A spherical harmonics domain far-field beamforming clear imaging method and system

Abstract

The application provides a spherical harmonic function domain far-field beam forming clear imaging method, system, device and medium, the method comprises the following steps: adopting a plane wave model as a sound source distribution model, and constructing a far-field SHB output model according to the sound source distribution model and a microphone measuring point model; the microphone measuring point model is a solid ball microphone array; according to the far-field SHB output model, the far-field sound source information is determined by iterative search; the far-field sound source information includes sound source direction and sound source intensity; according to each far-field sound source information, the clear imaging of each sound source is processed in turn, and the corresponding SHB reconstruction output is obtained; the SHB reconstruction outputs corresponding to all far-field sound sources are accumulated to obtain the total output of beam forming, and the clear imaging result is obtained according to the total output of beam forming. The method can effectively reduce the main lobe width in sound source identification, eliminate side lobe pollution, thereby improving the acoustic imaging clarity and enhancing the weak source identification ability.

Description

Technical Field

[0001] This invention relates to the field of sound source identification technology, and in particular to a method and system for sharpening imaging using far-field beamforming in the spherical harmonic domain. Background Technology

[0002] Beamforming based on solid spherical microphone arrays offers advantages such as 360° panoramic acoustic imaging and suitability for medium- to long-distance measurements, making it a promising candidate for applications in noisy outdoor environments such as large converter substations. Spherical Harmonics Beamforming (SHB) is a classic processing algorithm that accurately locates sound sources when matched with spherical microphone arrays. However, its sound source identification results suffer from problems such as a wide low-frequency main lobe (weak spatial resolution) and numerous high-frequency side lobes (many false sound sources), resulting in limited acoustic imaging clarity.

[0003] Therefore, there is an urgent need to provide a sharpening imaging method based on far-field beamforming in the spherical harmonic domain, which can focus on the noise source identification problem in large converter substations. Summary of the Invention

[0004] The purpose of this invention is to provide a far-field beamforming method for enhanced imaging in the spherical harmonic domain, which addresses the shortcomings of existing SHB methods. This method can effectively reduce the main lobe width and eliminate side lobe contamination, thereby improving acoustic imaging clarity and enhancing weak source identification capabilities.

[0005] To achieve the above objectives, it is necessary to provide a method, system, computer device, and storage medium for sharpening imaging using far-field beamforming in the spherical harmonic domain, addressing the aforementioned technical problems.

[0006] In a first aspect, embodiments of the present invention provide a method for sharpening imaging using far-field beamforming in the spherical harmonic domain, the method comprising the following steps:

[0007] A plane wave model is used as the sound source distribution model, and a far-field SHB output model is constructed based on the sound source distribution model and the microphone measurement point model; the microphone measurement point model is a solid sphere microphone array.

[0008] Based on the far-field SHB output model, the information of each far-field sound source is determined through iterative search; the far-field sound source information includes the sound source direction and the sound source intensity.

[0009] Based on the information of each far-field sound source, the images of each sound source are sequentially processed to sharpen them, and the corresponding SHB reconstruction output is obtained.

[0010] The SHB reconstruction outputs corresponding to all far-field sound sources are summed to obtain the total beamforming output, and the sharpened imaging result is obtained based on the total beamforming output.

[0011] Furthermore, the step of constructing the far-field SHB output model based on the sound source distribution model and the microphone measurement point model includes:

[0012] Based on the focusing direction and intensity of the monopole point sound source, the classical SHB theoretical output model corresponding to the microphone measurement point model is obtained; the classical SHB theoretical output model is expressed as:

[0013]

[0014] Where b(k,Ω) F ) indicates that the classic SHB is focused in the direction Ω. F The theoretical output is given by: s represents the sound source intensity; k represents the wave number, and k = 2πf / c, where f represents the frequency and c represents the speed of sound. Let represent the spherical harmonic function along the focusing direction Ω, where n and m represent the order and degree, respectively; (·) * This indicates finding the conjugate;

[0015] Based on the spherical Fourier transform coefficients of the surface sound pressure signal of the solid spherical microphone array, the classical SHB theoretical output model is transformed to obtain the transformed SHB output model; the transformed SHB output model is expressed as:

[0016]

[0017] In the formula,

[0018]

[0019]

[0020] in, Represents the spherical Fourier transform coefficients; S represents the spherical Fourier transform; Q represents the total number of microphones; α q p(ka,Ω) represents the weight of the q-th microphone. Mq ) represents Ω on a solid sphere of radius a. Mq Directional sound pressure signal; R n (ka) represents the nth order radial function, and a represents the radius of the sphere array; and Let j represent the nth-order spherical Hankel function of the second kind and its corresponding first derivative; n (·) and j' n (·) denotes the nth-order spherical Bessel function of the first kind and its corresponding first derivative, respectively; j represents the imaginary unit;

[0021] Based on the classical SHB theoretical output model for a single sound source, the far-field SHB output correction coefficient is obtained.

[0022] The far-field SHB output model is obtained based on the far-field SHB output correction coefficient and the transformed SHB output model.

