近场反卷积聚焦波束形成声图测量

Near-field focused beamforming acoustic image measurement based on deconvolution

为了提高声图测量中对多个声源的分辨能力和定位精度,给出了一种近场二维反卷积聚焦波束形成声图测量方法。推导了水下声图测量的广义卷积模型,根据声图测量中点传播函数移变但可预测的特点,通过预存点传播函数字典的方式,将波束形成过程中的卷积问题转化成叠加积分问题,并应用二维Richardson-Lucy迭代算法实现了二维移变模型情况下的近场二维反卷积求解,从而实现高分辨声图测量。通过仿真和海试对比了反卷积、常规声图测量和MVDR声图测量的性能,结果表明反卷积算法在500次迭代情况下聚焦峰尺度小于另外两种算法的1/2,旁瓣级下降超过6 dB.

In order to improve the resolution and positioning accuracy of multiple sound sources in acoustic image measurement,a near-field deconvolution focus beamforming acoustic image measurement method is presented.Firstly,the near-field two-dimensional generalized convolution mathematical model among the conventional acoustic image measurement results,the objective function and the array directivity function is established.According to the point spread function shift-variant but predictable characteristics in the acoustic map measurement,the convolution problem in the beamforming process is transformed into the original superposition integral problem by pre-storing the point spread function dictionary.The two-dimensional Richardson-Lucy iterative algorithm is used to solve the deconvolution in the case of two-dimensional shift-variant model,and the high-resolution acoustic image measurement processing is realized.The performance of deconvolution,conventional and MVDR acoustic image measurement is compared by simulation and sea test.The results show that the deconvolution algorithm has a focus peak size less than 1/2 of the other two algorithms in 500 iterations.The sidelobe level drops by more than 6 dB.