Sound source positioning method and device based on improved generalized cross-correlation algorithm and medium

Abstract

The application discloses a sound source positioning method and device based on an improved generalized cross-correlation algorithm and a medium, and belongs to the sound source positioning field. The method is as follows: sound source signals are collected through a four-element cross sensor array; an improved weighting function is constructed according to the sound source signals; wherein the improved weighting function is obtained by weighted summation of a phase transformation weighting function and a maximum likelihood weighting function; a generalized cross-correlation operation is performed according to the sound source signals and the improved weighting function to obtain a time delay; and a geometric operation is performed in a three-dimensional rectangular coordinate system based on the time delay to obtain a sound source position. Therefore, by implementing the application, the accuracy and robustness of sound source positioning can be improved in a low signal-to-noise ratio environment and a strong reverberation environment.

Description

Technical Field

[0001] This invention relates to the field of sound source localization, and in particular to a sound source localization method, apparatus and medium based on an improved generalized cross-correlation algorithm. Background Technology

[0002] Hydrogen, as a clean energy source, is widely used in industry, energy, and transportation, especially in fuel cells, hydrogen storage, and hydrogen-powered systems. However, hydrogen is highly flammable in air and is colorless and odorless, making leaks extremely prone to causing fires or explosions, posing a serious threat to human safety and the environment. Therefore, accurate and timely detection of hydrogen leaks is crucial for ensuring safety. To improve the accuracy of hydrogen leak detection, acoustic source localization technology has been introduced into leak detection systems. When hydrogen leaks, the pressure difference causes the gas to be ejected at high speed, generating sound waves. Turbulence and shock waves, especially near the leak point, generate detectable ultrasonic signals. By collecting these sound signals using a multi-sensor array, the sound source can be located, thus pinpointing the exact location of the leak.

[0003] Sound source localization primarily relies on methods based on time difference of arrival (TDOA) or direction of arrival (DOA). Among these, the time difference of arrival method is currently the most widely used technique. It determines the location of the sound source by calculating the time delay of the ultrasonic signal traveling from the sound source to different sensors. Since ultrasonic signals require different amounts of time to travel to sensors at different locations, the time delay can be calculated using the following formula:

[0004]

[0005] Where r1 and r2 are the distances from the sound source to the two sensors, respectively; v is the sound wave propagation speed; and τ is the time delay of the ultrasonic signal between the two sensors. By processing the received ultrasonic signal, the location of the sound source can be determined based on the geometric relationship between the target and the primitive position. To accurately estimate the time delay and achieve high-precision sound source localization, the Generalized Cross-Correlation (GCC) algorithm is widely used. The GCC algorithm determines the time delay by calculating the cross-correlation function of two ultrasonic signals and extracting the time difference corresponding to the peak values. However, the GCC algorithm is susceptible to noise interference in low signal-to-noise ratio environments, leading to blurred cross-correlation peaks and affecting the accuracy of time delay estimation. To address this, researchers invented the Generalized Quadratic Cross-Correlation (GQCC) algorithm, which adds a quadratic correlation operation to the GCC algorithm to enhance the ultrasonic signal components while suppressing noise. The quadratic correlation operation utilizes the autocorrelation characteristics of the signal to improve peak clarity, making the time delay estimation more accurate. However, although the GQCC algorithm improves noise resistance, it still faces the problem of high computational complexity and may still have estimation errors in strong reverberation environments. Summary of the Invention

[0006] This invention provides a sound source localization method, device, and medium based on an improved generalized cross-correlation algorithm, which can improve the accuracy and robustness of sound source localization in low signal-to-noise ratio and strong reverberation environments.

[0007] This invention provides a sound source localization method based on an improved generalized cross-correlation algorithm, comprising:

[0008] Sound source signals are acquired through a four-element cross sensor array; wherein the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the four sensors are all equidistant from the origin;

[0009] An improved weighting function is constructed based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transformation weighting function and a maximum likelihood weighting function;

[0010] The time delay is obtained by performing a generalized cross-correlation operation based on the sound source signal and the improved weighting function.

[0011] Based on the time delay, geometric calculations are performed in the three-dimensional Cartesian coordinate system to obtain the location of the sound source.

[0012] This invention employs a four-element cross-shaped sensor array to acquire sound source signals. The sensor's position and layout ensure that the sound source's location can be calculated using a certain number of time delays. An improved weighting function is obtained by weighting the signals using a phase transformation weighting function and a maximum likelihood weighting function. This improved weighting function combines the robustness of the phase transformation weighting function in strong reverberation environments with the robustness of the maximum likelihood weighting function in low signal-to-noise ratio environments. Generalized cross-correlation calculations using the improved weighting function suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation. A system of equations is constructed using a certain number of time delays and the geometric relationship between the four-element cross-shaped sensor array and the sound source, allowing for the accurate determination of the sound source location. Compared to existing technologies that are susceptible to interference in low signal-to-noise ratio and strong reverberation environments, this application improves the accuracy and robustness of sound source localization.

[0013] Furthermore, the acquisition of sound source signals via a four-element cross-shaped sensor array includes:

[0014] The four sensors of the four-element cross sensor array each collect a simulated signal from a first sound source; wherein the simulated signal from the first sound source is emitted by the same sound source.