[0023] Furthermore, the far-field SHB output model is expressed as:

[0024]

[0025] In the formula,

[0026] w = 4π / (N+1) 2

[0027]

[0028]

[0029]

[0030] Γ=Diag([α1 α2…α Q ])∈R Q×Q

[0031]

[0032]

[0033] in, This represents the corrected far-field SHB output with the sound pressure cross-spectrum matrix as input; w represents the far-field SHB output correction coefficient; C represents the far-field SHB output model of the cross-spectrum matrix corresponding to the sound pressure measured by the solid sphere microphone array; Γ represents the diagonal matrix composed of the weights of each microphone in the solid sphere microphone array; y FN Indicates the focusing direction Ω F The vector composed of the corresponding spherical harmonic functions of each order; Y MN B represents the matrix consisting of the spherical harmonic functions of each order corresponding to all microphone directions in the solid sphere microphone array; FN p represents a diagonal matrix composed of radial functions of various orders; ★ This represents the sound pressure vector measured by the solid sphere microphone array; (·) -1 Indicates the inverse; (·) H Indicates transpose and conjugate; This indicates the expectation; R represents the set of real numbers; C represents the set of complex numbers; Diag(·) represents a diagonal matrix with the vector inside the parentheses as its diagonal; N represents the truncation length.

[0034] Furthermore, the step of determining the information of each far-field sound source through iterative search based on the far-field SHB output model includes:

[0035] Based on the far-field SHB output model, the initial SHB output peak value is obtained by searching.

[0036] The sound source corresponding to the initial SHB output peak is taken as the strongest sound source, and the initial SHB output peak and the corresponding position are taken as the sound source intensity and sound source direction of the strongest sound source, respectively.

[0037] Based on the sound source intensity and direction of the strongest sound source, the sound pressure cross spectrum matrix of the strongest sound source is obtained;

[0038] The acoustic pressure cross-spectrum matrix is ​​removed from the initial cross-spectrum matrix corresponding to the far-field SHB output model to obtain the residual acoustic pressure cross-spectrum matrix.

[0039] Based on the residual sound pressure cross spectrum matrix, the corresponding residual SHB output is obtained;

[0040] The residual SHB output peak value is obtained by searching, and the residual SHB output peak value and the corresponding position are respectively used as the source intensity and source direction of the secondary strong sound source.

[0041] Determine whether the preset iteration termination condition is met. If not, update the residual SHB output based on the second strongest sound source and continue the iterative search. Otherwise, stop the iterative search and obtain the corresponding far-field sound source information. The preset iteration termination condition is reaching the preset maximum number of iterations, or the peak value of the SHB output in the current iteration is not less than the peak value of the SHB output in the previous iteration.

[0042] Furthermore, the residual SHB output is represented as:

[0043]

[0044] In the formula,

[0045]

[0046] in, Ω represents the focusing direction after the (γ+1)th iteration. F SHB residual output on; This represents the vector composed of the residual SHB outputs of all focal points after the (γ+1)th iteration; This represents the residual peak value of the SHB output after the γth iteration; Represents the set of focal point directions; C (r+1) Let represent the residual acoustic pressure cross spectrum matrix of the (γ+1)th iteration; express The focusing vector from the far-field sound source in the direction of the solid sphere microphone array.

[0047] Furthermore, the step of sequentially performing sharpening processing on the imaging of each sound source based on the information of each far-field sound source to obtain the corresponding SHB reconstruction output includes:

[0048] Based on the sound source intensity and direction of each far-field sound source, the corresponding sound pressure cross-spectrum matrix is ​​obtained; the sound pressure cross-spectrum matrix is ​​expressed as:

[0049]

[0050] Among them, G (γ+1) Let G represent the sound pressure cross-spectrum matrix corresponding to the far-field sound source identified in the (γ+1)th iteration, and G (γ+1) ∈C Q×Q ; express The focusing vector from the far-field sound source in the direction of the solid sphere microphone array; This represents the residual peak value of the SHB output after the γth iteration;

[0051] Based on the cross-spectral matrix of each far-field sound source, the corresponding SHB reconstruction output is obtained; the SHB reconstruction output is expressed as:

[0052]

[0053] in, This indicates that the far-field sound source identified in the (γ+1)th iteration is located in the focusing direction Ω. F SHB reconstruction output; Ψ represents the vector composed of the SHB reconstruction outputs of all focal points after the (γ+1)th iteration; Ψ represents the beamwidth function.

[0054] Furthermore, the total output of the beamforming is expressed as:

[0055]

[0056] in, Indicates the total output of beamforming; This represents the vector composed of the SHB reconstruction outputs of all focal points after the γth iteration.

[0057] Secondly, embodiments of the present invention provide a spherical harmonic domain far-field beamforming system for enhanced imaging, the system comprising:

[0058] The model building module is used to construct a far-field SHB output model by using a plane wave model as the sound source distribution model and the microphone measurement point model; the microphone measurement point model is a solid sphere microphone array.

[0059] An iterative search module is used to determine the information of each far-field sound source through iterative search based on the far-field SHB output model; the far-field sound source information includes the sound source direction and the sound source intensity.

[0060] The reconstruction output module is used to sequentially sharpen the images of each sound source based on the information of each far-field sound source, and obtain the corresponding SHB reconstruction output.

