GPR Estimation of the Geometrical Features of Buried Metallic Targets in Testing Conditions

Soldovieri, Francesco; Catapano, Ilaria; Barone, Pier Matteo; Lauro, Sebastian; Mattei, Elisabetta; Pettinelli, Elena; Valerio, Guido; Comite, Davide; Galli, A.

doi:10.2528/pierb12120508

Cited by 18 publications

(13 citation statements)

References 53 publications

(69 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Direct knowledge of the velocity of the host medium also allows the usage of hyperbola fitting, with the radius of the rebar as the only unknown [37]. In [38] it is shown that using a Kirchoff-based linearized inversion can effectively recover some information regarding the shape of the targets regardless of the medium velocity. In the same context, a FWI approach is suggested in [39] to estimate the host medium properties and the size and depth of a steel rebar.…”

Section: Introductionmentioning

confidence: 99%

A Machine Learning-Based Fast-Forward Solver for Ground Penetrating Radar With Application to Full-Waveform Inversion

Giannakis

Giannopoulos

Warren

2019

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

The simulation, or forward modeling, of Ground Penetrating Radar (GPR) is becoming a more frequently used approach to facilitate interpretation of complex real GPR data, and as an essential component of full-waveform inversion (FWI). However, general full-wave 3D electromagnetic (EM) solvers, such as ones based on the Finite-Difference Time-Domain (FDTD) method, are still computationally demanding for simulating realistic GPR problems. We have developed a novel near realtime, forward modeling approach for GPR that is based on a machine learning (ML) architecture. The ML framework uses an innovative training method which combines a predictive principal component analysis technique, a detailed model of the GPR transducer, and a large dataset of modeled GPR responses from our FDTD simulation software. The ML-based forward solver is parameterized for a specific GPR application, but the framework can be applied to many different classes of GPR problems. To demonstrate the novelty and computational efficiency of our ML-based GPR forward solver, we used it to carry out FWI for a common infrastructure assessment application-determining the location and diameter of reinforcement bars in concrete. We tested our FWI with synthetic and real data, and found a good level of accuracy in determining the rebar location, size, and surrounding material properties from both datasets. The combination of the near real-time computation, which is orders of magnitude less than what is achievable by traditional full-wave 3D EM solvers, and the accuracy of our ML-based forward model is a significant step towards commercially-viable applications of FWI of GPR.

show abstract

Section: Introductionmentioning

confidence: 99%

A Machine Learning-Based Fast-Forward Solver for Ground Penetrating Radar With Application to Full-Waveform Inversion

Giannakis

Giannopoulos

Warren

2019

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Let us observe that the characteristics of the input signal, in terms of waveform and relevant frequency spectrum, strongly influence the imaging capabilities, especially in critical conditions [ 7,8].…”

Section: Synthetic Setupmentioning

confidence: 99%

“…In this frame, an EM computer aided-design (CAD) simulation of the scenario under test is exploited [ 7]- [ 9]. This numerical setup allows us to accurately estimate the incident and the scattered fields to be considered in the inverse problem solution.…”

Section: Introductionmentioning

confidence: 99%

An improved tomographic approach for accurate target reconstruction from GPR numerical data

Comite

Galli

Catapano

et al. 2014

Proceedings of the 15th International Conference on Ground Penetrating Radar

View full text Add to dashboard Cite

This study aims at investigating the advances in terms of the reconstruction capabilities offered by an ad-hoc imaging procedure capable of accounting for the actual radiation features of the Ground Penetrating Radar (GPR) antennas. The study is carried out by using a 'synthetic setup', which is implemented with a numerical tool able to simulate the characteristics of the antenna system as well as the signals used for the actual illumination. This enables us to obtain an accurate determination of both the incident and the backscattered electromagnetic (EM) fields in the investigated scenarios. GPR data are then processed by means of a consolidated frequency-domain microwave tomographic approach, which is reformulated in such a way to consider the actual EM field radiated by the adopted transmitting antenna. A representative example referred to synthetic data is provided to manage and discuss the improvements in the imaging process achievable with the proposed approach

show abstract

“…GPR data consist of target and clutter(direct wave, underground clutter and other noise signals) [3][4][5]. It is difficult to extract target signal especially in the shallow subsurface.…”

Section: Introductionmentioning

confidence: 99%

A Clutter Suppression Method Based on Improved Principal Component Selection Rule for Ground Penetrating Radar

Zhu¹,

Xue²,

Xia³

et al. 2017

PIER M

View full text Add to dashboard Cite

Abstract-Principal component analysis is usually used for clutter suppression of ground penetrating radar, but its performance is influenced by the selection of main components of target signal. In the paper, an improved principal component selection rule is proposed for selecting the main components of target signal. In the method, firstly difference spectrum of singular value is used to extract direct wave and strong target signal, and then, Fuzzy-C means clustering algorithm is used to determine the weights of principal component of weak target signal. Finally, the principal components of strong target signal and weak target signal are reconstructed to obtain target signal. Experimental results show that the proposed method can effectively remove the clutter signals and reserve more target information.

show abstract

GPR Estimation of the Geometrical Features of Buried Metallic Targets in Testing Conditions

Cited by 18 publications

References 53 publications

A Machine Learning-Based Fast-Forward Solver for Ground Penetrating Radar With Application to Full-Waveform Inversion

A Machine Learning-Based Fast-Forward Solver for Ground Penetrating Radar With Application to Full-Waveform Inversion

An improved tomographic approach for accurate target reconstruction from GPR numerical data

A Clutter Suppression Method Based on Improved Principal Component Selection Rule for Ground Penetrating Radar

Contact Info

Product

Resources

About