CN114429769A - Glass breaking sound detection method and related equipment - Google Patents

Abstract

The application is suitable for the technical field of audio processing, and provides a glass breaking sound detection method and related equipment_pAnd if the attenuation quantity of the energy of each audio frame relative to the peak value of the popping sound is within a preset attenuation range, judging that the sound signal is the glass breaking sound. Thereby eliminating the burst sound as short burst sound, improving the accuracy of detecting the broken sound of the glass, and determining the audio frame according to the frequency spectrum parameters of the audio frameThe energy attenuation amount and the operation process are simple.

Description

Glass breaking sound detection method and related equipment

technical field

The present application belongs to the technical field of audio processing, and in particular relates to a method for detecting the sound of broken glass and related equipment.

Background technique

The traditional glass breaking detection method is generally to install a piezoelectric film module on the glass, and determine whether the glass is broken according to the current generated by the piezoelectric film module. Although this method can effectively detect the sound of glass breaking, the piezoelectric film module Installation is troublesome and affects the appearance. In order to avoid installing the piezoelectric film module, it is possible to detect whether the glass is broken by judging whether the obtained sound signal contains the sound of glass breaking. Existing methods for detecting glass breaking based on sound signals are generally processing methods based on neural networks, or determine whether it is the sound of glass breaking according to the strength of the sound signal. However, the processing method based on the neural network has the problem of complicated operation, and the method of determining whether it is the sound of glass breaking according to the strength of the sound signal has the problem of inaccurate detection.

SUMMARY OF THE INVENTION

In view of this, the embodiments of the present application provide a method for detecting the sound of glass breaking and related equipment, which can improve the accuracy of detecting the sound of glass breaking, and the calculation is simple.

A first aspect of the embodiments of the present application provides a method for detecting the sound of broken glass, including:

Acquiring a sound signal, performing frame-by-frame processing on the sound signal, and obtaining M audio frames, where M represents an integer greater than 0;

If it is determined that the current audio frame is a crackling sound according to the spectrum parameters of the current audio frame, then according to the spectrum parameters of the current audio frame and the spectrum parameters of the audio frames after the current audio frame, the crackling sound is determined peak;

If the energy of the N _p audio frames after the current audio frame is within the preset attenuation range relative to the attenuation of the peak value of the crackling sound, it is determined that the sound signal is a glass breaking sound, and the N _p represents an integer greater than 0, and M>N _p .

In a possible implementation manner of the first aspect, if the current audio frame is determined to be a crackling sound according to the frequency spectrum parameter of the current audio frame, then according to the frequency spectrum parameter of the current audio frame, and the Spectral parameters of the audio frame following the current audio frame, determining the crackling peak, including:

If it is determined according to the spectral parameters of the current audio frame that the current audio frame is a crackling sound, and white noise is detected in N _B audio frames after the current audio frame, then according to the current audio frame The frequency spectrum parameter of , and the frequency spectrum parameter of the audio frame after the current audio frame, determine the crackle peak value, the NB represents an integer greater than 0, and _M > _NB .

In a possible implementation manner of the first aspect, the audio frame includes at least one frequency band, the frequency spectrum parameter includes energy of the frequency band, and the audio frames in N _B audio frames following the current audio frame White noise detected, including:

If there is a target audio frame in N _B audio frames after the current audio frame, then determine the calculation formula of the first count value, and the target audio frame is the audio frequency that satisfies the first preset relational expression between the energies of the frequency bands frame, the calculation formula of the first calculation value corresponds to the first preset relational formula;

The first count value is determined according to the calculation formula of the first count value, and if the first count value is greater than a preset first threshold, it is determined that the target audio frame is white noise.

In a possible implementation manner of the first aspect, the audio frame includes at least one subframe, the spectrum parameter includes energy of the subframe, the audio frame includes at least one frequency band, and the spectrum parameter includes the The energy of the frequency band, and determining that the current audio frame is a crackling sound according to the frequency spectrum parameter of the current audio frame, including:

If the maximum value of the energy of the subframes of the current audio frame is greater than the preset second threshold, then:

Determine the energy ratio of the current audio frame according to the energy of the subframes of the current audio frame;

According to the energy of the frequency band of the current audio frame and the energy of the frequency band of the previous audio frame of the current audio frame, determine the energy ratio of the current audio frame to the previous audio frame;

Determine the energy difference between the current audio frame and the previous audio frame according to the energy of the subframe of the current audio frame and the energy of the subframe of the previous audio frame;

Determine the energy ratio of the previous audio frame according to the energy of the subframe of the previous audio frame;

If the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the current audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the previous audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the current audio frame is greater than a fourth preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the previous audio frame is greater than a fourth preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, And the energy ratio of the current audio frame is greater than the seventh preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, and the energy ratio of the previous audio frame is greater than the seventh preset value;

Then it is determined that the current audio frame is a crackling sound.

In a possible implementation manner of the first aspect, the audio frame includes at least one frequency band, the frequency spectrum parameter includes energy of the frequency band, the frequency spectrum parameter according to the current audio frame, and the current frequency The spectral parameters of the audio frame following the audio frame, determining the crackling peak, including:

Taking the energy sum of each frequency band of the current audio frame as the peak of the crackling sound;

Taking the next audio frame of the current audio frame as the first audio frame;

calculating the energy summation of each frequency band of the first audio frame;

If the sum of the energy of each frequency band of the first audio frame is greater than the peak value of the crackling sound, updating the peak value of the crackling sound to the sum of the energy of each frequency band of the first audio frame;

The following audio frame of the first audio frame is used as the first audio frame, and the step of calculating the sum of the energy of each frequency band of the first audio frame and the subsequent steps are returned to until the preset end condition is satisfied.

In a possible implementation manner of the first aspect, the audio frame includes at least one frequency band, the spectrum parameter includes energy of the frequency band, and the frequency of the audio frames after the current _audio frame If the attenuation of the energy relative to the peak value of the crackling sound is all within the preset attenuation range, it is determined that the sound signal is the sound of glass breaking, including:

If the energy of the N _p audio frames following the current audio frame is within the preset attenuation range relative to the attenuation of the crackling peak, and the sound signal satisfies the first preset condition, it is determined that the The sound signal is glass breaking sound; the first preset condition includes any one or more of the following: the duration of the sound signal is within the preset duration, and the spectral parameters of the sound signal conform to preset spectral characteristics , the energy difference between the frequency bands of the sound signal is within a first preset range, the difference between the target frequencies of the sound signal is within a second preset range, and the target frequency is the energy peak corresponding to the frequency band frequency.

In a possible implementation manner of the first aspect, the duration of the sound signal is within a preset duration, including:

calculating the decay speed and/or decay time of the energy of the frequency band of the sound signal relative to the crackle peak;

If the decay speed is within the preset speed range, and/or the decay time is within the preset time range, it is determined that the duration of the sound signal is within the preset duration.

In a possible implementation manner of the first aspect, the spectral parameters of the sound signal conform to preset spectral characteristics, including:

determining the number of audio frames in the sound signal whose energy between frequency bands satisfies the second preset relational expression;

If the number of audio frames satisfying the second preset relational expression and the duration of the sound signal satisfy the third predetermined relational expression, it is determined that the energy of the first preset frequency band of the sound signal conforms to the preset spectral characteristics.

In a possible implementation manner of the first aspect, the energy difference between frequency bands of the sound signal is within a first preset range, including:

If in the sound signal, the energy of the frequency band of the n-th audio frame and the energy of the frequency band of the n-1-th audio frame satisfy the fourth preset relational expression,

Or the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame satisfy the fifth preset relational expression, then update the set second count value, and the n represents an integer greater than 2 , and M≥n;

If a sixth preset relational expression is satisfied between the second count value and the duration of the sound signal, it is determined that the energy difference between the frequency bands of the sound signal is within a first preset range.

In a possible implementation manner of the first aspect, the difference between the target frequencies of the sound signals is within a second preset range, including:

If in the sound signal, the difference between the target frequency of the ith frequency band of the n th audio frame and the target frequency of the ith frequency band of the n-1 th audio frame is within the preset difference range,

Or the difference between the target frequency of the ith frequency band of the nth audio frame and the target frequency of the ith frequency band of the n-2th audio frame is within the preset difference range, then update the set third count value, so The n represents an integer greater than 2, and M≥n, and the i represents an integer greater than 0;

If a seventh preset relational expression is satisfied between the third count value and the duration of the sound signal, it is determined that the difference between the target frequencies of the sound signal is within a second preset range.

A second aspect of the embodiments of the present application provides a glass breaking sound detection device, characterized in that it includes:

an acquisition module, configured to acquire a sound signal, perform frame-by-frame processing on the sound signal, and obtain M audio frames, where M represents an integer greater than 0;

The calculation module is used to determine that the current audio frame is a crackling sound according to the frequency spectrum parameter of the current audio frame, then according to the frequency spectrum parameter of the current audio frame, and the frequency spectrum of the audio frame after the current audio frame parameter to determine the crackling peak;

A determination module for determining that the sound signal is a glass breaking sound if the energy of the N _p audio frames after the current audio frame is within a preset attenuation range relative to the attenuation of the crackling sound peak , the N _p represents an integer greater than 0, and M>N _p .

In a possible implementation manner of the second aspect, the computing module includes a first computing unit, and the first computing unit is configured to:

If it is determined according to the spectral parameters of the current audio frame that the current audio frame is a crackling sound, and white noise is detected in N _B audio frames after the current audio frame, then according to the current audio frame The frequency spectrum parameter of , and the frequency spectrum parameter of the audio frame after the current audio frame, determine the crackle peak value, the NB represents an integer greater than 0, and _M > _NB .

In a possible implementation manner of the second aspect, the audio frame includes at least one frequency band, the spectrum parameter includes energy of the frequency band, and the first calculation unit is specifically configured to: if the current audio frame There is a target audio frame in the following N _B audio frames, then the calculation formula of the first count value is determined, and the target audio frame is an audio frame that satisfies the first preset relational expression between the energies of the frequency bands, and the first calculation The calculation formula of the value corresponds to the first preset relational expression;

The first count value is determined according to the calculation formula of the first count value, and if the first count value is greater than a preset first threshold, it is determined that the target audio frame is white noise.