[0015] The four analog signals from the first sound source are amplified to obtain four analog signals from the second sound source.

[0016] The four analog signals from the second sound source are subjected to analog-to-digital conversion to obtain four sound source signals; each sound source signal corresponds to a noise signal.

[0017] The embodiments of the present invention can enhance the quality of the sound source signal and improve the computational efficiency by performing signal amplification and analog-to-digital conversion on the analog signal of the sound source.

[0018] Furthermore, the step of constructing the improved weighting function based on the sound source signal includes:

[0019] Fourier transforms are performed on the four sound source signals and their corresponding noise signals to obtain four frequency domain sound source signals and their corresponding frequency domain noise signals.

[0020] An improved weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal; wherein, an improved weighting function corresponds to any two sensors.

[0021] This invention provides a data foundation for constructing an improved weighting function by performing Fourier transform on the sound source signal and its corresponding noise signal.

[0022] Further, the step of constructing the improved weighting function based on the frequency domain sound source signal and the frequency domain noise signal includes:

[0023] A phase transformation weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0024]

[0025] in, This is the phase transformation weighting function between the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor; Let be the cross-power spectral density function between the frequency domain noise signals of the i-th sensor and the j-th sensor; ρ is the noise parameter factor. Let be the signal coherence function between the i-th sensor and the j-th sensor. Let be the self-power spectrum function of the frequency domain sound source signal of the i-th sensor. Let be the autopower spectrum function of the frequency domain sound source signal of the j-th sensor;

[0026] The maximum likelihood weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0027]

[0028] in, X is the maximum likelihood weighting function between the i-th sensor and the j-th sensor; i (ω) represents the frequency domain sound source signal of the i-th sensor; X j (ω) represents the frequency domain sound source signal of the j-th sensor; N i (ω) represents the frequency domain noise signal of the i-th sensor; N j (ω) represents the frequency domain noise signal of the j-th sensor;

[0029] The phase transform weighting function and the maximum likelihood weighting function are weighted and summed to obtain the improved weighting function, specifically:

[0030]

[0031] in, This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; is the weighting factor, Q is the directional factor, R is the spatial constant, and D is the prior distance.

[0032] The embodiments of the present invention obtain an improved weighting function by weighting with a phase transformation weighting function and a maximum likelihood weighting function, so that the improved weighting function can combine the good robustness of the phase transformation weighting function in a strong reverberation environment and the good robustness of the maximum likelihood weighting function in a low signal-to-noise ratio environment.

[0033] Further, the step of performing a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay includes:

[0034] A generalized cross-correlation function is constructed based on the sound source signal and the improved weighting function; wherein, a generalized cross-correlation function corresponds to any two sensors;

[0035] The time delay is obtained by performing absolute value calculation and peak detection on the generalized cross-correlation function, specifically as follows:

[0036]

[0037] Where, τ ij The time delay for acquiring sound source signals between the i-th sensor and the j-th sensor; The operation to find the maximum value of the objective function corresponding to τ; R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor.

[0038] The embodiments of the present invention perform generalized cross-correlation calculation by improving the weighting function, which can suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation.

[0039] Furthermore, the step of constructing a generalized cross-correlation function based on the sound source signal and the improved weighting function includes:

[0040] In the frequency domain, the sound source signal is weighted using the improved weighting function;

[0041] Performing an inverse Fourier transform on the weighted result yields the generalized cross-correlation function, specifically:

[0042]

[0043] Among them, R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor; This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor.

[0044] This invention provides a data foundation for subsequent time delay calculation by constructing a generalized cross-correlation function.

[0045] Further, the step of performing geometric calculations based on the time delay in the three-dimensional Cartesian coordinate system to obtain the sound source location includes:

[0046] Based on the aforementioned time delay, a system of equations is constructed in the three-dimensional Cartesian coordinate system, specifically as follows:

[0047]

[0048] Where (x, y, z) are the coordinates of the sound source; R is the distance from the sound source to the origin of the three-dimensional rectangular coordinate system; r1, r2, r3, r4 are the distances from the sound source to the four sensors, respectively; d is the distance from any sensor to the origin; Δτ 12 ,Δτ 13 ,Δτ 14 denoted as , where represents the time delay between any sensor and the other three sensors in acquiring the sound source signal; c represents the speed of sound in air.

[0049] Solving the system of equations yields the location of the sound source, specifically:

[0050]

[0051] The embodiments of the present invention construct a set of equations by using a certain number of time delays and the geometric relationship between the four-element cross sensor array and the sound source, which can solve for the accurate location of the sound source.

[0052] Another embodiment of the present invention provides a sound source localization device based on an improved generalized cross-correlation algorithm, comprising: a signal acquisition module, a weighting function module, a generalized cross-correlation module, and a geometric operation module;

[0053] The signal acquisition module is used to acquire sound source signals through a four-element cross sensor array; wherein, the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the four sensors are all equidistant from the origin;

[0054] The weighting function module is used to construct an improved weighting function based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transform weighting function and a maximum likelihood weighting function;

[0055] The generalized cross-correlation module is used to perform a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay;

[0056] The geometric calculation module is used to perform geometric calculations in the three-dimensional rectangular coordinate system based on the time delay to obtain the location of the sound source.

