Microwave imagery: analytical method and maximum entropy method

Abstract

The principle of active microwave imagery is as follows: the medium and the object are traversed by an electromagnetic wave of known characteristics and the image of the object is then reconstituted from the scattered field measured beyond the object. Unfortunately the problem is ill-posed and the results are not unique. The analytical method uses the assumption div J=0; this assumption is discussed and compared with other methods and the results of numerical simulation are given. The maximum entropy method is used to obtain the image whose structure is the nearest figure. The images obtained by this method have a good accuracy and show proof of the high standing of this method of image reconstruction.

Three-dimensional reconstruction of non-homogeneous dielectric objects by the coupled-dipole method

“…Electromagnetic scattering phenomena have been shown to be of great importance because of their various applications in many areas [1][2][3][4]. The ability of electromagnetic waves to penetrate into objects without making alterations of their structure is an important property and permits to obtain information about the studied scattering object, i.e., it is possible to make a non-destructive imaging analysis.…”

A neural network approach to microwave imaging

“…However, the problem is still an ill-posed and nonlinear one. In order to improve the ill-posedness of the problem, some regularization procedures must be applied, such as the Tikhonov (Francois and Pichot, 1997;Tikhonov and Arsenin, 1977;Qin and Ciric, 1994), maximum entropy (Baribaud, 1990), and pseudoinversion algorithms (Ney et al, 1984). To deal with the nonlinearity of the problem, some nonlinear numerical methods were used extensively (e.g., BIM, Wang and Chew, 1989;DBIM, Chew and Wang, 1990;Newton-Kantorovich, Joachimowicz et al, 1998;and Levenberg-Marquardt, Francois and Pichot, 1997).…”