In a possible implementation manner of the second aspect, the audio frame includes at least one subframe, the spectrum parameter includes energy of the subframe, the audio frame includes at least one frequency band, and the spectrum parameter includes the the energy of the frequency band, the calculation module further includes a second calculation unit, and the second calculation unit is specifically used for:

If the maximum value of the energy of the subframes of the current audio frame is greater than the preset second threshold, then:

Determine the energy ratio of the current audio frame according to the energy of the subframes of the current audio frame;

According to the energy of the frequency band of the current audio frame and the energy of the frequency band of the previous audio frame of the current audio frame, determine the energy ratio of the current audio frame to the previous audio frame;

Determine the energy difference between the current audio frame and the previous audio frame according to the energy of the subframe of the current audio frame and the energy of the subframe of the previous audio frame;

Determine the energy ratio of the previous audio frame according to the energy of the subframe of the previous audio frame;

If the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the current audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the previous audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the current audio frame is greater than a fourth preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the previous audio frame is greater than a fourth preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, And the energy ratio of the current audio frame is greater than the seventh preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, and the energy ratio of the previous audio frame is greater than the seventh preset value;

Then it is determined that the current audio frame is a crackling sound.

In a possible implementation manner of the second aspect, the audio frame includes at least one frequency band, the spectral parameter includes energy of the frequency band, and the calculation unit further includes a third calculation unit, the third calculation unit Specifically for:

Taking the energy sum of each frequency band of the current audio frame as the peak of the crackling sound;

Taking the next audio frame of the current audio frame as the first audio frame;

calculating the energy summation of each frequency band of the first audio frame;

If the sum of the energy of each frequency band of the first audio frame is greater than the peak value of the crackling sound, updating the peak value of the crackling sound to the sum of the energy of each frequency band of the first audio frame;

The following audio frame of the first audio frame is used as the first audio frame, and the step of calculating the sum of the energy of each frequency band of the first audio frame and the subsequent steps are returned to until the preset end condition is satisfied.

In a possible implementation manner of the second aspect, the audio frame includes at least one frequency band, the spectrum parameter includes energy of the frequency band, and the determining module is specifically configured to:

If the energy of the N _p audio frames following the current audio frame is within the preset attenuation range relative to the attenuation of the crackling peak, and the sound signal satisfies the first preset condition, it is determined that the The sound signal is glass breaking sound; the first preset condition includes any one or more of the following: the duration of the sound signal is within the preset duration, and the spectral parameters of the sound signal conform to preset spectral characteristics , the energy difference between the frequency bands of the sound signal is within a first preset range, the difference between the target frequencies of the sound signal is within a second preset range, and the target frequency is the energy peak corresponding to the frequency band frequency.

In a possible implementation manner of the second aspect, the determining module is further configured to:

calculating the decay speed and/or decay time of the energy of the frequency band of the sound signal relative to the crackle peak;

If the decay speed is within the preset speed range, and/or the decay time is within the preset time range, it is determined that the duration of the sound signal is within the preset duration.

In a possible implementation manner of the second aspect, the determining module is further configured to:

determining the number of audio frames in the sound signal whose energy between frequency bands satisfies the second preset relational expression;

If the number of audio frames satisfying the second preset relational expression and the duration of the sound signal satisfy the third predetermined relational expression, it is determined that the energy of the first preset frequency band of the sound signal conforms to the preset spectral characteristics.

In a possible implementation manner of the second aspect, the determining module is further configured to:

If in the sound signal, the energy of the frequency band of the n-th audio frame and the energy of the frequency band of the n-1-th audio frame satisfy the fourth preset relational expression,

Or the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame satisfy the fifth preset relational expression, then update the set second count value, and the n represents an integer greater than 2 , and M≥n;

If a sixth preset relational expression is satisfied between the second count value and the duration of the sound signal, it is determined that the energy difference between the frequency bands of the sound signal is within a first preset range.

In a possible implementation manner of the second aspect, the determining module is further configured to:

If in the sound signal, the difference between the target frequency of the ith frequency band of the n th audio frame and the target frequency of the ith frequency band of the n-1 th audio frame is within the preset difference range,

Or the difference between the target frequency of the ith frequency band of the nth audio frame and the target frequency of the ith frequency band of the n-2th audio frame is within the preset difference range, then update the set third count value, so The n represents an integer greater than 2, and M≥n, and the i represents an integer greater than 0;

If a seventh preset relational expression is satisfied between the third count value and the duration of the sound signal, it is determined that the difference between the target frequencies of the sound signal is within a second preset range.

A third aspect of the embodiments of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program The method for detecting the sound of glass breaking according to the first aspect is realized.

A fourth aspect of the embodiments of the present application provides a glass breaking sound detection system, including: a sound collection device, an alarm device, and the electronic device according to the third aspect, the electronic device and the sound collection device and The alarm device is communicatively connected.

A fifth aspect of the embodiments of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the glass breaking according to the above first aspect is realized Sound detection method.

A sixth aspect of the embodiments of the present application provides a computer program product that, when the computer program product runs on an electronic device, causes the electronic device to execute the method for detecting the sound of glass breaking according to the first aspect.

Compared with the prior art, the embodiment of the present application has the following beneficial effects: by acquiring the sound signal, the sound signal is framed to obtain M audio frames, if the current audio frame is determined according to the frequency spectrum parameter of the current audio frame as crackling sound, then according to the spectrum parameters of the current audio frame and the spectrum parameters of the audio frame after the current audio frame, determine the peak value of the crackling sound, if the energy of the N _p audio frames after the current audio frame is relative to the If the attenuation of the peak value of the crackling sound is all within the preset attenuation range, it is determined that the sound signal is the sound of glass breaking. Since some short pops such as clapping, knocking, raining, etc. are obtained, the popping sound will also be detected, and the energy of the short popping sound is attenuated faster than the glass breaking sound. Therefore, when the popping sound is detected , further calculate the attenuation of the energy of N _p audio frames after the current audio frame relative to the peak value of the crackling sound, and determine that the sound signal is glass breaking sound when the attenuation is within a preset range. That is, since the situation where the popping sound is a short popping sound can be excluded, the accuracy of detecting the sound of glass breaking can be improved, and since the attenuation amount of the energy of the audio frame can be determined according to the spectral parameter of the audio frame, the The neural network algorithm determines whether the glass is broken or not, and the operation process is simpler.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the accompanying drawings that are required to be used in the description of the embodiments or the prior art.

1 is a schematic diagram of a glass breaking sound detection system provided by an embodiment of the present application;

2 is a schematic diagram of a sound collection device provided by an embodiment of the present application;

FIG. 3 is a schematic flowchart of the realization of the method for detecting the sound of broken glass provided by an embodiment of the present application;

4 is a flowchart of detecting whether the sound signal is a short pop provided by an embodiment of the present application;

5 is a flowchart of detecting whether the current audio frame is a crackling sound provided by an embodiment of the present application;

Fig. 6 is the detection flow chart of the white noise provided by the embodiment of the present application;

7 is a flow chart of a method for detecting a sound of broken glass provided by another embodiment of the present application;

8 is a flowchart of determining whether a sound signal is within a preset duration provided by an embodiment of the present application;

FIG. 9 is a flow chart of judging whether a spectral parameter of a sound signal conforms to a preset spectral characteristic provided by an embodiment of the present application;

10 is a flowchart of determining that the energy difference between frequency bands of a sound signal is within a first preset range provided by an embodiment of the present application;

11 is a flowchart for judging that the difference between target frequencies of sound signals is within a second preset range provided by an embodiment of the present application;

12 is a schematic diagram of a glass breaking sound detection device provided by an embodiment of the present application;

FIG. 13 is a schematic diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

In the following description, for the purpose of illustration rather than limitation, specific details such as a specific system structure and technology are set forth in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

In order to illustrate the technical solutions described in the present application, the following specific embodiments are used for description.

It is to be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described feature, integer, step, operation, element and/or component, but does not exclude one or more other features , whole, step, operation, element, component and/or the presence or addition of a collection thereof.

It should also be understood that the terminology used in the specification of the application herein is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural unless the context clearly dictates otherwise.

It should also be further understood that, as used in this specification and the appended claims, the term "and/or" refers to and including any and all possible combinations of one or more of the associated listed items .

As used in this specification and the appended claims, the term "if" may be contextually interpreted as "when" or "once" or "in response to determining" or "in response to detecting" . Similarly, the phrases "if it is determined" or "if the [described condition or event] is detected" may be interpreted, depending on the context, to mean "once it is determined" or "in response to the determination" or "once the [described condition or event] is detected. ]" or "in response to detection of the [described condition or event]".

In addition, in the description of the present application, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

Existing methods for detecting glass breaking based on sound signals are generally processing methods based on neural networks, or determine whether it is the sound of glass breaking according to the strength of the sound signal. However, the processing method based on the neural network has the problem of complicated operation, and the method of determining whether it is the sound of glass breaking according to the strength of the sound signal has the problem of inaccurate detection. To this end, the embodiments of the present application provide a method for detecting the sound of broken glass, which can improve the accuracy of detecting the sound of broken glass, and the operation is simple.

The glass breaking sound detection method provided by the embodiment of the present application is applied to a glass breaking sound detection system. As shown in FIG. 1 , the glass breaking sound detection system provided by the embodiment of the present application includes a sound collection device 100, an electronic device 200, and an alarm. The device 300 and the electronic device 200 are connected in communication with the sound collection device 100 and the alarm device 300 . The sound collection device 100 , the electronic device 200 and the alarm device 300 may be integrated on the same device, or may be independent devices. If the electronic device 200 is an independent device, the electronic device 200 may be a computing device such as a desktop computer, a notebook computer, a palmtop computer, and a cloud server.

As shown in FIG. 2 , in a possible implementation manner, the sound collection device 100 includes a microphone 11 , an amplifier 12 , a filter 13 and an analog-to-digital converter 14 . The microphone 11 is used to collect sound, convert the collected sound into a current signal, and input the current signal to the amplifier 12 . The amplifier 12 is used to adjust the current signal according to the set amplification ratio, and input the adjusted current signal to the filter 13 . The filter 13 is used to perform signal enhancement, signal equalization and filtering processing on the adjusted current signal to obtain a filtered signal, and input the filtered signal to the analog-to-digital converter 14 . The filtered signal is an analog signal, and the analog-to-digital converter 14 is used to convert the filtered signal into a digital signal according to the set sampling frequency and number of bits, and output the digital signal to the electronic device 200, and the electronic device 200 then converts the filtered signal into a digital signal. The digital signal is converted into an analog signal, and the analog signal is the sound signal described below.

The electronic device 200 is used to determine whether the received sound signal is a glass breaking sound, and if it is determined that the received sound signal is a glass breaking sound, a glass breaking alarm is generated, and the glass breaking alarm is sent to the alarm device 300, and the alarm device 300 is used for Sound or light is emitted according to the glass breaking alarm to remind the user that the glass is broken.

As shown in FIG. 3 , the method for detecting the sound of broken glass provided by an embodiment of the present application includes:

S101: Acquire a sound signal, perform frame segmentation processing on the sound signal, and obtain M audio frames, where M represents an integer greater than 0.

