Sound source localization method, device, medium and equipment

Abstract

The present invention provides a sound source localization method, device, medium and equipment, the method comprising the following steps: dividing an original speech signal into M frames of sub-speech signals; calculating the signal-to-noise ratio of each frame of the sub-speech signal; calculating the DOA result of each frame of the sub-speech signal, constructing a first Gaussian mixture model based on the DOA result, and calculating the weight of the first Gaussian mixture model; adjusting the DOA result based on the signal-to-noise ratio to obtain a DOA adjustment result, and constructing a second Gaussian mixture model based on the DOA adjustment result; calculating the similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculating the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model; locating the sound source direction of each frame of the sub-speech signal based on the weight of the second Gaussian mixture model. The present invention solves the problem of low accuracy of sound source localization in a complex speech environment.

Description

Technical Field

[0001] The present invention relates to the technical field of signal gain, and in particular to a sound source localization method, device, medium and equipment. Background Art

[0002] Traditional signal processing sound source localization technology uses an array of multiple microphones to receive sound signals, and uses signal processing algorithms to estimate the delay and amplitude difference of the microphone array signal to determine the position and direction of the sound source. This method has high positioning accuracy in theory, but faces some challenges in practical applications.

[0003] Environmental noise, echoes, and voice activity can all have a negative impact on the accuracy of sound source localization. In real environments, there are various background noises, including coherent noise, incoherent noise, and diffuse noise. These noises will mix with the signals of the target sound source, making it difficult to accurately extract and analyze the signals of the target sound source. The presence of noise makes it difficult for the positioning algorithm to accurately estimate the delay and amplitude difference of the sound source, resulting in positioning errors; when sound is reflected in the environment, echoes are generated. These reflected signals are mixed with the direct signal, resulting in additional interference in the signal processing process. The presence of echoes makes it difficult for the positioning algorithm to accurately estimate the delay and amplitude difference of the sound source, thus affecting the accuracy of positioning; in an environment where multiple people are communicating, there are multiple people talking at the same time. These cross-talking sounds will introduce additional sound components, making it difficult for the sound source localization algorithm to distinguish and locate each sound source, resulting in inaccurate positioning results.

[0004] Therefore, it is necessary to provide a sound source localization method, system, transposition and device to solve the above problems. Summary of the invention

[0005] The object of the present invention is to provide a sound source localization method, system, transposition and device for solving the problem of low accuracy of sound source localization in a complex speech environment.

[0006] In a first aspect, an embodiment of the present invention provides a sound source localization method, which includes the following steps: dividing an original speech signal into M frames of sub-speech signals, where M is a positive integer; calculating the signal-to-noise ratio of each frame of the sub-speech signal; calculating the response energy of each frame of the sub-speech signal in different directions, constructing a first Gaussian mixture model based on the response energy, and calculating the weight of the first Gaussian mixture model; adjusting the response energy based on the signal-to-noise ratio to obtain an adjusted response energy, and constructing a second Gaussian mixture model based on the adjusted response energy; calculating the similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculating the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model; and locating the sound source direction of each frame of the sub-speech signal based on the weight of the second Gaussian mixture model.

[0007] The beneficial effect of the sound source localization method provided by the embodiment of the present invention is: by dividing the original speech signal into frames, constructing a first Gaussian mixture model according to the DOA result of the sub-speech signal of each frame, adjusting the DOA result by the signal-to-noise ratio to obtain the DOA adjustment result, constructing a second Gaussian mixture model according to the DOA adjustment result, calculating the similarity between the first Gaussian mixture model and the second Gaussian mixture model, calculating the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model, and locating the sound source direction of the sub-speech signal of each frame based on the weight of the second Gaussian mixture model. The method of the present invention can obtain more accurate sound source localization results in a complex speech environment.

[0008] In a possible implementation manner, calculating the target gain corresponding to each frame of the frequency domain signal based on the spectrum energy corresponding to each frame of the frequency domain signal specifically includes:

[0009] Calculating the intermediate gain corresponding to each frame of the frequency domain signal based on the spectrum energy corresponding to each frame of the frequency domain signal;

[0010] The intermediate gain is adjusted to obtain a target gain corresponding to each frame of the frequency domain signal.

[0011] In a possible implementation manner, the relationship between the response energy and the adjusted response energy satisfies the following formula:

[0012]

[0013] in, is the adjusted response energy of the current frame, SNR(l) is the signal-to-noise ratio of the current frame, F(l) is the response energy of the current frame, is the adjusted response energy of the previous frame.