[0061] The sharpening imaging module is used to accumulate the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and to obtain the sharpening imaging result based on the total beamforming output.

[0062] Thirdly, embodiments of the present invention also provide a computer device, including 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 the steps of the above-described method.

[0063] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0064] This application provides a method, system, computer device, and storage medium for sharpening imaging using far-field beamforming in the spherical harmonic domain. The method employs a plane wave model as the sound source distribution model and, based on a far-field SHB output model constructed from the sound source distribution model and microphone measurement point model, iteratively searches to determine the information of each far-field sound source. Then, based on this information, it sequentially sharpens the imaging of each sound source to obtain the corresponding SHB reconstruction output. Finally, it sums the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and obtains the sharpened imaging result based on this total beamforming output. Compared with existing technologies, this spherical harmonic domain far-field beamforming sharpening imaging method not only effectively reduces the main lobe width and improves the spatial resolution of sound source identification, but also eliminates sidelobe contamination, sharpens the acoustic imaging results, improves the dynamic display range, and enhances the ability to identify weak sources. Attached Figure Description

[0065] Figure 1 This is a schematic diagram illustrating the application scenario of the far-field beamforming sharpening imaging method in the spherical harmonic domain according to an embodiment of the present invention;

[0066] Figure 2 This is a flowchart illustrating the method for sharpening imaging using far-field beamforming in the spherical harmonic domain, as described in this embodiment of the invention.

[0067] Figure 3 This is a schematic diagram of the layout of the solid sphere array in an embodiment of the present invention;

[0068] Figure 4 This is a schematic diagram of the plane wave model in an embodiment of the present invention;

[0069] Figure 5 This is a schematic diagram of the simulation process of the far-field beamforming sharpening imaging method in the spherical harmonic domain in an embodiment of the present invention;

[0070] Figure 6 This is a schematic diagram comparing the simulation results of the spherical harmonic domain far-field beamforming sharpening imaging method in this embodiment of the invention with existing SHB imaging.

[0071] Figure 7 This is a schematic diagram of the arithmetic mean of the weighted autospectral of each channel A of the solid sphere microphone array in an embodiment of the present invention;

[0072] Figure 8 This is a schematic diagram comparing the actual application results of the spherical harmonic domain far-field beamforming sharpening imaging method in this embodiment of the invention with existing SHB imaging.

[0073] Figure 9 This is a schematic diagram of the structure of the far-field beamforming system for sharpening imaging in the spherical harmonic domain, as described in this embodiment of the invention.

[0074] Figure 10 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation

[0075] To make the objectives, technical solutions, and beneficial effects of this application clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention and are used to illustrate the present invention, but are not intended to limit the scope of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0076] The spherical harmonic domain far-field beamforming sharpening imaging method provided by this invention can be applied to... Figure 1The terminal and server shown are described. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be a standalone server or a server cluster consisting of multiple servers. The server can use the spherical harmonic domain far-field beamforming sharpening imaging method of this invention to focus on the far-field sound source identification problem in outdoor open test environments (e.g., noise source identification in large converter substations). Based on the output results of the far-field SHB, sharpening acoustic imaging processing is performed to obtain sharpening imaging results. The obtained sharpening imaging results are then used for subsequent research on the server or sent to the terminal for users to view and analyze. The following embodiments will provide a detailed description of the spherical harmonic domain far-field beamforming sharpening imaging method of this invention.

[0077] In one embodiment, such as Figure 2 As shown, a method for sharpening imaging using far-field beamforming in the spherical harmonic domain is provided, comprising the following steps:

[0078] S11. A plane wave model is used as the sound source distribution model, and a far-field SHB output model is constructed based on the sound source distribution model and the microphone measurement point model; the microphone measurement point model is... Figure 3 The solid sphere microphone array shown; wherein, the plane wave model is as follows Figure 4 The image shows a wavefront model used in the far field, distinct from the spherical wave model. Array measurement of noise in substations falls under the far-field measurement scenario. Considering that the sound pressure amplitude in far-field sound source identification scenarios attenuates slowly with propagation distance, in... Figure 3 In the solid sphere microphone array layout diagram shown, sound source propagation is described by parallel arrows. This embodiment adopts a plane wave model, which assumes that the sound pressure amplitude does not change with the propagation distance, and constructs the SHB output model based on the plane wave assumption.

[0079] Specifically, the step of constructing the far-field SHB output model based on the sound source distribution model and the microphone measurement point model includes:

[0080] Based on the focusing direction and intensity of the monopole point sound source, the classical SHB theoretical output model corresponding to the microphone measurement point model is obtained; the classical SHB theoretical output model is expressed as:

[0081]

[0082] Where b(k,Ω) F ) indicates that the classic SHB is focused in the direction Ω. F The theoretical output is given by: s represents the sound source intensity; k represents the wave number, and k = 2πf / c, where f represents the frequency and c represents the speed of sound. Let represent the spherical harmonic function along the focusing direction Ω, where n and m represent the order and degree, respectively; (·) * This indicates the search for conjugates; based on the orthogonality of spherical harmonics, when the focusing direction Ω... F Ω with respect to the true sound source direction S When the signals are consistent, the output of equation (1) is infinite; otherwise, it is zero, thus achieving sound source localization. However, the sound source direction Ω in equation (1) is not constant. S The actual unknown leads to the spherical harmonic function of the sound source direction. It cannot be solved directly. Therefore, it needs to be transformed using the following method to facilitate the solution.

