Shape Reconstruction of Mixed Boundary Objects by Linear Sampling Method

Abstract

Geometric reconstruction of dielectric objects in the presence of a perfect electric conducting (PEC) object is investigated with the application of linear sampling method (LSM). The scattered fields from the cylindrical objects of circular cross sections are represented as a multipole expansion, and a numerical analysis is carried out to explore the behavior and limitations of LSM. By considering various numerical examples, the imaging capabilities of LSM for mixed cases (both dielectric and PEC objects) are studied in terms of frequency of operation and permittivity of dielectric objects. To overcome the drawbacks of standard LSM indicator, a modified LSM indicator based on a multifrequency approach is introduced. Apart from the synthetic data, the analysis is also validated by the experimental data provided by the Fresnel Institute.Index Terms-Linear sampling method (LSM), mixed boundary problem, multipole expansion, reconstruction.

Reconstruction of Mixed Boundary Objects and Classification Using Deep Learning and Linear Sampling Method

“…Linear sampling method was employed in all of these works, where the targets were treated as either dielectric or perfect electric conducting (PEC) objects. In [11], the analysis of LSM behavior in mixed boundary cases is presented, taking into account parameters such as operating frequency and permittivity. The study reported the following findings: 1) Regarding frequency, LSM exhibits dissimilar characteristics compared to dielectrics and perfect electric conductors (PEC).…”