[0014] In a possible implementation manner, the relationship between the response energy and the adjusted response energy satisfies the following formula:

[0015]

[0016] in, is the adjusted response energy of the current frame, F(l) is the response energy of the current frame, is the adjustment response energy of the previous frame, and the value range of α is 0-1.

[0017] In a possible implementation, calculating the similarity between the first Gaussian mixture model and the second Gaussian mixture model specifically includes: calculating the similarity based on a first probability and a second probability, and calculating the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model, the first probability being the probability of the first Gaussian mixture model at a first angle, and the second probability being the probability of the second Gaussian mixture model at the first angle.

[0018] In a possible implementation manner, the similarity satisfies the following formula:

[0019]

[0020] in, is the similarity, p l (x) is the first probability, is the second probability.

[0021] In a possible implementation manner, the weight of the second Gaussian mixture model and the sound source direction of each frame of the sub-speech signal satisfy the following formula:

[0022]

[0023] in, is the weight of the second Gaussian mixture model, ω k is the weight of the first Gaussian mixture model, For similarity.

[0024] In a second aspect, an embodiment of the present invention further provides a sound source localization device, which includes modules / units for executing any possible design method of the first aspect. These modules / units can be implemented by hardware, or by hardware executing corresponding software implementations.

[0025] In a third aspect, an embodiment of the present invention provides a terminal device, including a processor and a memory, wherein the memory is used to store one or more computer programs; when the one or more computer programs stored in the memory are executed by the processor, the terminal device can implement any possible design method of the first aspect.

[0026] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a computer program. When the computer program runs on an electronic device, the electronic device executes any possible design method of any of the above aspects.

[0027] In a fifth aspect, an embodiment of the present invention further provides a method comprising a computer program product, which, when executed on a terminal, enables the electronic device to execute any possible design of any of the above aspects.

[0028] For the beneficial effects of the second to fifth aspects, reference may be made to the description of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A schematic diagram of a communication scenario architecture provided by an embodiment of the present invention;

[0031] Figure 2 A schematic diagram of a sound source localization method provided by an embodiment of the present invention;

[0032] Figure 3 A schematic diagram of a sound source localization structure provided by an embodiment of the present invention;

[0033] Figure 4 A schematic diagram of a device structure provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be understood by people with general skills in the field to which the present invention belongs. "Including" and similar words used in this article mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0035] The technical solutions in the embodiments of the present invention are described below in conjunction with the accompanying drawings in the embodiments of the present invention. Among them, in the description of the embodiments of the present invention, the terms used in the following embodiments are only for the purpose of describing specific embodiments, and are not intended to be used as limitations on the present application. As used in the specification and the appended claims of the present application, the singular expressions "a", "said", "above", "the" and "this" are intended to also include expressions such as "one or more", unless there is a clear contrary indication in the context. It should also be understood that in the following embodiments of the present application, "at least one", "one or more" refer to one or more than two (including two). The term "and / or" is used to describe the association relationship of associated objects, indicating that three relationships can exist; for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship.

[0036] References to "one embodiment" or "some embodiments" etc. described in this specification mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Thus, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways. The term "connection" includes direct connection and indirect connection, unless otherwise specified. "First" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated.

[0037] In the embodiments of the present invention, words such as "exemplarily" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplarily" or "for example" in the embodiments of the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplarily" or "for example" is intended to present related concepts in a specific way.

[0038] The sound source localization method provided in the present application can be applied to a computer device, which can be a terminal or a server. It is understandable that the sound source localization method provided in the present application can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server.

[0039] Taking the sound source localization method provided by the present application as an example, the sound source localization method provided by the present application can also be applied to a system including a terminal and a server. Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The terminal 102 uploads the collected original voice signal to the server 104, and the server 104 divides the original voice signal into frames, adjusts the DOA result by the signal-to-noise ratio of each frame of the sub-voice signal to obtain the DOA adjustment result, constructs a first Gaussian mixture model based on the DOA result, constructs a second Gaussian mixture model based on the DOA adjustment result, calculates the similarity between the first Gaussian mixture model and the second Gaussian mixture model, calculates the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model, and then locates the sound source direction of each frame of the sub-voice signal according to the weight of the second Gaussian mixture model. Among them, the terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0040] In one embodiment, Figure 2 As shown, a sound source localization method is provided, referring to Figure 2 , including the following steps:

[0041] S201, dividing an original speech signal into M frames of sub-speech signals, where M is a positive integer.

