CN105534480B - Sound of snoring detection method and device - Google Patents
Sound of snoring detection method and device Download PDFInfo
- Publication number
- CN105534480B CN105534480B CN201610005669.9A CN201610005669A CN105534480B CN 105534480 B CN105534480 B CN 105534480B CN 201610005669 A CN201610005669 A CN 201610005669A CN 105534480 B CN105534480 B CN 105534480B
- Authority
- CN
- China
- Prior art keywords
- signal
- band signal
- energy value
- low
- frequency band
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Power Engineering (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Telephone Function (AREA)
Abstract
The invention discloses a kind of sound of snoring detection methods, and detection method includes the following steps for the sound of snoring：Obtain the low-band signal and high frequency band signal in collected audio signal；Calculate the energy value of the low-band signal and the high frequency band signal；Judge the energy value of the low-band signal with the ratio of the energy value of the corresponding high frequency band signal whether in the first preset range；If so, determining that the audio signal is sound of snoring audio signal.The invention also discloses a kind of sound of snoring detection devices.The sound of snoring detection method of the present invention improves the accuracy rate of sound of snoring detection.
Description
Technical field
The present invention relates to technical field of medical equipment more particularly to a kind of sound of snoring detection method and device.
Background technology
In recent years, snoring is gradually taken seriously the harm of human body, long-term to snore, the diseases such as meeting cause insomnia, fatigue, depression
Shape influences the normal life of people.Therefore, the detection quantitative to the sound of snoring, analysis are realized, and then proposes the method that treatment improves,
It is a significantly thing.
At present the common sound of snoring detection method of personal consumption electron-like be by detect sound periodicity to determine whether
It is the sound of snoring, this method is at low cost, but poor anti jamming capability.Under quiet environment, the sound of snoring can be detected, but working as has voice
Or other with similar in the sound of snoring frequency when sound interference, can voice or other be with sound, error detection similar in sound of snoring frequency
The sound of snoring, therefore, the Detection accuracy of existing sound of snoring detection method are relatively low.
Invention content
The main purpose of the present invention is to provide a kind of sound of snoring detection method and device, it is intended to solve existing sound of snoring detection
The relatively low problem of the Detection accuracy of method.
To achieve the above object, sound of snoring detection method described in a kind of sound of snoring detection method provided by the invention includes following step
Suddenly：
Obtain the low-band signal and high frequency band signal in collected audio signal；
Calculate the energy value of the low-band signal and the high frequency band signal；
Judge the energy value of the low-band signal and the ratio of the energy value of the corresponding high frequency band signal whether
In first preset range；
If so, determining that the audio signal is sound of snoring audio signal.
Optionally, the step of low-band signal and high frequency band signal obtained in collected audio signal includes：
Sampling and quantification treatment are carried out to collected audio signal, obtain lasting digital signal, and to the number
Signal carries out windowing process；
Bandpass filtering is carried out to the digital signal in window, obtains the low-band signal and high band letter of the audio signal
Number, wherein the low-band signal and high frequency band signal include multiple frame signals.
Optionally, the step of energy value for calculating the low-band signal and the high frequency band signal includes：
The energy value of each frame signal of the low-band signal and the high frequency band signal is calculated, and the low frequency
Energy value of the energy value of each frame signal of segment signal as the low-band signal, each frame letter of the high frequency band signal
Number energy value of the energy value as the high frequency band signal.
Optionally, the energy value and the energy value of the corresponding high frequency band signal for judging the low-band signal
Whether the step in the first preset range includes ratio：
The energy wave of low-band signal is formed according to the energy value of each frame signal of the low-band signal；
The default frame met corresponding to the adjacent peak of preset condition is chosen from the energy wave of the low-band signal
The energy value of signal, and the energy value with the frame signal of the high frequency band signal corresponding to the energy value for the frame signal chosen is obtained,
The preset condition is the time difference of adjacent peak in the second preset range；
Calculate the frame signal of the corresponding high frequency band signal of energy value of the frame signal of the low-band signal of each selection
Energy value ratio；
Judge each ratio being calculated whether in first preset range.
Optionally, the energy value and the energy value of the corresponding high frequency band signal for judging the low-band signal
Whether the step in the first preset range includes ratio：
The energy wave of high frequency band signal is formed according to the energy value of each frame signal of the high frequency band signal；
The default frame met corresponding to the adjacent peak of preset condition is chosen from the energy wave of the high frequency band signal
The energy value of signal, and the energy value with the frame signal of the low-band signal corresponding to the energy value for the frame signal chosen is obtained,
The preset condition is the time difference of adjacent peak in the second preset range；
The frame for calculating the energy value of the frame signal of the low-band signal of each acquisition and the high frequency band signal of selection is believed
Number energy value ratio；
Judge each ratio being calculated whether in first preset range.
In addition, to achieve the above object, the present invention also provides a kind of sound of snoring detection device, the sound of snoring detection device packet
It includes：
Acquisition module, for obtaining low-band signal and high frequency band signal in collected audio signal；
Computing module, the energy value for calculating the low-band signal and the high frequency band signal；
Judgment module, the energy value of energy value and the corresponding high frequency band signal for judging the low-band signal
Ratio whether in the first preset range；
Determining module, if for the ratio in the first preset range, it is determined that the audio signal is sound of snoring audio
Signal.
Optionally, the acquisition module includes：
Processing unit, for collected audio signal carry out sampling and quantification treatment, obtain lasting digital signal,
And windowing process is carried out to the digital signal；
Acquiring unit obtains the low-frequency range of the audio signal for carrying out bandpass filtering to the digital signal in window
Signal and high frequency band signal, wherein the low-band signal and high frequency band signal include multiple frame signals.
Optionally, the computing module is additionally operable to calculate each of the low-band signal and the high frequency band signal
The energy value of frame signal, and using the energy value of each frame signal of the low-band signal as the energy of the low-band signal
Value, the energy value of the energy value of each frame signal of the high frequency band signal as the high frequency band signal.