[0057] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the steps of a sound source localization method based on an improved generalized cross-correlation algorithm as described in the present invention.

[0058] Another embodiment of the present invention also provides a computer-readable storage medium item, including: a stored computer program, which, when the computer program is running, controls the device where the computer-readable storage medium is located to perform the steps of a sound source localization method based on an improved generalized cross-correlation algorithm as described in the present invention. Attached Figure Description

[0059] Figure 1 This is a flowchart illustrating an embodiment of the sound source localization method based on the improved generalized cross-correlation algorithm provided by the present invention.

[0060] Figure 2 A schematic diagram of the structure of one embodiment of the four-element cross sensor array provided by the present invention;

[0061] Figure 3 A flowchart illustrating one embodiment of the generalized cross-correlation algorithm provided by the present invention;

[0062] Figure 4 This is a flowchart illustrating an embodiment of the sound source localization system based on the improved generalized cross-correlation algorithm provided by the present invention.

[0063] Figure 5 This is a schematic diagram of an embodiment of the sound source localization device based on the improved generalized cross-correlation algorithm provided by the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0066] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0067] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0068] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0069] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0070] See Figure 1 To address the issue that existing technologies are susceptible to interference in low signal-to-noise ratio and strong reverberation environments, an embodiment of the present invention provides a sound source localization method based on an improved generalized cross-correlation algorithm, comprising steps S101 to S104:

[0071] Step S101: Acquire sound source signals through a four-element cross sensor array; wherein, the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the distances of the four sensors from the origin are all equal.

[0072] Specifically, the sensor can be an ultrasonic sensor, capable of converting the acquired sound signal into an analog electrical signal. Time delay refers to the time difference between the signal's propagation from the sound source to different sensors. To calculate the coordinates of the sound source in three-dimensional space, three different time delays are needed to construct a system of equations; therefore, the number of sensors cannot be less than four. In one embodiment of the invention, a quadrilateral four-element cross sensor array is used to acquire the sound source signal. The structure of the four-element cross sensor array is as follows... Figure 2As shown. The four-element cross sensor array includes four sensors (M1, M2, M3, M4) located in the xy coordinate plane of a three-dimensional rectangular coordinate system. M1 is located on the positive x-axis, M3 is located on the negative x-axis, M4 is located on the positive y-axis, and M2 is located on the negative y-axis. The distance from each of the four sensors to the origin O is d.

[0073] Step S102: Construct an improved weighting function based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transformation weighting function and a maximum likelihood weighting function.

[0074] Specifically, a generalized cross-correlation algorithm can be used to calculate the cross-correlation function of two signals, and the time difference corresponding to the peak value can be extracted to determine the time delay. The weighting function can significantly improve the clarity of the cross-correlation peak by adjusting the frequency domain characteristics of the cross-correlation function, thereby improving the accuracy of sound source localization. Phase transform weighting function and maximum likelihood weighting function are two weighting functions used in the generalized cross-correlation algorithm. The phase transform weighting function has good robustness to reverberation, while the maximum likelihood weighting function has good robustness to environmental noise. By jointly weighting the phase transform weighting function and the maximum likelihood weighting function, an improved weighting function is obtained, which simultaneously exhibits good robustness to both environmental noise and reverberation.

[0075] Step S103: Perform a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay.

[0076] Specifically, the process of solving for the time delay through generalized cross-correlation is as follows: Figure 3 As shown: First, the cross-power spectrum function of the received signal is obtained by Fourier transform of the sound source signals collected by the two sensors; second, the cross-power spectrum function is weighted by an improved weighting function to suppress the influence of reverberation and environmental noise on the accuracy of time delay estimation; then, the weighted cross-power spectrum function is subjected to inverse Fourier transform to obtain the generalized cross-correlation function; finally, the peak value of the absolute value of the cross-correlation function is detected, and the peak value is used as the time delay of the sound source signals collected by the two sensors.

[0077] Step S104: Perform geometric calculations in the three-dimensional Cartesian coordinate system based on the time delay to obtain the sound source location.

[0078] Specifically, there are eight unknowns: the three-dimensional coordinates (x, y, z) of the sound source, the distances from the sound source to the origin of the three-dimensional Cartesian coordinate system, and the distances to the four sensors. A system of eight equations can be constructed using the geometric relationship between the sound source and the four sensors, as well as three different time delays, to solve for these unknowns. In one embodiment of the invention, the geometric relationship between the sound source and the four sensors is as follows: Figure 2As shown, the three-dimensional coordinates of the sound source S are (x, y, z); based on the geometric relationship between the sound source S and the origin of the three-dimensional rectangular coordinate system, the equation x can be established. 2 +y 2 +z 2 =R 2 Where R is the distance from the sound source S to the origin of the three-dimensional rectangular coordinate system; based on the geometric relationship between the sound source S and the four sensors, equations (xd) can be established respectively. 2 +y 2 +z 2 =r1 2 x 2 +(yd) 2 +z 2 =r2 2 (x+d) 2 +y 2 +z 2 =r3 2 and x 2 +(y+d) 2 +z 2 =r4 2 Where r1, r2, r3, and r4 are the distances from the sound source S to the four sensors, respectively; based on the three different time delays, the equation r2-r1=Δτ can be constructed respectively. 12 c, r3-r1=Δτ 13 c and r4-r1=Δτ 14 c, where Δτ 12 ,Δτ 13 ,Δτ 14 Let be the time delay between any sensor and the other three sensors in acquiring the sound source signal, and let c be the speed of sound in the air. By constructing and solving the system of equations based on the above eight equations, the three-dimensional coordinates (x, y, z) of the sound source S can be obtained.

