Ground detection by a single electromagnetic far-field measurement
Introduction
This paper concerns locating objects on a ground by using the electromagnetic (EM) scattering measurement made from a height. In Fig. 1, we give a schematic illustration of our study, where one wants to detect the multiple objects on the ground . To that end, one sends certain detecting wave fields and then measures the scattered wave fields from a height, from which to infer knowledge about the target objects. A practical scenario for our study is the scoutplane detection in the battlefield.
In what follows, we present the mathematical formulation for the current study. The detecting waves are chosen to be the time-harmonic electromagnetic plane waves of the following form where , denotes the frequency, denotes the impinging direction, and denotes the polarization with . and are entire solutions to the Maxwell equations in the free space The ground is assumed to be perfectly electric conducting (PEC). The EM waves cannot penetrate inside the ground and propagate only in the space above the ground. If there is no object presented on the ground, one would have a reflected wave field such that the total wave field satisfies the following PEC boundary condition where ν is the unit upward normal vector to . If is flat, the reflected wave field is well-understood through the work [22], [23], and if is non-flat/rough, the reflection would be much more complex. Throughout the present work, we assume that is flat and shall leave the rough case for a future study. Furthermore, without loss of generality, we assume that . Denote and . Moreover, we let Π denote the usual reflection with respect to , i.e., for a generic 3-vector . Then, we have that (cf. [22], [23]) Next, we consider that there are EM objects presented on the ground. Let , , be a non-negative Lipschitz continuous function such that for , where R is a large enough positive constant. Let , where , denote the simply connected components of . Define Each , , represents an EM object on the ground, and will be referred to as a scatterer in the sequel. Let and be the EM parameters for the object supported in , respectively, representing the electric permittivity, magnetic permeability and electric conductivity. It is assumed that , and are all constants, satisfying , and . Furthermore, it is assumed that for . If , then is taken to be a PEC obstacle, disregarding and . In the free space, and . We set The presence of the scatterer on the ground would further perturb the propagation of the EM field , inducing the so-called scattered wave field in . The scattered wave field is radiating in nature, characterized by the Silver–Müller radiation condition which holds uniformly for all directions . The total electric wave field , together with the corresponding magnetic wave field H, is governed by the following Maxwell system where and Σ are given in (1.5). Similar to (1.2), we have that We seek a pair of solutions to the scattering system (1.6), (1.7), (1.8). Particularly, the radiating wave field has the following asymptotic expansion as , with is known as the electric far-field pattern, which encodes the scattering measurement illustrated in Fig. 1.
The ground detection problem can be abstractly formulated as where is the operator sending the scatterer to the corresponding far-field pattern, defined by the Maxwell system (1.6), (1.7), (1.8). It is easily verified that is nonlinear and moreover it is ill-conditioned since is completely continuous (cf. [12]). In what follows, is always assumed to be given with all . Furthermore, if and are all fixed, then is called a single EM measurement; otherwise, it is called multiple EM measurements. In practice, a single EM measurement can be obtained by sending a single incident plane wave, and then collecting the scattered electric wave in all the observation angles. Throughout the present study, we shall take a single EM measurement for the ground detection. Moreover, our study shall be conducted in a very general and practical setting. The number of the target scatterers is not required to be known in advance, and each scatterer could be either an inhomogeneous medium or an impenetrable perfectly conducting (PEC) obstacle. Furthermore, there might be multiscale components of small-size and extended-size (compared to the detecting wavelength) presented simultaneously. Some realistic a priori information is required on scatterers of extended-size. We propose a “direct” locating method without any inversion involved. To our best knowledge, both the direct scattering model and the inverse scattering schemes are new to the literature. The results extend those obtained in [17], [18] for locating multiscale EM scatterers located in a homogeneous space. The present study is closely related to the inverse electromagnetic scattering problems from rough surfaces; see, e.g., [6], [7], [8], [9], [10], [11], [16]. We also refer to [3], [4], [5], [6], [12], [13], [14], [19], [20], [21], [25], [26], [27] for the recent progress on the inverse scattering theory and numerical study. It remarked that for highly conductive objects with large Σ (1.7) reduces to the eddy current model, which is justified in [1]. A similar asymptotic formalism for detecting and characterizing small conductive inclusions from electromagnetic induction data is developed in a recent work [2]. The mechanisms proposed in this work could be migrated to the eddy current model to detect complicated settings, e.g., multiple regular-size or even multiscale conductive inclusions.