[0083] Based on the spherical Fourier transform coefficients of the surface sound pressure signal of the solid spherical microphone array, the classical SHB theoretical output model is transformed to obtain the transformed SHB output model; the transformed SHB output model is expressed as:

[0084]

[0085] In the formula,

[0086]

[0087]

[0088] in, The spherical Fourier transform coefficients represent the surface acoustic pressure signals of the solid spherical microphone array; S represents the spherical Fourier transform; Q represents the total number of microphones; α q p(ka,Ω) represents the weight of the q-th microphone. Mq ) represents Ω on a solid sphere of radius a. Mq Directional sound pressure signal; R n (ka) represents the nth order radial function of the solid sphere array, and a represents the radius of the sphere array; and Let j represent the nth-order spherical Hankel function of the second kind and its corresponding first derivative; n (·) and j' n (·) denote the nth-order spherical Bessel function of the first kind and its corresponding first derivative, respectively; j represents the imaginary unit.

[0089] spherical Fourier transform coefficients Theoretically, it is necessary to consider the entire sphere surface p(ka,Ω) Mq Integral, where p(ka,Ω) Mq ) represents a solid sphere of radius a with direction Ω. Mq The sound pressure signal. In actual measurement, the number of microphones cannot be infinite, so it is necessary to use the sum of the signals from a finite number of microphones to approximate the integral result, as shown in equation (3):

[0090]

[0091] Where, "S" represents the spherical Fourier transform, "S 2 " represents the surface area of ​​a unit sphere with radius 1; Q represents the total number of microphones; α" represents the total number of microphones. q This represents the weight of the q-th microphone. Furthermore, in practice, the order n of the spherical harmonic function cannot be ∞; usually, a truncation length N is used to replace ∞, thus obtaining the SHB output as shown in equation (4), and converting it into a matrix expression.

[0092]

[0093] Among them, y FN Indicates the focusing direction Ω F The vector composed of the corresponding spherical harmonic functions of each order; Y MN B represents the matrix composed of the spherical harmonics of all orders corresponding to the directions of all microphones in the solid sphere array; FN Γ represents the diagonal matrix composed of radial functions of each order; Γ represents the diagonal matrix composed of the weights of each microphone in the solid sphere array; p ★ The sound pressure vector is obtained from measurements of the microphone array; (·) -1 Indicates the inverse; (·) H This indicates transpose and conjugate. The specific definitions of each variable are as follows:

[0094] ① Spherical harmonic function vector / matrix:

[0095]

[0096]

[0097] Where C represents the set of complex numbers.

[0098] ②Radial function matrix:

[0099]

[0100] Diag(·) represents constructing a diagonal matrix with the vector inside the parentheses as its diagonal.

[0101] ③ Microphone weight vector:

[0102] Γ=Diag([α1 α2…α Q ])∈R Q×Q (8)

[0103] Where R represents the set of real numbers.

[0104] ④ Array measurement of sound pressure vector:

[0105] p★ =[p(ka,Ω) M1 p(ka,Ω) M2 )…p(ka,Ω MQ )] H ∈C Q×1 (9)

[0106] Considering that the SHB output calculated according to equation (4) can only realize the direction of sound source localization and cannot accurately quantify the intensity of sound source, this embodiment preferably introduces a correction coefficient w through the following method for analyzing SHB output in the case of a single sound source, and further corrects the SHB output model given by equation (4).

[0107] Based on the classical SHB theoretical output model for a single sound source, the far-field SHB output correction coefficient is obtained; where, the single sound source case can be understood as aligning the focusing direction with the sound source direction (i.e., Ω). F =Ω S At this point, according to equation (1), the classical SHB in the direction of the sound source Ω can be obtained. S The theoretical output is:

[0108]

[0109] Based on equation (10), the correction coefficient can be obtained as follows:

[0110] w = 4π / (N+1) 2 (11)

[0111] Based on the far-field SHB output correction coefficients and the transformed SHB output model, the far-field SHB output model is obtained; wherein, the far-field SHB output model can be understood as an SHB output model constructed with the cross-spectral matrix of sound pressure measured by the solid sphere array as input, and the specific derivation process is as follows:

[0112] First, the transformed SHB output model given in equation (4) is corrected using the far-field SHB output correction coefficient given in equation (11), resulting in the corrected far-field SHB output given in equation (12):

[0113]

[0114] Since both equations (4) and (10) use the sound pressure vector measured by the solid sphere array as input, considering the requirements of subsequent de-clarification processing, this embodiment converts equation (12) into an SHB output model that uses the cross-spectrum matrix of the sound pressure measured by the solid sphere array as input. That is, the matrix expression of the SHB output model is:

[0115]

[0116] in, This indicates the corrected far-field SHB output with the acoustic pressure cross-spectrum matrix as input. Indicates the expectation. This represents the cross-spectral matrix corresponding to the sound pressure measured by the solid microphone array.

