Abstract— Electromagnetic (EM) scattering is effectively used in the detection of buried objects. However, the most challenging problem is represented by the discrimination of targets and clutters. The electromagnetic signature of known objects is often used to this purpose, but generally it is limited either to the r.m.s. value of the scattered field or to one component. In this work a technique based on the analysis of all the components of the EM field is presented and applied to the detection of landmines. I. RATIONALE The scattering of the electromagnetic (EM) field is commonly used for the detection of buried objects either metallic or, with some more difficulties, non-metallic [1]. The main issue is represented by the recognition of detected objects, and thus, by the design of a source system, detection sensors and the development of algorithms allowing the classification of any detected object as to be a potential known target or a clutter. This goal is facilitated by the availability of a number of information (i.e. the so-called EM signatures) on known targets as greatest as possible. Despite this obvious consideration, usually only one EM field vector is analyzed in most popular techniques, the electric or the magnetic field, in the high- or in the low- frequency range, respectively [1]. Moreover, quite often only either the r.m.s. value or the amplitude of one vector component, depending upon the characteristics of the sensor system, resulting in a loss of information on the EM scattering behavior of the detected objects. In this work, the numerical features of a detection system based upon a wide frequency-spectrum analysis of the EM signatures in terms of the single electric and magnetic field spatial components are presented and discussed. In particular, it is worthy noting that the design of such a detection system is a formidable challenge which requires an accurate numerical analysis of the EM scattering [2] from known targets in correspondence of actual sources and sensors [3], and the analysis of the performance of algorithms for the comparison of measured data with the signatures library. The numerical tool is useful also for a sensitivity analysis aimed at assessing the uncertainties to be associated to the various EM signatures because of possible small variations in materials, shapes or positions of the targets. From a numerical point of view, the wide-band analysis may pose several problems as concerns discretization criteria, boundary conditions and accuracy of results. II. NUMERICAL RESULTS In this contribution, various typical metallic landmines bur-ied in earth with conductivity and relative permittivity of 0.001 S/m and 2.55, respectively, are considered and their EM signa-tures evaluated in terms of scattered field components. Then, the signatures are compared by means of a suitable cross-correlation with those deriving from a clutter located at various positions with respect to the field sensors. As an example, the magnetic field scattered by the three objects shown in fig. 1 is reported, having considered the source to be a small loop an-tenna located at an height of 15 cm on the earth surface and oriented at 30° with respect to the normal in the (x-y) plane, at a frequency of 30 MHz. Results, shown in fig. 2, are obtained by means of a MoM - based numerical code. The objects are located at x = - 10 cm, 5 cm below the earth surface. 1: 2: 3: Figure 1 – Two typical landmines of different shape and considered clutter. 1e-7 1e-6 1e-5 1e-4 1e-3 0.01 0.1 1 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 Hx 1 Hz 1 Hy 1 Hx 2 Hz 2 Hy 2 Hx 3 Hy 3 Hz 3 X-axis [m] Magnitude [A/m] Figure 2 –Magnetic field scattered by the three considered objects at a given position, at a frequency of 30 MHz. III. REFERENCES [1] D. J. Daniels, Ground Penetrating Radar 2 nd Edition. IEE Radar, Sonar, Navigation and Avionics Series 15, Institution of Electrical Engineers,London 2004 [2] R. Araneo, S. Celozzi, “Numerical Analysis of Subsurface Objects Discrimination Systems”, Proceedings of CEFC, 10 th Biennial IEEE Conference on Electromagnetic Field Computation, June 16-19, 2002, Perugia, Italy, p. 146. [2] M. Kanda, “An Electromagnetic Near-Field Sensor for the Simultaneous Electric and Magnetic-Field Measurements”, IEEE Transactions on Electromagnetic Compatibility, vol. 26, 1984, pp. 102-110. Recognition of Buried Objects by Their EM Scattering R. Araneo, F. Gentili, G. Lovat, F. Maradei Dept. of Electrical Engineering – University of Rome “La Sapienza” Via Eudossiana 18, 00184, Rome, Italy rodolfo.araneo, francesca.maradei@uniroma1.it PE3-6 1-4244-0320-0/06/$20.00 ©2006 IEEE 380