[0079] This invention employs a four-element cross-shaped sensor array to acquire sound source signals. The sensor's position and layout ensure that the sound source's location can be calculated using a certain number of time delays. An improved weighting function is obtained by weighting the signals using a phase transformation weighting function and a maximum likelihood weighting function. This improved weighting function combines the robustness of the phase transformation weighting function in strong reverberation environments with the robustness of the maximum likelihood weighting function in low signal-to-noise ratio environments. Generalized cross-correlation calculations using the improved weighting function suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation. A system of equations is constructed using a certain number of time delays and the geometric relationship between the four-element cross-shaped sensor array and the sound source, allowing for the accurate determination of the sound source location. Compared to existing technologies that are susceptible to interference in low signal-to-noise ratio and strong reverberation environments, this application improves the accuracy and robustness of sound source localization.

[0080] To enhance the quality of the analog sound source signal and improve computational efficiency, it is necessary to amplify and convert the analog sound source signal. Optionally, in this embodiment of the invention, the acquisition of the sound source signal through a four-element cross sensor array includes:

[0081] The four sensors of the four-element cross sensor array each collect a simulated signal from a first sound source; wherein the simulated signal from the first sound source is emitted by the same sound source.

[0082] The four analog signals from the first sound source are amplified to obtain four analog signals from the second sound source.

[0083] The four analog signals from the second sound source are subjected to analog-to-digital conversion to obtain four sound source signals; each sound source signal corresponds to a noise signal.

[0084] Specifically, the sound source signal acquired by the sensor is usually a weak analog signal, which needs to be amplified by the signal amplification section to ensure that its amplitude meets the analog-to-digital conversion range of the subsequent signal acquisition section. Specifically:

[0085] x i ′ (t)=Gx i (t);

[0086] Where, x i ′ (t) represents the amplified analog signal from the sound source; G is the amplification gain; x i (t) represents the simulated signal from the sound source.

[0087] The amplified analog signal from the sound source is sent to an analog-to-digital converter (ADC) and discretized at the sampling rate to obtain the digital signal from the sound source, specifically:

[0088] x i [n] = x i (nT s );

[0089] Where, x i [n] represents the digital signal from the sound source; n is the number of sampling periods; T s =1 / f s The sampling period.

[0090] The embodiments of the present invention can enhance the quality of the sound source signal and improve the computational efficiency by performing signal amplification and analog-to-digital conversion on the analog signal of the sound source.

[0091] To run the generalized cross-correlation algorithm, a weighting function needs to be constructed in advance. Optionally, in this embodiment of the invention, constructing an improved weighting function based on the sound source signal includes:

[0092] Fourier transforms are performed on the four sound source signals and their corresponding noise signals to obtain four frequency domain sound source signals and their corresponding frequency domain noise signals.

[0093] An improved weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal; wherein, an improved weighting function corresponds to any two sensors.

[0094] Specifically, since the generalized cross-correlation algorithm introduces a weighting function in the frequency domain to suppress the effects of reverberation and environmental noise, the sound source signal and its corresponding noise signal in the time domain need to be Fourier transformed before the improved weighting function can be constructed.

[0095] This invention provides a data foundation for constructing an improved weighting function by performing Fourier transform on the sound source signal and its corresponding noise signal.

[0096] To simultaneously suppress the effects of reverberation and environmental noise, it is necessary to combine a phase transformation weighting function and a maximum likelihood weighting function. Optionally, in this embodiment of the invention, constructing an improved weighting function based on the frequency domain sound source signal and the frequency domain noise signal includes:

[0097] A phase transformation weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0098]

[0099] in, This is the phase transformation weighting function between the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor; Let be the cross-power spectral density function between the frequency domain noise signals of the i-th sensor and the j-th sensor; ρ is the noise parameter factor. Let be the signal coherence function between the i-th sensor and the j-th sensor. Let be the self-power spectrum function of the frequency domain sound source signal of the i-th sensor. Let be the autopower spectrum function of the frequency domain sound source signal of the j-th sensor;

[0100] The maximum likelihood weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0101]

[0102] in, X is the maximum likelihood weighting function between the i-th sensor and the j-th sensor; i (ω) represents the frequency domain sound source signal of the i-th sensor; X j (ω) represents the frequency domain sound source signal of the j-th sensor; N i (ω) represents the frequency domain noise signal of the i-th sensor; N j (ω) represents the frequency domain noise signal of the j-th sensor;

[0103] The phase transform weighting function and the maximum likelihood weighting function are weighted and summed to obtain the improved weighting function, specifically:

[0104]

[0105] in, This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; is the weighting factor, Q is the directional factor, R is the spatial constant, and D is the prior distance.

