Abnormal sound decomposition method and device

Abstract

The invention discloses an abnormal sound decomposition method and device, and relates to the technical field of noise analysis, and the method comprises the following steps: obtaining the amplitude spectrum of background noise; obtaining an amplitude spectrum of each known abnormal sound problem; obtaining an amplitude spectrum of the abnormal sound problem set to be decomposed; constructing an abnormal sound decomposition model; and performing abnormal sound decomposition based on the abnormal sound decomposition model. According to the method, the background noise, each known abnormal sound problem and the to-be-decomposed abnormal sound problem set are subjected to spectral analysis, and the synthetic power spectrum sequence and the pure power spectrum of the to-be-decomposed abnormal sound problem set are subjected to correlation analysis, so that abnormal sound decomposition is realized, the process is effectively simplified, and the efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of noise analysis, and in particular to a method and device for decomposing abnormal noise. Background Art

[0002] With the development of the times, the abnormal noise performance of vehicles has received more and more attention from customers. In order to improve the abnormal noise performance of vehicles, various advanced automatic data processing systems have been introduced into more and more abnormal noise detection scenarios. For a certain thing, such as the dynamic abnormal noise performance test of a mass-produced model, when the vehicle is driving on a certain road at a certain speed, there may be only one abnormal noise problem, or there may be two or more abnormal noise problems, and these abnormal noise problems are all known (they have occurred before and there are mature solutions). In this case, for the abnormal noise detection automatic data processing system, it is easy to identify the occurrence of one abnormal noise problem. When two or more abnormal noise problems occur, it is difficult to identify which abnormal noises have occurred because these abnormal noises are mixed together, that is, these abnormal noises cannot be correctly decomposed.

[0003] Therefore, in order to meet actual needs, an abnormal noise decomposition technology is now provided. Summary of the Invention

[0004] In response to the defects existing in the prior art, the purpose of the present invention is to provide a method and device for decomposing abnormal noise, which performs spectral analysis on background noise, various known abnormal noise problems and a set of abnormal noise problems to be decomposed, and performs correlation analysis on the synthesized power spectrum sequence and the pure power spectrum of the abnormal noise problem set to be decomposed, thereby achieving the decomposition of abnormal noise, effectively simplifying the process and improving efficiency.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] In a first aspect, the present application provides a method for decomposing abnormal noise, the method comprising the following steps:

[0007] Obtain the amplitude spectrum of background noise;

[0008] Obtain the amplitude spectrum of each known abnormal noise problem;

[0009] Obtain the amplitude spectrum of the abnormal noise problem set to be decomposed;

[0010] Construct abnormal noise decomposition model;

[0011] Based on the abnormal noise decomposition model, abnormal noise decomposition is performed.

[0012] Based on the above technical solution, the method of obtaining the amplitude spectrum of background noise includes the following steps:

[0013] Based on the set sampling frequency corresponding to the background noise, a background noise signal is collected and obtained;

[0014] Analyzing the background noise signal based on a set frequency resolution of spectrum analysis to obtain a spectrum corresponding to the background noise;

[0015] Based on the frequency spectrum corresponding to the background noise, an amplitude spectrum of the background noise is obtained.

[0016] Based on the above technical solution, the method of obtaining the amplitude spectrum of each known abnormal sound problem includes the following steps:

[0017] Based on the set sampling frequency corresponding to each of the known abnormal sound problems, a noise signal of the known abnormal sound problem is collected;

[0018] Analyzing the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each of the known abnormal sound problems to obtain the spectrum corresponding to the known abnormal sound problem;

[0019] Based on the frequency spectrum corresponding to the known abnormal sound problem, an amplitude spectrum of the known abnormal sound problem is obtained.

[0020] Based on the above technical solution, the method of obtaining the amplitude spectrum of the abnormal noise problem set to be decomposed includes the following steps:

[0021] Perform abnormal sound monitoring based on a set sampling frequency to collect and obtain noise signals of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed;

[0022] A spectrum analysis is performed based on the noise signal of the abnormal sound problem set to be decomposed to obtain an amplitude spectrum of the abnormal sound problem set to be decomposed.

