Virtualization-based acoustic imaging for low-frequency source localization

Acoustic imaging is a powerful technique for localizing acoustic sources with spherical microphone arrays (SMAs) capable of scanning a 360°environment. However, conventional imaging algorithms require a large SMA (d>20 cm) with many microphones to accurately localize low-frequency sources (<600 Hz). This paper proposes a parametrized method to improve localization at low frequencies by using virtualization methods. Based on the measurements of a given SMA, the number of microphones is virtually increased using the simple Kriging interpolation method, and the size of the SMA is then virtually expanded using an acoustic holography approach based on spherical harmonics expansion. A parametric study allows the identification of rules to set the virtual parameters, i.e., the number of virtual microphones, the virtual radius and the spherical harmonics truncation order. These rules are defined with the aim of locating the source as accurately as possible while minimizing side-lobes in the acoustic image. An application case is presented, where the performance of the proposed method is compared to two conventional imaging algorithms (conventional beamforming and spherical harmonics beamforming), using simulated pressure signals from two small SMAs (r = 9.75 cm) with 16 and 30 microphones, respectively. The results highlight the advantages of the proposed method in locating low-frequency sources and improving acoustic image quality over a wide frequency range, for both single and multiple source configurations. Finally, the 30 microphones SMA is 3D printed and experimental validation is performed under anechoic conditions.