[0117] Considering that stronger sound sources may mask weaker sound sources when imaging the far-field SHB output model obtained through the above steps, this embodiment preferably avoids the masking of weak sources by iteratively searching for the strongest sound source and removing it one by one, thereby making the weak source recognition ability stronger and thus ensuring a larger dynamic range.

[0118] S12. Based on the far-field SHB output model, determine the information of each far-field sound source through iterative search; the far-field sound source information includes the sound source direction and sound source intensity; wherein, the information of each far-field sound source can be understood as the relevant information of the strongest sound source obtained in each round of iterative search.

[0119] Specifically, the step of determining the information of each far-field sound source through iterative search based on the far-field SHB output model includes:

[0120] Based on the far-field SHB output model, the initial SHB output peak value is obtained by searching.

[0121] The sound source corresponding to the initial SHB output peak is taken as the strongest sound source, and the initial SHB output peak and the corresponding position are taken as the sound source intensity and sound source direction of the strongest sound source, respectively.

[0122] Based on the sound source intensity and direction of the strongest sound source, the sound pressure cross spectrum matrix of the strongest sound source is obtained;

[0123] The acoustic pressure cross-spectrum matrix is ​​removed from the initial cross-spectrum matrix corresponding to the far-field SHB output model to obtain the residual acoustic pressure cross-spectrum matrix.

[0124] Based on the residual sound pressure cross spectrum matrix, the corresponding residual SHB output is obtained;

[0125] The residual SHB output peak value is obtained by searching, and the residual SHB output peak value and the corresponding position are respectively used as the source intensity and source direction of the secondary strong sound source.

[0126] Determine whether the preset iteration termination condition is met. If not, update the residual SHB output based on the second strongest sound source and continue the iterative search. Otherwise, stop the iterative search and obtain the corresponding far-field sound source information. The preset iteration termination condition can be understood as stopping when the preset maximum number of iterations is reached. Alternatively, if the peak value of the SHB output in the current iteration is greater than or equal to the peak value of the SHB output in the previous iteration in an iteration that has not reached the maximum number of iterations, the iteration can also be stopped. In practical applications, the condition can be set according to the specific situation.

[0127] The iterative search process described above can be understood as determining the location and intensity of multiple different sound sources through multiple iterations: First, the first iteration searches for the maximum output location and corresponding amplitude of the SHB, and considers this location as the location of the strongest sound source, with the amplitude being the intensity of the strongest sound source; then, in the second iteration, the sound pressure cross-spectrum matrix corresponding to the strongest sound source identified in the first iteration is obtained by inversely solving the focusing vector corresponding to the sound source intensity and direction, and this matrix is ​​removed from the initial cross-spectrum input of the SHB, i.e., the strongest sound source identified in the first iteration is removed (including the entire main lobe of the sound source), thus obtaining the residual sound pressure cross-spectrum matrix; finally, the "residual SHB output" is calculated and imaged based on the residual sound pressure cross-spectrum matrix, and the location with the largest amplitude in the residual SHB output is continued to be searched, considered as the second strongest source after the strongest source; by repeating this iterative search, the influence of strong sources on weak sources can be removed, enabling the identification of multiple sound sources, even weak sources with intensities much smaller than the strongest source, thereby effectively improving the weak source identification capability, increasing the dynamic range of sound source identification, and enhancing the weak source identification capability. The search process described above will be explained in detail below using the iteration process from the γth to the γ+1th iteration:

[0128] (1) Search for the peak output of SHB after the γth iteration. and its location Where b (γ) Let be the vector composed of the outputs of all focused grid points after the γth iteration. This represents the maximum output value obtained in the (γ+1)th iteration, where max(·) indicates taking the maximum value.

[0129] (2) Reconstruct the residual acoustic pressure cross spectrum matrix C of the (γ+1)th iteration. (r+1) :

[0130]

[0131] Among them, G (γ+1) ∈C Q×Q This represents the sound pressure cross-spectral matrix corresponding to the peak sound source identified in the (γ+1)th iteration. for The focusing vector from the sound source in a specific direction to the microphone array is expressed as follows:

[0132]

[0133] in, express The transfer function from the directional sound source to the q-th microphone array;

[0134] (3) Calculate the residual output of SHB after the (γ+1)th iteration:

[0135]

[0136] in, The focusing direction Ω after the (γ+1)th iteration F SHB residual output on This is a vector composed of the residual SHB outputs of all focal points after the (γ+1)th iteration. C is the set of focal point directions; (r+1) This represents the input of the residual sound pressure cross spectrum matrix after removing the influence of all sound sources identified in the first γ iterations;

[0137] (4) This clarification method can use the conditions given in equation (17) as the criterion for iteration termination, and at the same time set the maximum number of iterations γ. max :

[0138]

[0139] After obtaining multiple sound source information with progressively decreasing intensity through step S12, the following methods can be used to process them for clarity, thereby effectively improving the clarity of acoustic imaging.

[0140] S13. Based on the information of each far-field sound source, the imaging of each sound source is processed in sequence to obtain the corresponding SHB reconstruction output.

