Non-free field sound source positioning method, storage medium and computer equipment

Abstract

The invention provides a non-free field sound source localization method, a storage medium and computer equipment. The method comprises the following steps: setting a measurement surface; constructing a virtual sound source; performing spectral analysis on the sound pressure data; establishing a mathematical model; calculating a sparse base; establishing a sparse base model; calculating the boundary vibration velocity of the target sound source; and forming an acoustic image. The technical scheme of the invention is easy to implement and low in cost, and can be widely applied to the technical field of sound field separation and noise positioning.

Description

Technical Field

[0001] The invention relates to a sound field separation and noise localization technology, in particular to a sound field separation and noise localization technology in a non-free field environment. Background Art

[0002] Mechanical equipment will generate noise during operation, and noise is also considered a source of environmental pollution. The premise for reducing noise in mechanical equipment is to identify the noise source. At present, acoustic imaging technology based on microphone arrays has been applied in the industrial field. The core idea of ​​this technology is to use the sound field data measured by the microphone array to infer the acoustic physical quantities of the area of ​​interest and obtain an acoustic image. The obtained acoustic image can be used for noise source identification and acoustic-based fault diagnosis. According to the imaging principle, the current sound source localization methods can be divided into two categories: beamforming and near-field acoustic holography.

[0003] In a noisy environment, due to the presence of interference noise, the data collected by the microphones of the microphone array includes not only the radiated sound pressure of the target sound source, but also the radiated sound, reflected sound and scattered sound of external sound sources. In the presence of external interference, the traditional near-field acoustic holography technology is no longer applicable. Therefore, sound field separation technology based on different methods has been proposed. The sound field separation technology based on the boundary element method is not only applicable to sound sources of arbitrary shapes, but also has high accuracy.

[0004] The existing sound field separation technology based on the boundary element method requires the acquisition of the sound pressure of a double-layer measurement surface or the sound pressure and vibration velocity of a single-layer measurement surface. At the same time, the measurement surface needs to completely surround the target sound source (that is, the measurement surfaces in different directions need to be connected as a whole to form a closed measurement surface for the target sound source), so that the boundary element method can be used to calculate the incident sound pressure and then the scattered sound pressure. The existing sound field separation technology based on the boundary element method not only leads to an increase in the number of measurement points, but also in the identification process in a non-free field environment, it is necessary to first use the sound field separation technology to obtain the sound pressure of the target sound source in the free field, and then use the traditional near-field acoustic holography technology to identify the sound source. This will cause an increase in measurement and calculation costs, and the arrangement of measurement points in engineering practice will also be difficult to achieve. Summary of the invention

[0005] In order to solve the problems of large number of measurement points required to be arranged, high cost and difficult arrangement of measurement points in the existing sound field separation technology based on boundary element method, the present invention provides a non-free field sound source localization method, storage medium and computer equipment.

[0006] The technical solution of the present invention is as follows:

[0007] The non-free-field sound source localization method comprises the following steps:

[0008] A. Setting a measurement surface around the target sound source, and setting a microphone array of M channels on the measurement surface, wherein the microphone array is used to collect sound pressure data;

[0009] B. constructing a virtual sound source, wherein the boundary Γ2 of the virtual sound source surrounds the boundary Γ of the target sound source, and the measurement surface is arranged between the boundary of the virtual sound source and the boundary of the target sound source;

[0010] C. performing spectrum analysis on the sound pressure data;

[0011] D. establishing a mathematical model, wherein the mathematical model includes a transfer model between the sound pressure data and the boundary velocity of the target sound source and the boundary velocity of the virtual sound source;

[0012] E. Calculate the sparse basis of the boundary velocity of the target sound source; Calculate the sparse basis of the boundary velocity of the virtual sound source;

[0013] F. establishing a sparse model of the sound pressure of the measurement surface;

[0014] G. Calculate the sparse coefficient of the boundary velocity of the target sound source by using a sparse algorithm, and then calculate the boundary velocity of the target sound source;

[0015] H. Generate an acoustic image using the boundary vibration velocity of the target sound source.