[0106] Specifically, in real-world environments, the reflection from walls causes the two noise signals received by the two sensors to be correlated. Therefore, the phase transformation weighting function described in this embodiment is the phase transformation weighting function after subtracting the correlated noise component. The robustness of the phase transformation weighting function to reverberation and noise can be further optimized using the maximum likelihood weighting function, thus resolving the noise total energy formula |N T (ω)| 2 =q|X(ω)| 2 +(1-q)|N(ω)| 2 Substituting into the maximum likelihood weighting function, we obtain the weighting function when noise and reverberation coexist, specifically:

[0107]

[0108] Since all sensors are identical and very close together in the same space, we can assume that q = q i =q j Simplifying the above equation yields the improved weighted function. The weighting factor q is used to balance the influence between the two weighting functions. When the ambient noise is large, the improved weighting function is close to the maximum likelihood weighting function. When the reverberation noise is large, the improved weighting function is close to the phase transformation weighting function.

[0109] The embodiments of the present invention obtain an improved weighting function by weighting with a phase transformation weighting function and a maximum likelihood weighting function, so that the improved weighting function can combine the good robustness of the phase transformation weighting function in a strong reverberation environment and the good robustness of the maximum likelihood weighting function in a low signal-to-noise ratio environment.

[0110] To obtain the time delay estimate, a generalized cross-correlation function needs to be constructed. Optionally, in this embodiment of the invention, the step of performing a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay includes:

[0111] A generalized cross-correlation function is constructed based on the sound source signal and the improved weighting function; wherein, a generalized cross-correlation function corresponds to any two sensors;

[0112] The time delay is obtained by performing absolute value calculation and peak detection on the generalized cross-correlation function, specifically as follows:

[0113]

[0114] Where, τ ij The time delay for acquiring sound source signals between the i-th sensor and the j-th sensor; The operation to find the maximum value of the objective function corresponding to τ; R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor.

[0115] The embodiments of the present invention perform generalized cross-correlation calculation by improving the weighting function, which can suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation.

[0116] To construct a generalized cross-correlation function, the cross-power spectrum functions of the two sound source signals need to be weighted in the frequency domain. Optionally, in this embodiment of the invention, constructing the generalized cross-correlation function based on the sound source signals and the improved weighting function includes:

[0117] In the frequency domain, the sound source signal is weighted using the improved weighting function;

[0118] Performing an inverse Fourier transform on the weighted result yields the generalized cross-correlation function, specifically:

[0119]

[0120] Among them, R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor; This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor.

[0121] This invention provides a data foundation for subsequent time delay calculation by constructing a generalized cross-correlation function.

[0122] To obtain the location of the sound source, geometric calculations based on time delay are required. Optionally, in this embodiment of the invention, the geometric calculations based on the time delay in the three-dimensional Cartesian coordinate system to obtain the sound source location include:

[0123] Based on the aforementioned time delay, a system of equations is constructed in the three-dimensional Cartesian coordinate system, specifically as follows:

[0124]

[0125] Where (x, y, z) are the coordinates of the sound source; R is the distance from the sound source to the origin of the three-dimensional rectangular coordinate system; r1, r2, r3, r4 are the distances from the sound source to the four sensors, respectively; d is the distance from any sensor to the origin; Δτ 12 ,Δτ 13 ,Δτ 14 denoted as , where represents the time delay between any sensor and the other three sensors in acquiring the sound source signal; c represents the speed of sound in air.

[0126] Solving the system of equations yields the location of the sound source, specifically:

[0127]

[0128] The embodiments of the present invention construct a set of equations by using a certain number of time delays and the geometric relationship between the four-element cross sensor array and the sound source, which can solve for the accurate location of the sound source.

[0129] As a preferred embodiment, this invention can be used to locate hydrogen leaks. After obtaining the hydrogen leak location data, it can be uploaded to the cloud via IoT technology for long-term trend analysis. Early warnings can be issued based on the distance to the leak point, its location area, and leak trend, combined with a set safety radius R. safe Alarms are classified into different levels based on critical thresholds, as detailed in Table 1.

[0130] Table 1 - Hydrogen Leakage Warning Level Table

[0131]

[0132] like Figure 4 As shown, based on the above method embodiments, a system embodiment for locating hydrogen leaks is provided, including steps S1 to S4:

[0133] S1, collect the simulated sound source signal of the hydrogen leak sound source area through the sensor array; this is equivalent to the action of each of the four sensors of the four-element cross sensor array collecting a simulated sound source signal in step S101.

[0134] S2, converting the acquired analog sound source signals into digital sound source signals through signal amplification and analog-to-digital conversion; this is equivalent to performing the actions in step S101 of amplifying the four first analog sound source signals to obtain four second analog sound source signals, and performing analog-to-digital conversion on the four second analog sound source signals to obtain four sound source signals.