[0042] The speech signal is preprocessed to obtain the original speech signal, and the preprocessing includes echo cancellation, or the M-frame speech signal is preprocessed to obtain the M-frame sub-speech signal, and the preprocessing includes echo cancellation.

[0043] This step specifically includes: the original voice signal is the signal flow after the device samples or the previous algorithm processing c=1,…,M, where c represents M microphone channels.

[0044] Each sampling point has a certain number of bits, such as 16-bit sampling, so it needs to be normalized so that x i ∈[-1,1],x i is the original speech signal at the i-th sampling point.

[0045] In real-time processing, every R sampling points is taken as a frame, that is, the speech signal of the nth frame is x(n)=[x nL-R+1 , …, x nL ] T , x(n) is the speech signal stream of the nth frame, x nL-R+1 is the original speech signal of the nL-R+1th sampling point of the nth frame, x nLis the original speech signal of the nLth sampling point of the nth frame, x i =0,x i is the original speech signal of the ith sampling point. Generally, for an audio file with a sampling rate of 16KHz, 32ms is taken as one frame, and the frame shift is 16ms, that is, R=512, L=256, where R is the number of sampling points per frame, and L is the number of repeated sampling points in each frame.

[0046] S202, calculating the signal-to-noise ratio of each frame of the sub-speech signal.

[0047] The noise and speech of the n-th frame sub-speech signal are separated by a noise reduction algorithm.

[0048] The noise signal of the n-th frame sub-speech signal satisfies the following formula:

[0049] N(n)=[n nL-R+1 ,…,n nL ] T

[0050] N(n) is the noise signal of the n-th frame sub-speech signal, n nL-R+1 is the noise signal of the nL-R+1th sampling point in the nth frame, n nL is the noise signal of the nLth sampling point in the nth frame.

[0051] The speech signal of the n-th frame sub-speech signal satisfies the following formula:

[0052] S(n)=[s nL-R+1 ,…,s nL ] T

[0053] S(n) is the speech signal of the n-th frame sub-speech signal, s nL-R+1 is the speech signal of the nL-R+1th sampling point in the nth frame, s nL is the speech signal of the nLth sampling point in the nth frame.

[0054] The signal-to-noise ratio of the n-th frame sub-speech signal satisfies the following formula:

[0055]

[0056] snr(n) is the signal-to-noise ratio of the sub-speech signal of the nth frame, S is the speech signal of the nth frame, and N is the noise signal of the nth frame.

[0057] S203, calculating the DOA result of the sub-speech signal of each frame, constructing a first Gaussian mixture model based on the DOA result, and calculating the weight of the first Gaussian mixture model.

[0058] In this step, the DOA result can be obtained by selecting the srp-path algorithm, the music algorithm or the gcc algorithm. DOA: The full name of English is: Direction of Arrival, and the full name of Chinese is: Direction of Arrival Estimation. The DOA result is obtained based on the Direction of Arrival (DOA) estimation method of the microphone.

[0059] The DOA result is obtained by using the srp-path algorithm within 0-180 degrees with a step length of two degrees. The DOA result is F(l) = [F0, F2…, F 180 ] T , divide the DOA result into G intervals, and the peak value of each interval is Each interval Build the first Gaussian mixture model in each interval G is a positive integer, and the value of G is determined according to the actual required angular resolution.

[0060] Assume that the first Gaussian mixture model in the kth interval has N peak points τ i , the value range of i is 1 to N.

[0061] The mean of the first Gaussian mixture model in the kth interval satisfies the following formula:

[0062]

[0063] Among them, μ k is the mean of the first Gaussian mixture model in the kth interval, τ i is the i-th peak point of the first Gaussian mixture model in the k-th interval.

[0064] The variance of the first Gaussian mixture model in the kth interval satisfies the following formula:

[0065]

[0066] in, is the variance of the first Gaussian mixture model in the kth interval, μ k is the mean of the first Gaussian mixture model in the kth interval, τ i is the i-th peak point of the first Gaussian mixture model in the k-th interval, F(τ i ) is the energy of the i-th peak point of the first Gaussian mixture model in the k-th interval.

[0067] The weight of the first Gaussian mixture model in the kth interval satisfies the following formula:

[0068]

[0069] Among them, ωk is the weight of the first Gaussian mixture model in the kth interval, F(τ k ) is the energy of the first Gaussian mixture model in the kth interval, τ i is the i-th peak point of the first Gaussian mixture model in the k-th interval, F(τ i ) is the energy of the i-th peak point of the first Gaussian mixture model in the k-th interval.

