MIC noise reduction method and system, electronic device and storage medium

Abstract

The present invention discloses a MIC noise reduction method and system, an electronic device and a storage medium, wherein the MIC noise reduction method comprises: using a main MIC to collect a main signal; using a secondary MIC to collect a noise signal; using a filter to filter the noise signal according to a weighted vector to obtain an estimated noise; removing the estimated noise from the main signal to obtain a denoised voice signal; updating the weighted vector according to the denoised voice signal, and returning to the step of filtering the noise signal according to the weighted vector using the filter until the estimated noise is the best estimated noise. In the filtering process, the present invention continuously updates the weighted vector of the filter, which has a significant effect on eliminating the noise of the range hood; in the absence of voice, the weighted vector and convergence coefficient of the range hood at different working gears are obtained by training, which are used as the initial value and convergence coefficient of the weighted vector when the filter is iterated in the actual denoising process, thereby improving the convergence speed of the filter.

Description

Technical Field

[0001] The present invention relates to the technical field of noise reduction of range hoods, and in particular to a MIC noise reduction method and system, electronic equipment and storage medium. Background Art

[0002] Product intelligence is the current trend in the home appliance and kitchen appliance industry, and products with voice recognition functions are more competitive. Convenient operation and humanization have become the two points of the product. With the development of technology, the voice recognition function on the range hood is becoming more and more mature.

[0003] However, in the actual use of the voice recognition function of the range hood, an urgent problem has been encountered: when the range hood is turned on, it will make a lot of noise, which will mask the human voice, thereby reducing the voice recognition rate and causing a decline in user experience. Therefore, how to remove the noise made by the range hood when it is running has become the key to improving the voice recognition rate of the range hood.

[0004] In the prior art, the commonly used noise reduction methods are:

[0005] 1. Use classic filters to reduce noise

[0006] The classical filter includes a high-pass filter and a low-pass filter. The disadvantage of using the classical filter for noise reduction is that it can only filter noise in a specific frequency band.

[0007] 2. Use spectral subtraction to reduce noise

[0008] Spectral subtraction is to subtract the spectrum of the noisy signal from the spectrum of the noisy signal. The disadvantage of using spectral subtraction to reduce noise is that there will be residual noise, which will produce "musical noise" with certain rhythmic fluctuations and sound similar to music.

[0009] 3. Use Wiener filter to reduce noise

[0010] Wiener filtering assumes that the input of the linear filter is the sum of the useful signal and the noise, both of which are wide-sense stationary processes and their second-order statistical characteristics are known. According to the minimum mean square error criterion (the mean square value of the difference between the output signal of the filter and the required signal is the smallest), the parameters of the optimal linear filter are obtained. The disadvantage of the Wiener filter is that it is difficult to meet the condition of obtaining all the observed data in the semi-infinite time interval. At the same time, it cannot be used in the case where the noise is a non-stationary random process, and it is not convenient to apply it to the vector case. Therefore, Wiener filtering is not widely used in practical problems. Summary of the invention

[0011] The technical problem to be solved by the present invention is to overcome the defect in the prior art that it is difficult to remove the noise generated by the range hood during operation, and to provide a MIC noise reduction method and system, an electronic device and a storage medium.

[0012] The present invention solves the above technical problems through the following technical solutions:

[0013] The present invention provides a MIC noise reduction method, the MIC noise reduction method comprising:

[0014] Using a main MIC to collect a main signal, the main signal includes a speech signal and a noise signal;

[0015] Use the secondary MIC to collect noise signals;

[0016] Using a filter to filter the noise signal according to a weighted vector to obtain estimated noise;

[0017] Removing the estimated noise from the main signal to obtain a denoised speech signal;

[0018] The weighting vector is updated according to the denoised speech signal, and the step of using a filter to filter the noise signal according to the weighting vector to obtain an estimated noise is returned until the above operation is performed a preset number of times, the estimated noise obtained is the optimal estimated noise, and the denoised speech signal obtained is the optimal denoised speech signal.