[0016] Optionally, the measurement surface set in step A includes two layers of measurement surfaces S1 and S2 at different distances from the target sound source.

[0017] Optionally, step D comprises the following sub-steps:

[0018] The sound pressure p(r) at any point r in the sound field is:

[0019]

[0020] Among them, ∫ Γ []dS represents the surface integral on Г, i is the imaginary unit, r0 is the point on the boundary of the target sound source, G(r,r0)=exp(-jk|r-r0|) / (4π|r-r0|) is the basic solution of the free field Green's function, k=ω / c is the wave number, ω is the angular frequency, c is the speed of sound, ρ is the medium density, p(r0) is the sound pressure on the boundary of the target sound source, v n (r0) is the vibration velocity on the boundary of the target sound source, C(r) is the solid angle coefficient at r, and for the external sound field of the target sound source, the expression of C(r) is:

[0021]

[0022] Among them, D +is the external part, D - For the interior of Г;

[0023] For the internal sound field of the virtual sound source, the expression of C(r) is:

[0024]

[0025] Among them, D 1+ For the interior of Г2, D 1- For the exterior of Г2;

[0026] The sound pressures on the measuring surface are respectively and

[0027]

[0028] Among them, p f is the radiated sound pressure of the target sound source in the free field, p s is the scattered sound pressure generated by the target sound source, p i is the incident sound pressure, and the subscripts indicate the corresponding measurement surfaces S1 and S2;

[0029] N discrete points are set on the boundary of the target sound source, and the equations of the discrete points are established using the sound pressure formula at any point r in the sound field, and combined into the following matrix form:

[0030] p s =H s -1 G s v ns =Hsg·v ns

[0031] Among them, p s is the column vector composed of the sound pressure at the discrete points, v ns is the column vector composed of the vibration velocities of the discrete points, H s and G s Both are coefficient matrices of order N×N, Hsg=H s -1 G s is the transfer matrix between the sound pressure and the vibration velocity at the discrete point, the superscript -1 indicates the inverse of the matrix, and the solid angle coefficient C has been included in the diagonal elements of the coefficient matrix;

[0032] When point r is located in the radiation sound field, the radiation sound pressure of the target sound source is:

[0033] p f =H f p s +G f v ns =(Hf Hsg+G f )v ns =Av ns

[0034] Among them, p f is the column vector of the sound pressure at point r, H f and G f are all coefficient matrices of order M×N, and A is a transfer matrix between the sound pressure of point r and the vibration velocity of the discrete point of order M×N;

[0035] In a free field, the radiated sound pressure of the target sound source on the measurement surface is and for:

[0036]

[0037] Among them, v nГ is the vibration velocity of the target sound source at the discrete point, the subscript of matrix A indicates the plane where the integral is located, and the superscript of matrix A indicates the plane where point r is located;

[0038] The incident sound pressure on the measuring surface and for:

[0039]

[0040] in, is the vibration velocity of the virtual sound source at the discrete point;

[0041] The boundary conditions of the target sound source are:

[0042]

[0043] Wherein, β is the specific admittance of the boundary of the target sound source, ρc is the specific impedance of air, and k is the wave number;

[0044] When the target sound source is fixed and the boundary is rigid, the sound pressure at any point outside the target sound source is:

[0045]

[0046] Among them, p b () is defined as the blocking sound pressure, p i () is the incident sound pressure; S represents one of the measurement surfaces, and the area on one side of S adjacent to the target sound source is D 2+ , the area on the side of S away from the target sound source is D 2- ,Ω Г for

[0047]

[0048] When point r∈Г, the blocking sound pressure p on the boundary Г of the target sound source is Г b The matrix form of is:

[0049]

[0050] Among them, p Г i is the incident sound pressure on the boundary Γ of the target sound source;

[0051] The scattered sound pressure of the target sound source on S is:

