Improving three-dimensional target reconstruction in the multiple scattering regime using the decomposition of the time-reversal operator

Abstract

The singular vectors of the time reversal operator (décomposition de l'opérateur de retournement temporel, time reversal operator decomposition (DORT) processing) are often used for localizing small echogeneous targets in a cluttered environment. In this work, we show that they can also improve the imaging of relatively large and contrasted targets in a homogeneous environment. It is observed that non-linear inversion schemes, minimizing iteratively the discrepancy between experimental data and simulated field scattered by target estimates, are more efficient when the illuminations correspond to the DORT singular vectors. In addition, DORT preprocessing permits a drastic diminution of the data load and computer burden. This study is conducted with experimental microwave data of targets with size comparable or greater than the wavelength.