[0141] Specifically, the step of sequentially performing sharpening processing on the imaging of each sound source based on the information of each far-field sound source to obtain the corresponding SHB reconstruction output includes:

[0142] Based on the sound source intensity and direction of each far-field sound source, the corresponding sound pressure cross-spectrum matrix is ​​obtained; the sound pressure cross-spectrum matrix is ​​expressed as:

[0143]

[0144] Among them, G (γ+1) Let G represent the sound pressure cross-spectrum matrix corresponding to the far-field sound source identified in the (γ+1)th iteration, and G (γ+1) ∈C Q×Q ; express The focusing vector from the far-field sound source in the direction of the solid sphere microphone array; This represents the residual peak value of the SHB output after the γth iteration;

[0145] Based on the cross-spectral matrix of each far-field sound source, the corresponding SHB reconstruction output is obtained; the SHB reconstruction output is expressed as:

[0146]

[0147] in, This indicates that the far-field sound source identified in the (γ+1)th iteration is located in the focusing direction Ω. F SHB reconstruction output; Ψ represents the vector composed of the SHB reconstruction outputs of all focal points after the (γ+1)th iteration; Ψ represents the beamwidth function.

[0148] This SHB reconstruction step effectively clarifies the identified far-field sound sources, significantly reducing the main lobe width (smaller main lobe width leads to more accurate localization) and eliminating side lobe contamination (fewer side lobes result in clearer imaging). This not only improves the spatial resolution of sound source identification but also makes weak sources more prominent. It should be noted that, to clarify the acoustic imaging results of each identified sound source, this embodiment preferably sets the beamwidth function Ψ specifically: when... When, 0 < Ψ ≤ 1 and Ψ(0) ≡ 1; when When, Ψ≡0; where, Indicates the focusing direction Ω F and the direction of the sound source peak The angle Φ represents the set beamwidth.

[0149] S14. Accumulate the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and obtain the sharpened imaging result based on the total beamforming output; wherein, the total beamforming output is expressed as:

[0150]

[0151] in, Indicates the total output of beamforming; This represents the vector composed of the SHB reconstruction outputs of all focal points after the γth iteration.

[0152] This application's embodiments effectively construct a far-field SHB output model based on the plane wave assumption and a solid sphere sensor array. Through iterative searching, information on various far-field sound sources with different intensities is sequentially determined. After sequentially performing de-sharpening processing on the imaging of each far-field sound source to obtain the corresponding SHB reconstruction output, the SHB reconstruction outputs corresponding to all far-field sound sources are accumulated to obtain the total beamforming output. The technical solution, based on the total beamforming output, yields the de-sharpened imaging result. This approach not only effectively reduces the main lobe width and improves the spatial resolution of sound source identification, but also eliminates sidelobe contamination, sharpenes acoustic imaging results, improves the dynamic display range, and enhances the ability to identify weak sources.

[0153] To verify the technical effectiveness of the method of the present invention in far-field sound source identification, this embodiment also conducts relevant simulation and experimental verification:

[0154] 1) Simulation verification

[0155] use Figure 3 The simulation verification was performed on a 36-channel solid spherical microphone array with a radius of 0.0975m, and arbitrary directions in space were described as... Where, θ and These represent the elevation angle and azimuth angle, respectively. The specific simulation process is as follows: Figure 5 As shown;

[0156] In forward sound field simulation, the sound source distribution model within the free field is first assumed, including the sound source type, frequency, intensity, number, and direction; then, a solid sphere array model is established, and the number and position of the microphones are set, such as... Figure 3 As shown; finally, the theoretical sound pressure signal p received by the array is calculated based on the spherical harmonic function theory. ★ The corresponding sound pressure cross-spectral matrix C is then calculated. When identifying the sound source in reverse, the directions of arrival around the array are first set according to elevation intervals Δθ = 5° and azimuth intervals. Discretize the signal; then iteratively calculate the residual output of the SHB according to Equation (16), locate the sound source one by one and identify the source strength; finally calculate the SHB reconstruction output corresponding to each sound source according to Equation (18), and sum the outputs of each sound source to obtain a clear imaging result containing all outputs.

[0157] During the simulation, three sound sources of unequal intensity were set up, with source intensities of 2 Pa, ... and (Corresponding to 100dB, 94dB, and 91dB respectively), with sound source directions of (80°, 140°), (80°, 230°), and (120°, 170°), and it is assumed that each sound source radiates as a plane wave. Maximum number of iterations γ max =50, beamwidth Φ=20°. Simulation results are as follows: Figure 6As shown, from top to bottom, the acoustic imaging results of the existing SHB and the method of this invention are presented ("+" represents the true sound source direction), corresponding to 500Hz, 1500Hz, and 4500Hz from left to right. It can be seen that at the low frequency of 500Hz, the existing SHB outputs a large main lobe width, resulting in low spatial resolution and the large main lobe also obscuring the location of weak sources. The method of this invention, however, can still effectively identify weak sources. At the high frequency of 4500Hz, the sidelobe amplitude of the existing SHB output increases significantly, with excessively large sidelobes even exceeding the weak source, making it difficult to identify. The method of this invention effectively eliminates sidelobe contamination and can still accurately identify the direction and intensity of weak sources. This demonstrates that compared to the existing SHB, the imaging clarity of the method of this invention is significantly improved by reducing the main lobe width and suppressing sidelobe contamination, thereby effectively improving the spatial resolution and weak source identification capabilities of sound sources.