[0135] S3, perform improved generalized cross-correlation time delay calculation based on the digital signal of the sound source to obtain the time delay; this is equivalent to executing steps S102 and S103.

[0136] S4, perform geometric calculations based on time delay to locate the hydrogen leak location; this is equivalent to executing step S104.

[0137] S5, set up a multi-level monitoring and alarm system based on the location of the hydrogen leak; this is equivalent to using the sound source location obtained in step S104 to set up the multi-level monitoring and alarm system.

[0138] This invention collects simulated sound source signals using a sensor array. The position and layout of the sensors ensure that the location of the sound source can be calculated using a certain number of time delays. By amplifying and converting the simulated sound source signals to digital, the quality of the sound source signal and the computational efficiency can be improved. By improving the generalized cross-correlation time delay calculation to estimate the time delay, the influence of reverberation and environmental noise on the accuracy of the time delay estimation can be suppressed. By constructing a system of equations using a certain number of time delays and the geometric relationship between the sensor array and the sound source, the accurate location of the hydrogen leak can be obtained. By setting up a multi-level monitoring and alarm system at the location of the hydrogen leak, the threat posed by the hydrogen leak to personnel safety and the environment can be reduced.

[0139] like Figure 5 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided;

[0140] An embodiment of the present invention provides a sound source localization device based on an improved generalized cross-correlation algorithm, comprising: a signal acquisition module 501, a weighting function module 502, a generalized cross-correlation module 503, and a geometric operation module 504;

[0141] The signal acquisition module 501 is used to acquire sound source signals through a four-element cross sensor array; wherein, the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the four sensors are all equidistant from the origin;

[0142] The weighting function module 502 is used to construct an improved weighting function based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transformation weighting function and a maximum likelihood weighting function;

[0143] The generalized cross-correlation module 503 is used to perform a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay;

[0144] The geometric operation module 504 is used to perform geometric operations in the three-dimensional rectangular coordinate system based on the time delay to obtain the sound source location.

[0145] In this embodiment of the invention, the signal acquisition module 501 includes: a signal acquisition submodule, a signal amplification submodule, and an analog-to-digital conversion submodule;

[0146] The signal acquisition submodule is used to acquire a first sound source analog signal through the four sensors of the four-element cross sensor array; wherein the first sound source analog signal is emitted by the same sound source;

[0147] The signal amplification submodule is used to amplify the four first sound source analog signals respectively to obtain four second sound source analog signals.

[0148] The analog-to-digital conversion submodule is used to perform analog-to-digital conversion processing on the four second sound source analog signals respectively to obtain four sound source signals; wherein, each sound source signal corresponds to a noise signal.

[0149] The embodiments of the present invention can enhance the quality of the sound source signal and improve the computational efficiency by performing signal amplification and analog-to-digital conversion on the analog signal of the sound source.

[0150] In this embodiment of the invention, the weighting function module 502 includes: a frequency domain transformation submodule and a weighting function submodule;

[0151] The frequency domain transformation submodule is used to perform Fourier transform on the four sound source signals and their corresponding noise signals respectively to obtain four frequency domain sound source signals and their corresponding frequency domain noise signals.

[0152] The weighting function submodule is used to construct an improved weighting function based on the frequency domain sound source signal and the frequency domain noise signal; wherein, an improved weighting function corresponds to any two sensors.

[0153] This invention provides a data foundation for constructing an improved weighting function by performing Fourier transform on the sound source signal and its corresponding noise signal.

[0154] In this embodiment of the invention, the weighting function submodule includes: a phase transform weighting function unit, a maximum likelihood weighting function unit, and a weighted summation unit;

[0155] The phase transform weighting function unit is used to construct a phase transform weighting function based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0156]

[0157] in, This is the phase transformation weighting function between the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor; Let be the cross-power spectral density function between the frequency domain noise signals of the i-th sensor and the j-th sensor; ρ is the noise parameter factor. Let be the signal coherence function between the i-th sensor and the j-th sensor. Let be the self-power spectrum function of the frequency domain sound source signal of the i-th sensor. Let be the autopower spectrum function of the frequency domain sound source signal of the j-th sensor;

[0158] The maximum likelihood weighting function unit is used to construct a maximum likelihood weighting function based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows:

[0159]

[0160] in, X is the maximum likelihood weighting function between the i-th sensor and the j-th sensor; i (ω) represents the frequency domain sound source signal of the i-th sensor; X j (ω) represents the frequency domain sound source signal of the j-th sensor; N i (ω) represents the frequency domain noise signal of the i-th sensor; N j (ω) represents the frequency domain noise signal of the j-th sensor;

[0161] The weighted summation unit is used to perform a weighted summation of the phase transform weighting function and the maximum likelihood weighting function to obtain an improved weighting function, specifically:

[0162]

[0163] in, This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; is the weighting factor, Q is the directional factor, R is the spatial constant, and D is the prior distance.

[0164] The embodiments of the present invention obtain an improved weighting function by weighting with a phase transformation weighting function and a maximum likelihood weighting function, so that the improved weighting function can combine the good robustness of the phase transformation weighting function in a strong reverberation environment and the good robustness of the maximum likelihood weighting function in a low signal-to-noise ratio environment.