[0052]

[0053] The matrix form of the scattered sound pressure radiated by the target sound source onto S is:

[0054]

[0055] The scattered sound pressure radiated by the target sound source onto S is obtained from the incident sound pressure at the boundary of the target sound source:

[0056]

[0057] The incident sound pressure generated by the virtual sound source at the boundary of the target sound source is:

[0058]

[0059] The scattered sound pressure on the measurement surface is:

[0060]

[0061] in,

[0062] The mathematical model described in step D:

[0063]

[0064] Optionally, step E comprises:

[0065] Perform singular value decomposition on Hsg in the following formula in the sub-step of step D,

[0066] p s =H S -1 G s v ns =Hsg·v ms

[0067] The obtained right singular matrix V is used as the sparse basis of the boundary vibration velocity corresponding to the sound source:

[0068] Hsg=UΣV H .

[0069] Optionally, the vibration velocity of the boundary of the target sound source is:

[0070] v nΓ =V Γ α Γ

[0071] The vibration velocity of the boundary of the virtual sound source is:

[0072]

[0073] Among them, V Г is the sparse basis of the boundary velocity of the target sound source, V Г2 is the sparse basis of the boundary velocity of the virtual sound source, α Г is the decomposition coefficient of the boundary velocity of the target sound source under the sparse basis, α Г2 is the decomposition coefficient of the boundary velocity of the virtual sound source under a sparse basis.

[0074] Optionally, the sparse model in step F:

[0075]

[0076] A storage medium stores a computer program, which, when executed by a processor, implements steps B to H in the non-free-field sound source localization method as described above.

[0077] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement steps B to H in the non-free-field sound source localization method as described above.

[0078] The technical effects of the present invention are as follows:

[0079] The existing sound field separation technology based on the boundary element method requires integration on the holographic surface during the solution process, which requires that the measurement surface must completely surround the target sound source. And as the frequency increases, in order to meet the test requirements, the distance between adjacent measurement points must be reduced accordingly, and the number of measurement points will increase further. The non-free field sound source localization method of the present invention can directly use the sound pressure of the mixed field to directly realize the sound source identification, and does not require the holographic surface to surround the target sound source, reducing the number of measurement points and low cost.

[0080] The further effects of the above optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 The figure is a flow chart of an embodiment of a non-free-field sound source localization method of the present invention.

[0082] Figure 2 It is a three-dimensional distribution diagram of an embodiment of random distribution of measurement points on the measurement surface in a simulation experiment.

[0083] Figure 3 for Figure 2 The plane distribution diagram of the measurement points of the illustrated embodiment.

[0084] Figure 4 This is the theoretical acoustic image of the target sound source at 800Hz in the simulation experiment.

[0085] Figure 5 This is the theoretical acoustic image of the target sound source at 1250Hz in the simulation experiment.

[0086] Figure 6 This is an acoustic image obtained by using the method of the present invention in a simulation experiment, with the target sound source at 800 Hz and the measurement points regularly distributed.

[0087] Figure 7 This is an acoustic image obtained by using the method of the present invention in a simulation experiment, with the target sound source at 1250 Hz and the measurement points regularly distributed.

[0088] Figure 8 This is an acoustic image obtained by using the method of the present invention in a simulation experiment, with the target sound source at 800 Hz and the measurement points randomly distributed.

[0089] Fig. 9 This is an acoustic image obtained by using the method of the present invention in a simulation experiment, with the target sound source at 1250 Hz and the measurement points randomly distributed. DETAILED DESCRIPTION

[0090] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0091] like Figure 1 The main steps of the non-free field sound source localization method of the present invention are shown below. Figure 1 Provide detailed explanation.