[0158] 2) Experimental verification

[0159] To further verify the effectiveness and correctness of the method of this invention in practical applications, it was used for noise source identification in a large converter substation. The experiment adopted Brüel & The company uses a solid sphere microphone array with a radius of 0.0975m, integrating 36 4958-type microphones (microphone coordinates consistent with simulation), and synchronously acquires sound pressure signals using a PULSE 3660C data acquisition system. The arithmetic mean of the A-weighted autospectral of each channel of the microphone array is as follows: Figure 7 As shown (50–1000 Hz is the main noise frequency band), the noise at 500 Hz, 600 Hz, and 700 Hz is most significant. Therefore, further acoustic imaging was performed. The corresponding experimental imaging results are as follows: Figure 8 As shown, it can be seen that the sound source localization of the existing SHB and the method of the present invention is basically the same, but the imaging clarity of the method of the present invention is significantly better than that of SHB, which is consistent with the above simulation conclusions.

[0160] In one embodiment, such as Figure 9 As shown, a spherical harmonic domain far-field beamforming system for enhanced imaging is provided, the system comprising:

[0161] Model building module 1 is used to construct a far-field SHB output model by using a plane wave model as the sound source distribution model and the microphone measurement point model; the microphone measurement point model is a solid sphere microphone array.

[0162] The iterative search module 2 is used to determine the information of each far-field sound source through iterative search based on the far-field SHB output model; the far-field sound source information includes the sound source direction and the sound source intensity.

[0163] The reconstruction output module 3 is used to sequentially sharpen the images of each sound source based on the information of each far-field sound source, and obtain the corresponding SHB reconstruction output.

[0164] The sharpening imaging module 4 is used to accumulate the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and to obtain the sharpening imaging result based on the total beamforming output.

[0165] Specific limitations regarding the spherical harmonic domain far-field beamforming sharpening imaging system can be found in the limitations of the spherical harmonic domain far-field beamforming sharpening imaging method described above, and will not be repeated here. Each module in the aforementioned spherical harmonic domain far-field beamforming sharpening imaging system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the corresponding operations of each module.

[0166] Figure 10 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 10 As shown, the computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a spherical harmonic domain far-field beamforming sharpening imaging method. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0167] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0168] In one embodiment, a computer device is provided, including 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 the steps of the method described above.

[0169] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0170] In summary, the present invention provides a method, system, computer device, and storage medium for sharpening imaging using far-field beamforming in the spherical harmonic domain. The method employs a plane wave model as the sound source distribution model and, based on the far-field SHB output model constructed from the sound source distribution model and the microphone measurement point model, iteratively searches to determine the information of each far-field sound source. Then, based on this information, it sequentially sharpens the imaging of each sound source to obtain the corresponding SHB reconstruction output. Finally, it sums the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and obtains the sharpened imaging result based on this total beamforming output. This method not only effectively reduces the main lobe width and improves the spatial resolution of sound source identification, but also eliminates sidelobe contamination, sharpens the acoustic imaging results, improves the dynamic display range, and enhances the ability to identify weak sources.

[0171] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0172] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A method for sharpening imaging using far-field beamforming in the spherical harmonic domain, characterized in that, The method includes the following steps: A plane wave model is used as the sound source distribution model, and a far-field SHB output model is constructed based on the sound source distribution model and the microphone measurement point model; the microphone measurement point model is a solid sphere microphone array. Based on the far-field SHB output model, the information of each far-field sound source is determined through iterative search; the far-field sound source information includes the sound source direction and the sound source intensity. Based on the information of each far-field sound source, the images of each sound source are sequentially processed to sharpen them, and the corresponding SHB reconstruction output is obtained. The SHB reconstruction outputs corresponding to all far-field sound sources are summed to obtain the total beamforming output, and the sharpened imaging result is obtained based on the total beamforming output. The step of constructing the far-field SHB output model based on the sound source distribution model and the microphone measurement point model includes: Based on the focusing direction and intensity of the monopole point sound source, the classical SHB theoretical output model corresponding to the microphone measurement point model is obtained. Based on the spherical Fourier transform coefficients of the surface acoustic pressure signal of the solid spherical microphone array, the classical SHB theoretical output model is transformed to obtain the transformed SHB output model. Based on the classical SHB theoretical output model for a single sound source, the far-field SHB output correction coefficient is obtained; the far-field SHB output correction coefficient is: In the formula, Indicates the far-field SHB output correction factor; N represents the cutoff length; The far-field SHB output model is obtained based on the far-field SHB output correction coefficient and the transformed SHB output model. The step of determining the information of each far-field sound source through iterative search based on the far-field SHB output model includes: Based on the far-field SHB output model, the initial SHB output peak value is obtained by searching. The sound source corresponding to the initial SHB output peak is taken as the strongest sound source, and the initial SHB output peak and the corresponding position are taken as the sound source intensity and sound source direction of the strongest sound source, respectively. Based on the sound source intensity and direction of the strongest sound source, the sound pressure cross spectrum matrix of the strongest sound source is obtained; The acoustic pressure cross-spectrum matrix is ​​removed from the initial cross-spectrum matrix corresponding to the far-field SHB output model to obtain the residual acoustic pressure cross-spectrum matrix. Based on the residual sound pressure cross spectrum matrix, the corresponding residual SHB output is obtained; The residual SHB output peak value is obtained by searching, and the residual SHB output peak value and the corresponding position are respectively used as the source intensity and source direction of the secondary strong sound source. Determine whether the preset iteration termination condition is met. If not, update the residual SHB output based on the second strongest sound source and continue the iterative search. Otherwise, stop the iterative search and obtain the corresponding far-field sound source information. The preset iteration termination condition is reaching the preset maximum number of iterations, or the peak value of the SHB output in the current iteration is not less than the peak value of the SHB output in the previous iteration.