[0023] On the basis of the above technical solution, the construction of the abnormal sound decomposition model includes the following steps:

[0024] Obtaining a power spectrum of the background noise based on the amplitude spectrum of the background noise;

[0025] Based on the power spectrum of the background noise and the amplitude spectrum of each of the known abnormal sound problems, a pure power spectrum of each of the known abnormal sound problems is obtained;

[0026] Obtaining a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed;

[0027] The abnormal noise decomposition model is constructed based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

[0028] In a second aspect, the present application further provides an abnormal sound decomposition device, the device comprising:

[0029] An amplitude spectrum acquisition module, which is used to obtain the amplitude spectrum of background noise;

[0030] The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of each known abnormal sound problem;

[0031] The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of the abnormal noise problem set to be decomposed;

[0032] A model building module, which is used to build an abnormal noise decomposition model;

[0033] The abnormal noise decomposition module is used to decompose the abnormal noise based on the abnormal noise decomposition model.

[0034] On the basis of the above technical solution, the amplitude spectrum acquisition module is further used to collect and obtain the background noise signal based on the set sampling frequency corresponding to the background noise;

[0035] The amplitude spectrum acquisition module is further configured to analyze the background noise signal based on a set frequency resolution of the spectrum analysis to obtain a spectrum corresponding to the background noise;

[0036] The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the background noise based on the frequency spectrum corresponding to the background noise.

[0037] On the basis of the above technical solution, the amplitude spectrum acquisition module is further used to collect and obtain the noise signal of the known abnormal sound problem based on the set sampling frequency corresponding to each known abnormal sound problem;

[0038] The amplitude spectrum acquisition module is further configured to analyze the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each known abnormal sound problem, and obtain the spectrum corresponding to the known abnormal sound problem;

[0039] The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the known abnormal sound problem based on the frequency spectrum corresponding to the known abnormal sound problem.

[0040] On the basis of the above technical solution, the amplitude spectrum acquisition module is further used to perform abnormal sound monitoring based on a set sampling frequency, and collect and obtain the noise signal of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed;

[0041] The amplitude spectrum acquisition module is further used to perform spectrum analysis based on the noise signal of the abnormal sound problem set to be decomposed, and obtain the amplitude spectrum of the abnormal sound problem set to be decomposed.

[0042] On the basis of the above technical solution, the model building module is further used to obtain the power spectrum of the background noise based on the amplitude spectrum of the background noise;

[0043] The model building module is further configured to obtain a pure power spectrum of each known abnormal sound problem based on the power spectrum of the background noise and the amplitude spectrum of each known abnormal sound problem;

[0044] The model building module is further configured to obtain a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed;

[0045] The model building module is further configured to build the abnormal noise decomposition model based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

[0046] Compared with the prior art, the advantages of the present invention are:

[0047] The present invention performs spectrum analysis on background noise, various known abnormal noise problems and a set of abnormal noise problems to be decomposed, and performs correlation analysis on the synthesized power spectrum sequence and the pure power spectrum of the set of abnormal noise problems to be decomposed, thereby achieving the decomposition of abnormal noise, effectively simplifying the process and improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 This is a flowchart of the steps of the abnormal noise decomposition method according to an embodiment of the present invention;

[0050] Figure 2 2 is a structural block diagram of an abnormal sound decomposition device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0052] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0053] The embodiments of the present application provide a method and apparatus for decomposing abnormal noises, which perform spectral analysis on background noise, various known abnormal noise problems, and a set of abnormal noise problems to be decomposed, and perform correlation analysis on the synthesized power spectrum sequence and the pure power spectrum of the set of abnormal noise problems to be decomposed, thereby achieving abnormal noise decomposition, effectively simplifying the process, and improving efficiency.

[0054] To achieve the above technical effects, the overall idea of this application is as follows:

[0055] A method for decomposing an abnormal sound, comprising the steps of:

[0056] S1. Obtain the amplitude spectrum of background noise;

[0057] S2. Obtain the amplitude spectrum of each known abnormal noise problem;

[0058] S3, obtaining the amplitude spectrum of the abnormal noise problem set to be decomposed;

[0059] S4, constructing an abnormal noise decomposition model;

[0060] S5. Decompose the abnormal noise based on the abnormal noise decomposition model.

