Deep-Learning-Based Calibration in Contrast Source Inversion Based Microwave Subsurface Imaging

Hanabusa, Takahiro; Morooka, T.; Kidera, Shouhei

doi:10.1109/lgrs.2022.3169799

Cited by 7 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is important to note that applying this method to a real scenario necessitates an effective calibration process, converting the experimental data into a simulation data. Various approaches have been developed for this calibration, including linear transfer function-based methods [37], [38] and deep learning-based processing [39], [40]. However, it is challenging to fully convert experimental data into ideal simulation data as it requires accurate parameters such as dielectric property, dimensions or structures of antennas, or other components such as cable or insertion loss of a vector network analyzer or radar module.…”

Section: E Further Discussionmentioning

confidence: 99%

Complex Permittivity Imaging by Incorporating Synthetic Aperture Radar and Inverse Scattering Method for Stratified Ground Medium

Yamauchi,

Kidera

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

This paper introduces the incorporation approach with synthetic aperture radar (SAR) and contrast source inversion (CSI) based non-linear inverse scattering (NIS) approach for quantitative permittivity imaging for buried object under multi-layered heterogeneous ground media. It is challenging issue to retrieve a complex permittivity from ground penetrating radar (GPR) data, since the NIS problem considerably suffers from inaccuracy due to severe ill-posed condition. To overcome this limitation, this paper introduces the SAR image based region of interest (ROI) limitation in the CSI optimization scheme, where the number of unknowns are massively reduced. Furthermore, the SAR image is also upgraded by the pre-CSI optimization, where the Green's function and background clutter for heterogeneous background (e.g., multi-layered medium) are accurately generated. The FDTD based numerical tests, assuming GPR observation model, show that our proposed scheme effectively reconstructs a dielectric property of buried object, even in severe condition.

show abstract

Section: E Further Discussionmentioning

confidence: 99%

Complex Permittivity Imaging by Incorporating Synthetic Aperture Radar and Inverse Scattering Method for Stratified Ground Medium

Yamauchi,

Kidera

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…This section describes the calibration procedure that transforms the experimental data into the corresponding simulation data. This was achieved using a linear transfer function model, as discussed in references [36], [43], [46], [47]. Initially, we measured the reflection responses assuming that the ROI contained only air (serving as the calibration object).…”

Section: B Calibration Proceduresmentioning

confidence: 99%

Complex Permittivity Retrieval Approach With Radar Enhanced Contrast Source Inversion for Microwave Nondestructive Road Evaluation

Suzuki,

Nakamura,

Kidera

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

An experimental investigation of microwave quantitative imaging of nondestructive testing applications for concrete road structures is presented. A radar prior-based contrast source inversion (CSI) scheme is introduced to achieve an accurate reconstruction for target shape and dielectric property by applying the promising radar approach-range points migration (RPM). Moreover, to address the local optimum problem, we introduce an appropriate initial estimate for complex permittivity that focuses on the cost function of the CSI. A numerical and real data from an actual concrete road model, where a thin-layer water-filled cavity is buried into the boundary area between the asphalt and the concrete floorboard, show that our proposed approach improves reconstruction accuracy for permittivity and conductivity in a real nondestructive testing model.

show abstract

“…For example, some frequency-domain calibration schemes have been proposed in [27][28][29]. Recently, neural networks are also emerging to perform this task [30][31][32]. In particular, [31] introduced time-domain preprocessing of quasi-monostatic data acquired in free space configurations using long short-term memory (LSTM) cells [33].…”

Section: Introductionmentioning

confidence: 99%

Quantitative Inversion of Multiantenna Ground-Penetrating Radar Data with Modeling Error Correction Based on Long Short-Term Memory Cells

Fedeli,

Schenone,

Randazzo

2024

Remote Sensing

View full text Add to dashboard Cite

Quantitative inversion of GPR data opens the door to precise characterization of underground environments. However, in order to make the inverse scattering problem solution easier from a computational viewpoint, simplifying assumptions are often applied, i.e., two-dimensional approximations or the consideration of idealized field probes and electromagnetic sources. These assumptions usually produce modeling errors, which can degrade the dielectric reconstruction results considerably. In this article, a processing step based on long short-term memory cells is proposed for the first time to correct the modeling error in a multiantenna GPR setting. In particular, time-domain GPR data are fed into a neural network trained with couples of finite-difference time-domain simulations, where a set of sample targets are simulated in both realistic and idealized configurations. Once trained, the neural network outputs an approximation of multiantenna GPR data as they are collected by an ideal two-dimensional measurement setup. The inversion of the processed data is then accomplished by means of a regularizing Newton-based nonlinear scheme with variable exponent Lebesgue space formulation. A numerical study has been conducted to assess the capabilities of the proposed inversion methodology. The results indicate the possibility of effectively compensating for modeling error in the considered test cases.

show abstract

Deep-Learning-Based Calibration in Contrast Source Inversion Based Microwave Subsurface Imaging

Cited by 7 publications

References 13 publications

Complex Permittivity Imaging by Incorporating Synthetic Aperture Radar and Inverse Scattering Method for Stratified Ground Medium

Complex Permittivity Imaging by Incorporating Synthetic Aperture Radar and Inverse Scattering Method for Stratified Ground Medium

Complex Permittivity Retrieval Approach With Radar Enhanced Contrast Source Inversion for Microwave Nondestructive Road Evaluation

Quantitative Inversion of Multiantenna Ground-Penetrating Radar Data with Modeling Error Correction Based on Long Short-Term Memory Cells

Contact Info

Product

Resources

About