The rest of the paper is organized as follows. In Section 2, we present some results concerning the direct scattering problem for our subsequent use. Section 3 is devoted to the inverse scattering scheme. Numerical results and discussion are presented in Section 4.
Section snippets
Scattering from multiscale ground objects
In this section, we consider the scattering from multiscale ground objects. In order to ease the exposition, throughout the rest of the paper, we assume that , and hence the size of an EM object can be interpreted in terms of its Euclidean diameter.
Let and be as introduced in (1.4) and (1.5). We further assume that there exists , such that . Let where by reordering if necessary, we assume that
Locating multiscale ground objects
With the preparations made earlier, we are ready to develop the inverse scattering scheme of detecting the multiscale ground objects introduced in (2.1). Our result extends those developed in [17], [18] for locating multiscale space objects to this interesting case of ground detection.
We first consider locating multiple small ground objects of as described in (2.22)–(2.23). Let , . The next theorem underlies the foundation for our first locating
Numerical experiments and discussions
We present extensive numerical experiments in this section to illustrate the salient features of our new locating schemes (Scheme S, R and M) for the inverse EM scattering problem with locally perturbed ground objects in three dimensions. An oblique EM plane wave with the incident direction of polar angle radian and azimuthal angle radian is employed as the detecting wave field incident on the ground objects and it yields a perturbed EM wave scattering off from ground objects to the
Acknowledgements
This work was supported by the NSF of China (Nos. 11371115, 11201453) and the NSF Grant DMS 1207784.
References (27)
- et al.
Target detection and characterization from electromagnetic induction data
J. Math. Pures Appl.
(2014) - et al.
Asymptotic formulas for perturbations in the electromagnetic fields due to the presence of inhomogeneities of small diameter. II. The full Maxwell equations
J. Math. Pures Appl.
(2001) - et al.
Inverse scattering by a continuation method with initial guesses from a direct imaging algorithm
J. Comput. Phys.
(2007) Fast numerical solution of the electromagnetic medium scattering problem and applications to the inverse problem
J. Comput. Phys.
(2006)An inverse cavity problem for Maxwell's equations
J. Differ. Equ.
(2012)- et al.
Enhanced multilevel linear sampling methods for inverse scattering problems
J. Comput. Phys.
(2014) - et al.
A justification of eddy currents model for the Maxwell equations
SIAM J. Appl. Math.
(2000) - et al.
Reconstruction of Small Inhomogeneities from Boundary Measurements
(2004) - et al.
Polarization and Moment Tensors. With Applications to Inverse Problems and Effective Medium Theory
(2007) - et al.
A MUSIC algorithm for locating small inclusions buried in a half-space from the scattering amplitude at a fixed frequency
Multiscale Model. Simul.
(2005)
MUSIC-type electromagnetic imaging of a collection of small three-dimensional inclusions
SIAM J. Sci. Comput.
Boundary integral formulae for the reconstruction of electric and electromagnetic inhomogeneities of small volume
ESAIM Control Optim. Calc. Var.
Mode matching for the electromagnetic scattering from three-dimensional large cavities
IEEE Trans. Antennas Propag.
Cited by (1)
A time domain sampling method for inverse acoustic scattering problems
2016, Journal of Computational PhysicsCitation Excerpt :Moreover, the sampling-type methods usually require very little a priori information of the unknown/inaccessible target scatterer. Due to their practical importance, the sampling-type methods have drawn a great deal of interest in the literature, and examples include the linear sampling method [9], the factorization method [20], the point source and probe method [27], the enclosure method [16], the MUSCI-type method [2,3] and the recent one-shot and orthogonality sampling methods [23,24,28], among others. Generally, the frequency-domain sampling methods could work with the data corresponding to only a single frequency.