Optionally, the judgment module includes：
Component units are used to form the energy of low-band signal according to the energy value of each frame signal of the low-band signal
Measure wave；
Selection unit, for choosing the default adjacent peak for meeting preset condition from the energy wave of the low-band signal
The energy value of the corresponding frame signal of value, and obtain and believe with the frame of the high frequency band signal corresponding to the energy value for the frame signal chosen
Number energy value, the preset condition be adjacent peak time difference in the second preset range；
Computing unit, the corresponding high band of the energy value of the frame signal of the low-band signal for calculating each selection
The ratio of the energy value of the frame signal of signal；
Judging unit, whether each ratio for judging to be calculated is in first preset range.
Optionally, the component units are additionally operable to be formed according to the energy value of each frame signal of the high frequency band signal
The energy wave of high frequency band signal；
The selection unit is additionally operable to default of the selection from the energy wave of the high frequency band signal and meets preset condition
The energy value of frame signal corresponding to adjacent peak, and obtain and the low-band signal corresponding to the energy value for the frame signal chosen
Frame signal energy value, the preset condition be adjacent peak time difference in the second preset range；
The computing unit is additionally operable to calculate energy value and the selection of the frame signal of the low-band signal of each acquisition
High frequency band signal frame signal energy value ratio；
Whether the judging unit is additionally operable to judge each ratio being calculated in first preset range.
The present invention calculates the low frequency by obtaining the low-band signal and high frequency band signal of collected audio signal
The ratio of the energy value of the energy value of segment signal and the high frequency band signal, due to the energy value of the low-band signal of sound of snoring signal
And energy value and the height of the ratio range of the energy value of the high frequency band signal with the low-band signal of normal speech signal
The ratio range of the energy value of frequency band signals is significantly different, by judging the ratio whether in the first preset range, to
It can determine whether the collected audio signal is sound of snoring audio signal, improve the accuracy rate of sound of snoring detection.
Description of the drawings
Fig. 1 is the flow diagram of the first embodiment of sound of snoring detection method of the present invention；
Fig. 2 is refinement stream the step of obtaining the low-band signal and high frequency band signal in collected audio signal in Fig. 1
Journey schematic diagram；
Fig. 3 is the flow diagram of the second embodiment of sound of snoring detection method of the present invention；
Fig. 4 is to judge whether are the energy value of low-band signal and the ratio of the energy value of corresponding high frequency band signal in Fig. 2
Step in the first preset range refines flow diagram；
Fig. 5 is the energy waveform schematic diagram of low-band signal of the present invention；
Fig. 6 is the high-level schematic functional block diagram of the first embodiment of sound of snoring detection device of the present invention；
Fig. 7 is the refinement high-level schematic functional block diagram of acquisition module in Fig. 6；
Fig. 8 is the refinement high-level schematic functional block diagram of judgment module in Fig. 6；
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific implementation mode
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Based on the above issues, the present invention provides a kind of sound of snoring detection method.
Referring to Fig.1, Fig. 1 is the flow diagram of the first embodiment of sound of snoring detection method of the present invention.
In the present embodiment, the sound of snoring detection method includes：
Step S10 obtains low-band signal and high frequency band signal in collected audio signal；
The present inventor the study found that the sound of snoring high band (600HZ~950HZ) and low-frequency range (50HZ~
The attenuation ratio of energy distributions 300HZ) has apparent feature, and base region is between 50~100.And the height of normal speech
The attenuation ratio of the energy distributions of frequency range (600HZ~950HZ) and low-frequency range (50HZ~300HZ) between 1~10, so
The attenuation ratio of the energy distributions of the sound of snoring and normal speech makes a big difference.In the present embodiment, the low-band signal
For signal component of the frequency within the scope of 50HZ~300HZ in collected audio signal, the high frequency band signal is to collect
Audio signal in signal component of the frequency within the scope of 600HZ~950HZ, obtaining the low of the collected audio signal
When frequency range and high frequency band signal, need the audio signal being filtered by two-way bandpass filter, the two-way band logical
Filter is respectively the bandpass filter of 50HZ~300HZ and the bandpass filter of 600HZ~950HZ, is passing through above-mentioned band logical
After the filtering of filter, the low-band signal and high frequency band signal of the audio signal can be obtained.
Step S20 calculates the energy value of the low-band signal and the high frequency band signal；
After getting the low-band signal and high frequency band signal of the audio signal, the energy of the low-band signal is calculated
The energy value of magnitude and the high frequency band signal, the energy value represent the amplitude of the low-band signal or high frequency band signal.
Specifically, can windowing process be carried out for the low-band signal and high frequency band signal, to obtain the low-band signal and
Then the frame signal of high frequency band signal calculates separately the energy value of every frame signal, calculating each of the signal of the low-frequency range
It, can be using the average value of the energy value of each frame signal of the low-band signal as the low-frequency range when energy value of frame signal
The energy value of signal, while using the average value of the energy value of each frame signal of the high frequency band signal as the high frequency band signal
Energy value, or can also from the energy value of each frame signal of the low-band signal form time domain waveform in choose it is adjacent
Default peak value energy value of the energy value as the low-band signal, while by the low-band signal with the selection
Energy value of the energy value of each frame signal of the corresponding high frequency band signal of peak value as the high frequency band signal, wherein described
Default preferably 5.
Step S30 judges the ratio of the energy value and the energy value of the corresponding high frequency band signal of the low-band signal
Whether value is in the first preset range；
In the present embodiment, first preset range is 50~100.The energy value of the low-band signal can be institute
It states the energy value of each frame signal of low-band signal, or is averaged for the energy value of each frame signal of the low-band signal
Value, either for default energy value being chosen in the energy value of each frame signal of the low-band signal or be the low frequency
The average value for the default energy value chosen in the energy value of each frame signal of segment signal, correspondingly, the high frequency band signal
Energy value is the energy value of each frame signal of the high frequency band signal, or the energy of each frame signal for the high frequency band signal
The default energy value chosen in the average value of value, or energy value for each frame signal of the high frequency band signal, Huo Zhewei
The average value for the default energy value chosen in the energy value of each frame signal of the high frequency band signal.
