A method, device, system and storage medium for controlling a smoke machine.

Abstract

This invention relates to the field of tobacco appliance control technology, and in particular to a method, device, system, and storage medium for controlling a tobacco appliance. The method includes: acquiring a first sound signal from the tobacco appliance under various operating conditions; processing the first sound signal according to a preset signal interception rule to obtain a second sound signal; and extracting features from the second sound signal to obtain sound pressure and time-frequency features. This application intercepts and extracts features from the acquired first sound signal to determine the presence of noise. When noise is present, the noise signal is preprocessed and input into a preset noise identification model to identify the noise problem and corresponding solutions. The analysis and processing using the preset model effectively shortens the time required for manual noise identification and improves the accuracy of noise identification.

Description

Technical Field

[0001] This invention relates to the field of tobacco appliance control technology, and in particular to tobacco appliance control methods, devices, systems and storage media. Background Technology

[0002] A range hood is a kitchen appliance designed to purify the kitchen environment. It works by using a centrifugal fan installed inside the hood to draw in and exhaust cooking fumes. The centrifugal fan consists of a volute, an impeller housed within the volute, and a motor that drives the impeller. When the impeller rotates, a negative pressure is generated at the center of the fan, drawing in the cooking fumes from below. After being accelerated by the fan, the volute collects the fumes and guides them outdoors. Therefore, range hoods generate considerable noise during operation, a significant noise source in homes that impacts user experience and has become an important factor for consumers when purchasing such products. R&D engineers at various manufacturers have conducted extensive research on vibration and noise reduction for range hoods, resulting in continuous noise reduction and improved user experience.

[0003] To control noise levels in range hoods, manufacturers currently set noise test thresholds for their products before shipment. However, some prototype units still fail to meet these standards during pre-shipment testing, requiring engineers to analyze the noise issues and provide solutions. Conversely, some prototype units pass factory tests but often experience noise problems during user operation, necessitating on-site analysis by engineers. Existing noise analysis and solutions are inefficient and costly. Providing a solution to a noise problem typically takes two hours, especially for noise issues originating in the user's home, where the time required is even longer. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a method, device, system and storage medium for controlling a smoke hood.

[0005] In a first aspect, embodiments of the present invention provide a method for controlling a range hood, the method comprising:

[0006] Acquire the first sound signal of the range hood under various operating conditions;

[0007] According to the preset signal extraction rules, the first sound signal is extracted and processed to obtain the second sound signal;

[0008] Feature extraction is performed on the second sound signal to obtain sound pressure features and time-frequency features;

[0009] If the sound pressure characteristic is greater than the corresponding set threshold, and / or the time-frequency characteristic indicates that the second sound signal has frequency superposition, it is determined that the second sound signal has a noise problem.

[0010] According to the preset preprocessing rules, the second sound signal is preprocessed to obtain the third sound signal;

[0011] The third sound signal is input into a pre-trained noise recognition model, which outputs the type of noise problem and the corresponding solution.

[0012] In conjunction with the first aspect, the step of processing the first audio signal according to a preset signal extraction rule to obtain the second audio signal includes:

[0013] Using the point corresponding to the peak value in the first sound signal as the signal center point, a sound signal of a set duration is extracted as the second sound signal.

[0014] In conjunction with the first aspect, sound pressure characteristics include: sound pressure frequency domain characteristics, which include the sound pressure frequency domain mean and the sound pressure frequency domain peak value;

[0015] The steps for feature extraction from the second audio signal include:

[0016] The second sound signal is subjected to Fourier transform to obtain the sound frequency domain signal;

[0017] Calculate the average value of all frequency bands in the acoustic frequency domain signal to obtain the average sound pressure level in the frequency domain;

[0018] Select the maximum sound value from all frequency bands and define the maximum sound value as the sound pressure level peak value in the frequency domain.

[0019] In conjunction with the first aspect, sound pressure characteristics also include: sound pressure time-domain characteristics; sound pressure time-domain characteristics include the sound pressure time-domain mean and the sound pressure time-domain peak value;

[0020] The steps for feature extraction from the second audio signal include:

[0021] Calculate the average sound pressure level of the second sound signal to obtain the time-domain average sound pressure level;

[0022] The maximum peak value of the sound pressure in the second sound signal is determined to be the sound pressure time domain peak value.