[0165] In this embodiment of the invention, the generalized cross-correlation module 503 includes: a generalized cross-correlation submodule and a time delay estimation submodule;

[0166] The generalized cross-correlation submodule is used to construct a generalized cross-correlation function based on the sound source signal and the improved weighting function; wherein, there is a generalized cross-correlation function between any two sensors;

[0167] The time delay estimation submodule is used to perform absolute value calculation and peak detection on the generalized cross-correlation function to obtain the time delay, specifically as follows:

[0168]

[0169] Where, τ ij The time delay for acquiring sound source signals between the i-th sensor and the j-th sensor; The operation to find the maximum value of the objective function corresponding to τ; R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor.

[0170] The embodiments of the present invention perform generalized cross-correlation calculation by improving the weighting function, which can suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation.

[0171] In this embodiment of the invention, the generalized cross-correlation submodule includes: a weighted processing unit and a generalized cross-correlation unit;

[0172] The weighting processing unit is used to perform weighting processing on the sound source signal in the frequency domain using the improved weighting function;

[0173] The generalized cross-correlation unit is used to perform an inverse Fourier transform on the weighted result to obtain the generalized cross-correlation function, specifically:

[0174]

[0175] Among them, R ij (τ) is the generalized cross-correlation function between the i-th sensor and the j-th sensor; This is the improved weighting function corresponding to the i-th sensor and the j-th sensor; Let be the cross-power spectrum function between the frequency domain sound source signal of the i-th sensor and the frequency domain sound source signal of the j-th sensor.

[0176] This invention provides a data foundation for subsequent time delay calculation by constructing a generalized cross-correlation function.

[0177] In this embodiment of the invention, the geometric operation module 504 includes: a submodule for constructing a system of equations and a submodule for solving a system of equations;

[0178] The equation system construction submodule is used to construct a system of equations in the three-dimensional Cartesian coordinate system based on the time delay, specifically as follows:

[0179]

[0180] Where (x, y, z) are the coordinates of the sound source; R is the distance from the sound source to the origin of the three-dimensional rectangular coordinate system; r1, r2, r3, r4 are the distances from the sound source to the four sensors, respectively; d is the distance from any sensor to the origin; Δτ 12 ,Δτ 13 ,Δτ 14 denoted as , where represents the time delay between any sensor and the other three sensors in acquiring the sound source signal; c represents the speed of sound in air.

[0181] The submodule for solving the system of equations is used to solve the system of equations to obtain the location of the sound source, specifically as follows:

[0182]

[0183] The embodiments of the present invention construct a set of equations by using a certain number of time delays and the geometric relationship between the four-element cross sensor array and the sound source, which can solve for the accurate location of the sound source.

[0184] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the sound source localization method based on the improved generalized cross-correlation algorithm provided by any of the above-described method embodiments of the present invention.

[0185] This invention employs a signal acquisition module to acquire sound source signals. The sensor's position and layout ensure that the sound source's location can be calculated using a certain number of time delays. An improved weighting function is constructed using a weighting function module, combining the robustness of a phase transformation weighting function in strong reverberation environments with the robustness of a maximum likelihood weighting function in low signal-to-noise ratio environments. A generalized cross-correlation module performs generalized cross-correlation calculations to suppress the impact of reverberation and environmental noise on the accuracy of time delay estimation. A geometric calculation module constructs a system of equations to solve for the accurate sound source location. Compared to existing technologies that are susceptible to low signal-to-noise ratio and strong reverberation environments, this application improves the accuracy and robustness of sound source localization.

[0186] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0187] Based on the above embodiment of a sound source localization method based on an improved generalized cross-correlation algorithm, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a sound source localization method based on an improved generalized cross-correlation algorithm according to any embodiment of the present invention.

[0188] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0189] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0190] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0191] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a sound source localization method based on an improved generalized cross-correlation algorithm as described in any of the above-described method embodiments of the present invention.

[0192] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0193] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A sound source localization method based on an improved generalized cross-correlation algorithm, characterized in that, include: Sound source signals are acquired through a four-element cross sensor array; wherein the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the four sensors are all equidistant from the origin; An improved weighting function is constructed based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transformation weighting function and a maximum likelihood weighting function; The time delay is obtained by performing a generalized cross-correlation operation based on the sound source signal and the improved weighting function. Based on the time delay, geometric calculations are performed in the three-dimensional Cartesian coordinate system to obtain the sound source location, including: Based on the aforementioned time delay, a system of equations is constructed in the three-dimensional Cartesian coordinate system, specifically as follows: ； in, These are the coordinates of the sound source; The distance from the sound source to the origin of the three-dimensional rectangular coordinate system; These represent the distances from the sound source to the four sensors; The distance from any sensor to the origin; These represent the time delays between any sensor and the other three sensors in acquiring sound source signals. The speed at which sound travels through the air; Solving the system of equations yields the location of the sound source, specifically: 。 2. The sound source localization method based on the improved generalized cross-correlation algorithm as described in claim 1, characterized in that, The acquisition of sound source signals via a four-element cross-shaped sensor array includes: The four sensors of the four-element cross sensor array each collect a simulated signal from a first sound source; wherein the simulated signal from the first sound source is emitted by the same sound source. The four analog signals from the first sound source are amplified to obtain four analog signals from the second sound source. The four analog signals from the second sound source are subjected to analog-to-digital conversion to obtain four sound source signals; each sound source signal corresponds to a noise signal.