Step S40, if so, determining that the audio signal is sound of snoring audio signal.
If the ratio is in the first preset range, it is determined that the audio signal is sound of snoring audio signal, the ratio
Not in the first preset range, it is determined that the audio signal is not sound of snoring audio signal, such as normal speech signal or one
A little noise signals etc..
The present embodiment is calculated described low by obtaining the low-band signal and high frequency band signal of collected audio signal
The ratio of the energy value of the energy value of frequency band signals and the high frequency band signal, due to the energy of the low-band signal of sound of snoring signal
The ratio range of energy value and the energy value of the low-band signal of normal speech signal of value and the high frequency band signal and described
The ratio range of the energy value of high frequency band signal is significantly different, by judging the ratio whether in the first preset range, from
And can determine whether the collected audio signal is sound of snoring audio signal, improve the accuracy rate of sound of snoring detection.
Further, the second embodiment based on first embodiment proposition sound of snoring detection method of the present invention, reference Fig. 2,
In the present embodiment, the step S10 includes：
Step S11 carries out sampling and quantification treatment to collected audio signal, obtains lasting digital signal, and right
The digital signal carries out windowing process；
Step S12, in window digital signal carry out bandpass filtering, obtain the audio signal low-band signal and
High frequency band signal, wherein the low-band signal and high frequency band signal include multiple frame signals.
After collecting original audio signal in microphone, the audio signal by collected simulation is needed to be converted to number
Audio signal, for example, using 12bit bit AD samples, sample rate is that 2Kbit/s is sampled, the signal that sampling is obtained into
Row quantification treatment obtains lasting digital signal.After obtaining lasting digital signal, adding window is carried out to lasting digital signal
Processing.In the present embodiment, during windowing process, preferably length of window is 1000 milliseconds, and it is 200 milliseconds that frame, which moves,
The filtering for passing through two-way bandpass filter to the signal in window is also needed to after windowing process, obtains the low-band signal
Frame signal x (1), x (2) ..., x (n), the frame signal y (1) of the high frequency band signal, y (2) ..., y (n), the two-way band logical
Filter be preferably Butterworth IIR (Infinite Impulse Response, infinite impulse response) quadravalence 50HZ~
The bandpass filter of 300HZ and the bandpass filter of Butterworth IIR quadravalences 600HZ~950HZ.
Then the present embodiment adds obtained digital signal by carrying out sampling and quantification treatment to the signal adopted
It window and is filtered, to finally obtain the frame signal of low-band signal and high frequency band signal, the audio is believed to convenient
Number energy value calculated.
Further, the 3rd embodiment based on second embodiment proposition sound of snoring detection method of the present invention, reference Fig. 3,
In the present embodiment, the step S20 includes：
Step S21 calculates the energy value of each frame signal of the low-band signal and the high frequency band signal, and handle
Energy value of the energy value of each frame signal of the low-band signal as the low-band signal, the high frequency band signal
Energy value of the energy value of each frame signal as the high frequency band signal.
After obtaining the signal of frame signal and the high band of the low-band signal, pass through preset energy arithmetic meter
The energy value of each frame signal is calculated, the frame signal energy arithmetic of the low-band signal is specially：
Wherein, x (n) is the time domain waveform of each frame signal of low-band signal, and N is frame length, i.e. contains in a frame signal
The number of sampled data, E (low) are the energy value of each frame signal of low-band signal.
Similarly, the frame signal energy arithmetic of the high frequency band signal is specially：
Wherein, y (n) is the time domain waveform of each frame signal of high frequency band signal, and N is frame length, i.e. contains in a frame signal
The number of sampled data, E (high) are the energy value of each frame signal of high-frequency segment signal.
The present embodiment calculates the energy value of each frame signal by energy arithmetic so as to the low-band signal and
The analysis of the high frequency band signal is more convenient, intuitively.
Further, the fourth embodiment that sound of snoring detection method of the present invention is proposed based on above-mentioned 3rd embodiment, with reference to figure
4, in the present embodiment, the step S30 includes：
Step S31 forms the energy wave of low-band signal according to the energy value of each frame signal of the low-band signal；
Step S32 chooses the default adjacent peak institute for meeting preset condition from the energy wave of the low-band signal
The energy value of corresponding frame signal, and obtain and the frame signal of the high frequency band signal corresponding to the energy value for the frame signal chosen
Energy value, the preset condition are the time difference of adjacent peak in the second preset range；
Step S33 calculates the corresponding high frequency band signal of the energy value of the frame signal of the low-band signal of each selection
Frame signal energy value ratio；
Whether step S34 judges each ratio being calculated in first preset range.
