Theoretical And Experimental Research On Planar Near-Field Acoustic Holography Based On Compressed Sensing

Near-field acoustic holography(NAH)is the reverse reconstruction process of acoustic radiation.The sound field of other positions can be deduced,and abundant sound field information can be obtained,which can provide a reliable basis for the evaluation of the sound source by measuring the sound field on the holography surface.The compressed sensing method is introduced into near-field acoustic holography.The combined method inherits the advantages of the compressed sensing and also avoids the inherent shortcomings of the traditional NAH methods.The new method improves the reconstruction precision of near-field holography,and has great potential on application to real sound source.In this thesis,the algorithm of spatial Fourier transform is analyzed in detail,at the same time,the "ill-posed" problem in sound field reconstruction and its suppression methods are discussed.Then,the key process of compressed sensing is analyzed from three aspects:sparse representation of signal,observation matrix construction,signal optimization and reconstruction.Based on the above theories and methods,the sparse representation of the particle velocity on the Fourier transform base,the realization of the observation and optimization reconstruction based on compressed sensing are studied with the planar NAH.Then,the data testing process and parameter setting methods of NAH are analyzed.Compared with spatial Fourier transform and regularization,the NAH based on compressed sensing has a good reconstruction effect,and the method is verified by the point source simulation and aluminum plate excitation experiment.Microphones can be arranged randomly with the planar NAH based on compressed sensing,so the reconstruction is studied by comparing the regular and random arrangement of microphones.The result shows that the random arrangement is better than the regular arrangement.A hybrid Gaussian distribution method is proposed,and the effectiveness of the method is proved by point source simulation and refrigerator experiment.Finally,a microphone array scanning system is designed,and a set of NAH noise source test and identification system is developed on the LabVIEW.