[0070] S204: Adjust the DOA result based on the signal-to-noise ratio to obtain a DOA adjustment result, and construct a second Gaussian mixture model based on the DOA adjustment result.

[0071] The relationship between the DOA result and the DOA adjustment result satisfies the following formula:

[0072]

[0073] in, is the DOA adjustment result of the current frame, SNR(l) is the signal-to-noise ratio of the current frame, F(l) is the DOA result of the current frame, It is the DOA adjustment result of the previous frame.

[0074] The relationship between the DOA result and the DOA adjustment result satisfies the following formula:

[0075]

[0076] in, is the DOA adjustment result of the current frame, F(l) is the DOA result of the current frame, It is the DOA adjustment result of the previous frame, and the value range of α is 0-1.

[0077] Divide the DOA adjustment result into G intervals and establish the second Gaussian mixture model in each interval G is a positive integer, and the value of G is determined according to the actual required angular resolution.

[0078] Assume that there is Peak Points The value range of i is 1 to

[0079] The mean of the second Gaussian mixture model of the kth interval satisfies the following formula:

[0080]

[0081] in, is the mean of the second Gaussian mixture model in the kth interval, is the i-th peak point.

[0082] The variance of the second Gaussian mixture model of the kth interval satisfies the following formula:

[0083]

[0084] in, is the variance of the second Gaussian mixture model in the kth interval, is the mean of the second Gaussian mixture model in the kth interval, is the i-th peak point of the second Gaussian mixture model in the k-th interval, is the energy of the i-th peak point of the second Gaussian mixture model in the k-th interval.

[0085] S205: Calculate the similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculate the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model.

[0086] In this step, optionally, the server 104 can calculate the similarity based on a first probability and a second probability, and calculate the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model, wherein the first probability is the probability of the first Gaussian mixture model at a first angle, and the second probability is the probability of the second Gaussian mixture model at the first angle.

[0087] The similarity satisfies the following formula:

[0088]

[0089] in, is the similarity, p l (x) is the first probability, is the second probability.

[0090] S206: Locate the sound source direction of each frame of the sub-speech signal based on the weight of the second Gaussian mixture model.

[0091] The weight of the second Gaussian mixture model and the sound source direction of each frame of the sub-speech signal satisfy the following formula:

[0092]

[0093] in, is the weight of the second Gaussian mixture model, ω k is the weight of the first Gaussian mixture model, is the similarity, ω i is the weight of the i-th peak point of the first Gaussian mixture model in the k-th interval.

[0094] In actual calculation, in order to save the amount of calculation, the calculation formula of the weight of the second Gaussian mixture model can be simplified as follows:

[0095]

[0096] in, is the weight of the second Gaussian mixture model, ω k is the weight of the first Gaussian mixture model, For similarity.

[0097] Finally, the sound source localization angle of each frame of the sub-speech signal satisfies the following formula:

[0098]

[0099] in, is the weight of the second Gaussian mixture model, θ is the sound source localization angle of each frame of the sub-speech signal, τ k is the i-th peak point of the first Gaussian mixture model in the k-th interval.

[0100] In some embodiments of the present application, the present invention also discloses a sound source localization device, such as Figure 3 As shown, the device is used to implement the methods recorded in the above method embodiments, which includes: a division unit 301, used to divide the original speech signal into M frames of sub-speech signals, M is a positive integer; a first calculation unit 302, used to calculate the signal-to-noise ratio of each frame of sub-speech signals; a first construction unit 303, used to calculate the response energy of each frame of sub-speech signals in different directions, construct a first Gaussian mixture model based on the response energy, and calculate the weight of the first Gaussian mixture model; a second construction unit 304, used to adjust the response energy based on the signal-to-noise ratio to obtain an adjusted response energy, and construct a second Gaussian mixture model based on the adjusted response energy; a second calculation unit 305, used to calculate the similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculate the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model; a positioning unit 306, used to locate the sound source direction of each frame of sub-speech signals based on the weight of the second Gaussian mixture model.

[0101] All relevant contents of each step involved in the above method embodiment can be referred to the functional description of the corresponding functional module, and will not be repeated here.

[0102] In other embodiments of the present application, the present invention discloses a terminal device, such as Figure 4As shown, the terminal device may include: one or more processors 401; a memory 402; a display 403; one or more applications (not shown); and one or more computer programs 404. The above components may be connected via one or more communication buses 405. The one or more computer programs 404 are stored in the above memory 402 and configured to be executed by the one or more processors 401. The one or more computer programs 404 include instructions, which may be used to execute the following: Figure 2 And each step in the corresponding embodiment.