[0061] The embodiments of the present application are further described in detail below with reference to the accompanying drawings.

[0062] First, see Figure 1 As shown, the embodiment of the present application provides a method for decomposing abnormal noise, which includes the following steps:

[0063] S1. Obtain the amplitude spectrum of background noise;

[0064] S2. Obtain the amplitude spectrum of each known abnormal noise problem;

[0065] S3, obtaining the amplitude spectrum of the abnormal noise problem set to be decomposed;

[0066] S4, constructing an abnormal noise decomposition model;

[0067] S5. Decompose the abnormal noise based on the abnormal noise decomposition model.

[0068] In the embodiment of the present application, spectrum analysis is performed on the background noise, each known abnormal sound problem, and the set of abnormal sound problems to be decomposed to obtain their amplitude spectra;

[0069] Based on each amplitude spectrum, background noise is eliminated to obtain the pure power spectrum of each known abnormal noise problem and the set of abnormal noise problems to be decomposed;

[0070] Based on the linear superposition principle of power, an abnormal noise decomposition model was constructed by introducing the concept of occurrence intensity of each known abnormal noise problem.

[0071] By discretizing the occurrence intensity variable, the synthetic power spectrum sequence is obtained;

[0072] The synthetic power spectrum sequence was correlated with the pure power spectrum of the abnormal noise problem set to be decomposed, and the optimal abnormal noise problem occurrence intensity was obtained. The participation degree of each known abnormal noise problem was obtained, and the decomposition of the abnormal noise was achieved.

[0073] The technical solution of the embodiment of the present application performs spectral analysis on background noise, various known abnormal noise problems and the set of abnormal noise problems to be decomposed, and performs correlation analysis on the synthetic power spectrum sequence and the pure power spectrum of the set of abnormal noise problems to be decomposed, thereby realizing the decomposition of abnormal noise, effectively simplifying the process and improving efficiency.

[0074] In summary, the technical principles of the embodiments of the present application are simple and clear, the calculation is simple, and the numerical analysis method is used to replace the complicated theoretical analysis, making the abnormal noise decomposition feasible, simple and efficient.

[0075] Furthermore, the obtaining of the amplitude spectrum of the background noise comprises the following steps:

[0076] Based on the set sampling frequency corresponding to the background noise, a background noise signal is collected and obtained;

[0077] Analyzing the background noise signal based on a set frequency resolution of spectrum analysis to obtain a spectrum corresponding to the background noise;

[0078] Based on the frequency spectrum corresponding to the background noise, an amplitude spectrum of the background noise is obtained.

[0079] Furthermore, obtaining the amplitude spectrum of each known abnormal sound problem includes the following steps:

[0080] Based on the set sampling frequency corresponding to each of the known abnormal sound problems, a noise signal of the known abnormal sound problem is collected;

[0081] Analyzing the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each of the known abnormal sound problems to obtain the spectrum corresponding to the known abnormal sound problem;

[0082] Based on the frequency spectrum corresponding to the known abnormal sound problem, an amplitude spectrum of the known abnormal sound problem is obtained.

[0083] Furthermore, the acquisition of the amplitude spectrum of the abnormal noise problem set to be decomposed includes the following steps:

[0084] Perform abnormal sound monitoring based on a set sampling frequency to collect and obtain noise signals of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed;

[0085] A spectrum analysis is performed based on the noise signal of the abnormal sound problem set to be decomposed to obtain an amplitude spectrum of the abnormal sound problem set to be decomposed.

[0086] Furthermore, the construction of the abnormal noise decomposition model includes the following steps:

[0087] Obtaining a power spectrum of the background noise based on the amplitude spectrum of the background noise;

[0088] Based on the power spectrum of the background noise and the amplitude spectrum of each of the known abnormal sound problems, a pure power spectrum of each of the known abnormal sound problems is obtained;

[0089] Obtaining a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed;

[0090] The abnormal noise decomposition model is constructed based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

[0091] Based on the technical solution of the embodiment of this application, during specific implementation, the situation is as follows:

[0092] Step 1: Obtain the amplitude spectrum of background noise.

