Multifrequency Linear Sampling Method on Experimental Datasets

Abstract

We investigate the use of the Linear Sampling Method (LSM) for determining the shape of a scatterer from multi-frequency experimental data. We study three multifrequency indicators for two 2D data sets available online: one is provided by the Institut Fresnel, and another by the Electromagnetic Imaging Laboratory of the University of Manitoba. We show that the multi-frequency LSM works exceptionally well on the 2D Fresnel database, and also acceptably well on the Manitoba one. In particular, a new multi-frequency indicator is tested, and data completion for the Fresnel data set is studied. We also test an adaptive technique to cut down on the number of evaluations of the indicator function for well resolved scatterers.

Exploring the performance of the topological energy method for object and damage detection from noisy and poor databases

“…We have also presented some numerical results comparing the performance of the topological energy versus the topological derivative. We plan to study the performance of some other qualitative methods in the future, for instance, variants of the linear sampling method, when processing limited aperture data from Fresnel databases (see [27] for recent results in the full-aperture case). We also plan to deal with more realistic elastic wave propagation models for the simulation of welding joints, improve the transmitter description and process real experimental data.…”