[0092] Set the measurement surface

[0093] In this step, it is necessary to set up several measurement surfaces around the target sound source. The target sound source refers to the object studied by the method of the present invention, that is, the exact position of the target sound source that produces sound waves of a specific frequency is analyzed and located by this method. A single measurement surface is a plane, on which an array consisting of multiple microphones is set. The microphone is used to collect sound pressure data of the target sound source. The measurement surfaces with the same distance as the target sound source are a group (or a layer) of measurement surfaces, and one or more groups of measurement surfaces can be set. There are M microphones on the set measurement surface. In this embodiment, 2 groups of measurement surfaces are set, which are defined as S1 and S2 respectively.

[0094] Constructing a virtual sound source

[0095] In this step, a virtual sound source is constructed. The boundary Г2 of the virtual sound source surrounds the boundary Г of the target sound source. The measurement surface in the previous step is located between the boundary of the virtual sound source and the boundary of the target sound source.

[0096] Spectrum Analysis

[0097] In this step, spectrum analysis is performed on the sound pressure data collected by the microphone array to determine the frequency of the sound wave to be located, such as the frequency of noise.

[0098] Building a mathematical model

[0099] In this step, a mathematical model is established, which includes a transfer model between the sound pressure data and the boundary velocity of the target sound source and the boundary velocity of the virtual sound source. The velocity mentioned in the present invention is the normal particle velocity. The specific process of establishing the mathematical model is described as follows.

[0100] The constructed virtual sound source can convert the incident sound of the external interfering sound source of the target sound source into the internal radiated sound of the virtual sound source.

[0101] For the external sound radiation of a target sound source in a free field, the sound pressure p(r) at any point r in the sound field is:

[0102]

[0103] Among them, ∫ Γ []dS represents the surface integral on the surface Г, i is the imaginary unit, r0 is the point on the boundary of the target sound source, G(r,r0) = exp(-jkr-r0) / (4πr-r0) is the basic solution of the free field Green's function, k = ω / c is the wave number, ω is the angular frequency, c is the speed of sound, ρ is the medium density, p(r0) is the sound pressure on the boundary Г of the target sound source, v n (r0) is the vibration velocity on the boundary Г of the target sound source, C(r) is the solid angle coefficient at r, and for the external sound field of the target sound source, the expression of C(r) is:

[0104]

[0105] Among them, D + is the external part, D - For the interior of Г;

[0106] For the internal sound field of a virtual sound source, the expression of C(r) is:

[0107]

[0108] Among them, D 1+ For the interior of Г2, D 1- For the outside of Г2.

[0109] The microphones collect sound pressure data on the measurement surfaces S1 and S2 respectively. The sound pressures on the measurement surfaces are and

[0110]

[0111] Among them, p f is the radiated sound pressure of the target sound source in the free field, p s is the scattered sound pressure generated by the target sound source, p i is the incident sound pressure, and the subscripts indicate the corresponding measurement surfaces S1 and S2. i It can be represented by the internal radiation sound pressure generated by the virtual sound source, that is, it can also be expressed as p f The same can be calculated by the boundary integral equation given in formula 1. The difference is that the solid angle coefficients between the two are different. Therefore, the following only considers the target sound source radiation sound pressure p f The matrix calculation process is explained below.

[0112] Set N discrete points on the boundary of the target sound source, use Formula 1 to establish the equations of the discrete points, and merge them into the following matrix form:

[0113] p s =H s -1 G s v ns =Hsg·v ns

[0114] (Formula 5)

[0115] Among them, p s is the column vector v composed of the sound pressure at the discrete points ns is the column vector composed of the vibration velocities of the discrete points, H s and G s Both are coefficient matrices of order N×N, Hsg=H s -1 G sis the transfer matrix between the discrete point sound pressure and the vibration velocity, the superscript -1 indicates the inverse of the matrix, and the solid angle coefficient C has been included in the diagonal elements of the coefficient matrix; the virtual sound source also has a corresponding transfer matrix between the discrete point sound pressure and the vibration velocity, and the difference between the two transfer matrices is only the difference in the solid angle coefficient C;

[0116] When point r is located in the radiation sound field, the radiation sound pressure of the target sound source is:

