SVM-based Classification Approach for Synthetic-Impulse Microwave Imaging

Viani, Federico and Benedetti, Manuel and Donelli, Massimo and Lizzi, Leonardo and Cresp, A. and Aliferis, L. and Pichot, Christian and Massa, Andrea (2008) SVM-based Classification Approach for Synthetic-Impulse Microwave Imaging. [Technical Report] (Unpublished)

[img]
Preview
Text
DISI-08-047.pdf

Download (256kB) | Preview

Abstract

Surface-Penetratic Radar (SPR) is aimed at providing an image of the objects buried in an inaccessible host medium. Such a scenario generally occurs in many applications, ranging from to the search of objects in the subsoil [Schubert 2001] to through-wall imaging (TWI)[Ferries 1998]. In order to effectively reconstruct an accurate image of the investigation area, many methodologies based on x-rays, ultrasonics, eddy currents, or microwaves have been investigated. In such a framework, microwave imaging techniques are very promising since electromagnetic fields at centimeter wavelenghts can penetrate non-ideal conductor materials and are usually low power and relatively inexpensive. Recently, thanks to the development of effective ultrawideband (UWB) hardware, wideband microwave imaging systems have been proposed. Compared to single-frequency approaches, such systems allow to collect a wider amount of information, since the scattered field can be measured at different frequencies in the working bandwidth and/or in the time-domain. In order to take advantage of multi-frequency data, signal processing plays a key role. Different techniques combining pre-processing treatments and microwave imaging are currently used to construct an image of a scene [Daniels 2004]. In addition, time-reversal (TR) data processing has been recently employed in order to detect positions of the targets in a Synthetic-Impulse Microwave Imaging System (SIMIS) [Cresp 2008]. In such a framework and within the GALILEO 2008 project, the research units, ELEDIA @ Univ. Trento and LEAT @ Univ. Nice, are currently involved in the integration of a SVM-based classification approach in the SIMIS for the reconstruction of the position and the shape of targets. Such a report is focused on the description of problem geometry as well as of the specification of the SIMIS.

Item Type: Technical Report
Subjects: Uncategorized > TU Technical Reports and Publications
URI: http://www.eledia.org/students-reports/id/eprint/395

Actions (login required)

View Item View Item