[0023] In conjunction with the first aspect, the preprocessing rules include: signal denoising, endpoint detection, frame-by-frame windowing, and wavelet transform.

[0024] In conjunction with the first aspect, the noise recognition model comprises a trainer, a classification model, and a neural network connected in sequence;

[0025] The training process for the noise recognition model is as follows:

[0026] Obtain the original sample set; the original sample set includes multiple noise signal samples under various operating conditions.

[0027] The original sample set is divided into a training sample set and a validation sample set;

[0028] The training sample set is input into the trainer. For each training sample in the training sample set, feature extraction is performed on the training sample to obtain the frequency and amplitude of the noise corresponding to the training sample.

[0029] Label the training samples and train a classification model based on the labeled training samples to obtain the noise problem category corresponding to the training samples;

[0030] The training samples that determine the category of noise problem are input into a pre-set neural network. The neural network is trained based on the CNN neural network convolution algorithm, and the noise recognition results and corresponding solutions are output.

[0031] The preset neural network is validated based on the validation sample set;

[0032] Determine whether the accuracy of the preset neural network reaches the preset accuracy threshold;

[0033] If not, calculate the loss function of the preset neural network and iteratively optimize the neural network based on the loss function until the accuracy of the neural network model reaches the preset accuracy threshold to obtain the noise recognition model.

[0034] In conjunction with the first aspect, the method also includes:

[0035] If the sound pressure characteristic is less than or equal to the corresponding set threshold, and the time-frequency characteristic indicates that there is no frequency superposition of the second sound signal, it is determined that the smoke machine is operating normally and without noise.

[0036] Secondly, this application provides a smoke machine control device, the device comprising:

[0037] The acquisition module is used to acquire the first sound signal of the range hood under various operating conditions.

[0038] The signal interception module is used to intercept and process the first sound signal according to the preset signal interception rules to obtain the second sound signal;

[0039] The feature extraction module is used to extract features from the second sound signal to obtain sound pressure features and time-frequency features;

[0040] The determination module is used to determine that there is a noise problem in the second sound signal when the sound pressure feature is greater than the corresponding set threshold and / or the time-frequency feature indicates that the second sound signal has frequency superposition.

[0041] The preprocessing module is used to preprocess the second sound signal according to preset preprocessing rules to obtain the third sound signal;

[0042] The output module is used to input the third sound signal into the pre-trained noise recognition model and output the noise problem type and the corresponding solution.

[0043] Thirdly, this application provides a smoke machine control system, which includes a smoke machine and a controller. The smoke machine includes a smoke collection chamber and a housing. A hole is opened in the smoke collection chamber, and a microphone is disposed in the hole. The microphone is wirelessly connected to the controller, and the controller is used to execute the above-mentioned method.

[0044] Fourthly, this application provides a storage medium storing computer program instructions, which are read and executed by a processor to perform the above-described method.

[0045] The embodiments of this invention bring the following beneficial effects: This application extracts features from the acquired first sound signal to determine whether noise signals exist. When noise is present, the noise signal is preprocessed and then input into a preset noise recognition model to obtain the noise problem and corresponding solutions. The analysis and processing using the preset model can effectively shorten the time required for manual noise recognition and improve the accuracy of noise recognition.

[0046] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0048] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0049] Figure 1 This is a flowchart of a smoke hood control method provided in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram illustrating the process of intercepting and processing the first sound signal in the range hood control method provided in this embodiment of the invention;

[0051] Figure 3 The image obtained during the extraction of time-frequency features in the smoke machine control method provided in this embodiment of the invention does not have frequency superposition.

[0052] Figure 4 The image obtained during the extraction of time-frequency features in the smoke machine control method provided in this embodiment of the invention contains superimposed frequencies.

[0053] Figure 5 This is a schematic diagram of the preprocessing process in the smoke machine control method provided in the embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram of the structure of the smoke hood control device provided in an embodiment of the present invention;

[0055] Figure 7 This is a schematic diagram showing the installation position of the microphone in the smoke hood control system provided in an embodiment of the present invention;

[0056] Figure 8 This is a schematic diagram illustrating the calculation of the threshold corresponding to the sound pressure characteristics in the smoke machine control system provided in an embodiment of the present invention.