[0093] The object causing the abnormal noise (i.e., the object producing the abnormal noise) is not experiencing any abnormal noise issues, but only produces its own sound signals. This sound is called background noise. For example, when testing the dynamic abnormal noise performance of a production vehicle on a certain road surface at a certain speed, the vehicle does not produce any abnormal noise, but only the sound generated by the vehicle under the influence of the road surface. The sound signals emitted by the vehicle at this time are considered background noise.

[0094] The embodiment of the present application performs abnormal noise decomposition based on the linear superposition principle of power, and the influence of background noise must be removed; to remove the influence of background noise, it is necessary to first obtain the amplitude spectrum of the background noise.

[0095] The specific process is: determine the sampling frequency of the background noise signal, such as 16000Hz;

[0096] Collect background noise signals;

[0097] Determine the frequency resolution of spectrum analysis, such as 1Hz;

[0098] Perform FFT (Fast Fourier Transform) analysis to obtain the spectrum of the background noise, discard the phase spectrum, and retain the amplitude spectrum (which describes the amplitude of the sound pressure at each frequency);

[0099] Determine the frequency cutoff range of the amplitude spectrum, such as 21-7000 Hz, and take out the sound pressure amplitude at the corresponding frequency. In this way, we get an amplitude spectrum row vector of 6800 (7000-21+1) dimensions, which is called the background noise amplitude spectrum row vector and is denoted as A0.

[0100] In the row vector A0, the jth (1≤j≤6800) element is recorded as A 0j , represents the sound pressure amplitude at the jth frequency in the background noise.

[0101] It should be noted that linear weighting (i.e., no weighting) is used when performing spectrum analysis on background noise. The amplitude of the amplitude spectrum is expressed as an actual physical quantity (i.e., sound pressure) in Pa, rather than dB.

[0102] Step 2: Obtain the amplitude spectrum of each known abnormal noise problem.

[0103] In this step, it is necessary to obtain the amplitude spectrum of each known abnormal noise problem.

[0104] Execute the process of step 1, collect the sound signal of each known abnormal sound problem (that is, the object to which the abnormal sound problem belongs has only a single known abnormal sound problem) separately, perform spectrum analysis, and obtain its amplitude spectrum.

[0105] Assume there are n known abnormal noise problems, and the amplitude spectrum row vector of the i-th (1≤i≤n) known abnormal noise problem is recorded as origin_A i The row vector origin_A i In the example, the jth (1≤j≤6800) element is recorded as origin_A ij , represents the sound pressure amplitude at the jth frequency of the i-th known abnormal sound problem.

[0106] It should be emphasized that when testing and performing spectrum analysis on each known abnormal noise problem, the same sampling frequency, frequency resolution, and frequency intercept range as in step one must be used.

[0107] It should be noted that during the testing and analysis process of this step, the measured sound signals of each known abnormal sound problem contain background noise signals, and the amplitude spectrum obtained is inevitably mixed with the amplitude spectrum of the background noise. In subsequent steps, the background noise will be removed.

[0108] Step 3: Obtain the amplitude spectrum of the abnormal noise problem set to be decomposed.

[0109] The so-called abnormal noise problem set to be decomposed refers to the set of abnormal noise problems that occurred within a certain period of time (the sounds of various abnormal noise problems are mixed together). In the subsequent steps, the abnormal noise problems in this set will be decomposed to identify each known abnormal noise problem.

[0110] Execute the process of step 1, collect the sound signal of the abnormal sound problem set to be decomposed, perform spectrum analysis, and obtain a 6800-dimensional amplitude spectrum row vector, recorded as origin_B, where the jth (1≤j≤6800) element is recorded as origin_B j, represents the sound pressure amplitude at the jth frequency of the abnormal sound problem set to be decomposed.

[0111] Naturally, when testing and performing spectrum analysis on the set of abnormal noise problems to be decomposed, the same sampling frequency, the same frequency resolution, and the same frequency intercept range as in step 1 must be used.