[0019] Preferably, the step of removing the estimated noise from the main signal to obtain a denoised speech signal comprises calculating the denoised speech signal using the following formula:

[0020] Y(n)=S(n)-D′(n)

[0021] Among them, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, Y(n) represents the value of the denoised speech signal when the main MIC is used for the nth collection of the main signal, and D′(n) represents the value of the estimated noise obtained by calculating the noise signal collected when the secondary MIC is used for the nth collection using the filter.

[0022] Preferably, the D′(n) is calculated by the following formula:

[0023]

[0024] Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, and M represents the order of the filter.

[0025] Preferably, the weighted vector is calculated by the following formula:

[0026] w(n)=w(n-1)+2μ*Y(n)*d(n)

[0027] Wherein, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, μ represents the convergence parameter, and

[0028] Preferably, the initial value of the weighted vector and the convergence parameter are determined by the following process:

[0029] Using the main MIC to collect an initial main signal, the initial main signal includes an initial speech signal and an initial noise signal;

[0030] Using the secondary MIC to collect the initial noise signal;

[0031] Set several alternative convergence parameters;

[0032] According to a preset rule, selecting one candidate convergence parameter from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter;

[0033] Using the filter, filtering the initial noise signal according to the initial weighting vector to obtain an initial estimated noise;

[0034] Removing the initial estimated noise from the initial main signal to obtain a denoised initial speech signal;

[0035] updating the initial weighting vector according to the denoised initial speech signal, and returning to the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain an initial estimated noise, until the above operation is performed a preset number of times, the obtained initial estimated noise is an optimal initial estimated noise, the obtained initial weighting vector is an optimal initial weighting vector, the obtained denoised initial speech signal is an optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter;

[0036] Returning to the step of sequentially selecting an alternative convergence parameter from the plurality of alternative convergence parameters as a to-be-determined convergence parameter, until all the alternative convergence parameters are traversed;

[0037] The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

[0038] Preferably, the device to which the MIC noise reduction method is applicable includes a range hood, and the range hood has a plurality of working gears, and different working gears correspond to different values ​​of the plurality of candidate convergence parameters.

[0039] The present invention also provides a MIC noise reduction system, the MIC noise reduction system comprising:

[0040] A main MIC module, used to collect a main signal using the main MIC, wherein the main signal includes a speech signal and a noise signal;

[0041] A secondary MIC module, used for collecting the noise signal using the secondary MIC;

[0042] A filter module, used for filtering the noise signal using a filter according to a weighted vector to obtain estimated noise;

[0043] A denoising module, used for removing the estimated noise from the main signal to obtain a denoised speech signal;

[0044] A weighted vector updating module is used to update the weighted vector according to the denoised speech signal and recall the filter module.

[0045] Preferably, the denoising module calculates the denoised speech signal using the following formula:

[0046] Y(n)=S(n)-D′(n)

[0047] Among them, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, Y(n) represents the value of the denoised speech signal when the main MIC is used for the nth collection of the main signal, and D′(n) represents the value of the estimated noise obtained by calculating the noise signal collected when the secondary MIC is used for the nth collection using the filter.

[0048] Preferably, the filter module calculates D′(n) by the following formula:

[0049]

[0050] Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, and M represents the order of the filter.

[0051] Preferably, the weighted vector updating module calculates the updated weighted vector using the following formula:

[0052] w(n)=w(n-1)+2μ*Y(n)*d(n)

[0053] Wherein, w(n) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the nth time, when n=1, w(0) represents the initial value of the weighted vector, μ represents the convergence parameter, and

[0054] Preferably, the initial value of the weighted vector and the convergence parameter are determined by the following process:

[0055] Using the main MIC to collect an initial main signal, the initial main signal includes an initial speech signal and an initial noise signal;

[0056] Using the secondary MIC to collect the initial noise signal;

[0057] Set several alternative convergence parameters;

[0058] According to a preset rule, selecting one candidate convergence parameter from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter;

[0059] Using the filter, filtering the initial noise signal according to the initial weighting vector to obtain an initial estimated noise;

[0060] Removing the initial estimated noise from the initial main signal to obtain a denoised initial speech signal;

[0061] updating the initial weighting vector according to the denoised initial speech signal, and returning to the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain an initial estimated noise, until the above operation is performed a preset number of times, the obtained initial estimated noise is an optimal initial estimated noise, the obtained initial weighting vector is an optimal initial weighting vector, the obtained denoised initial speech signal is an optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter;

[0062] Returning to the step of sequentially selecting an alternative convergence parameter from the plurality of alternative convergence parameters as a to-be-determined convergence parameter, until all the alternative convergence parameters are traversed;

[0063] The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

[0064] Preferably, the MIC noise reduction system is applicable to a device including a range hood, wherein the range hood has a plurality of working gears, and different working gears correspond to different values ​​of the plurality of candidate convergence parameters.

[0065] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the aforementioned MIC noise reduction method when executing the computer program.

[0066] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the aforementioned MIC noise reduction method is implemented.

[0067] The positive progressive effect of the present invention is that: during the filtering process, the present invention continuously updates the weighted vector of the filter, and returns to the step of using the filter to filter the noise signal according to the weighted vector to obtain the estimated noise, until the estimated noise is the optimal estimated noise, and finally the optimal estimated noise is removed from the main signal, which has a significant effect on eliminating the noise of the range hood; in the absence of speech, this scheme trains the filter weighted vector and convergence coefficient of the range hood at different working gears, which are used as the initial value and convergence coefficient of the weighted vector of the first step when the filter performs standard iteration in the actual noise reduction process, thereby improving the convergence speed of the filter, improving the practicability of the filter, improving the recognition rate, and greatly enhancing the intelligent experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a flow chart of the MIC noise reduction method provided in Example 1 of the present invention.

[0069] Figure 2 A schematic diagram of the modules of the MIC noise reduction system provided in Embodiment 2 of the present invention.

[0070] Figure 3 This is a schematic diagram of the structure of an electronic device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0071] The present invention is further described below by way of examples, but the present invention is not limited to the scope of the examples.

[0072] Example 1

[0073] This embodiment provides a MIC noise reduction method, such as Figure 1 As shown, the MIC noise reduction method includes the following steps:

[0074] S11. Use a main MIC to collect a main signal, where the main signal includes a speech signal and a noise signal.

[0075] S12: Use the secondary MIC to collect the noise signal.

[0076] Specifically, in this embodiment, the secondary MIC can be expanded to multiple secondary MICs, and multiple secondary MICs are used to collect multiple independent noise signals, which is applied to the scenario where multiple independent noises need to be eliminated.

[0077] S13: Using a filter to filter the noise signal according to the weighting vector to obtain estimated noise.

[0078] Specifically, in S13, the estimated noise can be calculated by the following formula:

[0079]

[0080] Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, and M represents the order of the filter.

[0081] S14. Removing the estimated noise from the main signal to obtain a denoised speech signal.

[0082] Specifically, in S14, the denoised speech signal can be calculated using the following formula:

[0083] Y(n)=S(n)-D′(n)

[0084] Among them, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, Y(n) represents the value of the denoised speech signal when the main MIC is used for the nth collection of the main signal, and D′(n) represents the value of the estimated noise obtained by calculating the noise signal collected when the secondary MIC is used for the nth collection using the filter.

[0085] S15. Update the weighting vector according to the denoised speech signal, and return to the step of using a filter to filter the noise signal according to the weighting vector to obtain an estimated noise, until the above operation is performed a preset number of times, the obtained estimated noise is the optimal estimated noise, and the obtained denoised speech signal is the optimal denoised speech signal.

[0086] Specifically, in this embodiment, the best estimation noise is finally removed from the main signal.

[0087] Specifically, in S15, the weighted vector is calculated by the following formula:

[0088] w(n)=w(n-1)+2μ*Y(n)*d(n)

[0089] Wherein, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, μ represents the convergence parameter, and

[0090] Specifically, the initial value of the weighted vector and the convergence parameter can be determined by the following process:

[0091] The main MIC is used to collect an initial main signal, where the initial main signal includes an initial speech signal and an initial noise signal.

[0092] The secondary MIC is used to collect the initial noise signal.

[0093] Sets several alternative convergence parameters.

[0094] According to a preset rule, one candidate convergence parameter is selected from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter.

[0095] The filter is used to filter the initial noise signal according to the initial weighting vector to obtain the initial estimated noise.

[0096] Specifically, the initial estimated noise can be calculated by the following formula:

[0097]

[0098] Specifically, the above formula only involves numerical operations and does not involve units, wherein d′(nm) represents the value of the initial noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d′(nm) is 0, w′(n-1) represents the initial weighting vector of the filter when the secondary MIC is used to collect the initial noise signal for the n-1th time, when n=1, w′(0)=0, and M represents the order of the filter.

[0099] The initial estimated noise is removed from the initial main signal to obtain a denoised initial speech signal.

[0100] Specifically, the denoised initial speech signal can be calculated using the following formula:

[0101] Y′(n)=S′(n)-D″(n)

[0102] Specifically, the above formula only involves numerical operations and does not involve units, wherein S′(n) represents the value of the initial main signal collected when the main MIC is used for the nth collection, Y′(n) represents the value of the denoised initial speech signal when the main MIC is used to collect the initial main signal for the nth time, and D″(n) represents the value of the initial estimated noise obtained by calculating the initial noise signal collected when the secondary MIC is used for the nth collection using the filter.

[0103] The initial weighting vector is updated according to the denoised initial speech signal, and the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain the initial estimated noise is returned, until the above operation is performed a preset number of times, the obtained initial estimated noise is the optimal initial estimated noise, the obtained initial weighting vector is the optimal initial weighting vector, the obtained denoised initial speech signal is the optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter.

[0104] Specifically, the initial weight vector is calculated by the following formula:

[0105] w′(n)=w″(n-1)+2μ′*Y′(n)*d′(n)

[0106] Specifically, the above formula only involves numerical operations and does not involve units, wherein w′(n-1) represents the initial weight vector of the filter when the secondary MIC is used to collect the initial noise signal for the n-1th time, when n=1, w′(0)=0, μ′ represents a to-be-determined convergence parameter, and

[0107] Return to the step of selecting one candidate convergence parameter from the plurality of candidate convergence parameters in turn as the to-be-determined convergence parameter according to the preset rule, until all candidate convergence parameters are traversed.

[0108] The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

[0109] Specifically, in this embodiment, the MIC noise reduction method is applicable to equipment with stable strong noise, and the equipment with stable strong noise includes a range hood, and the range hood has multiple working gears, and different working gears correspond to different values ​​of the several alternative convergence parameters.

[0110] The MIC noise reduction method of this embodiment continuously updates the weighted vector of the filter, and returns to the step of using the filter to filter the noise signal according to the weighted vector to obtain the estimated noise, until the above operation is performed a preset number of times, the estimated noise obtained is the optimal estimated noise, and finally the optimal estimated noise is removed from the main signal, and the denoised speech signal obtained is the optimal denoised speech signal, which has a significant effect on eliminating the noise of the range hood; in the absence of speech, this embodiment trains to obtain the filter weighted vector and convergence coefficient of the range hood at different working gears, which are used as the initial value and convergence coefficient of the weighted vector of the first step when the filter performs standard iteration in the actual noise reduction process, thereby improving the convergence speed of the filter, improving the practicability of the filter, improving the recognition rate, and greatly enhancing the intelligent experience.

[0111] Example 2

[0112] This embodiment provides a MIC noise reduction system, such as Figure 2 As shown, the MIC noise reduction system includes: a main MIC module 1, a secondary MIC module 2, a filter module 3, a denoising module 4, and a weighted vector updating module 5.

[0113] The main MIC module 1 is used to collect a main signal using the main MIC, where the main signal includes a speech signal and a noise signal.

[0114] The secondary MIC module 2 is used to collect the noise signal using the secondary MIC.

[0115] Specifically, in this embodiment, the secondary MIC module 2 can use multiple secondary MICs to collect multiple independent noise signals, which is applied to the scenario where multiple independent noises need to be eliminated.

[0116] The filter module 3 is used to use a filter to filter the noise signal according to the weighting vector to obtain estimated noise.

[0117] Specifically, the filter module 3 calculates D′(n) by the following formula:

[0118]

[0119] Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, and M represents the order of the filter.

[0120] The denoising module 4 is used to remove the estimated noise from the main signal to obtain a denoised speech signal.

[0121] Specifically, the denoising module 4 calculates the denoised speech signal using the following formula:

[0122] Y(n)=S(n)-D′(n)

[0123] Among them, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, Y(n) represents the value of the denoised speech signal when the main MIC is used for the nth collection of the main signal, and D′(n) represents the value of the estimated noise obtained by calculating the noise signal collected when the secondary MIC is used for the nth collection using the filter.

[0124] The weighted vector updating module 5 updates the weighted vector according to the denoised speech signal and calls the filter module 3 again until the estimated noise is the optimal estimated noise. Finally, the optimal estimated noise is removed from the main signal, and the denoised speech signal obtained is the optimal denoised speech signal.

[0125] Specifically, the weighted vector updating module 5 calculates the updated weighted vector using the following formula:

[0126] w(n)=w(n-1)+2μ*Y(n)*d(n)

[0127] Wherein, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighted vector, μ represents the convergence parameter, and

[0128] Specifically, the initial value of the weighted vector and the convergence parameter are determined by the following process:

[0129] The secondary MIC is used to collect the initial noise signal.

[0130] Sets several alternative convergence parameters.

[0131] According to a preset rule, one candidate convergence parameter is selected from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter.

[0132] The filter is used to filter the initial noise signal according to the initial weighting vector to obtain the initial estimated noise.

[0133] The initial estimated noise is removed from the initial main signal to obtain a denoised initial speech signal.

[0134] The initial weighting vector is updated according to the denoised initial speech signal, and the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain the initial estimated noise is returned, until the above operation is performed a preset number of times, the obtained initial estimated noise is the optimal initial estimated noise, the obtained initial weighting vector is the optimal initial weighting vector, the obtained denoised initial speech signal is the optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter.

[0135] Return to the step of sequentially selecting an alternative convergence parameter from the plurality of alternative convergence parameters as a to-be-determined convergence parameter until all the alternative convergence parameters are traversed.

[0136] The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

[0137] Specifically, in this embodiment, the MIC noise reduction method is applicable to equipment with stable strong noise, and the equipment with stable strong noise includes a range hood, and the range hood has multiple working gears, and different working gears correspond to different values ​​of the several alternative convergence parameters.

[0138] The MIC noise reduction system of this embodiment continuously updates the weighted vector of the filter by calling the weighted vector update module 5, and re-calls the filter module 3 until the estimated noise is the optimal estimated noise. Finally, the optimal estimated noise is removed from the main signal, which has a significant effect on eliminating the noise of the range hood. In the absence of speech, this embodiment trains the filter weighted vector and convergence coefficient of the range hood at different working gears, which are used as the initial value and convergence coefficient of the weighted vector of the first step of standard iteration in the actual noise reduction process of the filter, thereby improving the convergence speed of the filter, improving the practicability of the filter, improving the recognition rate, and greatly enhancing the intelligent experience.

