基于压缩感知的平面近场声全息方法及其应用  

METHOD FOR PLANE NEAR-FIELD ACOUSTIC HOLOGRAPHY BASED ON COMPRESSIVE SAMPLING AND ITS APPLICATION

作　　者：杜宝[1] 罗健[2] 胡飞[2] 柳小勤[1] 伍星[1] 

机构地区：[1]昆明理工大学机电工程学院,昆明650500 [2]四川长虹电器股份有限公司,绵阳621000

出　　处：《机械强度》2016年第3期485-489,共5页Journal of Mechanical Strength

基　　金：国家自然科学基金(51465022);云南省自然科学基金面上项目(2013FB014)资助~~

摘　　要：传统的平面近场声全息在其重建过程中,无论采用波数域滤波或是正则化方法,都无法解决其本身的不适定性,进而无法提高重建精度。基于压缩感知的平面近场声全息,有效地利用了质点振速信号的稀疏性,将传统地直接求解质点振速转化为间接求解稀疏系数,避免了重建过程中的不适定问题。结合基于Tikhonov正则化方法的平面近场声全息,从数值仿真角度分析了两种方法的重建精度,结果表明,基于压缩感知的平面近场声全息方法具有很好的重建精度,尤其是在重建面的边界区域。最后,应用该方法有效地识别出了冰箱的噪声源,为其噪声控制提供有效依据。In traditional plane near-field acoustic holography reconstruction process,regardless of wave number domain filter or regularization method,they cannot solve their own ill-posed problem,and thus cannot improve the reconstruction accuracy. Plane near-field acoustic holography based on Compressive Sampling,translates traditional solved particle velocity into solving sparse coefficient,by effective using of the particle velocity sparsity,avoiding the reconstruction process of ill-posed problem. Combined with plane near-field acoustic holography based on Tikhonov regularization method, we analyze the reconstruction accuracy of numerical simulation. The results show that, plane near-field acoustic holography based on Compressive Sampling has a good accuracy in reconstruction process,especially in the border region of reconstructive surface.Finally,this method can effectively identify the source of the noise of the refrigerator,to provide an effective basis for noise control.

关 键 词：平面近场声全息 压缩感知 TIKHONOV正则化 

分 类 号：TB53[理学—物理] O429[理学—声学]