2. The spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 1, characterized in that, The classic SHB theory output model is expressed as follows: in, This indicates that the classic SHB is in the focus direction. The theoretical output is given above; s represents the sound source intensity; k represents the wave number, and... , Indicates frequency, Indicates the speed of sound; Indicates the direction of focus spherical harmonics on, and and These represent the order and degree, respectively. This indicates finding the conjugate; The transformed SHB output model is represented as follows: In the formula, in, denoted by s; S denotes the spherical Fourier transform coefficients; Indicates the total number of microphones; Indicates the first The weight of the microphone; Indicates radius as On the solid sphere The direction of the sound pressure signal; Indicates the first A radial function of order 1, and Indicates the radius of the sphere array; and They represent the first The second-order sphere Hankel function and its corresponding first derivative; and They represent the first Bessel functions of the first kind of sphere and their corresponding first derivatives; It represents the imaginary unit.

3. The spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 2, characterized in that, The far-field SHB output model is represented as follows: In the formula, in, This indicates the corrected far-field SHB output with the acoustic pressure cross-spectrum matrix as input. This represents the far-field SHB output model of the cross-spectral matrix corresponding to the sound pressure measurement of a solid sphere microphone array. This represents a diagonal matrix composed of the weights of each microphone in a solid sphere microphone array. Indicates the direction of focus The vector consisting of the corresponding spherical harmonic functions of each order; This represents the matrix composed of the spherical harmonic functions of each order corresponding to all microphone directions in the solid sphere microphone array. This represents a diagonal matrix composed of radial functions of various orders; This represents the sound pressure vector measured by the solid sphere microphone array. This indicates finding the inverse; Indicates transpose and conjugate; This indicates a demand for expectation; Represents the set of real numbers; Represents the set of complex numbers; This represents a diagonal matrix with the vector within the parentheses forming its diagonal.

4. The spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 1, characterized in that, The residual SHB output is represented as follows: In the formula, in, Indicates the first Focus direction after the next iteration SHB residual output on; Indicates the first The vector composed of the SHB residual outputs of all focal points after the iteration; Indicates the first The residual peak value of SHB output after the next iteration; Represents the set of focal point directions; Indicates the first The residual acoustic pressure cross-spectral matrix of the next iteration; express The focusing vector from the far-field sound source in the direction of the solid sphere microphone array.

5. The spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 3, characterized in that, The step of sequentially performing image sharpening processing on each sound source based on the information of each far-field sound source to obtain the corresponding SHB reconstruction output includes: Based on the sound source intensity and direction of each far-field sound source, the corresponding sound pressure cross-spectrum matrix is ​​obtained; the sound pressure cross-spectrum matrix is ​​expressed as: in, Indicates the first The sound pressure cross-spectrum matrix corresponding to the far-field sound source identified in the next iteration, and ; express The focusing vector from the far-field sound source in the direction of the solid sphere microphone array; Indicates the first The residual peak value of SHB output after the next iteration; Based on the cross-spectral matrix of each far-field sound source, the corresponding SHB reconstruction output is obtained; the SHB reconstruction output is expressed as: in, Indicates the first The far-field sound source identified in the next iteration is in the focusing direction SHB reconstruction output; Indicates the first The vector composed of the SHB reconstruction outputs of all focal points after each iteration; This represents the beamwidth function.

6. The spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 5, characterized in that, The total output of the beamforming is expressed as follows: in, Indicates the total output of beamforming; Indicates the first After each iteration, the vector is composed of the SHB reconstruction outputs of all focal points.

7. A far-field beamforming system for sharpening imaging in the spherical harmonic domain, characterized in that, The system employing the spherical harmonic domain far-field beamforming sharpening imaging method as described in claim 1 comprises: The model building module is used to construct a far-field SHB output model by using a plane wave model as the sound source distribution model and the microphone measurement point model; the microphone measurement point model is a solid sphere microphone array. An iterative search module is used to determine the information of each far-field sound source through iterative search based on the far-field SHB output model; the far-field sound source information includes the sound source direction and the sound source intensity. The reconstruction output module is used to sequentially sharpen the images of each sound source based on the information of each far-field sound source, and obtain the corresponding SHB reconstruction output. The sharpening imaging module is used to accumulate the SHB reconstruction outputs corresponding to all far-field sound sources to obtain the total beamforming output, and to obtain the sharpening imaging result based on the total beamforming output.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images