[0117] p f =H f p s +G f v ns =(H f Hsg+G f )v ns =Av ns

[0118] (Formula 6)

[0119] Among them, p f is a column vector composed of the sound pressure of the measuring point, A is a transfer matrix between the sound pressure of the point r and the vibration velocity of the discrete point of M×N order;

[0120] In a free field, the radiated sound pressure of the target sound source on the measurement surface and for:

[0121]

[0122] Among them, v nГ is the vibration velocity of the discrete point of the target sound source, the subscript of the matrix A indicates the plane where the integral is located, and the superscript of the matrix A indicates the plane where the point r is located;

[0123] Measuring the incident sound pressure on surfaces S1 and S2 and It can be regarded as generated by a virtual sound source and expressed as:

[0124]

[0125] in, is the vibration velocity of the virtual sound source at its discrete point.

[0126] The scattered sound pressure on the measurement surfaces S1 and S2 is calculated below.

[0127] Calculate the scattered sound pressure of the target sound source. The boundary conditions of the target sound source are:

[0128]

[0129] Where β is the specific admittance of the boundary of the target sound source, ρc is the specific impedance of air, and k is the wave number. In practice, most sound sources have a lower admittance than air, so β ​​is approximately 0. Assume that the target sound source is fixed and the boundary of the target sound source is rigid, that is, p f (r)=0 and v nГ =0, at this time, the sound pressure radiated by the target sound source is the scattered sound pressure generated by the incident sound on the surface of the target sound source. The sound pressure at any point in the space outside the target sound source is:

[0130]

[0131] Among them, p b () is defined as the blocking sound pressure, p i () is the incident sound pressure. S represents one of the measurement surfaces, and the area on one side of S adjacent to the target sound source is D 2+ , the area on the side of S away from the target sound source is D 2- ,Ω Г for

[0132]

[0133] When point r∈Г, the blocking sound pressure p on the boundary Г of the target sound source Г b The matrix form of is:

[0134]

[0135] Among them, p Г i is the incident sound pressure on the boundary Г of the target sound source.

[0136] The scattered sound pressure p of the target sound source on S caused by the incident sound s (r) is:

[0137]

[0138] The matrix form of the scattered sound pressure radiated by the target sound source onto S caused by the incident sound is:

[0139]

[0140] Substituting Equation 12 into Equation 14, the scattered sound pressure radiated by the target sound source onto S is obtained from the incident sound pressure at the boundary of the target sound source:

[0141]

[0142] The incident sound pressure generated by the virtual sound source at the boundary of the target sound source for:

[0143]

[0144] Substituting equation 16 into equation 15, we can obtain the scattered sound pressure on the measurement surfaces S1 and S2 respectively: for:

[0145]

[0146] in,

[0147] Substituting equations 7, 8, and 17 into equation 4, the mathematical model of the measurement surface in a noisy environment can be established, that is, the mathematical model described in step D:

[0148]

[0149] Equation 18 can be simplified to:

[0150] p=Bv n

[0151] (Formula 19)

[0152] Compute sparse basis

[0153] Perform singular value decomposition on the transfer matrix Hsg between the boundary sound pressure and the vibration velocity of the target sound source in equation 5, and use the obtained right singular matrix V as the sparse basis of the boundary vibration velocity of the target sound source

[0154] Hsg=UΣV H

[0155] (Formula 20)

[0156] Similarly, the transfer matrix between the sound pressure and the vibration velocity of the virtual sound source boundary is subjected to singular value decomposition, and the obtained right singular matrix can be used as a sparse basis of the vibration velocity of the virtual sound source boundary.

[0157] The boundary vibration velocity of the target sound source and the virtual sound source can be expressed as:

[0158]

[0159] Among them, V Г is the sparse basis of the boundary velocity of the target sound source, V Г2 is the sparse basis of the boundary velocity of the virtual sound source, α Г is the decomposition coefficient of the boundary velocity of the target sound source under the sparse basis, α Г2 is the decomposition coefficient of the boundary velocity of the virtual sound source under the sparse basis.