[0057] Figure label:

[0058] 10 - Acquisition module, 20 - Signal interception module, 30 - Feature extraction module, 40 - Determination module, 50 - Preprocessing module, 60 - Output module;

[0059] 100 - Smoke collection chamber, 200 - Chassis, 300 - Microphone. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] To facilitate understanding of this embodiment, the application scenarios and design concepts of this application embodiment will be briefly introduced below.

[0062] Kitchen range hoods generate significant noise during operation, making it a major noise source in homes and impacting user experience. This noise has become a crucial factor for consumers when purchasing the product. Currently, when noise issues arise during actual use of the range hood after purchase, engineers are required to visit the user's home to analyze the problem and provide solutions. However, existing noise analysis methods and solution development processes are inefficient and costly.

[0063] Based on this, embodiments of this application provide a method, apparatus, system, and storage medium for controlling a range hood, to improve the efficiency and accuracy of noise analysis. The method provided in this application is applied to a controller in a range hood control system. The range hood control system also includes a microphone for collecting sound signals from the range hood at various operating levels. The microphone is connected to the controller to transmit the collected sound signals to the controller. The controller includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method provided in this application.

[0064] Example 1

[0065] This application provides a method for controlling a smoke hood, combined with... Figure 1 As shown, the method specifically includes:

[0066] S110, the processor acquires the first sound signal of the range hood under various operating conditions.

[0067] S120, the processor performs interception processing on the first sound signal according to the preset signal interception rules to obtain the second sound signal.

[0068] S130, the processor extracts features from the second sound signal to obtain sound pressure features and time-frequency features.

[0069] S140, if the processor determines that the second sound signal has a noise problem when the sound pressure characteristic is greater than the corresponding set threshold and / or the time-frequency characteristic indicates that the second sound signal has frequency superposition.

[0070] S150, the processor preprocesses the second sound signal according to the preset preprocessing rules to obtain the third sound signal.

[0071] S160: The processor inputs the third sound signal into the pre-trained noise recognition model and outputs the noise problem type and the corresponding solution.

[0072] In this embodiment, the first sound signal of the range hood under various operating conditions is acquired, and the first sound signal is truncated and its features are extracted to obtain characteristic parameters. The presence of noise in the first sound signal is determined almost entirely based on these characteristic parameters. If noise is present, the first sound signal is preprocessed and input into a preset noise recognition model. The model then outputs the noise problem type and corresponding solution. This method of analysis and processing using a preset model effectively shortens the time required for manual noise identification and improves the accuracy of noise recognition.

[0073] In step S110, each operating mode should include at least three modes: low, high, and high-power (or "stir-fry"), respectively. These three modes are the three commonly set modes for range hoods. However, as the number of preset modes increases, the number of first sound signals acquired also increases. The component used to acquire the first sound signal can be any type of sound sensor, such as a microphone or noise sensor; in this embodiment, the component is a microphone 300. In this embodiment, according to the new national standard noise testing method, the first sound signals of the range hood are collected under the low, high, and high-power (or "stir-fry") modes, and the collected first sound signals are transmitted to the controller every t0 seconds. As an example, in this embodiment, t0 = 30 seconds.

[0074] In conjunction with the first aspect, S120 includes:

[0075] Using the point corresponding to the peak value in the first sound signal as the signal center point, a sound signal of a set duration is extracted as the second sound signal.

[0076] Specifically, the duration t is set in the first sound signal with duration t0. x The sound signal is used as the second sound signal. That is, t x ≤t0.

[0077] Based on the above example, let t be taken. x =10s. In the first sound signal with a duration of t0 = 30s, the total number of sample points N is obtained. The sample point N corresponding to the first sound peak is selected as the signal center point, and samples are extracted from both sides of point N. The number of sample points. This allows us to extract t. x The first audio signal, with a length of 10 seconds, is used as the second audio signal. The extraction process is as follows: Figure 2 As shown.