Among them, the duration of each audio frame can be 30ms, and a Hamming window can be used to process the sound signal in frames. Since the Hamming window has the characteristics of reducing information loss, the Hamming window is used to process the sound signal. Framing processing can reduce the spectral distortion of the sound signal.

In a possible implementation manner, after the sound signal is processed by framing, spectrum compensation processing is performed on the audio frame to compensate the distortion of the sound signal, so that the amplitude response of the processed sound signal in each frequency band tends to be close to the state .

S102: If it is determined that the current audio frame is a crackling sound according to the frequency spectrum parameter of the current audio frame, then according to the frequency spectrum parameter of the current audio frame and the frequency spectrum parameter of the audio frame after the current audio frame, determine the peak value of the crackling sound.

Specifically, Fourier transform processing and sub-framing processing are performed on the audio frame to obtain spectral parameters of the audio frame. Among them, the audio frame is a time-domain signal, and by performing Fourier transform processing on the audio frame, the frequency-domain signal corresponding to the audio frame can be obtained. The frequency-domain signal is divided into several frequency bands, and the amplitude of each frequency in each frequency band is averaged. band energy. According to the divided frequency bands, the peak value of each frequency band and the frequency corresponding to the peak value can also be determined. The audio frame is divided into frames again to obtain at least one subframe, and the root mean square of the waveform data of the time domain signal of the subframe is obtained to obtain the energy of each subframe. The energy of the frequency band, the peak value of the frequency band, the frequency corresponding to the peak value, and the energy of the subframe are all spectral parameters.

In a possible implementation manner, if it is detected that the maximum value of the energy of the subframes of the current audio frame is greater than a preset value, it is determined that the current audio frame is a crackling sound. After it is determined that the current audio frame is a crackling sound, the energy sum of the frequency bands of the current audio frame may be determined as a crackling sound peak value. Alternatively, after determining the total energy of the frequency bands of the current audio frame as the peak of the crackling sound, take the next audio frame of the current audio frame as the first audio frame, and calculate the sum of the energy of each frequency band of the first audio frame. The sum of the energy of each frequency band of an audio frame is greater than the peak value of the crackling sound, then the peak value of the crackling sound is updated to the sum of the energy of each frequency band of the first audio frame, and then the next audio frame of the first audio frame is used as the first audio frame, Return to perform the step of calculating the energy sum of each frequency band of the first audio frame and the subsequent steps until the preset end condition is satisfied. Wherein, the preset end condition may be when the preset duration is reached, or the sum of the energy of each frequency band of the first audio frame reaches the preset energy range.

S103: If the energy of the N _p audio frames after the current audio frame is within the preset attenuation range relative to the attenuation of the crackling sound peak, then determine that the sound signal is the sound of glass breaking, and N _p represents an integer greater than 0, and M>N _p .

Specifically, the attenuation amount may be calculated according to the ratio or difference between the energy of N _p audio frames following the current audio frame and the peak value of the crackling sound. Wherein, N _p is a preset empirical value, for example, N _p =4.

In a possible implementation, starting from the next audio frame of the current audio frame, calculate the ratio of the energy sum of the frequency bands of each audio frame to the peak value of the crackling sound. If the ratio is greater than the preset ratio, it indicates that the corresponding audio frequency The attenuation of the energy of the frame is within the preset attenuation range. If the ratio of the sum of the energy of each audio frame to the peak value of the crackling sound in the N _p audio frames after the current audio frame is greater than the preset ratio, it is determined that the sound signal For the sound of glass breaking.

In the above-mentioned embodiment, when some short pops such as clap, knocking, bottle pinching, mahjong, coin landing, and rain are obtained, popping sounds are also detected, and the short popping sounds are relative to glass. The energy of the broken sound decays faster, therefore, when the crackling sound is detected, the attenuation of the energy of the N _p audio frames after the current audio frame relative to the peak value of the crackling sound is further calculated. Within the preset range, it is determined that the sound signal is the sound of breaking glass. Otherwise, the sound signal is not the sound of breaking glass, which can eliminate the situation that the popping sound is a short popping sound, and improve the accuracy of detecting the sound of breaking glass. The attenuation of the energy of the audio frame can be determined. Compared with the method of detecting the sound of glass breaking through the neural network algorithm, the calculation process is relatively simple.

In a possible implementation, the process of detecting whether the sound signal is a short pop is shown in Figure 4. First, the sum of the energy of the frequency bands of the current audio frame and the maximum value of the energy of the frequency bands are calculated, and the audio frame The sum of the energies of the frequency bands, or the maximum of the energies of the frequency bands is set as the crackle peak. After setting the peak value of the crackling sound, the first audio frame after the current audio frame is taken as the first audio frame, and the frame count value of the audio frame is set to FrameCount, and the initial value of FrameCount is 1. For the first audio frame, first determine whether FrameCount satisfies FrameCount≤N _p . If FrameCount does not satisfy FrameCount≤N _p , continue to the next stage of detection, for example, determine the sound of broken glass, or continue to detect other spectral parameters of the sound signal; if FrameCount satisfies FrameCount≤N _p , then update FrameCount to FrameCount +1, calculate the sum Efs of the energies of the frequency bands of the first audio frame, or the maximum value Efmax among the energies of the frequency bands. For example, if the set pop peak is the sum of the energy of the frequency band of the current audio frame, then calculate the sum of the energy of the frequency band of the first audio frame, if the set pop peak is the energy of the frequency band of the current audio frame is the maximum value, then the maximum value of the energy of the frequency band of the first audio frame is calculated.

Taking the set peak value of the crackling sound as the sum of the energy of the frequency bands of the current audio frame as an example, if the sum of the energy of the frequency bands of the first audio frame is greater than the peak value of the crackling sound, it means that a new peak value of the crackling sound is detected, and it will burst. The sound peak value is updated to the energy sum of the frequency bands of the first audio frame, and the position corresponding to the new crackling sound peak value, that is, the current FrameCount, is recorded. After the popping sound peak value update is completed, the next audio frame (ie, the second audio frame) of the first audio frame is used as the first audio frame, and the above-mentioned detection process of the first audio frame is repeated, that is, the execution returns to determine whether FrameCount satisfies FrameCount≤N Steps of _p and subsequent steps. If the sum of the energy of the frequency bands of the second audio frame is greater than the peak value of the crackling sound, update the peak value of the crackling sound again to the sum of the energy of the frequency bands of the second audio frame, and use the next audio frame of the second audio frame as the first audio frame , and repeat the above-mentioned detection process of the first audio frame.

In another possible implementation manner, in order to prevent the abnormal signal from being identified as the crackle peak, it is also possible to determine whether to update the crackle peak according to the formula Efs>Ef_max*Ar and Efs*As>Ef_min, where "*" represents Multiplication operation, Efs is the energy sum of the frequency bands of the first audio frame, Ef_max is the currently set crackle peak value, Ef_min is the minimum value of the set crackle peak value, Ar and As are the set constants, for example, you can Set 0.8≤Ar<1, As≤0.7. If Efs satisfies the formula Efs>Ef_max*Ar and Ef_max*As>Ef_min, it means that a new crackle peak is detected.

If the sum of the energy of the frequency bands of the first audio frame is less than or equal to the peak of the crackling sound, it means that no new peak of crackling sound is detected. Calculate the attenuation of the sum of the energy of the frequency bands of the first audio frame relative to the peak of the crackling sound, If the attenuation is within the preset attenuation range, it means that the first audio frame is not attenuated quickly, the next audio frame is taken as the first audio frame, and the above-mentioned detection process of the first audio frame is repeated, that is, the execution returns to determine whether FrameCount satisfies FrameCount≤N _p steps and subsequent steps. If the attenuation is not within the preset attenuation range, it means that the energy of the first audio frame attenuates too fast, determine that the first audio frame is a short pop, and return to the set initial detection state, that is, wait for the sound signal to be acquired again.

Among them, it can be judged whether the energy of the first audio frame decays too fast according to the formula Efs*Decay_i<Ef_max, wherein, Decay_i is the energy decay coefficient corresponding to the first audio frame, i=FrameCount, Decay_i increases with the increase of FrameCount Large, eg Decay_i=FrameCount-2. If Efs satisfies the formula Efs*Decay_i<Ef_max, it means that the energy decay of the first audio frame is too fast; otherwise, it means that the energy decay of the first audio frame is not fast.

In the above embodiment, by calculating the energy of the frequency band of each audio frame, it is judged whether a new crackling sound peak occurs, and if a new crackling sound peak occurs, the crackling sound peak value is updated, thereby improving the accuracy of the determined crackling sound peak value. Therefore, the accuracy of the calculated energy attenuation is improved, and the accuracy of the short pop detection is further improved.

In a possible implementation manner, the process of detecting whether the current audio frame is a crackling sound is shown in Figure 5. First, the maximum value E_Max(n) of the energy of the subframes of the current audio frame is calculated. If the maximum value E_Max(n) of the energy of the subframes of the frame is smaller than the preset second threshold value E_MAX, the method returns to detect the next audio frame. If the maximum value E_Max(n) of the energy of the subframes of the current audio frame is greater than the preset second threshold E_MAX, then determine the energy ratio E_Ratio(n) of the current audio frame according to the energy of the subframes of the current audio frame, according to The energy of the subframe of the previous audio frame determines the energy ratio E_Ratio(n-1) of the previous audio frame, according to the energy of the frequency band of the current audio frame and the energy of the frequency band of the previous audio frame of the current audio frame, determine the current The energy ratio Ef_Ratio(n) of the audio frame and the previous audio frame; Determine the energy difference E_Dist of the current audio frame and the previous audio frame according to the energy of the subframe of the current audio frame and the energy of the subframe of the previous audio frame (n). Taking the audio frame including four subframes as an example, the calculation formula of the energy ratio E_Ratio(n) of the current audio frame is E_Ratio(n)=MAX(Ej(n)/Ej-1(n))|j=1～4 , where Ej(n) represents the energy of the jth subframe of the current audio frame, Ej(n-1) represents the energy of the jth subframe of the previous audio frame of the current audio frame, and Ej(n) /Ej-1(n) represents the ratio of the energy of the current subframe to the energy of the previous subframe, "MAX" represents the maximum value, and Ej-1(n)>0. The calculation method of the energy ratio E_Ratio(n-1) of the previous audio frame is the same as the calculation method of the energy ratio of the current audio frame. The formula for calculating the energy ratio Ef_Ratio(n) between the current audio frame and the previous audio frame is Ef_Ratio(n)=Efh(n)/Efh(n-1), where Efh(n) is the highest in the current audio frame The energy of the frequency band, Efh(n-1) is the energy of the highest frequency band in the previous audio frame, Efh(n-1)>0. The formula for calculating the energy difference E_Dist(n) between the current audio frame and the previous audio frame is E_Dist(n)=MAX(Ej(n)-Ej(n-1))|j=1˜4.