[0112] Same as origin_A in step 2 i In the same situation, origin_B also contains the amplitude spectrum of background noise. In subsequent steps, the background noise will be removed.

[0113] Step 4: Construction of abnormal noise decomposition model.

[0114] In this step, an abnormal noise decomposition model is constructed based on the linear superposition principle of power.

[0115] As described in step 1, A 0j is the sound pressure amplitude of the background noise at the jth frequency;

[0116] Then, A 0j The result of the square operation is A 0j 2 is the sound power of the background noise at the jth frequency;

[0117] Perform square operation on each element in the row vector A0 to form a row vector, denoted as A 02 , which can be called the background noise power spectrum.

[0118] Perform the same operation to form the power spectrum of each known abnormal sound problem, which is a row vector, denoted as origin_A i 2 (1≤i≤n);

[0119] Perform the same operation to form the power spectrum of the abnormal noise problem set to be decomposed, which is a row vector, denoted as origin_B 2 .

[0120] As mentioned above, origin_A i , origin_B is mixed with the amplitude spectrum of background noise, then origin_A i 2 、origin_B 2 The power spectrum of the background noise is also mixed in. Obviously, the power spectrum of the background noise must be removed.

[0121] Will origin_A i 2 Subtract A0 2 , origin_A i 2 Subtract A0 from each element in2 The corresponding elements in form a row vector, which is called the pure power spectrum of each known abnormal sound problem, denoted by A i 2 (1≤i≤n). Similarly, origin_B 2 Subtract A0 2 , which will be origin_B 2 Subtract A0 from each element in 2 The corresponding elements in form a row vector, which is called the pure power spectrum of the abnormal noise problem set to be decomposed, denoted by B 2 This method will be in A i 2 (1≤i≤n), B 2 The abnormal noise decomposition model is constructed based on the

[0122] Due to the influence of various related factors, the occurrence intensity of each known abnormal noise problem is not the same every time it occurs. Assume that the occurrence intensity of the i-th (1≤i≤n) abnormal noise problem in the set of abnormal noise problems to be decomposed is a i , the following abnormal noise decomposition model can be constructed.

[0123]

[0124] Among them, 1≤i≤n;

[0125] A i 2 is the pure power spectrum of the i-th known abnormal noise problem;

[0126] B 2 is the pure power spectrum of the abnormal noise problem set to be decomposed.

[0127] In this step, A has been determined i (1≤i≤n) and the value of B, only a i is the unknown quantity. Next, we need to calculate a i , so that Equation 1 holds true.

[0128] Step 5: Decomposition of abnormal noise.

[0129] In formula 1, there are n unknowns a i (1≤i≤n), but there is only one equation. Therefore, its analytical solution cannot be found.

[0130] In the embodiment of the present application, a numerical solution is used to solve the problem. i The range of its value is [0,1.5], with an interval of 0.1, so its value sequence is {0,0.1,0.2……1.5}, a total of 16 values. Construct a row vector sth=[a1,…a i …a n], is called the occurrence intensity vector. Therefore, Equation 1 can be rewritten as follows.

[0131] sth×E_matrix=B 2 Formula 2

[0132] in It can be called the power spectrum matrix.

[0133] As mentioned above, a i (1≤i≤n) has 16 values, so sth has 16 values. n These values are sequentially organized into a matrix called the occurrence intensity matrix, denoted as sth_value, where each row is a value of the occurrence intensity vector sth. Let the elements of the first row of sth_value be all 0, that is, the first value of the occurrence intensity vector sth, and all its elements are 0. The k-th row of the occurrence intensity matrix (the k-th value of sth, 1≤k≤16 n ), denoted as sth_value(k). Substituting it into the left side of Equation 2, the result can be called the synthetic power spectrum, denoted as E_syn(k), that is,

[0134] E_syn(k)=sth_value(k)×E_matrix Formula 3

[0135] The correlation coefficient between E_syn(k) and B is calculated using the following formula.

[0136]

[0137] Where γ(k) is the correlation coefficient, 2≤k≤16 n , used to describe the similarity between E_syn(k) and B. The larger its value, the more similar the two are;

[0138] E_syn(k,j) is the j-th element of the row vector E_syn(k), 2≤k≤16 n , 1≤j≤6800.

