Dynamic target identification and classification based on resonance topography grouping

We address the long-standing challenge of sonar target identification against weak ground truths and interfering scatter components by harnessing robust topographic elements of target resonance profiles. Our goal is to achieve unknown target discovery when supervised learning is not practical due to ground truth uncertainties. Specifically, we examine topographic manifolds to localize resonance components, and thus discover target sub-signatures that may be combined wholly or partially to enable robust classification. Numerical results as well as field data-driven evidence are presented.