If the energy ratio Ef_Ratio(n) of the current audio frame and the previous audio frame is greater than the first preset value Ef_RATIO_THRD_1, and the energy ratio of the current audio frame is greater than the second preset value E_RATIO_THRD_1;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than the first preset value Ef_RATIO_THRD_1, and the energy ratio of the previous audio frame is greater than the second preset value E_RATIO_THRD_1;

Or, if the energy ratio Ef_Ratio(n) of the current audio frame and the previous audio frame is greater than the third preset value Ef_RATIO_THRD_2, and the energy ratio E_Ratio(n) of the current audio frame is greater than the fourth preset value E_RATIO_THRD_2;

Or, if the energy ratio Ef_Ratio(n) of the current audio frame and the previous audio frame is greater than the third preset value Ef_RATIO_THRD_2, and the energy ratio of the previous audio frame is greater than the fourth preset value E_RATIO_THRD_2;

Or, if the energy difference E_Dist(n) between the current audio frame and the previous audio frame is greater than the fifth preset value ED_MAX, and the energy ratio Ef_Ratio(n) between the current audio frame and the previous audio frame is greater than the sixth preset value Ef_RATIO_THRD_3, and the energy ratio of the current audio frame is greater than the seventh preset value E_RATIO_THRD_3;

Or, if the energy difference E_Dist(n) between the current audio frame and the previous audio frame is greater than the fifth preset value ED_MAX, and the energy ratio Ef_Ratio(n) between the current audio frame and the previous audio frame is greater than the sixth preset value Ef_RATIO_THRD_3, and the energy ratio of the previous audio frame is greater than the seventh preset value E_RATIO_THRD_3;

Then it is determined that the current audio frame is a crackling sound. Otherwise, the audio frame next to the current audio frame is taken as the current audio frame, and the above-mentioned detection process of the current audio frame is repeated. That is, it returns to perform the step of calculating the maximum value of the energy of the subframes of the current audio frame and the subsequent steps.

Among them, E0(n)=E4(n-1), that is, the energy of the 0th subframe of the current audio frame is the energy of the 4th subframe of the previous audio frame, E_RATIO_THRD_1, E_RATIO_THRD_2, E_RATIO_THRD_3, Ef_RATIO_THRD_1, Ef_RATIO_THRD_2, Ef_RATIO_THRD_3, ED_MAX is a parameter preset according to experience. For example, E_MAX=5000, E_RATIO_THRD_1=10, E_RATIO_THRD_2=20, E_RATIO_THRD_3=5, Ef_RATIO_THRD_1=40, Ef_RATIO_THRD_2=20, Ef_RATIO_THRD_3=10.

In the above-described embodiment, when it is detected that the maximum value of the energy of the subframe of the current audio frame is greater than the preset second threshold, the energy ratio of the current audio frame, the energy ratio of the previous audio frame, the current audio The energy ratio between the frame and the previous audio frame, the energy difference between the current audio frame and the previous audio frame, the energy ratio of the current audio frame, the energy ratio of the previous audio frame, the current audio frame and the previous audio frame. The energy ratio and the energy difference between the current audio frame and the previous audio frame further determine whether the current audio frame is a crackling sound, which improves the accuracy of crackling sound detection.

In a possible implementation manner, after determining that the current audio frame is a crackling sound according to the spectral parameters of the current audio frame, the electronic device determines whether white noise exists in N _B audio frames following the current audio frame, where N _B is a preset empirical value, for example, N _B =3. If white noise is detected in N _B audio frames after the current audio frame, the attenuation of the energy of the audio frames after the current audio frame relative to the peak of the crackling sound is further calculated, thereby further improving the detection of glass breakage. Accuracy of sound.

In a possible implementation, the white noise detection process is shown in Figure 6. The frame count value FrameCount of the audio frame is set, and the initial value of FrameCount is 1, and the first audio frame after the current audio frame is set. As the first audio frame, for the first audio frame, first determine whether FrameCount satisfies FrameCount≤N _B , if it satisfies FrameCount≤N _B , update FrameCount to FrameCount+1, and calculate the spectral parameter of the first audio frame. In the embodiment of the present application, the spectrum parameter of the first audio frame is the energy of each frequency band of the first audio frame. Specifically, the first audio frame is divided into several frequency bands according to a preset frequency band division rule. For example, set the highest frequency to 8000Hz and the sampling frequency to 16000Hz, and divide the first audio frame into 10 frequency bands: Ef0, Ef1, Ef2, Ef3, Ef4, Ef5, Ef6, Ef7, Ef8, and Ef9. Ef0: 0～500Hz, Ef1: 500～1000Hz, Ef2: 1000～1500Hz, Ef3: 1500～2000Hz, Ef4: 2000～2500Hz, Ef5: 2500～3500Hz, Ef6: 3500～4500Hz, Ef7: 4500～5500Hz, Ef8: 5500～6500Hz, Ef9: 6500～8000Hz, Ef9 is the highest frequency band Efh. After the energy of each frequency band is calculated, the first count value WhiteNoiseDecisionCnt is set, and the initial value of WhiteNoiseDecisionCnt is 0. If the energy of the frequency bands of the first audio frame satisfies the first preset relational expression, the first audio frame is used as the target audio frame, and the first count value is determined according to the calculation formula of the first calculation value. If the first count value is less than the first threshold WNDC, the next audio frame of the first audio frame is taken as the first audio frame, and the above-mentioned detection process of the first audio frame is repeated, that is, the step of judging whether FrameCount satisfies FrameCount≤N _B is returned to be executed and subsequent steps. If the first count value is greater than the first threshold WNDC, record WhiteNoiseDetected=WhiteNoiseDetected+1, and then judge whether WhiteNoiseDetected satisfies WhiteNoiseDetected>0, if it satisfies WhiteNoiseDetected>0, then the first audio frame is white noise, and the next stage of detection is continued. That is, the attenuation of the energy of the audio frame following the current audio frame relative to the peak of the crackling sound is calculated. If WhiteNoiseDetected>0 is not satisfied, return to the initial detection state. If the FrameCount corresponding to the first audio frame does not satisfy FrameCount≤N _B , it is directly judged whether WhiteNoiseDetected satisfies WhiteNoiseDetected>0.

Wherein, the first preset relational expression is any one or more of the following relational expressions: Efi>Efi+1*WN_VAR) and (Efi+1>Efi*WN_VAR), i=0,1,2,… 8; Efi>Efi+2*WN_VAR) and (Efi+2>Efi*WN_VAR), i=0,1,2,…7; Efi>Efi+3*WN_VAR) and (Efi+3>Efi*WN_VAR) , i=0,1,2,...6; (Ef1+Ef2+Ef3+Ef4)*WN_VAR>Ef0; (Ef5+Ef6+Ef7+Ef8)*WN_VAR>Efh.

In a possible implementation, the calculation formula of the first calculated value WhiteNoiseDecisionCnt is: if Efi>Efi+1*WN_VAR) and (Efi+1>Efi*WN_VAR), then WhiteNoiseDecisionCnt is increased by N1; if Efi>Efi+2 *WN_VAR) and (Efi+2>Efi*WN_VAR), then WhiteNoiseDecisionCnt increases by N2; if Efi>Efi+3*WN_VAR) and (Efi+3>Efi*WN_VAR), then WhiteNoiseDecisionCnt increases by N3; if (Ef1+Ef2+ Ef3+Ef4)*WN_VAR>Ef0, then WhiteNoiseDecisionCnt increases by N4; if (Ef5+Ef6+Ef7+Ef8)*WN_VAR>Efh), then WhiteNoiseDecisionCnt increases by N5. Among them, WN_VAR, N1, N2, N3, N4, N5, WNDC are all preset empirical values, for example, WN_VAR=0.5, N1=3, N2=2, N3=1, N4=3, N5=3, WNDC=7. Finally, the value of WhiteNoiseDecisionCnt is determined according to the increment of WhiteNoiseDecisionCnt (the accumulated value of N1, the accumulated value of N2, the accumulated value of N3, the accumulated value of N4 and the sum of the accumulated value of N5).

It should be noted that, in other possible implementation manners, any one or more of the above calculation formulas can also be used as the calculation formula of the first calculation value WhiteNoiseDecisionCnt.

In the above embodiment, since the relationship between the spectral energies reflects the energy distribution of the audio frame, whether the audio frame is white noise is detected according to the relationship between the spectral energies, which improves the detection accuracy.

In other possible implementation manners, it may also be determined whether there is white noise in audio frames following the current audio frame according to whether the energy of the frequency band of the current audio frame is within a preset energy range.

In another embodiment, after determining that the current audio frame is a crackling sound according to the spectral parameters of the current audio frame, the electronic device determines whether white noise exists in N _B audio frames following the current audio frame. If white noise is detected in N _B audio frames after the current audio frame, further calculate the attenuation of the energy of the audio frames after the current audio frame relative to the peak of the crackling sound, if the N B audio frames after the current audio frame The attenuation of the energy of the _p audio frames relative to the peak value of the crackling sound is all within the preset attenuation range, and then it is further judged whether the sound signal satisfies the first preset condition. If the sound signal satisfies the first preset condition, it is determined that the sound signal is Glass breaking sound to further improve the accuracy of glass breaking sound detection. Wherein, the first preset condition includes any one or more of the following: the duration of the sound signal is within the preset duration, the spectral parameters of the sound signal conform to the preset spectral characteristics, the energy difference between the frequency bands of the sound signal is within the first Within a preset range, the difference between the target frequencies of the sound signals is within a second preset range, and the target frequency is the frequency corresponding to the energy peak of the frequency band.

Among them, detecting whether the spectral parameters of the sound signal conform to the preset spectral characteristics is to exclude the sound signal as a special audio sound signal, such as human voice, wood percussion, drum sound, musical instrument sound, metal percussion sound, high-heeled shoes walking sound, rain, etc. The purpose of detecting whether the energy difference between the frequency bands of the sound signal is within the first preset range is to exclude sound signals with a fixed frequency spectrum, such as air-conditioning sound, fan sound, vacuum cleaner sound, and motor sound. The purpose of detecting whether the difference between the target frequencies of the sound signals is within the second preset range is to exclude sound signals whose sound signals are resonant audio, such as bells, alarms, horns, and cup knocks.