[0139] For formula 4, traverse k (2≤k≤16 n ), get all the correlation coefficients γ (γ can be regarded as a vector);

[0140] Take the largest one in γ and record its serial number as max_val_seq, that is, the largest correlation coefficient is γ(max_val_seq), and its value is recorded as γ_max;

[0141] Its corresponding occurrence intensity vector is sth_value(max_val_seq), and its value is denoted as sth_value_result, which is the value to be finally obtained. Obviously, γ_max is a numerical value, and sth_value_result is a row vector.

[0142] Define a constant, called the correlation coefficient limit value, denoted as corr_limit, and its typical value is 0.75.

[0143] If γ_max < corr_limit, it means that the similarity between E_syn(max_val_seq) and B is low. The reason may be that unknown abnormal sound problems are concentrated in the abnormal sound problems to be decomposed, and the decomposition of the abnormal sound problem set to be decomposed fails. At this time, it is necessary to first analyze the unknown abnormal sound problems mixed in, then execute step two, and finally decompose again.

[0144] If γ_max ≥ corr_limit, it means that the similarity between E_syn(max_val_seq) and B is high, and the decomposition of the abnormal sound problem set to be decomposed is successful. At this time, the magnitudes of the elements in sth_value_result describe the participation degrees of the known abnormal sound problems in the abnormal sound problem set to be decomposed; the larger the value of the element, the higher the participation degree of the corresponding abnormal sound problem; the smaller the value of the element, the lower the participation degree of the corresponding abnormal sound problem.

[0145] For Equation 4, when traversing k, the value range of k is 2 ≤ k ≤ 16 n , because when k = 1, all elements of sth_value(1) are 0. Substituting sth_value(1) into Equation 3, all elements of E_syn(1) are 0. Substituting E_syn(1) into Equation 4, an overflow (a calculation error of 0 / 0) will occur on the right side of Equation 4. Therefore, k does not start from 1.

[0146] sth_value(1) indicates that the occurrence intensities of n known abnormal sound problems are all 0, that is, no known abnormal sound problem occurs. That is, for Equation 1, sth_value(1) is the exact solution corresponding to the situation where no known abnormal sound problem occurs.

[0147] Define a constant called the occurrence intensity limit, denoted as sth_limit, with a typical value of 0.2. If the value of an element in sth_value_result is less than sth_limit, it indicates that the corresponding abnormal noise issue has a very low level of involvement and activity, and it can be considered that the abnormal noise has not occurred. In particular, if the values of all elements in sth_value_result are less than sth_limit, it indicates that no abnormal noise issue has occurred. In this case, for Equation 1, sth_value_result is an approximate solution (or alternative solution) to the absence of any known abnormal noise issue.

[0148] In summary, it should be noted that the sampling frequency (16000 Hz), frequency resolution (1 Hz), and amplitude spectrum frequency cutoff range (21-7000 Hz) in step 1, and the occurrence intensity value range ([0,1.5]) and occurrence intensity value interval (0.1) in step 5 are not fixed and can be set according to actual scenarios and actual needs.

[0149] Second, see Figure 2 As shown, an embodiment of the present application provides an abnormal sound decomposition device, which includes:

[0150] An amplitude spectrum acquisition module, which is used to obtain the amplitude spectrum of background noise;

[0151] The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of each known abnormal sound problem;

[0152] The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of the abnormal noise problem set to be decomposed;

[0153] A model building module, which is used to build an abnormal noise decomposition model;

[0154] The abnormal noise decomposition module is used to decompose the abnormal noise based on the abnormal noise decomposition model.

[0155] In the embodiment of the present application, spectrum analysis is performed on the background noise, each known abnormal sound problem, and the set of abnormal sound problems to be decomposed to obtain their amplitude spectra;

[0156] Based on each amplitude spectrum, background noise is eliminated to obtain the pure power spectrum of each known abnormal noise problem and the set of abnormal noise problems to be decomposed;

[0157] Based on the linear superposition principle of power, an abnormal noise decomposition model was constructed by introducing the concept of occurrence intensity of each known abnormal noise problem.