In the energy value for judging the low-band signal with the ratio of the energy value of corresponding high frequency band signal whether
When within one preset range, specifically, the energy value of each frame signal for the low-band signal being calculated can be formed low
The energy wave of frequency band signals is the energy waveform figure of the low-band signal of the present embodiment for example, referring to Fig. 5, Fig. 5.Will be described
After the energy value composition energy wave of each frame signal of low-band signal, default is chosen from the energy wave and meets default item
The energy value of frame signal corresponding to the adjacent peak of part, while being obtained from the energy value of each frame signal of the high frequency band signal
Default energy value corresponding with the frame signal of low-band signal chosen is taken, the preset condition is the low-band signal
Energy waveform in adjacent peak time difference in the second preset range；It in another embodiment of the invention, can be by institute
State the energy wave of the energy value composition high frequency band signal of each frame signal of high frequency band signal, then the energy from the high frequency band signal
Choose the default energy value for meeting the frame signal corresponding to the adjacent peak of preset condition in amount wave, while from the low-frequency range
Default energy value corresponding with the frame signal of high frequency band signal chosen, the default item are obtained in each frame signal of signal
Part be the high frequency band signal energy waveform in adjacent peak time difference in the second preset range.Specifically, with described
It, can be by continuing default adjacent peak value in energy waveform for energy waveform is the energy waveform of low-band signal
It is analyzed, obtains the periodic quantity and energy value of default peak value.For example, dividing 5 adjacent peak values in energy waveform
Analysis, obtains 5 energy peak E1 (low), E2 (low) ... E5 (low), 5 periodic quantity Tp1, Tp2 ... Tp5, wherein the week
Time value is the time difference of two adjacent peak values in energy waveform.After obtaining periodic quantity, judge whether above-mentioned 5 periodic quantities are full
It is enough lower rule：(1)、(Tp1,Tp2…Tp5)>3 seconds, i.e., above-mentioned five periodic quantities are both greater than 3 seconds；(2)、(Tp1,Tp2…
Tp5)<8 seconds, i.e., above-mentioned five periodic quantities are both less than 8 seconds；(3), adjacent periods difference t1=Tp2-Tp1 ..., t5=Tp5-
Tp4, (t1...t5)<1 second, i.e., adjacent period difference is both less than 1 second, it is to be understood that each threshold value that the above-mentioned period meets
May be other values.If the energy value for the default frame signal chosen meets above-mentioned rule, obtain and the selection
The energy value of the frame signal of high frequency band signal corresponding to the frame signal of low-band signal, such as E1 (high), E2 (high) ...
E5(high).In the present embodiment, the energy value of the low-band signal and the frame signal corresponding to high frequency band signal, refers to
The energy value of the low-band signal and the high frequency band signal in the frame signal of same time.It is corresponding, calculate the low frequency
The ratio of the energy value of the energy value of segment signal and the high frequency band signal is available such as under type：P1=E1 (low)/E1
(high), P2=E2 (low)/E2 (high) ... P5=E5 (low)/E5 (high).After obtaining each ratio, each ratio is judged
Whether value is all in the first preset range, i.e. whether all in the first preset range, described first is default by P1, P2, P3, P4, P5
50~100 are ranged preferably from, when above-mentioned condition meets, determines that the audio signal is sound of snoring audio signal.Similarly, when described
When energy waveform is the energy waveform of high frequency band signal, a peak value of presetting for continuing adjacent in the energy waveform is divided
Analysis obtains the periodic quantity and energy value of default peak value.For example, analyzing 5 adjacent peak values in energy waveform, obtain
5 energy peak E1 (high1), E2 (high1) ... E5 (high1), 5 periodic quantity Tp1, Tp2 ... Tp5, correspondingly, when choosing
The energy value of the frame signal preset when meeting above-described rule, obtain and believe with the frame of the high frequency band signal of the selection
The energy value of the frame signal of low-band signal corresponding to number, such as E1 (low1), E2 (low1) ... E5 (low1), the low frequency
The calculation formula of the ratio of the energy value of the energy value of segment signal signal and the high frequency band signal is as described below：P1=E1
(low1) (high1 ... P5=E5 (low1)/E5 (high1) are obtaining each ratio by/E1 (high1), P2=E2 (low1)/E2
Afterwards, judge each ratio whether all in the first preset range, i.e. P1, P2, P3, P4, P5 whether all in the first preset range,
First preset range is preferably 50~100, when above-mentioned condition meets, determines that the audio signal is believed for sound of snoring audio
Number.In another embodiment of the invention, it can also then judge the average value by the way that above-mentioned 5 ratio is averaged
Whether in the first preset range, which is preferably 60~100.
The present embodiment by obtain multiple frame signals of low-band signal energy value and corresponding high frequency band signal it is multiple
The ratio of the energy value of frame signal, and by judging the multiple ratios obtained whether in the first preset range, so that it is determined that adopting
Whether the audio signal integrated further improves the accuracy of sound of snoring detection as audio signal.
The present invention further provides a kind of sound of snoring detection devices.
It is the high-level schematic functional block diagram of the first embodiment of apparatus of the present invention with reference to Fig. 6, Fig. 6.
In the present embodiment, the sound of snoring detection device includes：Acquisition module 10, computing module 20, judgment module 30 and
Determining module 40.
The acquisition module 10, for obtaining low-band signal and high frequency band signal in collected audio signal；
The present inventor the study found that the sound of snoring high band (600HZ~950HZ) and low-frequency range (50HZ~
The attenuation ratio of energy distributions 300HZ) has apparent feature, and base region is between (50~100).And normal speech
The attenuation ratio of high band (600HZ~950HZ) and the energy distributions of low-frequency range (50HZ~300HZ) between (1~10),
So the attenuation ratio of the energy distributions of the sound of snoring and normal speech makes a big difference.In the present embodiment, the low-frequency range
Signal is signal component of the frequency within the scope of 50HZ~300HZ in collected audio signal, and the high frequency band signal is to adopt
Signal component of the frequency within the scope of 600HZ~950HZ in the audio signal collected is obtaining the collected audio signal
Low-frequency range and when high frequency band signal, need the audio signal being filtered by two-way bandpass filter, the two-way
Bandpass filter is respectively the bandpass filter of 50HZ~300HZ and the bandpass filter of 600HZ~950HZ, by above-mentioned
After the filtering of bandpass filter, the low-band signal and high frequency band signal of the audio signal can be obtained.
The computing module 20, the energy value for calculating the low-band signal and the high frequency band signal；
After getting the low-band signal and high frequency band signal of the audio signal, the energy of the low-band signal is calculated
The energy value of magnitude and the high frequency band signal, the energy value represent the amplitude of the low-band signal or high frequency band signal.
Specifically, can windowing process be carried out for the low-band signal and high frequency band signal, to obtain the low-band signal and
Then the frame signal of high frequency band signal calculates separately the energy value of every frame signal, calculating each of the signal of the low-frequency range
It, can be using the average value of the energy value of each frame signal of the low-band signal as the low-frequency range when energy value of frame signal
The energy value of signal, while using the average value of the energy value of each frame signal of the high frequency band signal as the high frequency band signal
Energy value, or can also from the energy value of each frame signal of the low-band signal form time domain waveform in choose it is adjacent
Default peak value energy value of the energy value as the low-band signal, while by the low-band signal with the selection
Energy value of the energy value of each frame signal of the corresponding high frequency band signal of peak value as the high frequency band signal, wherein described
Default preferably 5.
