Ultrasonic broadband time-reversal with multiple signal classification imaging using full matrix capture

In this paper, time-reversal with multiple signal classification (TR-MUSIC) is generalised to its multi-frequency form, broadband TR-MUSIC over a given frequency bandwidth. Both methods are tested with experimental ultrasonic array data collected using the so-called full matrix capture (FMC) process and their performance is evaluated in terms of spatial resolution and robustness to structural noise, two important indicators for all imaging methods. Our study demonstrates that both methods have the capacity to resolve targets spaced closer than the Rayleigh limit, the minimum resolvable distance of an imaging system, leading to super-resolution imaging. A parameter τ (–6 dB main lobe width) is introduced to estimate the resolved capacity of both methods. The parameter τ calculated from broadband TR-MUSIC is larger than that of TR-MUSIC because more frequency components are employed for imaging. On the other side, in noisy cases such as copper, TR-MUSIC fails to locate the targets due to the presence of false image peaks that can be misinterpreted as scatterers; however, broadband TR-MUSIC can diminish the effect of strong noise and provide a stable imaging result.