[0158] By discretizing the occurrence intensity variable, the synthetic power spectrum sequence is obtained;

[0159] The synthetic power spectrum sequence was correlated with the pure power spectrum of the abnormal noise problem set to be decomposed, and the optimal abnormal noise problem occurrence intensity was obtained. The participation degree of each known abnormal noise problem was obtained, and the decomposition of the abnormal noise was achieved.

[0160] The technical solution of the embodiment of the present application performs spectral analysis on background noise, various known abnormal noise problems and the set of abnormal noise problems to be decomposed, and performs correlation analysis on the synthetic power spectrum sequence and the pure power spectrum of the set of abnormal noise problems to be decomposed, thereby realizing the decomposition of abnormal noise, effectively simplifying the process and improving efficiency.

[0161] In summary, the technical principles of the embodiments of the present application are simple and clear, the calculation is simple, and the numerical analysis method is used to replace the complicated theoretical analysis, making the abnormal noise decomposition feasible, simple and efficient.

[0162] Furthermore, the amplitude spectrum acquisition module is further configured to acquire a background noise signal based on a set sampling frequency corresponding to the background noise;

[0163] The amplitude spectrum acquisition module is further configured to analyze the background noise signal based on a set frequency resolution of the spectrum analysis to obtain a spectrum corresponding to the background noise;

[0164] The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the background noise based on the frequency spectrum corresponding to the background noise.

[0165] Furthermore, the amplitude spectrum acquisition module is further configured to acquire noise signals of known abnormal sound problems based on a set sampling frequency corresponding to each of the known abnormal sound problems;

[0166] The amplitude spectrum acquisition module is further configured to analyze the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each known abnormal sound problem, and obtain the spectrum corresponding to the known abnormal sound problem;

[0167] The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the known abnormal sound problem based on the frequency spectrum corresponding to the known abnormal sound problem.

[0168] Furthermore, the amplitude spectrum acquisition module is further configured to perform abnormal sound monitoring based on a set sampling frequency, and acquire noise signals of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed;

[0169] The amplitude spectrum acquisition module is further used to perform spectrum analysis based on the noise signal of the abnormal sound problem set to be decomposed, and obtain the amplitude spectrum of the abnormal sound problem set to be decomposed.

[0170] Furthermore, the model building module is further used to obtain the power spectrum of the background noise based on the amplitude spectrum of the background noise;

[0171] The model building module is further configured to obtain a pure power spectrum of each known abnormal sound problem based on the power spectrum of the background noise and the amplitude spectrum of each known abnormal sound problem;

[0172] The model building module is further configured to obtain a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed;

[0173] The model building module is further configured to build the abnormal noise decomposition model based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

[0174] It should be noted that the abnormal noise decomposition device mentioned in the second aspect is similar to the technical principles of the abnormal noise decomposition method mentioned in the first aspect in terms of technical issues, technical means and technical effects, and will not be elaborated here.

[0175] In the description of this application, it should be noted that the terms "upper" and "lower" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. Unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be internal communication between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to the specific circumstances.

[0176] It should be noted that, in this application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0177] The foregoing is merely a list of specific embodiments of the present application, intended to enable those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the broadest scope consistent with the principles and novel features of the present application.

Claims

1. A method for decomposing abnormal noise, characterized in that: The method comprises the following steps: Obtain the amplitude spectrum of background noise; Obtain the amplitude spectrum of each known abnormal noise problem; Obtain the amplitude spectrum of the abnormal noise problem set to be decomposed; Construct abnormal noise decomposition model; Based on the abnormal noise decomposition model, abnormal noise decomposition is performed.

2. The abnormal noise decomposition method according to claim 1, characterized in that: The obtaining of the amplitude spectrum of the background noise comprises the following steps: Based on the set sampling frequency corresponding to the background noise, a background noise signal is collected and obtained; Analyzing the background noise signal based on a set frequency resolution of spectrum analysis to obtain a spectrum corresponding to the background noise; Based on the frequency spectrum corresponding to the background noise, an amplitude spectrum of the background noise is obtained.