The computing module 20 is additionally operable to calculate each frame signal of the low-band signal and the high frequency band signal
Energy value, and using the energy value of each frame signal of the low-band signal as the energy value of the low-band signal, institute
State energy value of the energy value of each frame signal of high frequency band signal as the high frequency band signal.
After obtaining the signal of frame signal and the high band of the low-band signal, pass through preset energy arithmetic meter
The energy value of each frame signal is calculated, the frame signal energy arithmetic of the low-band signal is specially：
Wherein, x (n) is the time domain waveform of each frame signal of low-band signal, and N is frame length, i.e. contains in a frame signal
The number of sampled data, E (low) are the energy value of each frame signal of low-band signal.
Similarly, the frame signal energy arithmetic of the high frequency band signal is specially：
Wherein, y (n) is the time domain waveform of each frame signal of high frequency band signal, and N is frame length, i.e. contains in a frame signal
The number of sampled data, E (high) are the energy value of each frame signal of high-frequency segment signal.
The judgment module 30, the energy value for judging the low-band signal and the corresponding high frequency band signal
Whether the ratio of energy value is in the first preset range；
In the present embodiment, first preset range is 50~100.The energy value of the low-band signal can be institute
It states the energy value of each frame signal of low-band signal, or is averaged for the energy value of each frame signal of the low-band signal
Value, either for default energy value being chosen in the energy value of each frame signal of the low-band signal or be the low frequency
The average value for the default energy value chosen in the energy value of each frame signal of segment signal, correspondingly, the high frequency band signal
Energy value is the energy value of each frame signal of the high frequency band signal, or the energy of each frame signal for the high frequency band signal
The default energy value chosen in the average value of value, or energy value for each frame signal of the high frequency band signal, Huo Zhewei
The average value for the default energy value chosen in the energy value of each frame signal of the high frequency band signal.
The determining module 40, if for the ratio in the first preset range, it is determined that the audio signal is snore
Sound audio signals.
If the ratio is in the first preset range, it is determined that the audio signal is sound of snoring audio signal, the ratio
Not in the first preset range, it is determined that the audio signal is not sound of snoring audio signal, such as normal speech signal or one
A little noise signals etc..
The present embodiment is calculated described low by obtaining the low-band signal and high frequency band signal of collected audio signal
The ratio of the energy value of the energy value of frequency band signals and the high frequency band signal, due to the energy of the low-band signal of sound of snoring signal
The ratio range of energy value and the energy value of the low-band signal of normal speech signal of value and the high frequency band signal and described
The ratio range of the energy value of high frequency band signal is significantly different, by judging the ratio whether in the first preset range, from
And can determine whether the collected audio signal is sound of snoring audio signal, improve the accuracy rate of sound of snoring detection.
Further, the second embodiment based on first embodiment proposition sound of snoring detection device of the present invention, reference Fig. 7,
In the present embodiment, the acquisition module 10 includes：Processing unit 11 and acquiring unit 12.
The processing unit 11 obtains lasting number for carrying out sampling and quantification treatment to collected audio signal
Word signal, and windowing process is carried out to the digital signal；
The acquiring unit 12 obtains the audio signal for carrying out bandpass filtering to the digital signal in window
Low-band signal and high frequency band signal, wherein the low-band signal and high frequency band signal include multiple frame signals；
After collecting original audio signal in microphone, the audio signal by collected simulation is needed to be converted to number
Audio signal, for example, using 12bit bit AD samples, sample rate is that 2Kbit/s is sampled, the signal that sampling is obtained into
Row quantification treatment obtains lasting digital signal.After obtaining lasting digital signal, adding window is carried out to lasting digital signal
Processing.In the present embodiment, during windowing process, preferably length of window is 1000 milliseconds, and it is 200 milliseconds that frame, which moves,
The filtering for passing through two-way bandpass filter to the signal in window is also needed to after windowing process, obtains the low-band signal
Frame signal x (1), x (2) ..., x (n), the frame signal y (1) of the high frequency band signal, y (2) ..., y (n), the two-way band logical
Filter be preferably Butterworth IIR (Infinite Impulse Response, infinite impulse response) quadravalence 50HZ~
The bandpass filter of 300HZ and the bandpass filter of Butterworth IIR quadravalences 600HZ~950HZ.
Then the present embodiment adds obtained digital signal by carrying out sampling and quantification treatment to the signal adopted
It window and is filtered, to finally obtain the frame signal of low-band signal and high frequency band signal, the audio is believed to convenient
Number energy value calculated.
Further, the 3rd embodiment that sound of snoring detection device of the present invention is proposed based on any of the above-described embodiment, with reference to figure
8, in the present embodiment, the judgment module 30 includes：Component units 31, selection unit 32, computing unit 33 and judging unit
34。
The component units 31, the energy value composition low-frequency range for each frame signal according to the low-band signal are believed
Number energy wave；
The selection unit 32 meets preset condition for choosing default from the energy wave of the low-band signal
The energy value of frame signal corresponding to adjacent peak, and obtain and the high frequency band signal corresponding to the energy value for the frame signal chosen
Frame signal energy value, the preset condition be adjacent peak time difference in the second preset range；
The energy value of the computing unit 33, the frame signal of the low-band signal for calculating each selection is corresponding
The ratio of the energy value of the frame signal of high frequency band signal；
The judging unit 34, whether each ratio for judging to be calculated is in first preset range.