[0160] Building a sparse model

[0161] Substituting equation 21 into equation 18, we can obtain:

[0162]

[0163] By simplifying the above formula, we can get the sparse model of the sound pressure on the measurement surface in a non-free field:

[0164] p=Bv n =BVα=Dα

[0165] (Formula 23)

[0166] Get the boundary velocity of the target sound source

[0167] After the sparse model is established, the sparse coefficient α is calculated using the sparse algorithm under the premise that the sound pressure of the measurement surface is known, and then the decomposition coefficient α of the boundary velocity of the target sound source under the corresponding sparse basis is extracted. Г , and then use formula 21 to calculate the boundary vibration velocity v of the target sound source nГ .

[0168] Generate sonogram

[0169] The boundary vibration velocity of the target sound source is used to generate an acoustic image, so that the part where the target sound source vibrates violently can be determined, thereby achieving noise positioning and identification.

[0170] The existing sound field separation technology based on the boundary element method requires integration on the holographic surface during the solution process, which requires that the measurement surface must completely surround the target sound source, that is, the measurement surfaces in different directions surrounding the target sound source need to be connected together. And with the increase of frequency, in order to meet the test requirements, it is necessary to ensure that the distance between adjacent measuring points on the measurement surface is reduced, and the number of measuring points will increase further. The above limitations of the prior art make it difficult to implement and costly when actually arranging the measurement surface. The non-free field sound source localization method of the present invention can directly use the sound pressure of the mixed field to directly realize the sound source identification, and does not require the holographic surface to surround the target sound source. At the same time, a sparse model is established, and the sparse algorithm is used to break through the limitations of the Nyquist sampling theorem, which can effectively reduce the number of measuring points and is low in cost.

[0171] The computer program stored on the storage medium of the embodiment of the present invention and the computer program executed by the computer device of the embodiment of the present invention can be executed as follows: Figure 1 The illustrated embodiment shows steps B to H of the non-free-field sound source localization method of the present invention.

[0172] The technical effect of the present invention is verified by simulation experiments as follows.

[0173] The model settings of the simulation experiment are as follows:

[0174] The cuboid is used as the target sound source. The shell size of the cuboid is 0.4m×0.4m×0.5m, and the thickness of the shell is 0.005m. The lower surface of the cuboid is fixed, and the center of the lower surface is taken as the coordinate origin. The excitation forces of 50N and 30N are applied at (0.03, 0.03, 0.5)m and (-0.02, 0.2, 0.265)m respectively. The shell surface of the cuboid is discretized into 1794 nodes and 1792 units, and the interval between adjacent nodes is 0.025m. For plane wave interference along the x direction, in order to ensure that the sound pressure amplitude of the interference wave changes with frequency, the sound pressure amplitude of a point (0.0275, 0.0275, 0.52)m on the shell surface in the free field environment is selected as the sound pressure amplitude of the plane wave interference.

[0175] Two layers of measuring surfaces S1 and S2 are set with a rectangular parallelepiped as the core. The measuring surface S1 is composed of 6 planes equidistant from the outer surface of the rectangular parallelepiped, and the plane dimensions are all 0.3m×0.3m. The 6 planes are not connected, and each plane is adjacent to and parallel to one outer surface of the rectangular parallelepiped. Multiple measuring points on the plane (i.e., the microphones in the microphone array arranged on the measuring surface) are evenly arranged, and the interval between adjacent measuring points is 0.075m. The distance between each plane of the measuring surface S1 and the outer surface of the rectangular parallelepiped is 0.02m. The measuring surface S2 is also composed of 6 planes equidistant from the outer surface of the rectangular parallelepiped, and the plane dimensions are all 0.3m×0.3m. The 6 planes are not connected, and each plane is parallel to one outer surface of the rectangular parallelepiped. The measuring points on the measuring surface S2 are evenly arranged, and the interval between adjacent measuring points is 0.075m. The distance between each plane of the measuring surface S2 and the outer surface of the rectangular parallelepiped is 0.07m. In the present invention, the measuring points on the measuring surface can be evenly distributed or randomly distributed. Figure 2 and Figure 3 An example of a random distribution is shown.