[0078] In conjunction with the first aspect, S130, the processor extracts features from the second sound signal to obtain sound pressure features and time-frequency features; wherein, the sound pressure features include: sound pressure frequency domain features and sound pressure time domain features.

[0079] The frequency domain characteristics of sound pressure level (SPL) include the SPL mean and the SPL peak value. The time domain characteristics of SPL include the SPL mean and the SPL peak value.

[0080] Step S130 involves feature extraction from the second sound signal, specifically including:

[0081] S131, the processor performs a Fourier transform on the second sound signal to obtain the sound frequency domain signal.

[0082] S132 calculates the average value of all frequency bands in the audio frequency domain signal to obtain the average sound pressure level in the frequency domain.

[0083] S133, select the maximum sound value in all frequency bands, and determine the maximum sound value as the sound pressure frequency domain peak value.

[0084] Step S131 is calculated according to the following formula:

[0085]

[0086] Where F(W) is the audio frequency domain signal, x(t) is the audio time domain signal, and e -iwt This is the projection function.

[0087] Next, in step S132, t is obtained based on the acoustic frequency domain signal F(W). x The average sound pressure level across all frequency bands of the second sound signal of duration is used to obtain the sound pressure frequency domain mean.

[0088] Step S133: Obtain t based on the acoustic frequency domain signal F(W). x The peak value of the sound pressure in the frequency domain is obtained by taking the peak value of the second sound signal across all frequency bands.

[0089] The sound pressure time-domain characteristics include the sound pressure time-domain mean and the sound pressure time-domain peak value; step S130 also includes:

[0090] S1300, the processor calculates the average sound pressure level of the second sound signal to obtain the time-domain average sound pressure level.

[0091] Similar to step S132, the second sound signal t is calculated. x The arithmetic mean of each sound pressure level within the specified time period is used to obtain the time-domain mean of the sound pressure.

[0092] S1301, The processor determines that the maximum peak value of the sound pressure in the second sound signal is the sound pressure time domain peak value.

[0093] In addition, step S130 further includes: the processor extracting features from the second audio signal to obtain time-frequency features, specifically:

[0094] Calculate using the following formula:

[0095]

[0096] Where H[x(t)] is the wavelet transform function, x(t) is the sound time-domain signal, and λ is a set constant.

[0097] This time-frequency feature is used to characterize whether there is frequency superposition in the second sound signal. Based on wavelet analysis, the time-frequency feature of the second sound signal is judged, and the image obtained when superposition exists is as follows. Figure 4 As shown, the image obtained without superposition is as follows: Figure 3As shown. When frequency superposition exists, there are at least two harmonics in the image (referring to M1 and M2 in the figure).

[0098] Then, step S140 compares the extracted sound pressure feature with the set threshold, and if the sound pressure feature is greater than the corresponding set threshold, and / or the time-frequency feature indicates that the second sound signal has a noise problem.

[0099] Among them, the time-domain mean and frequency-domain mean of sound pressure correspond to the mean threshold; the time-domain peak value and frequency-domain peak value of sound pressure correspond to the peak threshold. That is to say, if any sound pressure feature is greater than the corresponding threshold, and / or if the time-frequency feature indicates that the second sound signal has frequency superposition, it is determined that the second sound signal has a noise problem.

[0100] In this way, after the first sound signal is collected, it is processed by interception and feature extraction, and the presence of noise is determined based on the features. Compared with the previous method, this can improve the efficiency and accuracy of noise identification.

[0101] The set threshold corresponding to the sound pressure characteristic is a noise threshold calculated from the noise value measured by national or enterprise standards. The noise value specified by national or enterprise standards is the working noise value. The test method involves placing the microphone on the center line of the range hood, 700mm away from the wall, at a height of 1500mm (combined with...). Figure 8 Point A in the diagram), and in this embodiment, the microphone 300 is integrated into the smoke collection cavity 100 (combined with point A in the diagram). Figure 8 (Point B). Therefore, the noise threshold in this scheme needs to be calculated using the noise values ​​obtained from national or enterprise standard tests. The calculation method is to first measure the noise transfer function H between position A and position B, and then measure the operating noise a at position A. Noise threshold = a × H. The transfer function is measured by using a volumetric sound source for excitation at position A, and then testing the noise response at position B, which is the frequency response function.