In the energy value for judging the low-band signal with the ratio of the energy value of corresponding high frequency band signal whether
When within one preset range, specifically, the energy value of each frame signal for the low-band signal being calculated can be formed low
The energy wave of frequency band signals is the energy waveform figure of the low-band signal of the present embodiment for example, referring to Fig. 5, Fig. 5.Will be described
After the energy value composition energy wave of each frame signal of low-band signal, default is chosen from the energy wave and meets default item
The energy value of frame signal corresponding to the adjacent peak of part, while being obtained from the energy value of each frame signal of the high frequency band signal
Default energy value corresponding with the frame signal of low-band signal chosen is taken, the preset condition is the low-band signal
Energy waveform in adjacent peak time difference in the second preset range；It in another embodiment of the invention, can be by institute
State the energy wave of the energy value composition high frequency band signal of each frame signal of high frequency band signal, then the energy from the high frequency band signal
Choose the default energy value for meeting the frame signal corresponding to the adjacent peak of preset condition in amount wave, while from the low-frequency range
Default energy value corresponding with the frame signal of high frequency band signal chosen, the default item are obtained in each frame signal of signal
Part be the high frequency band signal energy waveform in adjacent peak time difference in the second preset range.Specifically, with described
It, can be by continuing default adjacent peak value in energy waveform for energy waveform is the energy waveform of low-band signal
It is analyzed, obtains the periodic quantity and energy value of default peak value.For example, dividing 5 adjacent peak values in energy waveform
Analysis, obtains 5 energy peak E1 (low), E2 (low) ... E5 (low), 5 periodic quantity Tp1, Tp2 ... Tp5, wherein the week
Time value is the time difference of two adjacent peak values in energy waveform.After obtaining periodic quantity, judge whether above-mentioned 5 periodic quantities are full
It is enough lower rule：(1)、(Tp1,Tp2…Tp5)>3 seconds, i.e., above-mentioned five periodic quantities are both greater than 3 seconds；(2)、(Tp1,Tp2…
Tp5)<8 seconds, i.e., above-mentioned five periodic quantities are both less than 8 seconds；(3), adjacent periods difference t1=Tp2-Tp1 ..., t5=Tp5-
Tp4, (t1...t5)<1 second, i.e., adjacent period difference is both less than 1 second, it is to be understood that each threshold value that the above-mentioned period meets
May be other values.If the energy value for the default frame signal chosen meets above-mentioned rule, obtain and the selection
The energy value of the frame signal of high frequency band signal corresponding to the frame signal of low-band signal, such as E1 (high), E2 (high) ...
E5(high).In the present embodiment, the energy value of the low-band signal and the frame signal corresponding to high frequency band signal, refers to
The energy value of the low-band signal and the high frequency band signal in the frame signal of same time.It is corresponding, calculate the low frequency
The ratio of the energy value of the energy value of segment signal and the high frequency band signal is available such as under type：P1=E1 (low)/E1
(high), P2=E2 (low)/E2 (high) ... P5=E5 (low)/E5 (high).After obtaining each ratio, each ratio is judged
Whether value is all in the first preset range, i.e. whether all in the first preset range, described first is default by P1, P2, P3, P4, P5
50~100 are ranged preferably from, when above-mentioned condition meets, determines that the audio signal is sound of snoring audio signal.Similarly, when described
When energy waveform is the energy waveform of high frequency band signal, a peak value of presetting for continuing adjacent in the energy waveform is divided
Analysis obtains the periodic quantity and energy value of default peak value.For example, analyzing 5 adjacent peak values in energy waveform, obtain
5 energy peak E1 (high1), E2 (high1) ... E5 (high1), 5 periodic quantity Tp1, Tp2 ... Tp5, correspondingly, when choosing
The energy value of the frame signal preset when meeting above-described rule, obtain and believe with the frame of the high frequency band signal of the selection
The energy value of the frame signal of low-band signal corresponding to number, such as E1 (low1), E2 (low1) ... E5 (low1), the low frequency
The calculation formula of the ratio of the energy value of the energy value of segment signal signal and the high frequency band signal is as described below：P1=E1
(low1) (high1 ... P5=E5 (low1)/E5 (high1) are obtaining each ratio by/E1 (high1), P2=E2 (low1)/E2
Afterwards, judge each ratio whether all in the first preset range, i.e. P1, P2, P3, P4, P5 whether all in the first preset range,
First preset range is preferably 50~100, when above-mentioned condition meets, determines that the audio signal is believed for sound of snoring audio
Number.In another embodiment of the invention, it can also then judge the average value by the way that above-mentioned 5 ratio is averaged
Whether in the first preset range, which is preferably 60~100.
The present embodiment by obtain multiple frame signals of low-band signal energy value and corresponding high frequency band signal it is multiple
The ratio of the energy value of frame signal, and by judging the multiple ratios obtained whether in the first preset range, so that it is determined that adopting
Whether the audio signal integrated further improves the accuracy of sound of snoring detection as audio signal.
The embodiments of the present invention are for illustration only, can not represent the quality of embodiment.Pass through above embodiment party
The description of formula, it is required general that those skilled in the art can be understood that above-described embodiment method can add by software
The mode of hardware platform is realized, naturally it is also possible to which by hardware, but the former is more preferably embodiment in many cases.It is based on
Such understanding, substantially the part that contributes to existing technology can be with software product in other words for technical scheme of the present invention
Form embody, which is stored in a storage medium (such as ROM/RAM, magnetic disc, CD), including
Some instructions are used so that a station terminal equipment (can be mobile phone, computer, server, air conditioner or the network equipment etc.)
Execute the method described in each embodiment of the present invention.
It these are only the preferred embodiment of the present invention, be not intended to limit the scope of the invention, it is every to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of sound of snoring detection method, which is characterized in that detection method includes the following steps for the sound of snoring：
Obtain the low-band signal and high frequency band signal in collected audio signal；
Calculate the energy value of the low-band signal and the high frequency band signal；
Judge the energy value of the low-band signal with the ratio of the energy value of the corresponding high frequency band signal whether first
In preset range；
If so, determining that the audio signal is sound of snoring audio signal.