In a possible implementation manner, the first preset condition includes that the duration of the sound signal is within the preset duration, the spectral parameter of the sound signal conforms to the preset spectral characteristics, and the energy difference between frequency bands of the sound signal is within the first preset duration. The range and the difference between the target frequencies of the sound signal are within the second preset range. Correspondingly, the process flow of the glass breaking sound detection method is shown in FIG. 7 . After the electronic device receives the sound signal, it is first set to the initial detection state, and then the sound signal is pre-processed, and the pre-processing is the aforementioned sound signal. Framing processing and processing of spectral compensation for audio frames. After the sound signal is pre-processed, the frequency spectrum parameters of the audio frame are obtained, and the first-stage crackling sound is detected first. The detection of crackling sound is to judge whether the current audio frame is crackling sound. If the current audio frame is not crackling sound, it will return to the initial detection state. If the current audio frame is crackling sound, the second stage of whitening will be performed. Noise detection. White noise detection is to determine whether white noise is detected in the N _B audio frames after the current audio frame. If no white noise is detected in the N _B audio frames after the current audio frame, return In the initial detection state, if white noise is detected, the third stage short pop detection will be performed. Short pop detection is to judge whether the sound signal is a short pop according to the attenuation of the audio frame after the current audio frame relative to the peak of the pop. If it is determined that the sound signal is a short pop, it will return to the initial detection state. For short pops, the fourth stage of sound length detection, special audio detection, fixed frequency spectrum detection, and resonance audio detection are performed. If the length of the sound signal is within the preset time length, and the sound signal is not a sound signal of special audio, and is not a sound signal of a fixed frequency spectrum, and is not a sound signal of resonant audio, it is determined that the sound signal is a glass breaking sound, and a glass breaking sound is generated. Alarm; otherwise, return to the initial detection state.

Wherein, the electronic device determines whether the sound signal is within the preset time length according to the decay speed and/or decay time of the energy of the frequency band of the sound signal relative to the peak of the crackling sound, if the decay speed is within the preset speed range, and/or the decay time is within Within the preset time range, it is determined that the duration of the sound signal is within the preset duration.

In a possible implementation manner, the process of judging whether the sound signal is within the preset duration is as shown in FIG. 8 , taking the next audio frame of the current audio frame as the first audio frame, and first calculating that the energy in the first audio frame is greater than The preset number of subframes E_cnt of the first detection energy E_MIN, for example, the first audio frame includes 4 subframes, the initial value of E_cnt is set to 0, and the energy Ej of each subframe is calculated sequentially, j=1～4 , if Ej>E_MIN is satisfied, then E_cnt is updated to E_cnt+1. Then calculate the number Emax_cnt of subframes whose energy is greater than the preset second detection energy E_MAX in the first audio frame, where E_MAX>E_MIN, for example, set the initial value of Emax_cnt to 0, and sequentially calculate the energy Ej of each subframe , if Ej>E_MAX is satisfied, then Emax_cnt is updated to Emax_cnt+1. Then calculate the maximum value Ej_Max of the energy of the subframes in the first audio frame. After obtaining E_cnt, Emax_cnt and Ej_Max, first determine whether the sound end condition is reached, that is, whether the sound is about to end. In a possible implementation, the average energy of the first audio frame may be calculated, and the average energy of the first audio frame is (E1+E2+E3+E4)/4, if the average energy of the first audio frame is less than the crackling sound 1/4 of the peak value, indicating that the sound end condition is reached. In another possible implementation manner, the maximum value Ej_Max of the energy of the subframe can be calculated. If Ej_Max<E_MIN, it means that the sound end condition is reached, wherein the crackling sound peak is the frequency band in the audio frame of the detected crackling sound. the maximum value of the energy. If the sound is about to end, judge whether it satisfies E_cnt≥T_MIN and Emax_cnt>0. If it does not satisfy E_cnt≥T_MIN and Emax_cnt>0, it means that the energy of the first audio frame is small, which further means that the energy of the sound signal decays faster, then return In the initial detection state, if E_cnt≥T_MIN and Emax_cnt>0, then judge whether E_cnt≤T_MAX is satisfied. If E_cnt≤T_MAX is satisfied, it means that the duration of the sound signal is within the preset duration, and the next stage of detection is continued, otherwise, return to to the initial detection state. Among them, T_MIN and T_MAX are preset constants, T_MAX>T_MIN.

If the sound end condition is not reached, judge whether the average energy of the first audio frame is less than 1/2 of the peak value of the crackle sound, and if the average energy of the first audio frame is less than 1/2 of the peak value of the crackle sound, judge whether it satisfies E_cnt>=T_DECAY2 , where T_DECAY2 is the frame count value of the corresponding audio frame when the preset energy decays to 1/2 of the peak value of the crackling sound. If it satisfies E_cnt≥T_DECAY2, it means that the energy decay time of the sound signal is long, and it returns to the initial detection state. If the average energy of the first audio frame is greater than or equal to 1/2 of the peak value of the crackling sound, or E_cnt<T_DECAY2, it means that the first audio frame is adjacent to the current audio frame, and the energy of the first audio frame has not yet begun to attenuate significantly. The next audio frame of an audio frame is used as the first audio frame, and the above-mentioned detection process of the first audio frame is repeated, that is, the number E_cnt of subframes whose energy is greater than the preset first detection energy E_MIN in the first audio frame is returned to be calculated. steps and subsequent steps.

In other possible implementations, when the first audio frame reaches the sound end condition, the duration of the sound signal can be calculated according to the size of the FrameCount of the first audio frame; or according to the ratio of the first audio frame to the peak value of the crackling sound, and the duration between the current audio frame and the first audio frame, to determine the decay speed of the energy of the first audio frame relative to the peak of the crackling sound.

In a possible implementation manner, the method for the electronic device to determine whether the spectral parameter of the sound signal conforms to the preset spectral characteristic includes: determining the number of audio frames in the sound signal whose energy between frequency bands satisfies the second preset relational expression, If the number of audio frames satisfying the second predetermined relational expression and the duration of the sound signal satisfy the third predetermined relational expression, it is determined that the energy of the first predetermined frequency band of the sound signal conforms to the predetermined spectral characteristic. Among them, the audio frame includes Ef0, Ef1, Ef2, Ef3, Ef4, Ef5, Ef6, Ef7, Ef8, Ef9 a total of 10 frequency bands, Ef9=Efh, Ef0 also includes Ef00 (0～250Hz) and Ef01 (250～500Hz) two frequency band. The second preset relationship is any one or more of the following formulas: Efh*5≤Ef0, Efh*5≤Ef1, Efh*8≤Ef1+Ef2, (Ef5+Ef6+Ef7+Ef8+Efh)* 3≤(Ef0+Ef1+Ef2+Ef3+Ef4), Ef00*5≤Ef01, Ef0*5≤Ef1, (Ef0+Ef1+Ef2)*12≤(Ef4+Ef5+Ef6+Ef7+Ef8+Efh).

If the energy between the frequency bands of an audio frame satisfies the second preset relational expression, it means that the feature of the audio frame conforms to the special audio feature. The duration of the sound signal can be represented by the frame count value FrameCount of the audio frame corresponding to the end of the sound signal, and the third preset relational expression can be used to represent the number of audio frames that satisfy the second preset relational expression, and the duration of the sound signal. , the size relationship or proportional relationship between the two.

In a possible implementation manner, the flow of the electronic device judging whether the spectral parameters of the sound signal conform to the preset spectral characteristics is shown in FIG. 9 , after determining that the current audio frame is a crackling sound, the next The audio frame is regarded as the first audio frame, and the parameters SpecialSpectrumCount(i)=0, i=1, 2, 3, . Whether an audio frame conforms to special audio characteristics. If the energy between the frequency bands of the first audio frame satisfies the second preset relational expression, it means that the characteristics of the first audio frame conform to special audio characteristics, and the number of audio frames whose energy between frequency bands satisfies the second predetermined relational expression is increased by 1 , in the embodiment of the present application, SpecialSpectrumCount(i) is updated to SpecialSpectrumCount(i)+Wi, Wi is a preset constant, which may not be 1, the special audio features corresponding to different second preset relational expressions are different, Wi value is also different.

In a possible implementation, the calculation formula of SpecialSpectrumCount(i) is as follows:

If Efh*5≤Ef0, then SpecialSpectrumCount(1) is updated to SpecialSpectrumCount(1)+1;

If Efh*5≤Ef1, then SpecialSpectrumCount(2) is updated to SpecialSpectrumCount(2)+1;

If Efh*8≤Ef1+Ef2, then SpecialSpectrumCount(3) is updated to SpecialSpectrumCount(3)+2;

If (Ef5+Ef6+Ef7+Ef8+Efh)*3≤(Ef0+Ef1+Ef2+Ef3+Ef4), then SpecialSpectrumCount(4) is updated to SpecialSpectrumCount(4)+3;

If Ef00*5≤Ef01, then SpecialSpectrumCount(5) is updated to SpecialSpectrumCount(5)+2;

If Ef0*5≤Ef1, then SpecialSpectrumCount(6) is updated to SpecialSpectrumCount(6)+2;

If (Ef0+Ef1+Ef2)*12≤(Ef4+Ef5+Ef6+Ef7+Ef8+Efh), then SpecialSpectrumCount(7) is updated to SpecialSpectrumCount(7)+3.

After calculating the SpecialSpectrumCount(i), the next audio frame of the first audio frame is used as the first audio frame, and the above-mentioned detection process of the first audio frame is repeated, that is, the step of judging whether the first audio frame conforms to the special audio feature is returned to be executed and next steps. If the energy between the frequency bands of the first audio frame satisfies the second preset relational expression, it means that the characteristics of the first audio frame do not conform to the special audio characteristics, and the next audio frame of the first audio frame is regarded as the first audio frame, repeating The above-mentioned detection process of the first audio frame is performed until the last audio frame of the sound signal is detected. After the last audio frame is detected, it is judged whether SpecialSpectrumCount(i) and the duration FrameCount of the sound signal satisfy the third relational expression. Meet the detection conditions of special audio, that is, the sound signal is a special audio, which further indicates that the sound signal is not the sound of glass breaking, and returns to the initial detection state. If the third relational expression is satisfied, it means that the spectral parameters of the sound signal do not meet the preset spectral characteristics, that is, the sound signal does not meet the detection conditions for special audio, and the next stage of detection is performed.