[0102] Based on the above method, as the microphone 300 is positioned differently, the corresponding H changes. Then, according to conventional threshold determination rules, the average sound pressure threshold and the peak sound pressure threshold can be determined. Alternatively, these average and peak sound pressure thresholds can be manually set and adjusted based on different range hoods or actual usage scenarios.

[0103] Then, in step S150, combined with Figure 5 As shown, the preprocessing rules include: signal denoising, endpoint detection, frame windowing, and wavelet transform.

[0104] Signal denoising involves bandpass filtering and smoothing to remove interference. The bandpass filter only allows signals within a specific frequency band to pass through, permitting frequencies above a low threshold and below a high threshold. In this application, the selected bandwidth is 100-10000Hz. Smoothing employs a moving average method, calculating the arithmetic mean (y) of data points near a given point (within ±5Hz). a = (y1+y2+……+y i ) / i), y a The arithmetic mean is y1, y2, ..., y3. i This represents the noise amplitude corresponding to a single frequency within the ±5Hz range.

[0105] Endpoint detection is used to identify the start and end points of the second audio signal, excluding non-second audio signals. Framing and windowing discretizes the frequency-domain second audio signal into framed signals by windowing. The purpose is to accurately locate the problematic frequency band during subsequent noise analysis. Endpoint detection uses the short-time energy method (E0). n =F(w)×F(w-λ)), where F(w) is the noise frequency domain signal and F(w-λ) is the conjugate spectrum of the noise signal. A value greater than the short-time energy threshold indicates a smoke machine noise segment.

[0106] The frequency superposition information of the second sound signal is calculated based on wavelet transform, which is the same as the time-frequency feature acquisition operation in step S130, thereby determining the frequency information superimposed on the second sound signal.

[0107] Based on the aforementioned preset preprocessing rules, the second audio signal is preprocessed to obtain the third audio signal.

[0108] In step S140, it is determined that the second sound signal has a noise problem. After the second sound signal with noise problem is preprocessed to obtain the third sound signal in step S150, the third sound signal is input into the pre-trained noise recognition model in step S160, and the noise problem type and the corresponding solution are output.

[0109] The noise recognition model consists of a trainer, a classification model, and a neural network connected in sequence.

[0110] The training process for the noise recognition model is as follows:

[0111] S210, the processor acquires the original sample set; wherein, the original sample includes multiple noise signal samples under various gear conditions.

[0112] S220, the processor divides the original sample set into a training sample set and a validation sample set.

[0113] S230, the processor inputs the training sample set to the trainer, and extracts features from each training sample in the training sample set to obtain the frequency and amplitude of the noise corresponding to the training sample.

[0114] S240, the processor adds labels to the training samples and trains a classification model based on the labeled training samples to obtain the noise category corresponding to the training samples.

[0115] In step S240, the label includes the frequency, amplitude, subjective score, and objective parameters of the noise obtained in step S230. The subjective score is used to characterize the artificially added noise label, such as the probability of noise occurring, and the objective parameters are used to characterize the noise. Thus, a trainer is constructed with multiple training samples and a database of frequency, amplitude, subjective score, and objective parameters corresponding to each training sample.

[0116] Then, according to the classification rules, a classification model is trained using multiple training samples to classify the training samples and obtain multiple noise as problem categories.

[0117] S250: The processor inputs training samples that determine the noise problem category into a preset neural network, trains the neural network based on the CNN neural network convolution algorithm, and outputs noise recognition results and corresponding solutions.

[0118] In this embodiment, as an implementable method, the step of training the neural network based on the CNN neural network convolution algorithm includes:

[0119] Calculate based on the following formula:

[0120]

[0121] Where x(t) is the noise input, h(t) is the convolution kernel, y(t) is the output, and p is the time step.

[0122] This is a fairly standard method for calculating convolution, and it is not limited here.

[0123] S260, the processor validates the preset neural network based on the validation sample set.

[0124] S270, the processor determines whether the accuracy of the preset neural network has reached the preset accuracy threshold.

[0125] If the accuracy of the preset neural network does not reach the preset accuracy threshold, proceed to step S280.

