Elimination of Gaussian Correlated Measurement Noise with Bispectrum in Near-field Acoustic Holography

Abstract

In the implementation of near-field acoustic holography (NAH), the measured hologram data and reference signal are always contaminated by other noise sources, and NAH reconstruction accuracy is degraded. In the convention NAH, the complex holographic pressure is calculated by cross-spectrum based method, and uncorrelated noise can be suppressed. However, this method is not appropriate for the correlated noises. It's assumed that the measurement noise obeys Gaussian distribution, which can be eliminated effectively by high order statistics, a bispectrum based NAH method was presented in this paper to eliminate the correlated noises. The numerical simulation results show that the influence of Gaussian correlated noise can be suppressed effectively, and the reconstruction accuracy is improved.