Wherein, the third relational expression may be the magnitude relation between SpecialSpectrumCount(i) and the duration FrameCount of the sound signal. In the embodiment of the present application, in order to exclude the sound signal of the special audio frequency, the spectral parameters of the sound signal can be set to conform to the relational expression of the preset spectral characteristics, that is, the SpecialSpectrumCount(i) and the duration FrameCount of the sound signal are set to not satisfy the third relationship The condition of the formula, for example, SpecialSpectrumCount(i) and the duration FrameCount of the sound signal do not satisfy the third relational formula including:

(SpecialSpectrumCount(1)>3) and (SpecialSpectrumCount(1)≥FrameCount*2-2)),

or (SpecialSpectrumCount(2)>3) and (SpecialSpectrumCount(2)≥FrameCount*2-2)),

or (SpecialSpectrumCount(3)>3) and (SpecialSpectrumCount(3)≥FrameCount*2-4)),

or (SpecialSpectrumCount(4)>3) and (SpecialSpectrumCount(4)≥FrameCount-2)),

or (SpecialSpectrumCount(5)>3) and (SpecialSpectrumCount(5)≥FrameCount*2-2)),

or (SpecialSpectrumCount(6)>3) and (SpecialSpectrumCount(6)≥FrameCount*2-2)),

or (SpecialSpectrumCount(7)>3) and (SpecialSpectrumCount(7)≥FrameCount-3)),

Or SpecialSpectrumCount>10 and SpecialSpectrumCount≥FrameCount*4-3, where SpecialSpectrumCount=SpecialSpectrumCount(1)+SpecialSpectrumCount(2)+...+SpecialSpectrumCount(7).

In the above-mentioned embodiment, by setting the second relational expressions corresponding to various special audio features, the sound signal of the special audio frequency can be excluded more comprehensively, and the accuracy of the sound detection of glass breakage can be improved.

In a possible implementation manner, the method for the electronic device to determine that the energy difference between frequency bands of the sound signal is within the first preset range includes: if in the sound signal, the energy of the frequency band of the n th audio frame is the same as the n th frequency band. The energy of the frequency band of the -1 audio frame satisfies the fourth preset relational expression, or the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame satisfy the fifth preset relational expression, then update The set second count value, n represents an integer greater than 2, and M≥n; if the sixth preset relational expression is satisfied between the second count value and the duration of the sound signal, then determine the energy between the frequency bands of the sound signal The difference is within a first preset range.

Wherein, the n-1 th audio frame is the previous audio frame of the n th audio frame, and the n-2 th audio frame is the previous two audio frames of the n th audio frame. The fourth preset relationship is satisfied between the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-1th audio frame, indicating that the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-1th audio frame are between within the preset variation range. The fifth preset relationship is satisfied between the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame, indicating that the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame are between within the preset variation range. The duration of the sound signal can be represented by the frame count value FrameCount of the audio frame corresponding to the end of the sound signal, and the sixth preset relational expression can be used to represent the magnitude relationship or ratio between the second count value and the duration of the sound signal relation. If the second count value and the duration of the sound signal do not satisfy the sixth preset relational expression, it means that in the sound signal, each audio frame and its adjacent audio frames have a non-changing frequency response, that is, the sound The signal is a sound signal with a fixed frequency spectrum.

In a possible implementation manner, the flow of the electronic device judging that the energy difference between the frequency bands of the sound signal is within the first preset range is shown in FIG. 10 . After determining that the current audio frame is a crackling sound, for the nth audio frame, set the second count value ConstantSpectrumCount=0, and then determine whether the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-1th audio frame satisfy the fourth preset relational expression, and the nth audio frame Whether the fifth preset relational expression is satisfied between the energy of the frequency band of and the energy of the frequency band of the n-2 th audio frame. If only the fourth preset relationship is satisfied, it means that the energy of the frequency band of the n-th audio frame and the energy of the frequency band of the n-1-th audio frame are within the preset variation range, NC=NC1, if only the fifth preset relationship is satisfied formula, indicating that the energy of the frequency band of the n th audio frame and the energy of the frequency band of the n-2 th audio frame are within a preset variation range, NC=NC2. If both the fourth preset relational expression and the fifth preset relational expression are satisfied, NC=NC1+NC2. Then the second count value ConstantSpectrumCount is updated to ConstantSpectrumCount+NC. Among them, NC1 and NC2 are preset constants, for example, NC1=2, NC2=1. After the ConstantSpectrumCount is calculated, repeat the above-mentioned detection process of the n-th audio frame for the n+1-th audio frame, that is, return to determine whether there is a difference between the energy of the frequency band of the n-th audio frame and the energy of the frequency band of the n-1-th audio frame. The steps of satisfying the fourth preset relational expression and subsequent steps.

If the energy difference between the frequency band of the nth audio frame and the frequency band of the n-1th audio frame is not within the preset variation range, and the energy difference between the frequency band of the nth audio frame and the frequency band of the n-2th audio frame If it is not within the preset variation range, for the n+1 th audio frame, the above-mentioned detection process of the n th audio frame is repeated until the last audio frame of the sound signal is detected. After the last audio frame is detected, it is determined whether ConstantSpectrumCount and FrameCount satisfy the sixth relational expression. If the sixth relational expression is not satisfied, it means that the energy difference between the frequency bands of the sound signal is not within the first preset range, that is, the sound signal conforms to the sixth relational expression. The detection condition of the fixed spectrum, that is, the sound signal is the sound signal of the fixed spectrum, not the sound of glass breaking, and returns to the initial detection state. If the sixth relational expression is satisfied, it means that the energy difference between the frequency bands of the sound signal is within the first preset range, that is, the sound signal does not meet the detection conditions of the fixed frequency spectrum, and the next stage of detection is performed.

In a possible implementation manner, the fourth preset relationship is (Ef(i)>Ef_Pre(i)*SP_VAR) and (Ef1_Pre(i)>Ef(i)*SP_VAR)), and the fifth preset relationship is The formula is (Ef(i)>Ef_Pre2(i)*SP_VAR2) and (Ef1_Pre2(i)>Ef(i)*SP_VAR2)), where i=1,2,3,..., Ef(i) represents the nth The energy of the ith frequency band of the audio frame, Ef_Pre(i) represents the energy of the ith frequency band of the n-1 th audio frame, Ef_Pre2(i) represents the energy of the ith frequency band of the n-2 th audio frame, SP_VAR and SP_VAR2 are pre- Set constants, such as SP_VAR=0.8, SP_VAR2=0.5. The sixth relational formula is ConstantSpectrumCount<FrameCount*0.6.

In other possible implementation manners, it may also be determined whether the energy difference between the audio frames is within the first preset range according to the relationship between the sums of the energy of the frequency bands of the audio frames.

In the above-described embodiment, by calculating the difference between the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-1th audio frame, and calculating the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame The difference can eliminate the sound signal of fixed frequency spectrum and improve the detection accuracy of glass breaking sound.

In a possible implementation manner, the method for the electronic device to determine that the difference between the target frequencies of the sound signal is within the second preset range includes: if in the sound signal, the target frequency of the i-th frequency band of the n-th audio frame is different from the target frequency of the i-th frequency band of the n-th audio frame. The difference between the target frequency of the ith frequency band of the n-1 audio frame is within the preset difference value range, or the target frequency of the ith frequency band of the n th audio frame and the target frequency of the ith frequency band of the n-2 th audio frame If the difference between the frequencies is within the preset difference range, update the set third count value, n represents an integer greater than 2, and M≥n, i represents an integer greater than 0; if the third count value is the same as the sound If the durations of the signals satisfy the seventh preset relational expression, it is determined that the difference between the target frequencies of the sound signals is within the second preset range.

Among them, since the target frequency is the frequency corresponding to the energy peak of the frequency band, if the difference between the target frequency of the ith frequency band of the n th audio frame and the target frequency of the ith frequency band of the n-1 th audio frame is within the preset difference Within the range, it means that the n-th audio frame and the n-1-th audio frame have the same frequency peak. If the difference between the target frequency of the i-th frequency band of the n-th audio frame and the target frequency of the i-th frequency band of the n-2-th audio frame is within the preset difference range, it means that the n-th audio frame and the Frames show the same frequency peaks. The duration of the sound signal can be represented by the frame count value FrameCount of the audio frame corresponding to the end of the sound signal, and the seventh preset relational expression can be used to represent the magnitude or ratio relationship between the third count value and the duration of the sound signal. If the seventh preset relational expression is not satisfied between the third count value and the duration of the sound signal, it means that there is a fixed resonance frequency in the sound signal for a long time, that is, the sound signal is a sound signal of resonant audio frequency.

In a possible implementation manner, the flow of the electronic device judging that the difference between the target frequencies of the sound signals is within the second preset range is shown in FIG. 11 . After determining that the current audio frame is a crackling sound, for the nth Audio frame, first perform frequency band division, divide the nth audio frame into NA frequency bands, NA is a constant set according to the empirical value, and set the third count value SameMaxPeakPosCnt(i)=0, correspondingly, i=1, 2,3,…,NA.

After the frequency band division is completed, the target frequency of each frequency band of the nth audio frame is calculated, that is, the frequency corresponding to the energy peak value of each frequency band. In a possible implementation manner, the amplitude of the frequency band i is set to be X(k), and the maximum value among the peaks of the amplitudes is the energy peak. Among them, k is the frequency index of the frequency band i, the frequency index corresponds to the frequency, and k is an integer greater than 0. For example, when the frequency band is 500-600 Hz, the corresponding k=1 when the frequency is 500 Hz, and the corresponding k=2 when the frequency is 505 Hz, etc. . For each amplitude X(k), if X(k)>X(k-1) and X(k)>X(k-2)*2 and X(k)>X(k-3)*3 and X(k)>X(k+1) and X(k)>X(k+2)*2 and X(k)>X(k+3)*3, then X(k) is an amplitude peak , that is, the peak value of the wave, and the position corresponding to the peak value of the wave is the frequency index k. Then calculate the maximum value of each amplitude peak value, take the maximum value of the amplitude peak value as the energy peak value MaxPeakPosi(n) of the frequency band, and the frequency corresponding to the energy peak value is the position corresponding to the energy peak value, that is, the frequency index k corresponding to the energy peak value.

After calculating the position corresponding to the energy peak of each frequency band of the nth audio frame, use the same method to calculate the energy peak MaxPeakPosi(n-1) of the n-1th audio frame and the corresponding position, and then determine the corresponding energy peak value MaxPeakPosi(n). Whether the position corresponding to the energy peak value MaxPeakPosi(n-1) is within the preset difference range, the preset difference range can be set to 1 or 2. If the position corresponding to the energy peak MaxPeakPosi(n) and the position corresponding to the energy peak MaxPeakPosi(n-1) are within the preset difference range, it means that the target frequency of the i-th frequency band of the n-th audio frame is different from the target frequency of the n-1-th audio frame. The difference between the target frequencies of the i-th frequency band is within the preset difference range, then NS=NS1, and NS1 is a preset constant, for example, NS1=1. Use the same method to calculate the energy peak MaxPeakPosi(n-2) and the corresponding position of the n-2 audio frame, and then determine whether the position corresponding to the energy peak MaxPeakPosi(n) and the energy peak MaxPeakPosi(n-2) are in within the preset difference range. If the position corresponding to the energy peak MaxPeakPosi(n) and the position corresponding to the energy peak MaxPeakPosi(n-2) are within the preset difference range, it means that the target frequency of the i-th frequency band of the n-th audio frame and the target frequency of the n-2-th audio frame The difference between the target frequencies of the i-th frequency band is within a preset difference range, then NS=NS2, and NS2 is a preset constant, for example, NS2=1. If the position corresponding to the energy peak MaxPeakPosi(n) and the position corresponding to the energy peak MaxPeakPosi(n-1), and the position corresponding to the energy peak MaxPeakPosi(n) and the energy peak MaxPeakPosi(n-2) are within the preset difference within the value range, then NS=NS1+NS2. Then the third count value SameMaxPeakPosCnt(i) is updated to SameMaxPeakPosCnt(i)+NS. After the SameMaxPeakPosCnt(i) is calculated, repeat the above-mentioned detection process of the nth audio frame for the n+1th audio frame, that is, return to the step of performing frequency band division for the nth audio frame and the subsequent steps.