[0176] The boundary of the virtual sound source is constructed with the target sound source as the center outside the measurement surface S2. The boundary of the virtual sound source is also in the shape of a cuboid. Each plane of the virtual boundary is adjacent to and parallel to a shell surface of the target sound source. The size of the boundary of the virtual sound source is 0.8m×0.8m×0.9m, and the distance between the plane of the boundary of the virtual sound source and the shell surface of the target sound source adjacent to it is 0.2m. 1794 nodes are discretely set on each plane of the boundary of the virtual sound source to form 1792 units.

[0177] Figure 4 and Figure 5 The theoretical result of the target sound source of the above simulation model, that is, the closer the subsequent experimental results are to this result, the smaller the error of the experimental results will be.

[0178] Figures 6 to 9 The results of simulation experiments are shown. Figures 6 to 9The results show that, using the method of the present invention, when the measuring points are evenly arranged, the reconstruction errors at 800 Hz and 1250 Hz are 26.4% and 21.3% respectively, and the target sound source vibrating violently is accurately identified; when the measuring points are randomly arranged, the reconstruction errors at 800 Hz and 1250 Hz are 29.3% and 23.4% respectively, and the target sound source vibrating violently can still be accurately identified.

[0179] It is worth noting that the above description is only a preferred embodiment of the present invention, and does not limit the scope of patent protection of the present invention. The present invention can also be replaced by equivalent technologies. Therefore, any equivalent changes made by using the description and illustrations of the present invention, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.

Claims

1. A non-free field sound source localization method, characterized in that: The following steps are involved: A. Setting a measurement surface around the target sound source, and setting a microphone array of M channels on the measurement surface, wherein the microphone array is used to collect sound pressure data; B. constructing a virtual sound source, wherein the boundary Γ2 of the virtual sound source surrounds the boundary Γ of the target sound source, and the measurement surface is arranged between the boundary of the virtual sound source and the boundary of the target sound source; C. performing spectrum analysis on the sound pressure data; D. establishing a mathematical model, wherein the mathematical model includes a transfer model between the sound pressure data and the boundary velocity of the target sound source and the boundary velocity of the virtual sound source; E. Calculate the sparse basis of the boundary velocity of the target sound source; Calculate the sparse basis of the boundary velocity of the virtual sound source; F. establishing a sparse model of the sound pressure of the measurement surface; G. Calculate the sparse coefficient of the boundary velocity of the target sound source by using a sparse algorithm, and then calculate the boundary velocity of the target sound source; H. Generate an acoustic image using the boundary vibration velocity of the target sound source.

2. The non-free field sound source localization method according to claim 1, characterized in that: The measurement surface set in step A includes two layers of measurement surfaces S1 and S2 at different distances from the target sound source.