3. The abnormal noise decomposition method according to claim 1, characterized in that: The method of obtaining the amplitude spectrum of each known abnormal sound problem includes the following steps: Based on the set sampling frequency corresponding to each of the known abnormal sound problems, a noise signal of the known abnormal sound problem is collected; Analyzing the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each of the known abnormal sound problems to obtain the spectrum corresponding to the known abnormal sound problem; Based on the frequency spectrum corresponding to the known abnormal sound problem, an amplitude spectrum of the known abnormal sound problem is obtained.

4. The abnormal noise decomposition method according to claim 1, characterized in that: The method of obtaining the amplitude spectrum of the abnormal noise problem set to be decomposed comprises the following steps: Perform abnormal sound monitoring based on a set sampling frequency to collect and obtain noise signals of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed; A spectrum analysis is performed based on the noise signal of the abnormal sound problem set to be decomposed to obtain an amplitude spectrum of the abnormal sound problem set to be decomposed.

5. The abnormal noise decomposition method according to claim 1, characterized in that: The construction of the abnormal noise decomposition model includes the following steps: Obtaining a power spectrum of the background noise based on the amplitude spectrum of the background noise; Based on the power spectrum of the background noise and the amplitude spectrum of each of the known abnormal sound problems, a pure power spectrum of each of the known abnormal sound problems is obtained; Obtaining a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed; The abnormal noise decomposition model is constructed based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

6. An abnormal noise decomposition device, characterized in that: The device comprises: An amplitude spectrum acquisition module, which is used to obtain the amplitude spectrum of background noise; The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of each known abnormal sound problem; The amplitude spectrum acquisition module is also used to obtain the amplitude spectrum of the abnormal noise problem set to be decomposed; A model building module, which is used to build an abnormal noise decomposition model; The abnormal noise decomposition module is used to decompose the abnormal noise based on the abnormal noise decomposition model.

7. The abnormal noise decomposition device according to claim 6, characterized in that: The amplitude spectrum acquisition module is further configured to acquire a background noise signal based on a set sampling frequency corresponding to the background noise; The amplitude spectrum acquisition module is further configured to analyze the background noise signal based on a set frequency resolution of the spectrum analysis to obtain a spectrum corresponding to the background noise; The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the background noise based on the frequency spectrum corresponding to the background noise.

8. The abnormal noise decomposition device according to claim 6, characterized in that: The amplitude spectrum acquisition module is further configured to acquire noise signals of known abnormal sound problems based on a set sampling frequency corresponding to each known abnormal sound problem; The amplitude spectrum acquisition module is further configured to analyze the noise signal of the known abnormal sound problem based on the set frequency resolution of the spectrum analysis corresponding to each known abnormal sound problem, and obtain the spectrum corresponding to the known abnormal sound problem; The amplitude spectrum acquisition module is further configured to obtain the amplitude spectrum of the known abnormal sound problem based on the frequency spectrum corresponding to the known abnormal sound problem.

9. The abnormal noise decomposition device according to claim 6, characterized in that: The amplitude spectrum acquisition module is further used to perform abnormal sound monitoring based on a set sampling frequency, and collect and obtain the noise signal of the abnormal sound problem set to be decomposed corresponding to the abnormal sound problem set to be decomposed; The amplitude spectrum acquisition module is further used to perform spectrum analysis based on the noise signal of the abnormal sound problem set to be decomposed, and obtain the amplitude spectrum of the abnormal sound problem set to be decomposed.

10. The abnormal noise decomposition device according to claim 6, characterized in that: The model building module is further configured to obtain a power spectrum of the background noise based on the amplitude spectrum of the background noise; The model building module is further configured to obtain a pure power spectrum of each known abnormal sound problem based on the power spectrum of the background noise and the amplitude spectrum of each known abnormal sound problem; The model building module is further configured to obtain a pure power spectrum of the abnormal sound problem set to be decomposed based on the power spectrum of the background noise and the amplitude spectrum of the abnormal sound problem set to be decomposed; The model building module is further configured to build the abnormal noise decomposition model based on the pure power spectrum of each of the known abnormal noise problems and the pure power spectrum of the set of abnormal noise problems to be decomposed.

Images