Achieving high-resolution 3D acoustic imaging in a large-aspect-ratio cabin by the non-synchronous measurements

Dingyu Hu, Jinjin Ding, Han Zhao, Liang Yu

Research output: Contribution to journalArticlepeer-review

8 Scopus citations

Abstract

The resolution of conventional acoustic imaging methods in a large-aspect-ratio cabin is reduced seriously by array parameters limitation and reverberation. The non-synchronous measurements technique breaks through the limitation of array shape and the number of microphones. This method can approximate the result of synthetic large-aperture array acoustic imaging. The increase in the number of microphones and low-rank constraint assumption in non-synchronous measurements suppress the reverberation interference. The low-rank constraint filters out the eigenvalues of the reverberation sound other than the main noise sources, which is acted as an eigenvalue filter. It has been demonstrated that non-synchronous measurements with 3D beamforming can improve the resolution of acoustic imaging in cubic space, but the related research on the large-aspect-ratio cabin is still missing. The non-synchronous measurements method is applied to achieve high-resolution acoustic imaging in a large-aspect-ratio cabin in this paper. The array measurement scheme is investigated comprehensively, including the number of non-synchronous measurements, the array step distance and appropriate spatial basis functions. The experimental results show that the non-synchronous measurements can achieve high-resolution 3D acoustic imaging in the cabin of an urban rail vehicle.

Original languageEnglish
Article number108953
JournalApplied Acoustics
Volume197
DOIs
StatePublished - Aug 2022
Externally publishedYes

Keywords

  • 3D beamforming
  • Acoustic imaging
  • Large-aspect-ratio cabin
  • Non-synchronous measurements

Fingerprint

Dive into the research topics of 'Achieving high-resolution 3D acoustic imaging in a large-aspect-ratio cabin by the non-synchronous measurements'. Together they form a unique fingerprint.

Cite this