[0126] S280, calculate the loss function of the preset neural network and iteratively optimize the neural network according to the loss function until the accuracy of the neural network model reaches the preset accuracy threshold, and obtain the noise recognition model.

[0127] The method provided in this embodiment inputs a third sound signal with noise into a pre-trained noise recognition model that has reached a preset accuracy, so as to identify the type of noise problem and provide a solution. In this way, the whole process is completed automatically, which can reduce the amount of manual labor and the difficulty of noise recognition and diagnosis, and improve the efficiency and accuracy of noise recognition.

[0128] In conjunction with the first aspect, the method provided in this embodiment further includes: determining that the range hood is operating normally and without noise when the sound pressure feature is less than or equal to the corresponding set threshold and the time-frequency feature indicates that the second sound signal does not have frequency superposition.

[0129] In other words, the range hood is considered to be operating normally and without noise only if all sound pressure characteristics do not exceed the corresponding set threshold and there is no frequency superposition in the time domain characteristic representation.

[0130] Secondly, this application provides a smoke machine control device, combined with Figure 6 As shown, the device includes: an acquisition module 10, a signal interception module 20, a feature extraction module 30, a determination module 40, a preprocessing module 50, and an output module 60.

[0131] The acquisition module 10 is used to acquire the first sound signal of the range hood under various operating conditions.

[0132] The signal interception module 20 is used to intercept the first sound signal according to the preset signal interception rules to obtain the second sound signal.

[0133] The feature extraction module 30 is used to extract features from the second sound signal to obtain sound pressure features and time-frequency features.

[0134] The determination module 40 is used to determine that there is a noise problem in the second sound signal when the sound pressure feature is greater than the corresponding set threshold and / or the time-frequency feature indicates that the second sound signal has frequency superposition.

[0135] The preprocessing module 50 is used to preprocess the second sound signal according to the preset preprocessing rules to obtain the third sound signal.

[0136] The output module 60 is used to input the third sound signal into the pre-trained noise recognition model and output the noise problem type and the corresponding solution.

[0137] Thirdly, embodiments of this application provide a range hood control system, which includes a range hood and a controller, combined with... Figure 7 As shown, the smoke machine includes a smoke collection chamber 100 and a casing 200. The smoke collection chamber 100 has a hole, and a microphone 300 is installed in the hole. The microphone 300 is wirelessly connected to the controller.

[0138] The controller includes a memory and a processor. The memory may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between the system network element and at least one other network element is achieved through at least one communication interface (wired or wireless), which can use the Internet, wide area network, local area network, metropolitan area network, etc. The processor may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above methods can be completed by integrated logic circuits in the processor or by software instructions. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The methods, steps, and logic block diagrams disclosed in the embodiments of this invention can be implemented or executed. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this invention can be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the method in the aforementioned embodiments.

[0139] Fourthly, embodiments of this application provide a readable storage medium storing computer program instructions, which are read and executed by a processor to perform the above-described method.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0141] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0142] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0143] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0144] Finally, it should be noted that the above embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for controlling a smoke hood, characterized in that, The method includes: Acquire the first sound signal of the range hood under various operating conditions; According to preset signal extraction rules, the first sound signal is extracted to obtain the second sound signal; Feature extraction is performed on the second sound signal to obtain sound pressure features and time-frequency features; If the sound pressure feature is greater than the corresponding set threshold, and / or the time-frequency feature indicates that the second sound signal has frequency superposition, it is determined that the second sound signal has a noise problem; According to the preset preprocessing rules, the second sound signal is preprocessed to obtain the third sound signal; The third sound signal is input into a pre-trained noise recognition model, which outputs the noise problem type and the corresponding solution. The sound pressure features include: sound pressure frequency domain features, which include the sound pressure frequency domain mean and the sound pressure frequency domain peak value; the step of extracting features from the second sound signal includes: Perform a Fourier transform on the second sound signal to obtain the sound frequency domain signal; Calculate the average value of all frequency bands in the acoustic frequency domain signal to obtain the average value of the sound pressure frequency domain; The maximum sound value in all frequency bands is selected, and the maximum sound value is determined as the peak value of the sound pressure frequency domain; The sound pressure feature further includes: sound pressure time-domain features; the sound pressure time-domain features include the sound pressure time-domain mean and the sound pressure time-domain peak value; the step of extracting features from the second sound signal includes: Calculate the average sound pressure level of the second sound signal to obtain the time-domain average sound pressure level; The maximum peak value of the sound pressure in the second sound signal is determined to be the time-domain peak value of the sound pressure.