[0103] Through the description of the above implementation methods, technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0104] Each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0105] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as flash memory, mobile hard disk, read-only memory, random access memory, disk or optical disk.

[0106] The above is only a specific implementation of the embodiment of the present invention, but the protection scope of the embodiment of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed in the embodiment of the present invention should be included in the protection scope of the embodiment of the present invention. Therefore, the protection scope of the embodiment of the present invention should be based on the protection scope of the claims.

[0107] Although the embodiments of the present invention are described in detail above, it is obvious to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations are within the scope and spirit of the present invention as described in the claims. Moreover, the present invention described herein may have other embodiments and may be implemented or realized in a variety of ways.

Claims

1. A sound source localization method, characterized in that: The following steps are involved: Divide the original speech signal into M frames of sub-speech signals, where M is a positive integer; Calculate the signal-to-noise ratio of each frame of the sub-speech signal; Calculating the DOA result of the sub-speech signal of each frame, constructing a first Gaussian mixture model based on the DOA result, and calculating the weight of the first Gaussian mixture model; Adjust the DOA result based on the signal-to-noise ratio to obtain a DOA adjustment result, and construct a second Gaussian mixture model based on the DOA adjustment result; Calculating a similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculating a weight of the second Gaussian mixture model based on the similarity and a weight of the first Gaussian mixture model; Positioning the sound source of each frame of the sub-speech signal based on the weight of the second Gaussian mixture model; The calculating the similarity between the first Gaussian mixture model and the second Gaussian mixture model specifically includes: The similarity is calculated based on a first probability and a second probability, and the weight of the second Gaussian mixture model is calculated based on the similarity and the weight of the first Gaussian mixture model, wherein the first probability is the probability of the first Gaussian mixture model at a first angle, and the second probability is the probability of the second Gaussian mixture model at the first angle.

2. The method according to claim 1, characterized in that The relationship between the DOA result and the DOA adjustment result satisfies the following formula: in, is the DOA adjustment result of the current frame, is the signal-to-noise ratio of the current frame, is the DOA result of the current frame, It is the DOA adjustment result of the previous frame.

3. The method according to claim 1, characterized in that The relationship between the DOA result and the DOA adjustment result satisfies the following formula: in, is the DOA adjustment result of the current frame, is the DOA result of the current frame, It is the DOA adjustment result of the previous frame, and the value range of α is 0-1.

4. The method according to claim 1, characterized in that: The similarity satisfies the following formula: in, is the similarity, is the first probability, is the second probability, is the first Gaussian mixture model, is the second Gaussian mixture model.

5. The method according to claim 1, characterized in that The weight of the second Gaussian mixture model and the sound source direction of each frame of the sub-speech signal satisfy the following formula: in, is the weight of the second Gaussian mixture model, is the weight of the first Gaussian mixture model, is the similarity, is the first Gaussian mixture model, is the second Gaussian mixture model.

6. A sound source localization device, characterized in that: include: A division unit, used for dividing the original speech signal into M frames of sub-speech signals, where M is a positive integer; A first calculation unit, used for calculating the signal-to-noise ratio of each frame of the sub-speech signal; A first construction unit, configured to calculate the response energy of each frame of the sub-speech signal in different directions, construct a first Gaussian mixture model based on the response energy, and calculate the weight of the first Gaussian mixture model; A second construction unit is configured to adjust the response energy based on the signal-to-noise ratio to obtain an adjusted response energy, and construct a second Gaussian mixture model based on the adjusted response energy; A second calculation unit, configured to calculate a similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculate a weight of the second Gaussian mixture model based on the similarity and a weight of the first Gaussian mixture model; A positioning unit, used for positioning the sound source direction of each frame of the sub-speech signal based on the weight of the second Gaussian mixture model; The second calculation unit is used to calculate the similarity between the first Gaussian mixture model and the second Gaussian mixture model, and calculate the weight of the second Gaussian mixture model based on the similarity and the weight of the first Gaussian mixture model, specifically for: The similarity is calculated based on a first probability and a second probability, and the weight of the second Gaussian mixture model is calculated based on the similarity and the weight of the first Gaussian mixture model, wherein the first probability is the probability of the first Gaussian mixture model at a first angle, and the second probability is the probability of the second Gaussian mixture model at the first angle.

7. A terminal device, characterized in that: include: A processor and a memory, wherein the memory is used to store a computer program; The processor is configured to execute the computer program stored in the memory so as to enable the terminal device to execute the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images