Emanuela Bermani

An innovative real-time technique for buried object detection

A new online inverse scattering methodology is proposed. The original problem is recast into a regression estimation and successively solved by means of a support vector machine (SVM). Although the approach can be applied to various inverse scattering applications, it is very suitable for dealing with buried object detection. The application of SVMs to the solution of such problems is firstly illustrated. Then some examples, concerning the localization of a given object from scattered field data acquired at a number of measurement points, are presented. The effectiveness of the SVM method is evaluated in comparison with classical neural network based approaches.

KERNELS EVALUATION OF SVM-BASED ESTIMATORS FOR INVERSE SCATTERING PROBLEMS

Buried object detection by means of microwave-based sensing techniques is faced in biomedical imaging, mine detection etc. Whereas conventional methods used for such a problem consist in solving nonlinear integral equations, this work considers a recently proposed approach based on Support Vector Machines, the techniques that proved to be theoretically justi?ed and effective in real world domains. Simulation is carried out on synthetic data generated by Finite Element code and a PML technique; noisy environments are considered as well. Results obtained for cases of polynomial and Gaussian kernels are presented and discussed.

Microwave detection and dielectric characterization of cylindrical objects from amplitude-only data by means of neural networks

In this paper, we propose a new microwave technique for the localization and the dielectric characterization of physically inaccessible cylindrical objects from amplitude-only data. By means of a neural network used to solve the inverse scattering problem; this technique allows to directly achieve the object retrieval, avoiding the drawbacks related to the measurement of the phase distribution of the field that generally represent a critical point, especially at high frequency. The efficiency of the proposed technique in the reconstruction of both the position and the dielectric properties of a circular cylindrical body from amplitude-only information is illustrated and compared with the reconstruction performances of a neural network imaging technique that makes use of both amplitude and phase of the scattered field. The presence of noisy data is also taken into account, showing the dependence of the reconstruction accuracy on the signal-to-noise-ratio.

A Multi-Source Strategy based on a Learning-by-Examples Technique for Buried Object Detection

In the framework of buried object detection and subsurface sensing, some of the main difficulties in the reconstruction process are certainly due to the aspect-limited nature of available measurement data and to the requirement of an on-line reconstruction. To limit these problems, a multi-source (MS) learning-by-example (LBE) technique is proposed in this paper. In order to fully exploit the more attractive features of the MS strategy, the proposed approach is based on a support vector machine (SVM). The effectiveness of the MS-LBE technique is evaluated by comparing the achieved results with those obtained by means of a previously developed single-source (SS) SVM-based procedure for an ideal as well as a noisy enviroment.

A finite-element procedure based on a boundary-value approach for the evaluation of the electromagnetic exposure in biological phantoms

In this paper, a finite-element method, based on a boundary-value approach, for the evaluation of the electric-field distribution in exposed biological phantoms is presented. Starting from the measurement of the electric field around the phantom, the field prediction is obtained by solving a boundary-value problem. This allows to avoid the description of the electromagnetic source and to estimate of the electric-field distribution also when the illuminating source is unknown or when its numerical model is not available. In order to show the effectiveness of the proposed approach, some numerical results, concerning a two-dimensional geometry, are provided. Firstly, the accuracy and validity of the electromagnetic prediction are assessed by comparing numerical with reference solutions (analytically computed). In order to demonstrate the efficiency, robustness, and capability of this technique, different measurement strategies, noisy environments, and errors in the data acquisition are then taken into account.

Performance of a causal anisotropic absorber for the truncation of finite element meshes in time domain electromagnetic unbounded problems

The application of a new causal anisotropic absorber for the truncation of time domain finite element meshes is presented. Numerical results are illustrated in order to show the performance of the absorber. Both homogeneous and stratified absorbers are considered and their performances compared. The computational efficiency of the finite element code implementing this absorber is also proven by comparing its computational cost with that one of a conventional time domain finite element code with boundary conditions given by perfect conductors and no absorber.