3. The sound source localization method based on the improved generalized cross-correlation algorithm as described in claim 2, characterized in that, The step of constructing the improved weighting function based on the sound source signal includes: Fourier transforms are performed on the four sound source signals and their corresponding noise signals to obtain four frequency domain sound source signals and their corresponding frequency domain noise signals. An improved weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal; wherein, an improved weighting function corresponds to any two sensors.

4. The sound source localization method based on the improved generalized cross-correlation algorithm as described in claim 3, characterized in that, The step of constructing an improved weighting function based on the frequency domain sound source signal and the frequency domain noise signal includes: A phase transformation weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows: ； in, For the first The sensor and the first The phase transformation weighting function corresponding to each sensor; For the first The frequency domain sound source signal of the first sensor and the first The cross-power spectrum function between the frequency domain sound source signals of the sensors; For the first The frequency domain noise signal of the first sensor and the first The cross-power spectrum function between the frequency domain noise signals of the sensors; This is a noise parameter factor; For the first The sensor and the first The signal coherence function of each sensor For the first The self-power spectrum function of the frequency domain sound source signal of each sensor. For the first The auto-power spectrum function of the frequency domain sound source signal of a sensor; The maximum likelihood weighting function is constructed based on the frequency domain sound source signal and the frequency domain noise signal, specifically as follows: ； in, For the first The sensor and the first The maximum likelihood weighting function corresponding to each sensor; For the first Frequency domain sound source signal of each sensor; For the first Frequency domain sound source signal of each sensor; For the first Frequency domain noise signal of each sensor; For the first Frequency domain noise signal of each sensor; The phase transform weighting function and the maximum likelihood weighting function are weighted and summed to obtain the improved weighting function, specifically: in, For the first The sensor and the first The improved weighting function corresponding to each sensor; As a weighting factor, As a directional factor, It is a space constant. This is the prior distance.

5. The sound source localization method based on the improved generalized cross-correlation algorithm as described in claim 4, characterized in that, The step of performing a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay includes: A generalized cross-correlation function is constructed based on the sound source signal and the improved weighting function; wherein, a generalized cross-correlation function corresponds to any two sensors; The time delay is obtained by performing absolute value calculation and peak detection on the generalized cross-correlation function, specifically as follows: ； in, For the first The sensor and the first The time delay between sensors in acquiring sound source signals; To find the maximum value of the objective function The operation; For the first The sensor and the first The generalized cross-correlation function between the sensors.

6. The sound source localization method based on the improved generalized cross-correlation algorithm as described in claim 5, characterized in that, The construction of the generalized cross-correlation function based on the sound source signal and the improved weighting function includes: In the frequency domain, the sound source signal is weighted using the improved weighting function; Performing an inverse Fourier transform on the weighted result yields the generalized cross-correlation function, specifically: ； in, For the first The sensor and the first The generalized cross-correlation function between the sensors; For the first The sensor and the first The improved weighting function corresponding to each sensor; For the first The frequency domain sound source signal of the first sensor and the first The cross-power spectrum function between the frequency domain sound source signals of the sensors.

7. A sound source localization device based on an improved generalized cross-correlation algorithm, characterized in that, include: Signal acquisition module, weighting function module, generalized cross-correlation module, and geometric operation module; The signal acquisition module is used to acquire sound source signals through a four-element cross sensor array; wherein, the four-element cross sensor array consists of four sensors located in the same coordinate axis plane of a three-dimensional rectangular coordinate system, with two sensors on each of the two coordinate axes of the coordinate axis plane and the two sensors located on both sides of the origin of the three-dimensional rectangular coordinate system, and the distances of the four sensors from the origin are all equal. The weighting function module is used to construct an improved weighting function based on the sound source signal; wherein the improved weighting function is obtained by weighted summation of a phase transform weighting function and a maximum likelihood weighting function; The generalized cross-correlation module is used to perform a generalized cross-correlation operation based on the sound source signal and the improved weighting function to obtain the time delay; The geometric calculation module is used to perform geometric calculations in the three-dimensional Cartesian coordinate system based on the time delay to obtain the sound source location, including: Based on the aforementioned time delay, a system of equations is constructed in the three-dimensional Cartesian coordinate system, specifically as follows: ； in, These are the coordinates of the sound source; The distance from the sound source to the origin of the three-dimensional rectangular coordinate system; These represent the distances from the sound source to the four sensors; The distance from any sensor to the origin; These represent the time delays between any sensor and the other three sensors in acquiring sound source signals. The speed at which sound travels through the air; Solving the system of equations yields the location of the sound source, specifically: 。 8. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a sound source localization method based on an improved generalized cross-correlation algorithm as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a sound source localization method based on an improved generalized cross-correlation algorithm as described in any one of claims 1-6.

Images