Asymptotic properties of MUSIC-type imaging in two-dimensional inverse scattering from thin electromagnetic inclusions

摘要

The main purpose of this paper is to study the structure of the well-knownnon-iterative MUltiple SIgnal Classification (MUSIC) algorithm for identifyingthe shape of extended electromagnetic inclusions of small thickness located ina two-dimensional homogeneous space. We construct a relationship between theMUSIC-type imaging functional for thin inclusions and the Bessel function ofinteger order of the first kind. Our construction is based on the structure ofthe left singular vectors of the collected multistatic response matrix whoseelements are the measured far-field pattern and the asymptotic expansionformula in the presence of thin inclusions. Some numerical examples are shownto support the constructed MUSIC structure.