Research On High-resolution Acoustic Image Measurement Method

Underwater acoustic image measurement technology is a precise passive positioning technology suitable for near-field,which is often used to locate noise sources.The most used algorithm for acoustic image measurement is focused beamforming algorithm,which is very robust.However,when the noise source frequency is low,its focus peak scale is large and its resolution is poor.Therefore,this paper has carried out research on high-resolution acoustic image measurement methods,including deconvolution algorithms and sparse Bayesian learning(SBL)algorithms,and proposed a fast algorithm based on spatial resampling Richardson-Lucy(RL)deconvolution algorithms.Through simulation and sea trial data processing,the feasibility of the algorithms is verified.Firstly,taking the horizontal linear array as an example,the acoustic image measurement model is established in the rectangular coordinate system,and the conventional acoustic image is obtained.The resolution of it is analyzed,it is directly related to the focus peak scale,which is inversely proportional to the focus peak scale.It is concluded that the focus peak scale decreases with the increase of array length and signal frequency,and decreases with the decrease of target distance.Then,study the high-resolution acoustic image measurement method.The convolution model of underwater acoustic image measurement is established,and the deconvolution acoustic image measurement algorithms based on RL and Deconvolution Approach for the Mapping of Acoustic Sources(DAMAS)are studied;The observation vector model is established,and a high-resolution acoustic image measurement algorithm based on SBL is studied;The linear array based acoustic image can be considered as deconvolution of a two-dimensional(2D)point spread function(PSF)shift-variant model.The original 2D RL method has a large computational burden,so a non-uniform spatial resampling RL fast algorithm is proposed to reduce the amount of calculation.Non-uniform sampling is performed on the scanning grid of an acoustic image to minimize the number of grid points.The high-resolution localization ability of the above algorithms is proved by simulation.Finally,the performance analysis of three acoustic image measurement algorithms,RL,spatial resampling RL,and SBL,is performed.Compare the relationship between the focus peak scale and SNR and iterations,limit resolution,sidelobe level,positioning error,and performance in the case of an error in the position of the array element.The simulation results show that the three algorithms are all convergent,high-resolution,and robust.And the spatial resampling RL algorithm can improve the computing speed and has the performance close to that of RL algorithm.By processing the sea trial data,the actual application performance of the algorithms is verified.In summary,in order to improve the resolution and positioning accuracy of low-frequency sound sources,the high-resolution acoustic image measurement algorithms are studied in this paper.The simulation and sea trial data processing results show that the deconvolution,spatial resampling RL and SBL acoustic image measurement algorithms are robust,and have high resolution and accurate positioning for low-frequency sound sources.