3. The non-free field sound source localization method according to claim 2, characterized in that: Step D contains the following sub-steps: The sound pressure p(r) at any point r in the sound field is: Among them, ∫ Γ [ ]dS represents the surface integral on Г, i is the imaginary unit, r0 is the point on the boundary of the target sound source, G(r,r0)=exp(-jk|r-r0|) / (4π|r-r0|) is the basic solution of the free field Green's function, k=ω / c is the wave number, ω is the angular frequency, c is the speed of sound, ρ is the medium density, p(r0) is the sound pressure on the boundary of the target sound source, v n (r0) is the vibration velocity on the boundary of the target sound source, C(r) is the solid angle coefficient at r, and for the external sound field of the target sound source, the expression of C(r) is: Among them, D + is the external part, D - For the interior of Г; For the internal sound field of the virtual sound source, the expression of C(r) is: Among them, D 1+ For the interior of Г2, D 1- For the exterior of Г2; The sound pressures on the measuring surface are respectively and Among them, p f is the radiated sound pressure of the target sound source in the free field, p s is the scattered sound pressure generated by the target sound source, p i is the incident sound pressure, and the subscripts indicate the corresponding measurement surfaces S1 and S2; N discrete points are set on the boundary of the target sound source, and the equations of the discrete points are established using the sound pressure formula at any point r in the sound field, and combined into the following matrix form: p s =H s -1 G s v ns =Hsg·v ns Among them, p s is the column vector composed of the sound pressure at the discrete points, v ns is the column vector composed of the vibration velocities of the discrete points, H s and G s Both are coefficient matrices of order N×N, Hsg=H s -1 G s is the transfer matrix between the sound pressure and the vibration velocity at the discrete point, the superscript -1 indicates the inverse of the matrix, and the solid angle coefficient C has been included in the diagonal elements of the coefficient matrix; When point r is located in the radiation sound field, the radiation sound pressure of the target sound source is: on f =H f on s +G f v ns =(H f Hsg+G f )v ns =Off ns Among them, p f is the column vector of the sound pressure at point r, H f and G f are all coefficient matrices of order M×N, and A is a transfer matrix between the sound pressure of point r and the vibration velocity of the discrete point of order M×N; In a free field, the radiated sound pressure of the target sound source on the measurement surface is and for: Among them, v nГ is the vibration velocity of the target sound source at the discrete point, the subscript of matrix A indicates the plane where the integral is located, and the superscript of matrix A indicates the plane where point r is located; The incident sound pressure on the measuring surface and for: in, is the vibration velocity of the virtual sound source at the discrete point; The boundary conditions of the target sound source are: Wherein, β is the specific admittance of the boundary of the target sound source, ρc is the specific impedance of air, and k is the wave number; When the target sound source is fixed and the boundary is rigid, the sound pressure at any point outside the target sound source is: Among them, p b () is the blocking sound pressure, p i () is the incident sound pressure; S represents one of the measurement surfaces, and the area on one side of S adjacent to the target sound source is D 2+ , the area on the side of S away from the target sound source is D 2- ,Ω Г for When point r∈Г, the blocking sound pressure p on the boundary Г of the target sound source is Г b The matrix form of is: Among them, p Г i is the incident sound pressure on the boundary Γ of the target sound source; The scattered sound pressure of the target sound source on S is: The matrix form of the scattered sound pressure radiated by the target sound source onto S is: The scattered sound pressure radiated by the target sound source onto S is obtained from the incident sound pressure at the boundary of the target sound source: The incident sound pressure generated by the virtual sound source at the boundary of the target sound source is: The scattered sound pressure on the measurement surface is: in, The mathematical model described in step D:

4. The non-free field sound source localization method according to claim 3, characterized in that: Step E includes: Perform singular value decomposition on Hsg in the following formula in the sub-step of step D, p s =H S -1 G s v ns =Hsg·v ns The obtained right singular matrix V is used as the sparse basis of the boundary vibration velocity corresponding to the sound source: Hsg=UΣV H 。 5. The non-free field sound source localization method according to claim 4, characterized in that: The vibration velocity of the boundary of the target sound source is: v nΓ =V Γ α r The vibration velocity of the boundary of the virtual sound source is: Among them, V Г is the sparse basis of the boundary velocity of the target sound source, V Г2 is the sparse basis of the boundary velocity of the virtual sound source, α Г is the decomposition coefficient of the boundary velocity of the target sound source under the sparse basis, α Г2 is the decomposition coefficient of the boundary velocity of the virtual sound source under a sparse basis.

6. The non-free-field sound source localization method according to claim 5, characterized in that: The sparse model described in step F:

7. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, steps B to H in the non-free-field sound source localization method according to any one of claims 1 to 6 are implemented.

8. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement steps B to H in the non-free-field sound source localization method according to any one of claims 1 to 6.

Images