If the difference between the target frequency of the ith band of the nth audio frame and the target frequency of the ith band of the n-1th audio frame, and the target frequency of the ith band of the nth audio frame and the n-2th audio frequency The difference between the target frequencies of the ith frequency band of the frame is not within the preset difference range. For the n+1 th audio frame, repeat the above-mentioned detection process of the n th audio frame until the last audio frequency of the sound signal is detected. frame. After the last audio frame is detected, determine whether SameMaxPeakPosCnt(i) and FrameCount satisfy the seventh relational expression. If the seventh relational expression is not satisfied, it means that the difference between the target frequencies of the sound signals is not within the second preset range, that is, The sound signal meets the detection conditions of the resonant audio, that is, the sound signal is the sound signal of the resonant audio, not the sound of glass breaking, and returns to the initial detection state. If the seventh relational expression is satisfied, it means that the difference between the target frequencies of the sound signals is within the second preset range, and the detection of the next stage is continued.

In the embodiment of the present application, in order to exclude the sound signal of the resonant audio frequency, it is possible to set the relational expression that the sound signal conforms to the detection condition of the resonant audio frequency, that is, to set the condition that SameMaxPeakPosCnt(i) and FrameCount do not satisfy the seventh relational expression, for example , the conditions under which SameMaxPeakPosCnt(i) and FrameCount do not satisfy the seventh relation include:

FrameCount>5 and SameMaxPeakPosCnt(0)≥FrameCount*1.5-2;

Or FrameCount>5 and SameMaxPeakPosCnt(1)≥FrameCount*1.5-2;

Or FrameCount>5 and SameMaxPeakPosCnt(2)≥FrameCount*1.5-2;

Or FrameCount>5 and SameMaxPeakPosCnt(3)≥FrameCount*1.5-2;

Or FrameCount>5 and SameMaxPeakPosCnt(0)+SameMaxPeakPosCnt(1)≥FrameCount*2-2;

Or FrameCount>5 and SameMaxPeakPosCnt(0)+SameMaxPeakPosCnt(1)+SameMaxPeakPosCnt(2)+SameMaxPeakPosCnt(3)≥FrameCount*3-2)).

In the above-described embodiment, by calculating the difference between the target frequency of the i-th frequency band of the n-th audio frame and the target frequency of the i-th frequency band of the n-1-th audio frame, and the target frequency of the i-th frequency band of the n-th audio frame The difference with the target frequency of the ith frequency band of the n-2 th audio frame can exclude the sound signal of the resonant audio, and improve the detection accuracy of the glass breaking sound.

It should be understood that the size of the sequence numbers of the steps in the above embodiments does not mean the sequence of execution, and the execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

Corresponding to the method for detecting the sound of glass breaking described in the above embodiment, FIG. 12 shows a structural block diagram of the glass breaking device provided by the embodiment of the present application. For the convenience of description, only the part related to the embodiment of the present application is shown. .

As shown in Figure 12, the glass breaking sound detection device includes,

an acquisition module 10, configured to acquire a sound signal, perform frame-by-frame processing on the sound signal, and obtain M audio frames, where M represents an integer greater than 0;

The calculation module 20 is used to determine that the current audio frame is a crackling sound according to the frequency spectrum parameter of the current audio frame, then according to the frequency spectrum parameter of the current audio frame, and the frequency of the audio frame after the current audio frame Spectral parameters to determine the crackling peak;

The determination module 30 is configured to determine that the sound signal is glass breakage if the energy of the N _p audio frames after the current audio frame is within a preset attenuation range relative to the attenuation of the crackling sound peak Sound, the N _p represents an integer greater than 0, and M>N _p .

In a possible implementation manner, the computing module 20 includes a first computing unit, and the first computing unit is used for:

If it is determined according to the spectral parameters of the current audio frame that the current audio frame is a crackling sound, and white noise is detected in N _B audio frames after the current audio frame, then according to the current audio frame The frequency spectrum parameter of , and the frequency spectrum parameter of the audio frame after the current audio frame, determine the crackle peak value, the NB represents an integer greater than 0, and _M > _NB .

In a possible implementation manner, the audio frame includes at least one frequency band, the spectrum parameter includes energy of the frequency band, and the first calculation unit is specifically configured to: if the _NB after the current audio frame There is a target audio frame in the audio frames, then determine the calculation formula of the first count value, the target audio frame is the audio frame that satisfies the first preset relational expression between the energies of the frequency bands, and the calculation formula of the first calculation value corresponding to the first preset relational expression;

The first count value is determined according to the calculation formula of the first count value, and if the first count value is greater than a preset first threshold, it is determined that the target audio frame is white noise.

In a possible implementation manner, the audio frame includes at least one subframe, the spectral parameter includes energy of the subframe, the audio frame includes at least one frequency band, and the spectral parameter includes energy of the frequency band , the computing module 20 further includes a second computing unit, and the second computing unit is specifically used for:

If the maximum value of the energy of the subframes of the current audio frame is greater than the preset second threshold, then:

Determine the energy ratio of the current audio frame according to the energy of the subframes of the current audio frame;

According to the energy of the frequency band of the current audio frame and the energy of the frequency band of the previous audio frame of the current audio frame, determine the energy ratio of the current audio frame to the previous audio frame;

Determine the energy difference between the current audio frame and the previous audio frame according to the energy of the subframe of the current audio frame and the energy of the subframe of the previous audio frame;

Determine the energy ratio of the previous audio frame according to the energy of the subframe of the previous audio frame;

If the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the current audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a first preset value, and the energy ratio of the previous audio frame is greater than a second preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the current audio frame is greater than a fourth preset value;

Or, if the energy ratio of the current audio frame and the previous audio frame is greater than a third preset value, and the energy ratio of the previous audio frame is greater than a fourth preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, And the energy ratio of the current audio frame is greater than the seventh preset value;

Or, if the energy difference between the current audio frame and the previous audio frame is greater than the fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than the sixth preset value, and the energy ratio of the previous audio frame is greater than the seventh preset value;

Then it is determined that the current audio frame is a crackling sound.

In a possible implementation manner, the audio frame includes at least one frequency band, the spectral parameter includes energy of the frequency band, and the calculation module 20 further includes a third calculation unit, and the third calculation unit is specifically used for :

Taking the energy sum of each frequency band of the current audio frame as the peak of the crackling sound;

Taking the next audio frame of the current audio frame as the first audio frame;

calculating the energy summation of each frequency band of the first audio frame;

If the sum of the energy of each frequency band of the first audio frame is greater than the peak value of the crackling sound, updating the peak value of the crackling sound to the sum of the energy of each frequency band of the first audio frame;

The following audio frame of the first audio frame is used as the first audio frame, and the step of calculating the sum of the energy of each frequency band of the first audio frame and the subsequent steps are returned to until the preset end condition is satisfied.

In a possible implementation manner, the audio frame includes at least one frequency band, the spectral parameter includes energy of the frequency band, and the determining module 30 is specifically configured to:

If the energy of the N _p audio frames following the current audio frame is within the preset attenuation range relative to the attenuation of the crackling peak, and the sound signal satisfies the first preset condition, it is determined that the The sound signal is glass breaking sound; the first preset condition includes any one or more of the following: the duration of the sound signal is within the preset duration, and the spectral parameters of the sound signal conform to preset spectral characteristics , the energy difference between the frequency bands of the sound signal is within a first preset range, the difference between the target frequencies of the sound signal is within a second preset range, and the target frequency is the energy peak corresponding to the frequency band frequency.

In a possible implementation manner, the determining module 30 is further configured to:

calculating the decay speed and/or decay time of the energy of the frequency band of the sound signal relative to the crackle peak;

If the decay speed is within the preset speed range, and/or the decay time is within the preset time range, it is determined that the duration of the sound signal is within the preset duration.

In a possible implementation manner, the determining module 30 is further configured to:

determining the number of audio frames in the sound signal whose energy between frequency bands satisfies the second preset relational expression;

If the number of audio frames satisfying the second preset relational expression and the duration of the sound signal satisfy the third predetermined relational expression, it is determined that the energy of the first preset frequency band of the sound signal conforms to the preset spectral characteristics.

In a possible implementation manner, the determining module 30 is further configured to:

If in the sound signal, the energy of the frequency band of the n-th audio frame and the energy of the frequency band of the n-1-th audio frame satisfy the fourth preset relational expression,

Or the energy of the frequency band of the nth audio frame and the energy of the frequency band of the n-2th audio frame satisfy the fifth preset relational expression, then update the set second count value, and the n represents an integer greater than 2 , and M≥n;

If a sixth preset relational expression is satisfied between the second count value and the duration of the sound signal, it is determined that the energy difference between the frequency bands of the sound signal is within a first preset range.

In a possible implementation manner, the determining module 30 is further configured to:

If in the sound signal, the difference between the target frequency of the ith frequency band of the n th audio frame and the target frequency of the ith frequency band of the n-1 th audio frame is within the preset difference range,

Or the difference between the target frequency of the ith frequency band of the nth audio frame and the target frequency of the ith frequency band of the n-2th audio frame is within the preset difference range, then update the set third count value, so The n represents an integer greater than 2, and M≥n, and the i represents an integer greater than 0;

If a seventh preset relational expression is satisfied between the third count value and the duration of the sound signal, it is determined that the difference between the target frequencies of the sound signal is within a second preset range.

It should be noted that the information exchange, execution process and other contents between the above-mentioned devices/units are based on the same concept as the method embodiments of the present application. For specific functions and technical effects, please refer to the method embodiments section. It is not repeated here.

