Leveraging recent analytical findings on the nature of the linear sampling method (LSM) equation for the design of virtual experiments (VEs), a new effective linear reconstruction strategy is presented and discussed. The latter allows one to outperform the related VE linear inversion (VELI) method, thus successfully solving the inverse scattering problem in the canonical 2-D scalar case for an extended range of dielectric scatterers, including both homogeneous and inhomogeneous ones. A numerical assessment is performed against experimental Fresnel Institute data, which are the most widely used benchmark to comparatively test inverse scattering methods. Interestingly, unlike mostly adopted nonlinear or deep-learning-based inversion techniques, this article exploits a linear method to process the Fresnel Institute dataset.
Extended Virtual Experiments Linear Inversion With Assessment on Experimental Fresnel Institute Data
Donato, Loreto Di
2026-01-01
Abstract
Leveraging recent analytical findings on the nature of the linear sampling method (LSM) equation for the design of virtual experiments (VEs), a new effective linear reconstruction strategy is presented and discussed. The latter allows one to outperform the related VE linear inversion (VELI) method, thus successfully solving the inverse scattering problem in the canonical 2-D scalar case for an extended range of dielectric scatterers, including both homogeneous and inhomogeneous ones. A numerical assessment is performed against experimental Fresnel Institute data, which are the most widely used benchmark to comparatively test inverse scattering methods. Interestingly, unlike mostly adopted nonlinear or deep-learning-based inversion techniques, this article exploits a linear method to process the Fresnel Institute dataset.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