2. sound of snoring detection method as described in claim 1, which is characterized in that low in the collected audio signal of acquisition
The step of frequency band signals and high frequency band signal includes：
Sampling and quantification treatment are carried out to collected audio signal, obtain lasting digital signal, and to the digital signal
Carry out windowing process；
Bandpass filtering is carried out to the digital signal in window, obtains the low-band signal and high frequency band signal of the audio signal,
Wherein, the low-band signal and high frequency band signal include multiple frame signals.
3. sound of snoring detection method as claimed in claim 2, which is characterized in that the calculating low-band signal and the height
The step of energy value of frequency band signals includes：
The energy value of each frame signal of the low-band signal and the high frequency band signal is calculated, and the low-frequency range is believed
Number each frame signal energy value of the energy value as the low-band signal, each frame signal of the high frequency band signal
Energy value of the energy value as the high frequency band signal.
4. sound of snoring detection method as claimed in claim 3, which is characterized in that the energy value for judging the low-band signal
Whether the step in the first preset range includes with the ratio of the energy value of the corresponding high frequency band signal：
The energy wave of low-band signal is formed according to the energy value of each frame signal of the low-band signal；
The default frame signal met corresponding to the adjacent peak of preset condition is chosen from the energy wave of the low-band signal
Energy value, and obtain with choose frame signal energy value corresponding to high frequency band signal frame signal energy value, it is described
Preset condition is the time difference of adjacent peak in the second preset range；
Calculate the energy of the frame signal of the corresponding high frequency band signal of energy value of the frame signal of the low-band signal of each selection
The ratio of magnitude；
Judge each ratio being calculated whether in first preset range.
5. sound of snoring detection method as claimed in claim 3, which is characterized in that the energy value for judging the low-band signal
Whether the step in the first preset range includes with the ratio of the energy value of the corresponding high frequency band signal：
The energy wave of high frequency band signal is formed according to the energy value of each frame signal of the high frequency band signal；
The default frame signal met corresponding to the adjacent peak of preset condition is chosen from the energy wave of the high frequency band signal
Energy value, and obtain with choose frame signal energy value corresponding to low-band signal frame signal energy value, it is described
Preset condition is the time difference of adjacent peak in the second preset range；
Calculate the frame signal of the energy value of the frame signal of the low-band signal of each acquisition and the high frequency band signal of selection
The ratio of energy value；
Judge each ratio being calculated whether in first preset range.
6. a kind of sound of snoring detection device, which is characterized in that the sound of snoring detection device includes：
Acquisition module, for obtaining low-band signal and high frequency band signal in collected audio signal；
Computing module, the energy value for calculating the low-band signal and the high frequency band signal；
Judgment module, the ratio of energy value and the energy value of the corresponding high frequency band signal for judging the low-band signal
Whether value is in the first preset range；
Determining module, if for the ratio in the first preset range, it is determined that the audio signal is sound of snoring audio signal.
7. sound of snoring detection device as claimed in claim 6, which is characterized in that the acquisition module includes：
Processing unit obtains lasting digital signal, and right for carrying out sampling and quantification treatment to collected audio signal
The digital signal carries out windowing process；
Acquiring unit obtains the low-band signal of the audio signal for carrying out bandpass filtering to the digital signal in window
And high frequency band signal, wherein the low-band signal and high frequency band signal include multiple frame signals.
8. sound of snoring detection device as claimed in claim 7, which is characterized in that the computing module is additionally operable to calculate described low
The energy value of each frame signal of frequency band signals and the high frequency band signal, and each frame signal of the low-band signal
Energy value of the energy value as the low-band signal, described in the energy value of each frame signal of the high frequency band signal is used as
The energy value of high frequency band signal.
9. sound of snoring detection device as claimed in claim 8, which is characterized in that the judgment module includes：
Component units are used to form the energy of low-band signal according to the energy value of each frame signal of the low-band signal
Wave；
Selection unit, for choosing the default adjacent peak institute for meeting preset condition from the energy wave of the low-band signal
The energy value of corresponding frame signal, and obtain and the frame signal of the high frequency band signal corresponding to the energy value for the frame signal chosen
Energy value, the preset condition are the time difference of adjacent peak in the second preset range；
Computing unit, the corresponding high frequency band signal of the energy value of the frame signal of the low-band signal for calculating each selection
Frame signal energy value ratio；
Judging unit, whether each ratio for judging to be calculated is in first preset range.
