good initial model, 3D to 2D transformation of data, effective SW estimation, and FWI.

Claims

1. A method for determining physical properties of a subsurface object disposed in a surrounding media, comprising:

acquiring measured ground penetrating data from a ground penetrating radar device, wherein the measured ground penetrating data includes at least one measured diffraction hyperbola;

estimating a plurality of parameters of an initial model of the subsurface object and the surrounding media, such that the initial model produces a simulated diffraction hyperbola that is offset less than one-half a wavelength from the measured diffraction hyperbola, wherein the plurality of parameters includes at least one of the following;

a velocity of electromagnetic waves through media in which the subsurface object is located;

a depth of the subsurface object in the surrounding media;

a horizontal location of the subsurface object;

a size of the subsurface object;

a conductivity of the surrounding media;

a permittivity of the surrounding media;

a permittivity of any material or fluid within an outer surface of the subsurface object;

acquiring a measured source wavelet and deconvolving the initial model from the measured ground penetrating data to produce a corrected source wavelet;

simulating ground penetrating data for the initial model with the corrected source wavelet; and

determining a cost function and adjusting at least one of the plurality of parameters to minimize the cost function.

2. The method of claim 1, wherein the step of adjusting at least one of the plurality of parameters to minimize the cost function includes adjusting every parameter in the plurality of parameters with the exception of the conductivity of the surrounding media.

3. The method of claim 1, further including a step of filtering out high frequency noise from the raw ground penetrating data.

4. The method of claim 1, further including a step of applying a dewow filter to the raw ground penetrating data.

5. The method of claim 1, further including a step of applying time-zero corrections to the raw ground penetrating data.

6. The method of claim 1, further including a step of transforming the initial model from three-dimensional data to two-dimensional data.

7. The method of claim 1, wherein the step of estimating the plurality of parameters of the initial model includes a least squares method for estimating the initial depth of the subsurface object in the media and the initial velocity of the electromagnetic waves through media in which the subsurface object is located.

8. The method of claim 1, wherein the step of creating an initial model of the subsurface object further includes, responsive to receiving two distinct hyperbolas from a top and a bottom of the subsurface object, estimating a diameter of the subsurface object from a root mean square average of a travel time difference between peak amplitudes of the two distinct hyperbolas.

9. The method of claim 1, wherein the step of producing a corrected source wavelet includes more than one iteration of deconvolving the initial model from the measured ground penetrating data.

10. The method of claim 1, wherein the step of adjusting the plurality of parameters to minimize the cost function is performed until the cost function can no longer be minimized.

11. The method of claim 1, further including a step of establishing boundary values for one or more of the plurality of parameters between which the parameters are adjusted in the step of adjusting the plurality of parameters to minimize the cost function.

12. A method for determining the cross-sectional size of an underground pipe, comprising:

operating a ground penetrating radar device with at least one antenna in a broadside orientation with respect to the underground pipe, wherein the ground penetrating radar device discharges and receives electromagnetic waves;

acquiring measured ground penetrating data from the ground penetrating radar device, wherein the measured ground penetrating data includes at least one measured diffraction hyperbola;

estimating a plurality of parameters of an initial model of the underground pipe and the surrounding media, such that the initial model produces a simulated diffraction hyperbola that is offset less than one-half a wavelength from the measured diffraction hyperbola, wherein the plurality of parameters includes;

a velocity of electromagnetic waves through media in which the underground pipe is located;

a depth of the underground pipe in the surrounding media;

a horizontal location of the underground pipe;

a size of the underground pipe;

a conductivity of the surrounding media;

a permittivity of the surrounding media;

a permittivity of any material or fluid within an outer surface of the underground pipe;

producing a corrected source wavelet by acquiring a measured source wavelet and deconvolving the initial model from the measured ground penetrating data;

simulating ground penetrating data for the initial model with the corrected source wavelet; and

determining a cost function by running a forward model and optimizing each of the plurality of parameters using full waveform inversion to minimize the cost function, exception for the conductivity of the surrounding media.

13. The method of claim 12, further including a step of filtering out high frequency noise from the raw ground penetrating data.

14. The method of claim 12, further including a step of transforming the initial model from three-dimensional data to two-dimensional data.

15. The method of claim 12, wherein the step of creating an initial model of the subsurface object includes a least squares method for estimating the initial depth of the subsurface object in the media and the initial velocity of electromagnetic waves through media in which the subsurface object is located.

16. The method of claim 12, wherein the step of creating an initial model of the subsurface object further includes, responsive to receiving two distinct hyperbolas from a top and a bottom of the subsurface object, estimating a diameter of the subsurface object from a root mean square average of a travel time difference between peak amplitudes of the two distinct hyperbolas.

17. The method of claim 12, wherein the step of producing a corrected source wavelet includes more than one iteration of deconvolving the initial model from the measured ground penetrating data.

18. The method of claim 12, wherein the step of adjusting the plurality of parameters to minimize the cost function is performed a predetermined number of times.

19. The method of claim 12, further including a step of establishing boundary values for one or more of the plurality of parameters between which the parameters are adjusted in the step of adjusting the plurality of parameters to minimize the cost function.

20. A method for determining the cross-sectional size of an underground pipe, comprising:

operating a ground penetrating radar device with at least one antenna in a broadside orientation with respect to the underground pipe, wherein the ground penetrating radar device discharges and receives electromagnetic waves;

acquiring measured ground penetrating data from the ground penetrating radar device, wherein the measured ground penetrating data includes at least one measured diffraction hyperbola;

estimating a plurality of parameters of an initial model of the underground pipe and the surrounding media, such that the initial model produces a simulated diffraction hyperbola that is offset less than one-half a wavelength from the measured diffraction hyperbola, wherein the plurality of parameters includes;

a velocity of electromagnetic waves through media in which the underground pipe is located;

a depth of the underground pipe in the surrounding media;

a horizontal location of the underground pipe;

a size of the underground pipe;

a conductivity of the surrounding media;

a permittivity of the surrounding media;

a permittivity of any material or fluid within an outer surface of the underground pipe;

transforming the initial model from three-dimensional data to two-dimensional data;

producing a corrected source wavelet by acquiring a measured source wavelet and deconvolving the initial model from the measured ground penetrating data;

simulating ground penetrating data for the initial model with the corrected source wavelet; and

determining a cost function and optimizing each of the plurality of parameters to minimize the cost function, exception for the conductivity of the surrounding media.