2. The method according to claim 1, characterized in that, The step of processing the first audio signal according to a preset signal truncation rule to obtain the second audio signal includes: Using the point corresponding to the peak value in the first sound signal as the signal center point, a sound signal of a set duration is extracted as the second sound signal.

3. The method according to claim 1, characterized in that, The preprocessing rules include: signal denoising, endpoint detection, frame windowing, and wavelet transform.

4. The method according to claim 1, characterized in that, The noise recognition model comprises a trainer, a classification model, and a neural network connected in sequence. The training process of the noise recognition model is as follows: Obtain the original sample set; wherein, the original sample set includes multiple noise signal samples under various gear conditions; The original sample set is divided into a training sample set and a validation sample set; The training sample set is input into the trainer, and for each training sample in the training sample set, feature extraction is performed on the training sample to obtain the frequency and amplitude of the noise corresponding to the training sample. Labels are added to the training samples, and the classification model is trained based on the labeled training samples to obtain the noise problem category corresponding to the training samples; The training samples that determine the noise problem category are input into a preset neural network, and the neural network is trained based on the CNN neural network convolution algorithm to output noise recognition results and corresponding solutions. The preset neural network is validated based on the validation sample set; Determine whether the accuracy of the neural network reaches a preset accuracy threshold; If not, calculate the loss function of the preset neural network and iteratively optimize the neural network according to the loss function until the accuracy of the neural network reaches the preset accuracy threshold to obtain the noise recognition model.

5. The method according to claim 1, characterized in that, The method further includes: If the sound pressure characteristic is less than or equal to the corresponding set threshold, and the time-frequency characteristic indicates that there is no frequency superposition of the second sound signal, it is determined that the smoke machine is operating normally and without noise.

6. A smoke machine control device, characterized in that, The device includes: The acquisition module is used to acquire the first sound signal of the range hood under various operating conditions. The signal interception module is used to intercept the first sound signal according to a preset signal interception rule to obtain the second sound signal; The feature extraction module is used to extract features from the second sound signal to obtain sound pressure features and time-frequency features; The determination module is used to determine that the second sound signal has a noise problem when the sound pressure feature is greater than the corresponding set threshold and / or the time-frequency feature indicates that the second sound signal has frequency superposition. The preprocessing module is used to preprocess the second sound signal according to a preset preprocessing rule to obtain the third sound signal; The output module is used to input the third sound signal into a pre-trained noise recognition model and output the noise problem type and the corresponding solution. The sound pressure features include: sound pressure frequency domain features, which include the sound pressure frequency domain mean and the sound pressure frequency domain peak value; the step of extracting features from the second sound signal includes: Perform a Fourier transform on the second sound signal to obtain the sound frequency domain signal; Calculate the average value of all frequency bands in the acoustic frequency domain signal to obtain the average value of the sound pressure frequency domain; The maximum sound value in all frequency bands is selected, and the maximum sound value is determined as the peak value of the sound pressure frequency domain; The sound pressure feature further includes: sound pressure time-domain features; the sound pressure time-domain features include the sound pressure time-domain mean and the sound pressure time-domain peak value; the step of extracting features from the second sound signal includes: Calculate the average sound pressure level of the second sound signal to obtain the time-domain average sound pressure level; The maximum peak value of the sound pressure in the second sound signal is determined to be the time-domain peak value of the sound pressure.

7. A smoke hood control system, characterized in that, The smoke machine control system includes a smoke machine and a controller. The smoke machine includes a smoke collection chamber and a housing. The smoke collection chamber has a hole, and a microphone is installed in the hole. The microphone is wirelessly connected to the controller. The controller is used to execute the method described in any one of claims 1-5.

8. A storage medium, characterized in that, The storage medium stores computer program instructions, which, when read and executed by a processor, perform the method described in any one of claims 1 to 5.

Images