10. sound of snoring detection device as claimed in claim 9, which is characterized in that the component units are additionally operable to according to the height
The energy wave of the energy value composition high frequency band signal of each frame signal of frequency band signals；
The selection unit is additionally operable to default of the selection from the energy wave of the high frequency band signal and meets the adjacent of preset condition
The energy value of frame signal corresponding to peak value, and obtain the frame with the low-band signal corresponding to the energy value for the frame signal chosen
The energy value of signal, the preset condition are the time difference of adjacent peak in the second preset range；
The computing unit is additionally operable to calculate the height of the energy value and selection of the frame signal of the low-band signal of each acquisition
The ratio of the energy value of the frame signal of frequency band signals；
Whether the judging unit is additionally operable to judge each ratio being calculated in first preset range.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610005669.9A CN105534480B (en) | 2016-01-05 | 2016-01-05 | Sound of snoring detection method and device |
| PCT/CN2016/101248 WO2017118126A1 (en) | 2016-01-05 | 2016-09-30 | Method and device for detecting snoring sound |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610005669.9A CN105534480B (en) | 2016-01-05 | 2016-01-05 | Sound of snoring detection method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105534480A CN105534480A (en) | 2016-05-04 |
| CN105534480B true CN105534480B (en) | 2018-08-14 |
Family
ID=55814435
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610005669.9A Active CN105534480B (en) | 2016-01-05 | 2016-01-05 | Sound of snoring detection method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105534480B (en) |
| WO (1) | WO2017118126A1 (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105534480B (en) * | 2016-01-05 | 2018-08-14 | 深圳和而泰智能控制股份有限公司 | Sound of snoring detection method and device |
| CN106725337B (en) * | 2017-01-05 | 2021-02-19 | 北京怡和嘉业医疗科技股份有限公司 | Snore detection method and device and positive pressure breathing machine |
| CN110547802B (en) * | 2019-09-11 | 2022-09-06 | 京东方科技集团股份有限公司 | Device for recognizing respiratory state |
| CN113938805B (en) * | 2020-07-14 | 2024-04-23 | 广州汽车集团股份有限公司 | Method and device for quantizing bass tone quality |
| CN116229983A (en) * | 2022-09-15 | 2023-06-06 | 深圳市倍轻松科技股份有限公司 | Method and device for determining snore signals, electronic equipment and storage medium |
| CN117892067B (en) * | 2024-03-15 | 2024-05-28 | 国网上海市电力公司 | Low-frequency current partial discharge monitoring anti-interference method, device, equipment and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001276020A (en) * | 2000-04-04 | 2001-10-09 | M I Labs:Kk | Biological information collecting instrument utilizing hermetically sealed air type sound sensor |
| CN201840571U (en) * | 2010-04-16 | 2011-05-25 | 金同(广州)医疗保健品有限公司 | Snore tester |
| CN103220384A (en) * | 2012-01-18 | 2013-07-24 | 黄魏宁 | Dangerous snore cell phone acquisition and self-rescue system |
| CN103251388A (en) * | 2013-04-25 | 2013-08-21 | 北京大学深圳研究生院 | Method and system of snoring monitoring and prevention and treatment based on smart phone platform |
| CN104622432A (en) * | 2015-02-06 | 2015-05-20 | 华南理工大学 | Sleep snore monitoring method and system based on bass ratio |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7246619B2 (en) * | 2001-10-10 | 2007-07-24 | Ric Investments, Llc | Snore detecting method and apparatus |
| CN102499637B (en) * | 2011-09-26 | 2013-11-06 | 大连理工大学 | Obstructive sleep apnea-hypopnea syndrome screening method and device thereof |
| CN102429662B (en) * | 2011-11-10 | 2014-04-09 | 大连理工大学 | Screening system for sleep apnea syndrome in family environment |
| JP2015188642A (en) * | 2014-03-28 | 2015-11-02 | パイオニア株式会社 | Respiratory sound analysis apparatus, respiratory sound analysis method, computer program, and recording medium |
| CN105147244B (en) * | 2015-06-24 | 2017-03-22 | 湖南明康中锦医疗科技发展有限公司 | Snoring detection control equipment and detection control method thereof |
| CN105534480B (en) * | 2016-01-05 | 2018-08-14 | 深圳和而泰智能控制股份有限公司 | Sound of snoring detection method and device |
-
2016
- 2016-01-05 CN CN201610005669.9A patent/CN105534480B/en active Active
- 2016-09-30 WO PCT/CN2016/101248 patent/WO2017118126A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001276020A (en) * | 2000-04-04 | 2001-10-09 | M I Labs:Kk | Biological information collecting instrument utilizing hermetically sealed air type sound sensor |
| CN201840571U (en) * | 2010-04-16 | 2011-05-25 | 金同(广州)医疗保健品有限公司 | Snore tester |
| CN103220384A (en) * | 2012-01-18 | 2013-07-24 | 黄魏宁 | Dangerous snore cell phone acquisition and self-rescue system |
| CN103251388A (en) * | 2013-04-25 | 2013-08-21 | 北京大学深圳研究生院 | Method and system of snoring monitoring and prevention and treatment based on smart phone platform |
| CN104622432A (en) * | 2015-02-06 | 2015-05-20 | 华南理工大学 | Sleep snore monitoring method and system based on bass ratio |
Non-Patent Citations (1)
| Title |
|---|
| 阻塞性睡眠呼吸暂停综合征的分级检测;张海秀 等;《现代生物医学进展》;20120331;第12卷(第7期);第1247-1250页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2017118126A1 (en) | 2017-07-13 |
| CN105534480A (en) | 2016-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105534480B (en) | Sound of snoring detection method and device | |
| CN103440869B (en) | Audio-reverberation inhibiting device and inhibiting method thereof | |
| US7508948B2 (en) | Reverberation removal | |
| DE602004001694T2 (en) | Device for suppressing wind noise | |
| CN103413547B (en) | A kind of method that room reverberation is eliminated | |
| CN104538011B (en) | A kind of tone adjusting method, device and terminal device | |
| CN108663115A (en) | A kind of car inside idle noise objective quantification evaluation method | |
| CN108766454A (en) | A kind of voice noise suppressing method and device | |
| CN104916288B (en) | The method and device of the prominent processing of voice in a kind of audio | |
| CN103544961B (en) | Audio signal processing method and device | |
| US20100329471A1 (en) | Ambient noise compensation system | |
| CN103730110B (en) | A kind of method and apparatus of detection sound end | |
| CN104658543A (en) | Method for eliminating indoor reverberation | |
| CN102475554A (en) | Method for guiding interior sound package by utilizing sound quality | |
| Gerkmann et al. | Empirical distributions of DFT-domain speech coefficients based on estimated speech variances | |
| CN109920439A (en) | The variable-speed motor that subtracts based on tone energy and human ear frequency selectivity is uttered long and high-pitched sounds evaluation method | |
| CN104916292A (en) | Method and apparatus for detecting audio signals | |
| CN103280225B (en) | Low-complexity silence detection method | |
| Sottek | Progress in calculating tonality of technical sounds | |
| CN110751959A (en) | Method for evaluating noise discomfort degree of automobile | |
| CN115700544A (en) | Chromatographic signal denoising method combining empirical mode decomposition and wavelet soft threshold | |
| CN106024017A (en) | Voice detection method and device | |
| CN1044293C (en) | Method and apparatus for encoding/decoding of background sounds | |
| Blommer et al. | Sound quality metric development for wind buffeting and gusting noise | |
| CN109600697A (en) | The outer playback matter of terminal determines method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |