CN109741759A - A kind of acoustics automatic testing method towards specific birds species - Google Patents
A kind of acoustics automatic testing method towards specific birds species Download PDFInfo
- Publication number
- CN109741759A CN109741759A CN201811566250.6A CN201811566250A CN109741759A CN 109741759 A CN109741759 A CN 109741759A CN 201811566250 A CN201811566250 A CN 201811566250A CN 109741759 A CN109741759 A CN 109741759A
- Authority
- CN
- China
- Prior art keywords
- potential
- singing
- signal
- segment
- channel
- 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.)
- Granted
Links
- 241000271566 Aves Species 0.000 title claims abstract description 22
- 238000012360 testing method Methods 0.000 title abstract 2
- 241000894007 species Species 0.000 claims abstract description 48
- 238000001514 detection method Methods 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims abstract description 27
- 230000003044 adaptive effect Effects 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000012706 support-vector machine Methods 0.000 claims abstract description 11
- 238000012544 monitoring process Methods 0.000 claims description 16
- 239000012634 fragment Substances 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000007613 environmental effect Effects 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000001228 spectrum Methods 0.000 claims description 5
- 230000000903 blocking effect Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 230000003068 static effect Effects 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 238000003491 array Methods 0.000 claims description 2
- 238000009432 framing Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 claims description 2
- 238000005070 sampling Methods 0.000 claims description 2
- 238000012549 training Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 abstract description 3
- 238000000605 extraction Methods 0.000 abstract description 2
- 238000012512 characterization method Methods 0.000 abstract 2
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000005236 sound signal Effects 0.000 description 3
- 241001465754 Metazoa Species 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 241000282412 Homo Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Catching Or Destruction (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
The invention discloses a kind of acoustics automatic testing methods towards specific birds species.This method first uses the potential song segment of the specific birds species of acoustic events detection processing procedure extraction based on gauss hybrid models, in conjunction with based on candidate sound event energy, the duration that pipes, frequency distribution last handling process, complete robust detection and automatic segmentation；Then self-adapting signal noise reduction is carried out to each sound clip, corresponding mel cepstrum coefficients characteristic parameter is extracted to enhanced sound clip, obtains potential song segment characterizations collection；Potential song segment characterizations collection and support vector machines is finally combined to complete specific birds species Acoustic detection.The present invention realizes that process is convenient; furthermore the signal-to-noise ratio of sound event can be obviously improved by using the adaptive signal enhancement processing based on microphone array to improve particular species identification accuracy rate; the Acoustic detection for realizing specific birds species under the natural environment of field, is of great significance for the ecological protection of field Rare Birds species.
Description
Technical Field
The invention belongs to the field of ecological monitoring and acoustic signal technology identification, and particularly relates to an automatic acoustic detection method for specific bird species.
Background
Birds are widely distributed sounding animals, the life habits of birds are easier to observe by humans than other animal communities, and the birds can sensitively sense small changes of the ecological environment, so that the birds are considered as ideal species for monitoring changes of the ecological environment by most ecological scholars.
However, in recent years, due to the continuous expansion of human urbanization, the continuous destruction of the ecological environment by human activities leads to the large-scale reduction of the number and variety of birds, which not only causes the loss of the diversity of species in the world, but also leads to the imbalance of the ecological environment caused by the sharp reduction of birds as the indicative species of forest vegetation communities. Therefore, the protection of birds is the focus of attention in the ecological field, and especially the supervision of rare birds is more important. Traditional pointcounting methods (pointcounts) are costly and invade the habitat of organisms when monitoring birds. There is an increasing need to effectively implement automatic bird detection in order to more rapidly and non-invasively assess bird activity.
Based on the singing signal of bird species, the characteristics of the field actual measurement bird sound signal are extracted by using an acoustic signal analysis means, and the method is the basis for carrying out large-scale data analysis and bird sound identification model establishment subsequently. In a real complex environment, considering that the sensitivity of the bird species identification and classification method based on the characteristic parameters to environmental noise and other interference sound source signals is high, many scholars actively explore the bird sound identification method in noise. Chinese patent CN102930870A discloses a bird voice recognition method using an anti-noise power normalization cepstrum coefficient, which includes first performing noise reduction processing on a voice power spectrum by using a multi-band spectrum subtraction method, then extracting the anti-noise power normalization cepstrum coefficient from the noise-reduced voice power spectrum, and finally performing recognition under different environments and signal-to-noise ratios on 34 kinds of bird voices by combining a Support Vector Machine (SVM). Chinese patent CN103489446A discloses a method for identifying birds singing based on self-adaptive energy detection in a complex environment, which comprises the steps of firstly carrying out energy detection on a sound signal, extracting Wavelet Packet decomposition Subband cepstrum coefficients (WPSCC) anti-noise characteristics based on Mel scale from a detected and screened sound signal frame, and carrying out classification identification on 15 types of birds singing in a noise environment by combining with an SVM. A bird sound recognition algorithm under a noise environment based on random forest and large-scale acoustic feature extraction is introduced by a fluidic system (fluidic environment bird sound recognition simulation [ J ] system simulation technology of random forest and large-scale acoustic features, 2017 (359-. Zhang Sai et al first extract potential birds singing segments using a Gaussian Mixture Model (GMM) -based acoustic event detection process, then extract Mel-subband-based parameterized features of each segment, and use SVM to classify and recognize 11 types of birds singing in the field environment (Zhang Sai Hua, Zhao Mega, xuanzhi, Zhang Yi. automatic birds singing recognition based on Mel-subband parameterized features [ J ] computer application, 2017,37(4): 1111-.
Most of the existing researches only add one kind of noise or do not suppress the noise in the experiment, however, the field acoustic environment is very complex and has various environmental interference sources, and the requirement of the acoustic monitoring task under the actual field complex acoustic environment cannot be met only by adopting single-channel audio enhancement front-end processing.
Disclosure of Invention
The invention aims to provide an automatic acoustic detection method for specific bird species.
Experiment the technical solution of the present invention is: an acoustic automatic detection method for specific bird species, comprising the steps of:
step 1, acquiring field continuous bird sound monitoring data signals, automatically segmenting, and then extracting potential sound fragments of specific bird species;
step 2, carrying out adaptive signal noise reduction enhancement processing on each potential singing segment obtained in the step 1;
step 3, extracting characteristic parameters of each potential singing segment subjected to noise reduction enhancement in the step 2, and constructing a feature set of the potential singing segment;
and 4, combining the potential song segment feature set obtained in the step 3 and an identification algorithm in machine learning to finish acoustic detection of specific bird species.
Compared with the prior art, the invention has the following remarkable advantages: 1) according to the invention, a multi-element three-dimensional microphone array is adopted to collect continuous bird sound monitoring data, and the collected data contains abundant time and space information, so that the monitoring of bird species on a large space-time scale can be realized; 2) the method can complete steady detection and automatic segmentation by detecting the potential chirping event of the specific bird species based on the Gaussian mixture model and combining the post-processing process based on the energy, chirping duration and frequency distribution of the candidate chirping event; 3) according to the method, the noise reduction enhancement processing of the self-adaptive signals based on the microphone array is adopted, so that the signal-to-noise ratio of a sound event is improved, and the identification accuracy of specific species is improved; 4) the method of the invention has convenient implementation process and easy implementation.
The present invention is described in further detail below with reference to the attached drawings.
Drawings
FIG. 1 is a flow chart of the acoustic automatic detection method for specific bird species of the present invention.
FIG. 2 is a flow chart of the detection of a potential chirping event for a particular avian species of the present invention.
Fig. 3 is a schematic diagram of adaptive delay estimation of a P-element stereo microphone array.
Fig. 4 is a block diagram of a generalized sidelobe canceller structure of a P-element stereo microphone array.
Fig. 5 is a schematic diagram of adaptive delay estimation of a 4-element stereo microphone array.
Fig. 6 is a block diagram of a 4-element stereo microphone array generalized sidelobe canceller structure.
Detailed Description
Referring to fig. 1, the automatic acoustic detection method for specific bird species of the present invention comprises the following steps:
step 1, collecting field continuous bird sound monitoring data signals, automatically segmenting, and then extracting potential singing segments of specific bird species.
Further, with reference to fig. 2, step 1 specifically includes:
step 1-1, collecting multi-channel field continuous bird sound monitoring data signals by using a multi-element three-dimensional microphone array, and performing pre-emphasis processing on the collected field continuous bird sound monitoring data signals to compensate for excessive attenuation of high-frequency signals and inhibit low-frequency wind noise;
step 1-2, performing framing, windowing and fast Fourier transform on the continuous bird sound monitoring data signal processed in the step 1-1 to obtain a power spectrogram;
1-3, setting the lower limit and the upper limit of the frequency to be f respectivelyLAnd fHDetermining the short-time logarithmic energy le (l) of each frame by the formula:
le(l)＝log10(e(l))
wherein,
where l is a frame number, i is a frequency number, S (i, l) represents a short-time Fourier transform result at a time-frequency point (i, l), and NLAnd NHRespectively represents fLAnd fHThe corresponding frequency point serial number, e (l) is the short-time energy of the l frame;
step 1-4, generating frame logarithmic energy distribution by using a Gaussian mixture model containing two Gaussian components, wherein the two Gaussian components respectively represent probability density functions of a potential chirping event frame set and an environmental noise frame set;
step 1-5, aiming at each frame, judging whether the frame belongs to a potential sound segment or an environmental noise segment by comparing the posterior probability to obtain a plurality of potential sound segments, specifically:
comparing the posterior probability of each frame belonging to the set of potential singing event frames with the posterior probability belonging to the set of environmental noise frames, if the posterior probability of the frame belonging to the set of potential singing event frames is greater than the posterior probability belonging to the set of environmental noise frames, the frame belongs to a certain potential singing segment, and other frames which are continuous with the frame in time and also meet the conditions are also belonging to the segment potential singing segment, thereby obtaining a plurality of potential singing segments, wherein all the potential singing segments form a set D ═ AE { (AE)1,AE2,…,AEKK is the number of potential chirping segments;
step 1-6, solving the logarithmic energy of each potential ringing segment obtained in step 1-5, wherein the formula is as follows:
and obtaining the maximum logarithmic energy ME:
for the kth potential chirp segment, if ME-EAEkIf the potential sound segment is more than or equal to q, the potential sound segment is considered to be an over-weak segment with low ecological research value, and the over-weak segment is removed, wherein q is a threshold value preset according to the actual situation, and the unit of q is dB;
step 1-7, based on the existing specific bird species singing database data, obtaining the upper and lower thresholds of the singing time of the specific bird species singing segment, namely the longest singing time t through statistical analysisHAnd the shortest sounding time tLAnd according to the signal sampling rate fsWill tHAnd tLConverted into a maximum chirp length nHAnd a minimum chirp length nL:
nH＝fs×tH
nL＝fs×tL
And acquiring the length T of each potential chirping segment obtained in the steps 1-6 as follows:
t-frame length x number of frames in a potential chirp segment
The length T of the potential sound segment is less than nLAnd is greater than nHEliminating potential sound segments;
step 1-8, acquiring the frequency distribution range of the specific bird species singing fragment through statistical analysis based on the existing specific bird species singing database data, and setting the data beyond the frequency range to zero aiming at the potential singing fragment acquired in the step 1-7.
Further, q in steps 1-6 is 20 dB.
And 2, carrying out adaptive signal noise reduction enhancement processing on each potential singing segment obtained in the step 1.
Further, step 2 performs adaptive signal noise reduction enhancement processing on each potential chirping segment obtained in step 1, specifically:
assuming that the multi-element stereo microphone array is a P-element stereo microphone array, sequentially numbering P-element stereo microphone array channels in a certain sequence to be 1,2,3 …, P;
step 2-1, with reference to fig. 3, performing sound source direction estimation on each potential chirping segment by using a self-adaptive filtering method, specifically:
step 2-1-1, aiming at P channel signal data of one potential singing segment, supposing that P channel signals are m respectively1(n)、m2(n)、m3(n)、…、mP(n)，n＝1,2,3,...,Lm，LmTaking the signal of the channel 1 as a reference signal for signal length, and constructing a snapshot x of the signal of the channel 2k：
T
xk＝[m2(k),m2(k+1),...,m2(k+L-1)]；
Wherein, subscript k is 1,2m-L +1 denotes the kth snapshot, L denotes the filter length, and superscript T denotes the transpose;
step 2-1-2, obtaining autocorrelation matrix RxxThe formula used is:
wherein K is Lm-L +1 is the number of snapshots;
step 2-1-3, obtaining cross correlation matrix rxdThe formula used is:
in the formula,is the filter center point;
step 2-1-4, solving the weight vector w1The formula used is:
w1＝Rxx -1rxd；
step 2-1-5, for the weight vector w obtained in step 2-1-41Detecting the peak value, recording the abscissa of the peak value as z, and counting the number d of the time delay points of the channel 1 signal and the channel 2 signal1＝z-D；
Step 2-1-6, repeating the step 2-1-1 to the step 2-1-5, and obtaining the number d of the time delay points of the channel 1 signal and the c channel signalc，c＝1,2,...,P-1；
Step 2-2, with reference to fig. 4, performs adaptive enhancement on the potential chirping segment by using a generalized sidelobe canceller, specifically:
step 2-2-1, obtaining a main channel signal d (k):
d(k)＝wc Tm(k)；
in the formula, wc＝[wc1,wc2,..,wcP]TIs a static weight vector, wc1、wc2、..、wcPA weight value w corresponding to each channelc1+wc2+...+wcP＝1；m(k)＝[m1(k),m2(k-d1),...,mP(k-dP-1)]T，k＝1,2,...,Lm；
Step 2-2-2, finding auxiliary channel signal e (k):
wherein, WSA blocking matrix of dimension P x (P-1);
step 2-2-3, obtaining an enhanced pure potential singing segment signal y (k):
y(k)＝d(k)-vT(k)e(k)；
where v (k) represents the dynamic weight vector of the adaptive interference canceller.
Further, step 2 performs adaptive signal noise reduction enhancement processing on each potential chirping segment obtained in step 1, specifically:
assuming that the multi-element stereo microphone array is a 4-element stereo microphone array; numbering 4-element three-dimensional microphone arrays in a certain sequence as 1,2,3 and 4;
step 2-1, with reference to fig. 5, performing sound source direction estimation on each potential chirping segment by using a self-adaptive filtering method, specifically:
step 2-1-1, aiming at 4-channel signal data of one potential chirping segment, assuming that the 4-channel signals are m respectively1(n)、m2(n)、m3(n)、m4(n)，n＝1,2,3,...,Lm，LmTaking the channel 4 signal as a reference signal for signal length, and constructing a snapshot x of the channel 1 signalk：
xk＝[m1(k),m1(k+1),...,m1(k+L-1)]T；
Wherein, subscript k is 1,2m-L +1 denotes the kth snapshot, L denotes the filter length, and superscript T denotes the transpose;
step 2-1-2, obtaining autocorrelation matrix RxxThe formula used is:
wherein K is Lm-L +1 is the number of snapshots;
step 2-1-3, obtaining cross correlation matrix rxdThe formula used is:
in the formula,is the filter center point;
step 2-1-4, solving the weight vector w1The formula used is:
w1＝Rxx -1rxd；
step 2-1-5, for the weight vector w obtained in step 2-1-41Detecting the peak value, recording the abscissa of the peak value as z, and counting the number d of the time delay points of the channel 4 signal and the channel 1 signal1＝z-D；
Step 2-1-6, repeating the step 2-1-1 to the step 2-1-5, and obtaining the number of time delay points of the channel 4 signal, the channel 2 signal and the channel 3 signal, wherein the number of the time delay points is d2、d3；
Step 2-2, with reference to fig. 6, performs adaptive enhancement on the potential chirping segment by using a generalized sidelobe canceller, specifically:
step 2-2-1, obtaining a main channel signal d (k):
d(k)＝wc Tm(k)；
in the formula, wc＝[wc4,wc1,wc2,wc3]TIn the form of a static weight vector, the weight vector,wc1、...、wc4is the weight value w corresponding to each channelc1+wc2+wc3+wc4＝1，m(k)＝[m4(k),m1(k-d1),m2(k-d2),m3(k-d3)]T，k＝1,2,...,Lm；
Step 2-2-2, finding auxiliary channel signal e (k):
wherein, WSA blocking matrix of dimension 4 x 3;
step 2-2-3, obtaining an enhanced pure potential singing segment signal y (k):
y(k)＝d(k)-vT(k)e(k)；
where v (k) represents the weight vector of the adaptive interference canceller.
And 3, extracting characteristic parameters of each potential singing segment subjected to noise reduction enhancement in the step 2, and constructing a feature set of the potential singing segment.
Further, step 3 specifically comprises:
step 3-1, calculating a power spectrogram of each potential singing segment after noise reduction and enhancement according to the step 1-2;
step 3-2, setting a Mel band-pass filter group in the frequency distribution range of the singing fragment of the specific bird species, and then enabling the power spectrum of the potential singing fragment to pass through the filter group to obtain the output of each filter;
3-3, taking logarithm of the output result, and performing discrete cosine transform to obtain a Mel cepstrum coefficient characteristic parameter;
and 3-4, combining the Mel cepstrum coefficient characteristic parameters corresponding to all potential singing segments to construct and obtain a feature set of the potential singing segments.
And 4, combining the potential song segment feature set obtained in the step 3 and an identification algorithm in machine learning to finish acoustic detection of specific bird species.
Further, the identification algorithm in the step 4 specifically adopts a support vector machine identification algorithm.
Further, step 4 specifically includes:
the method comprises the steps of taking Mel cepstrum coefficient features in an existing specific bird species singing feature database as training samples, taking potential singing segment feature sets as input samples of a support vector machine, and automatically detecting specific bird species through decision of the support vector machine.
In summary, the acoustic automatic detection method for specific bird species of the present invention employs GMM to detect potential whistling events of specific bird species in combination with a post-processing process based on energy, frame length, and frequency distribution of candidate sound events, thereby completing robust detection and automatic segmentation. In addition, the invention can obviously improve the signal-to-noise ratio of the sound event by adopting the self-adaptive noise reduction enhancement treatment based on the microphone, thereby improving the identification accuracy of the specific species, realizing the acoustic detection of the specific bird species under the field natural environment and having important significance for the ecological protection and the related ecological research of the field rare bird species.
Claims (8)
1. An acoustic automatic detection method for specific bird species, characterized by comprising the following steps:
step 1, acquiring field continuous bird sound monitoring data signals, automatically segmenting, and then extracting potential sound fragments of specific bird species;
step 2, carrying out adaptive signal noise reduction enhancement processing on each potential singing segment obtained in the step 1;
step 3, extracting characteristic parameters of each potential singing segment subjected to noise reduction enhancement in the step 2, and constructing a feature set of the potential singing segment;
and 4, combining the potential song segment feature set obtained in the step 3 and an identification algorithm in machine learning to finish acoustic detection of specific bird species.
2. The acoustic automatic detection method for specific bird species according to claim 1, wherein the step 1 of acquiring field continuous bird sound monitoring data signals, automatically segmenting the signals, and then extracting potential song segments of the specific bird species specifically comprises the following steps:
step 1-1, collecting multi-channel field continuous birdsound monitoring data signals by using a multi-element three-dimensional microphone array, and performing pre-emphasis processing on the collected field continuous birdsound monitoring data signals;
step 1-2, performing framing, windowing and fast Fourier transform on the continuous bird sound monitoring data signal processed in the step 1-1 to obtain a power spectrogram;
1-3, setting the lower limit and the upper limit of the frequency to be f respectivelyLAnd fHDetermining the short-time logarithmic energy le (l) of each frame by the formula:
le(l)＝log10(e(l))
wherein,
where l is a frame number, i is a frequency number, S (i, l) represents a short-time Fourier transform result at a time-frequency point (i, l), and NLAnd NHRespectively represents fLAnd fHThe corresponding frequency point serial number, e (l) is the short-time energy of the l frame;
step 1-4, generating frame logarithmic energy distribution by using a Gaussian mixture model containing two Gaussian components, wherein the two Gaussian components respectively represent probability density functions of a potential chirping event frame set and an environmental noise frame set;
step 1-5, aiming at each frame, judging whether the frame belongs to a potential sound segment or an environmental noise segment by comparing the posterior probability to obtain a plurality of potential sound segments, specifically:
comparing the posterior probability of each frame belonging to the set of potential singing event frames with the posterior probability belonging to the set of environmental noise frames, if the posterior probability of the frame belonging to the set of potential singing event frames is greater than the posterior probability belonging to the set of environmental noise frames, the frame belongs to a certain potential singing segment, and other frames which are continuous with the frame in time and also meet the conditions are also belonging to the segment potential singing segment, thereby obtaining a plurality of potential singing segments, wherein all the potential singing segments form a set D ═ AE { (AE)1,AE2,…,AEKK is the number of potential chirping segments;
step 1-6, solving the logarithmic energy of each potential ringing segment obtained in step 1-5, wherein the formula is as follows:
and obtaining the maximum logarithmic energy ME:
for the kth potential chirp segment, if ME-EAEkIf the number of the potential sound segments is larger than or equal to q, eliminating the potential sound segments, wherein q is a threshold preset according to the actual situation, and the unit of q is dB;
step 1-7, based on the existing specific bird species singing database data, obtaining the upper and lower thresholds of the singing time of the specific bird species singing segment, namely the longest singing time t through statistical analysisHAnd the shortest sounding time tLAnd according to the signal sampling rate fsWill tHAnd tLConverted into a maximum chirp length nHAnd a minimum chirp length nL:
nH＝fs×tH
nL＝fs×tL
And acquiring the length T of each potential chirping segment obtained in the steps 1-6 as follows:
t-frame length x number of frames in a potential chirp segment
The length T of the potential sound segment is less than nLAnd is greater than nHEliminating potential sound segments;
step 1-8, acquiring the frequency distribution range of the specific bird species singing fragment through statistical analysis based on the existing specific bird species singing database data, and setting the data beyond the frequency range to zero aiming at the potential singing fragment acquired in the step 1-7.
3. The method for acoustic automatic detection of specific avian species according to claim 2, characterized in that q in steps 1-6 is taken to be 20 dB.
4. The method for acoustic automatic detection facing specific bird species according to claim 1 or 2, wherein step 2 is to perform adaptive signal noise reduction enhancement processing on each potential singing segment obtained in step 1, and specifically comprises the following steps:
assuming that the multi-element stereo microphone array is a P-element stereo microphone array, sequentially numbering P-element stereo microphone array channels in a certain sequence to be 1,2,3 …, P;
step 2-1, performing sound source direction estimation on each potential singing segment by adopting a self-adaptive filtering method, which specifically comprises the following steps:
step 2-1-1, aiming at P channel signal data of one potential singing segment, supposing that P channel signals are m respectively1(n)、m2(n)、m3(n)、…、mP(n)，n＝1,2,3,...,Lm，LmTaking the signal of the channel 1 as a reference signal for signal length, and constructing a snapshot x of the signal of the channel 2k：
xk＝[m2(k),m2(k+1),...,m2(k+L-1)]T；
Wherein, subscript k is 1,2m-L +1 denotes the kth snapshot, L denotes the filter length, and superscript T denotes the transpose;
step 2-1-2, obtaining autocorrelation matrix RxxThe formula used is:
wherein K is Lm-L +1 is the number of snapshots;
step 2-1-3, obtaining cross correlation matrix rxdThe formula used is:
in the formula,is the filter center point;
step 2-1-4, solving the weight vector w1The formula used is:
w1＝Rxx -1rxd；
step 2-1-5, for the weight vector w obtained in step 2-1-41Detecting the peak value, recording the abscissa of the peak value as z, and counting the number d of the time delay points of the channel 1 signal and the channel 2 signal1＝z-D；
Step 2-1-6, repeating the step 2-1-1 to the step 2-1-5, and obtaining the number d of the time delay points of the channel 1 signal and the c channel signalc，c＝1,2,...,P-1；
Step 2-2, adopting a generalized sidelobe canceller to perform self-adaptive enhancement on the potential sound fragment, specifically:
step 2-2-1, obtaining a main channel signal d (k):
d(k)＝wc Tm(k)；
in the formula, wc＝[wc1,wc2,..,wcP]TIs a static weight vector, wc1、wc2、..、wcPA weight value w corresponding to each channelc1+wc2+...+wcP＝1；m(k)＝[m1(k),m2(k-d1),...,mP(k-dP-1)]T，k＝1,2,...,Lm；
Step 2-2-2, finding auxiliary channel signal e (k):
wherein, WSA blocking matrix of dimension P x (P-1);
step 2-2-3, obtaining an enhanced pure potential singing segment signal y (k):
y(k)＝d(k)-vT(k)e(k)；
where v (k) represents the dynamic weight vector of the adaptive interference canceller.
5. The method of claim 4, wherein the step 2 of performing adaptive signal noise reduction enhancement processing on each potential chirping segment obtained in the step 1 specifically comprises the following steps:
assuming that the multi-element stereo microphone array is a 4-element stereo microphone array; numbering 4-element three-dimensional microphone arrays in a certain sequence as 1,2,3 and 4;
step 2-1, performing sound source direction estimation on each potential singing segment by adopting a self-adaptive filtering method, which specifically comprises the following steps:
step 2-1-1, aiming at 4-channel signal data of one potential chirping segment, assuming that the 4-channel signals are m respectively1(n)、m2(n)、m3(n)、m4(n)，n＝1,2,3,...,Lm，LmTaking the channel 4 signal as a reference signal for signal length, and constructing a snapshot x of the channel 1 signalk：
xk＝[m1(k),m1(k+1),...,m1(k+L-1)]T；
Wherein, subscript k is 1,2m-L +1 denotes the kth snapshot, L denotes the filter length, and superscript T denotes the transpose;
step 2-1-2, obtaining autocorrelation matrix RxxThe formula used is:
wherein K is Lm-L +1 is the number of snapshots;
step 2-1-3, obtaining cross correlation matrix rxdThe formula used is:
in the formula,is the filter center point;
step 2-1-4, solving the weight vector w1The formula used is:
w1＝Rxx -1rxd；
step 2-1-5, for the weight vector w obtained in step 2-1-41Detecting the peak value, recording the abscissa of the peak value as z, and counting the number d of the time delay points of the channel 4 signal and the channel 1 signal1＝z-D；
Step 2-1-6, repeating the step 2-1-1 to the step 2-1-5, and obtaining the number of time delay points of the channel 4 signal, the channel 2 signal and the channel 3 signal, wherein the number of the time delay points is d2、d3；
Step 2-2, adopting a generalized sidelobe canceller to perform self-adaptive enhancement on the potential sound fragment, specifically:
step 2-2-1, obtaining a main channel signal d (k):
d(k)＝wc Tm(k)；
in the formula, wc＝[wc4,wc1,wc2,wc3]TIs a static weight vector, wc1、...、wc4Is the weight value w corresponding to each channelc1+wc2+wc3+wc4＝1，m(k)＝[m4(k),m1(k-d1),m2(k-d2),m3(k-d3)]T，k＝1,2,...,Lm；
Step 2-2-2, finding auxiliary channel signal e (k):
wherein, WSA blocking matrix of dimension 4 x 3;
step 2-2-3, obtaining an enhanced pure potential singing segment signal y (k):
y(k)＝d(k)-vT(k)e(k)；
where v (k) represents the weight vector of the adaptive interference canceller.