[0139] Example 3

[0140] Figure 3 This is a structural diagram of an electronic device provided in Embodiment 3 of the present invention, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the MIC noise reduction method of Embodiment 1 is implemented when the processor executes the computer program. Figure 3 The electronic device 30 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0141] The electronic device 30 may be in the form of a general-purpose computing device, for example, it may be a server device. The components of the electronic device 30 may include, but are not limited to: at least one processor 31, at least one memory 32, and a bus 33 connecting different system components (including the memory 32 and the processor 31).

[0142] The bus 33 includes a data bus, an address bus, and a control bus.

[0143] The memory 32 may include a volatile memory, such as a random access memory (RAM) 321 and / or a cache memory 322 , and may further include a read-only memory (ROM) 323 .

[0144] The memory 32 may also include a program / utility 325 having a set (at least one) of program modules 324, such program modules 324 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0145] The processor 31 executes various functional applications and data processing by running the computer programs stored in the memory 32, such as the MIC noise reduction method of embodiment 1 of the present invention.

[0146] The electronic device 30 may also communicate with one or more external devices 34 (e.g., keyboards, pointing devices, etc.). Such communication may be performed via an input / output (I / O) interface 35. Furthermore, the model generating device 30 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 36. Figure 3 As shown, the network adapter 36 communicates with other modules of the model generation device 30 via the bus 33. It should be understood that although Figure 3 Not shown, other hardware and / or software modules may be used in conjunction with the model generated device 30, including but not limited to microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0147] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.

[0148] Example 4

[0149] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the MIC noise reduction method of the aforementioned embodiment 1 is implemented.

[0150] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0151] In a possible implementation manner, the present invention may also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the MIC noise reduction method of embodiment 1.

[0152] The program code for executing the present invention may be written in any combination of one or more programming languages, and may be executed entirely on a user device, partially on a user device, as an independent software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0153] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that this is only for illustration and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.

[0154] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that this is only for illustration and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.

Claims

1. A MIC noise reduction method, characterized in that: The MIC noise reduction method comprises: Using a main MIC to collect a main signal, the main signal includes a speech signal and a noise signal; Use the secondary MIC to collect noise signals; Using a filter to filter the noise signal according to a weighted vector to obtain estimated noise; The estimated noise is removed from the main signal to obtain a denoised speech signal using the following formula: Wherein, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, and Y(n) represents the value of the denoised speech signal when the main signal is collected for the nth time using the main MIC. represents the value of the estimated noise obtained by calculating the noise signal collected by the secondary MIC when the filter is used to collect the noise for the nth time; updating the weighting vector according to the denoised speech signal, and returning to the step of filtering the noise signal using a filter according to the weighting vector to obtain an estimated noise, until the above operation is performed a preset number of times, the obtained estimated noise is the best estimated noise, and the obtained denoised speech signal is the best denoised speech signal; Said Calculated by the following formula: Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighting vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighting vector, and M represents the order of the filter; The weighted vector is calculated by the following formula: Wherein, w(n-1) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time. When n=1, w(0) represents the initial value of the weighted vector. represents the convergence parameter, and ; The initial value of the weighted vector and the convergence parameter are determined by the following process: Using the main MIC to collect an initial main signal, the initial main signal includes an initial speech signal and an initial noise signal; Using the secondary MIC to collect the initial noise signal; Set several alternative convergence parameters; According to a preset rule, selecting one candidate convergence parameter from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter; Using the filter, filtering the initial noise signal according to the initial weighting vector to obtain an initial estimated noise; Removing the initial estimated noise from the initial main signal to obtain a denoised initial speech signal; updating the initial weighting vector according to the denoised initial speech signal, and returning to the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain an initial estimated noise, until the above operation is performed a preset number of times, the obtained initial estimated noise is an optimal initial estimated noise, the obtained initial weighting vector is an optimal initial weighting vector, the obtained denoised initial speech signal is an optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter; Returning to the step of sequentially selecting an alternative convergence parameter from the plurality of alternative convergence parameters as a to-be-determined convergence parameter, until all the alternative convergence parameters are traversed; The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

2. The MIC noise reduction method according to claim 1, characterized in that: The MIC noise reduction method is applicable to a device including a range hood, wherein the range hood has a plurality of working gears, and different working gears correspond to different values ​​of the plurality of candidate convergence parameters.