FIG. 13 is a schematic diagram of an electronic device provided in Embodiment 3 of the present application. As shown in FIG. 13 , the electronic device of this embodiment includes: a processor 21 , a memory 22 , and a computer program 23 stored in the memory 22 and executable on the processor 21 . When the processor 21 executes the computer program 23 , the steps in the above-mentioned embodiment of the method for detecting the sound of broken glass, such as steps S101 to S103 shown in FIG. 3 , are implemented. Alternatively, when the processor 21 executes the computer program 23, the functions of the modules/units in the foregoing device embodiments are implemented, for example, the functions of the acquisition module 10 to the determination module 30 shown in FIG. 12 .

Exemplarily, the computer program 23 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 22 and executed by the processor 21 to complete the this application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 23 in the electronic device.

Those skilled in the art can understand that FIG. 13 is only an example of an electronic device, and does not constitute a limitation to the electronic device, and may include more or less components than those shown in the figure, or combine some components, or different components, such as The electronic device may also include an input and output device, a network access device, a bus, and the like.

The processor 21 may be a central processing unit (Central Processing Unit, CPU), or other general-purpose processors, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 22 may be an internal storage unit of the electronic device, such as a hard disk or a memory of the electronic device. The memory 22 may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card equipped on the electronic device, Flash card (Flash Card) and so on. Further, the memory 12 may also include both an internal storage unit of the electronic device and an external storage device. The memory 22 is used to store the computer program and other programs and data required by the electronic device. The memory 22 can also be used to temporarily store data that has been output or is to be output.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned functions can be allocated to different functional units, Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit, and the above-mentioned integrated units may adopt hardware. It can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present application. For the specific working processes of the units and modules in the above-mentioned system, reference may be made to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

In the foregoing embodiments, the description of each embodiment has its own emphasis. For parts that are not described or described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

In the embodiments provided in this application, it should be understood that the disclosed apparatus/electronic device and method may be implemented in other manners. For example, the above-described embodiments of the apparatus/electronic device are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units. Or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.

The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer-readable storage medium. Based on this understanding, the present application can implement all or part of the processes in the methods of the above embodiments, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer When the program is executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, and the like. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory) , Random Access Memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the above-mentioned embodiments, those of ordinary skill in the art should understand that: it can still be used for the above-mentioned implementations. The technical solutions described in the examples are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions in the embodiments of the application, and should be included in the within the scope of protection of this application.

Claims (14)

1. A method for detecting a glass breaking sound includes:

acquiring a sound signal, and performing framing processing on the sound signal to obtain M audio frames, wherein M represents an integer greater than 0;

if the current audio frame is determined to be the pop sound according to the spectrum parameters of the current audio frame, determining a peak value of the pop sound according to the spectrum parameters of the current audio frame and the spectrum parameters of the audio frame after the current audio frame;

if N is after the current audio frame_pThe attenuation quantity of the energy of each audio frame relative to the peak value of the popping sound is within a preset attenuation range, the sound signal is judged to be glass breaking sound, and N is_pRepresents an integer greater than 0, and M>N_p。

2. The method of claim 1, wherein if the current audio frame is determined to be pop according to the spectral parameters of the current audio frame, determining a pop peak according to the spectral parameters of the current audio frame and the spectral parameters of the audio frame following the current audio frame comprises:

if the current audio frame is determined to be a pop sound according to the spectral parameters of the current audio frame and N is behind the current audio frame_BIf white noise is detected in each audio frame, determining a pop sound peak value according to the spectral parameters of the current audio frame and the spectral parameters of the audio frame after the current audio frame, wherein N is_BRepresents an integer greater than 0, and M>N_B。

3. The method of claim 2, wherein the audio frame comprises at least one frequency band, the spectral parameter comprises an energy of the frequency band, and the N after the current audio frame is detected_BDetecting white noise in an audio frame, comprising:

if N is after the current audio frame_BDetermining a calculation formula of a first counting value if a target audio frame exists in the audio frames, wherein the target audio frame is an audio frame with the energy of frequency bands meeting a first preset relation, and the calculation formula of the first counting value corresponds to the first preset relation;

and determining the first count value according to a calculation formula of the first count value, and if the first count value is greater than a preset first threshold, determining that the target audio frame is white noise.

4. The method of claim 1, wherein the audio frame comprises at least one sub-frame, the spectral parameter comprises an energy of the sub-frame, the audio frame comprises at least one frequency band, the spectral parameter comprises an energy of the frequency band, and the determining that the current audio frame is pop sound according to the spectral parameter of the current audio frame comprises:

if the maximum value of the energy of the sub-frame of the current audio frame is greater than a preset second threshold value, then:

determining the energy ratio of the current audio frame according to the energy of the subframe of the current audio frame;

determining the energy ratio of the current audio frame to the previous audio frame according to the energy of the frequency band of the current audio frame and the energy of the frequency band of the previous audio frame of the current audio frame;

determining the energy difference between the current audio frame and the previous audio frame according to the energy of the subframe of the current audio frame and the energy of the subframe of the previous audio frame;

determining an energy ratio of the previous audio frame according to the energy of the sub-frame of the previous audio frame;

if the energy ratio of the current audio frame to the previous audio frame is greater than a first preset value, and the energy ratio of the current audio frame is greater than a second preset value;

or, if the energy ratio of the current audio frame to the previous audio frame is greater than a first preset value, and the energy ratio of the previous audio frame is greater than a second preset value;

or, if the energy ratio of the current audio frame to the previous audio frame is greater than a third preset value, and the energy ratio of the current audio frame is greater than a fourth preset value;

or, if the energy ratio of the current audio frame to the previous audio frame is greater than a third preset value, and the energy ratio of the previous audio frame is greater than a fourth preset value;

or, if the energy difference between the current audio frame and the previous audio frame is greater than a fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than a sixth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than a seventh preset value;

or, if the energy difference between the current audio frame and the previous audio frame is greater than a fifth preset value, and the energy ratio between the current audio frame and the previous audio frame is greater than a sixth preset value, and the energy ratio between the previous audio frame is greater than a seventh preset value;

the current audio frame is determined to be a pop sound.

5. The method of claim 1, wherein the audio frame comprises at least one frequency band, the spectral parameter comprises an energy of the frequency band, and the determining the pop peak according to the spectral parameter of the current audio frame and the spectral parameter of the audio frame following the current audio frame comprises:

taking the energy sum of each frequency band of the current audio frame as a popping sound peak value;

taking a next audio frame of the current audio frame as a first audio frame;

calculating the energy sum of each frequency band of the first audio frame;

if the energy sum of each frequency band of the first audio frame is larger than the pop sound peak value, updating the pop sound peak value to the energy sum of each frequency band of the first audio frame;

and taking the audio frame after the first audio frame as the first audio frame, and returning to the step of calculating the energy sum of each frequency band of the first audio frame and the subsequent steps until a preset ending condition is met.

6. The method as claimed in claim 1, wherein the audio frame comprises at least one frequency band, the spectral parameter comprises energy of the frequency band, and the current audio frame is followed by N_pIf the attenuation amount of the energy of each audio frame relative to the pop sound peak value is within a preset attenuation range, the sound signal is judged to be glass breaking sound, and the method comprises the following steps:

if N is after the current audio frame_pThe attenuation quantity of the energy of each audio frame relative to the peak value of the popping sound is within a preset attenuation range, and if the sound signal meets a first preset condition, the sound signal is judged to be glass breaking sound; the first preset condition comprises the following stepsAny one or more of: the time length of the sound signal is within a preset time length, the frequency spectrum parameters of the sound signal accord with preset frequency spectrum characteristics, the energy difference between frequency bands of the sound signal is within a first preset range, the difference between target frequencies of the sound signal is within a second preset range, and the target frequencies are frequencies corresponding to energy peak values of the frequency bands.

7. The method of claim 6, wherein the detecting the sound signal has a duration within a predetermined duration comprises:

calculating a decay speed and/or a decay time of an energy of a frequency band of the sound signal with respect to the pop sound peak;

and if the attenuation speed is within a preset speed range and/or the attenuation time is within a preset time range, determining that the duration of the sound signal is within a preset duration.

8. The method of claim 6, wherein the detecting the sound signal with the spectral parameter according to a predetermined spectral characteristic comprises:

determining the number of audio frames which meet a second preset relational expression between the energies of frequency bands in the sound signals;

and if the number of the audio frames meeting the second preset relational expression and the time length of the sound signal meet a third preset relational expression, determining that the energy of the first preset frequency band of the sound signal meets preset spectral characteristics.

9. The method of claim 6, wherein the energy difference between the frequency bands of the sound signal is within a first predetermined range, and comprises:

if the energy of the frequency band of the nth audio frame and the energy of the frequency band of the (n-1) th audio frame in the sound signal satisfy a fourth preset relational expression,

or if the energy of the frequency band of the nth audio frame and the energy of the frequency band of the (n-2) th audio frame meet a fifth preset relational expression, updating a set second counting value, wherein n represents an integer greater than 2, and M is greater than or equal to n;

and if the second counting value and the duration of the sound signal meet a sixth preset relational expression, determining that the energy difference between the frequency bands of the sound signal is within a first preset range.

10. The method of claim 6, wherein the difference between the target frequencies of the sound signals is within a second predetermined range, comprising:

if the difference between the target frequency of the ith frequency band of the nth audio frame and the target frequency of the ith frequency band of the (n-1) th audio frame in the sound signal is within a preset difference range,

or if the difference value between the target frequency of the ith frequency band of the nth audio frame and the target frequency of the ith frequency band of the (n-2) th audio frame in the sound signal is within a preset difference value range, updating a set third counting value, wherein n represents an integer greater than 2, M is greater than or equal to n, and i represents an integer greater than 0;

and if the third counting value and the duration of the sound signal meet a seventh preset relational expression, determining that the difference between the target frequencies of the sound signals is within a second preset range.

11. A glass breakage sound detection device, comprising:

the device comprises an acquisition module, a frame processing module and a frame processing module, wherein the acquisition module is used for acquiring a sound signal and performing frame processing on the sound signal to obtain M audio frames, and M represents an integer larger than 0;

the computing module is used for determining a pop sound peak value according to the spectrum parameter of the current audio frame and the spectrum parameter of the audio frame after the current audio frame if the current audio frame is determined to be the pop sound according to the spectrum parameter of the current audio frame;

a determination module for determining if N is after the current audio frame_pAttenuation of energy of an audio frame relative to the pop sound peakDecrement, if all within the preset attenuation range, the sound signal is judged to be glass breaking sound, and N is_pRepresents an integer greater than 0, and M>N_p。

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any of claims 1 to 10 when executing the computer program.

13. A glass break sound detection system, comprising: sound collection apparatus, alarm apparatus and an electronic device as claimed in claim 12, the electronic device being communicatively connected to the sound collection apparatus and the alarm apparatus.

14. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 10.

Images