A Bayesian sparse inference approach in near-field wideband aeroacoustic imaging

Ning Chu (1) , Ali Mohammad-Djafari (1) , José Picheral (2)

1 L2S - Laboratoire des signaux et systèmes
2 E3S - Supélec Sciences des Systèmes

Recently improved deconvolution methods using sparse regularization achieve high spatial resolution in aeroacoustic imaging in the low Signal-to-Noise Ratio (SNR), but sparse prior and model parameters should be optimized to obtain super resolution and be robust to sparsity constraint. In this paper, we propose a Bayesian Sparse Inference Approach in Aeroacoustic Imaging (BSIAAI) to reconstruct both source powers and positions in poor SNR cases, and simultaneously estimate background noise and model parameters. Double Exponential prior model is selected for source spatial distribution and hyper-parameters are estimated by Joint Maximized A Posterior criterion and Bayesian Expectation and Minimization algorithm. On simulated and wind tunnel data, proposed approach is well applied for near-field wideband monopole and extended source imaging. Comparing to several classical methods, proposed approach is robust to noise, super resolution, wide dynamic range, and source number and SNR are not needed.

Domains: Engineering Sciences [physics]; Signal and Image processing; Computer Science [cs]; Signal and Image Processing

Ning Chu, Ali Mohammad-Djafari, José Picheral. A Bayesian sparse inference approach in near-field wideband aeroacoustic imaging. 2012 19th IEEE International Conference on Image Processing (ICIP 2012), Sep 2012, Orlando, United States. pp.2529-2532, ⟨10.1109/ICIP.2012.6467413⟩. ⟨hal-00770947⟩