3. A MIC noise reduction system, characterized in that: The MIC noise reduction system comprises: A main MIC module, used to collect a main signal using the main MIC, wherein the main signal includes a speech signal and a noise signal; A secondary MIC module, used for collecting the noise signal using the secondary MIC; A filter module, used for filtering the noise signal using a filter according to a weighted vector to obtain estimated noise; The denoising module is used to remove the estimated noise from the main signal to obtain a denoised speech signal using the following formula: Wherein, S(n) represents the value of the main signal collected when the main MIC is used for the nth collection, and Y(n) represents the value of the denoised speech signal when the main signal is collected for the nth time using the main MIC. (n) represents the value of the estimated noise obtained by calculating the noise signal collected by the secondary MIC when the filter is used to collect the noise for the nth time; A weighted vector updating module, used for updating the weighted vector according to the denoised speech signal and re-calling the filter module; The filter module is calculated by the following formula (n): Wherein, d(nm) represents the value of the noise signal collected when the secondary MIC is used for the nmth collection, when n= or <m, the value of d(nm) is 0, w(n-1) represents the weighting vector of the filter when the secondary MIC is used to collect the noise signal for the n-1th time, when n=1, w(0) represents the initial value of the weighting vector, and M represents the order of the filter; The weighted vector updating module calculates the updated weighted vector using the following formula: Wherein, w(n) represents the weighted vector of the filter when the secondary MIC is used to collect the noise signal for the nth time, and when n=1, w(0) represents the initial value of the weighted vector. represents the convergence parameter, and ; The initial value of the weighted vector and the convergence parameter are determined by the following process: Using the main MIC to collect an initial main signal, the initial main signal includes an initial speech signal and an initial noise signal; Using the secondary MIC to collect the initial noise signal; Set several alternative convergence parameters; According to a preset rule, selecting one candidate convergence parameter from the plurality of candidate convergence parameters in turn as a to-be-determined convergence parameter; Using the filter, filtering the initial noise signal according to the initial weighting vector to obtain an initial estimated noise; Removing the initial estimated noise from the initial main signal to obtain a denoised initial speech signal; updating the initial weighting vector according to the denoised initial speech signal, and returning to the step of using the filter to filter the initial noise signal according to the initial weighting vector to obtain an initial estimated noise, until the above operation is performed a preset number of times, the obtained initial estimated noise is an optimal initial estimated noise, the obtained initial weighting vector is an optimal initial weighting vector, the obtained denoised initial speech signal is an optimal denoised initial speech signal, and the optimal denoised initial speech signal corresponds to the undetermined convergence parameter; Returning to the step of sequentially selecting an alternative convergence parameter from the plurality of alternative convergence parameters as a to-be-determined convergence parameter, until all the alternative convergence parameters are traversed; The values ​​of all the best denoised initial speech signals obtained are compared, and the undetermined convergence parameter corresponding to the best denoised initial speech signal with the largest value is used as the convergence parameter, and the value of the best initial weighting vector corresponding to the best denoised initial speech signal with the largest value is used as the initial value of the weighting vector.

4. The MIC noise reduction system according to claim 3, characterized in that: The MIC noise reduction system is applicable to equipment including a range hood, wherein the range hood has a plurality of working gears, and different working gears correspond to different values ​​of the plurality of candidate convergence parameters.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the MIC noise reduction method according to any one of claims 1 to 2 is implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the MIC noise reduction method according to any one of claims 1 to 2 is implemented.

Images