6. The automatic acoustic detection method for specific bird species according to claim 4 or 5, wherein step 3 is to extract feature parameters from the noise-reduced and enhanced potential singing segments in step 2 to construct a feature set of the potential singing segments, and specifically comprises the following steps:
step 3-1, calculating a power spectrogram of each potential singing segment after noise reduction and enhancement according to the step 1-2;
step 3-2, setting a Mel band-pass filter group in the frequency distribution range of the singing fragment of the specific bird species, and then enabling the power spectrum of the potential singing fragment to pass through the filter group to obtain the output of each filter;
3-3, taking logarithm of the output result, and performing discrete cosine transform to obtain a Mel cepstrum coefficient characteristic parameter;
and 3-4, combining the Mel cepstrum coefficient characteristic parameters corresponding to all potential singing segments to construct and obtain a feature set of the potential singing segments.
7. The acoustic automatic detection method for specific bird species according to claim 6, characterized in that the identification algorithm of step 4 specifically adopts a support vector machine identification algorithm.
8. The method for acoustic automatic detection of specific bird species according to claim 6, wherein the step 4 combines the feature set of potential song fragments obtained in step 3 and the recognition algorithm in machine learning to complete the acoustic detection of specific bird species, specifically:
the method comprises the steps of taking Mel cepstrum coefficient features in an existing specific bird species singing feature database as training samples, taking potential singing segment feature sets as input samples of a support vector machine, and automatically detecting specific bird species through decision of the support vector machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811566250.6A CN109741759B (en) | 2018-12-21 | 2018-12-21 | Acoustic automatic detection method for specific bird species |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811566250.6A CN109741759B (en) | 2018-12-21 | 2018-12-21 | Acoustic automatic detection method for specific bird species |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109741759A true CN109741759A (en) | 2019-05-10 |
| CN109741759B CN109741759B (en) | 2020-07-31 |
Family
ID=66360805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811566250.6A Active CN109741759B (en) | 2018-12-21 | 2018-12-21 | Acoustic automatic detection method for specific bird species |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109741759B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110335613A (en) * | 2019-05-28 | 2019-10-15 | 广东工业大学 | A kind of birds recognition methods using sound pick-up real-time detection |
| CN111276151A (en) * | 2020-01-20 | 2020-06-12 | 北京正和恒基滨水生态环境治理股份有限公司 | Bird sound identification system and identification method |
| CN113314127A (en) * | 2021-04-23 | 2021-08-27 | 广州大学 | Space orientation-based bird song recognition method, system, computer device and medium |
| CN117724042A (en) * | 2024-02-18 | 2024-03-19 | 百鸟数据科技(北京)有限责任公司 | Method and system for positioning bird song sound source based on acoustic bispectrum |
| CN118173102A (en) * | 2024-05-15 | 2024-06-11 | 百鸟数据科技(北京)有限责任公司 | Bird voiceprint recognition method in complex scene |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101447190A (en) * | 2008-06-25 | 2009-06-03 | 北京大学深圳研究生院 | Voice enhancement method employing combination of nesting-subarray-based post filtering and spectrum-subtraction |
| CN102708860A (en) * | 2012-06-27 | 2012-10-03 | 昆明信诺莱伯科技有限公司 | Method for establishing judgment standard for identifying bird type based on sound signal |
| CN102800322A (en) * | 2011-05-27 | 2012-11-28 | 中国科学院声学研究所 | Method for estimating noise power spectrum and voice activity |
| US20140226838A1 (en) * | 2013-02-13 | 2014-08-14 | Analog Devices, Inc. | Signal source separation |
| CN106504762A (en) * | 2016-11-04 | 2017-03-15 | 中南民族大学 | Bird community quantity survey system and method |
| CN107393549A (en) * | 2017-07-21 | 2017-11-24 | 北京华捷艾米科技有限公司 | Delay time estimation method and device |
| CN107393542A (en) * | 2017-06-28 | 2017-11-24 | 北京林业大学 | A kind of birds species identification method based on binary channels neutral net |
| CN107644650A (en) * | 2017-09-29 | 2018-01-30 | 山东大学 | A kind of improvement sound localization method based on progressive serial orthogonalization blind source separation algorithm and its realize system |
| CN108694953A (en) * | 2017-04-07 | 2018-10-23 | 南京理工大学 | A kind of chirping of birds automatic identifying method based on Mel sub-band parameter features |
-
2018
- 2018-12-21 CN CN201811566250.6A patent/CN109741759B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101447190A (en) * | 2008-06-25 | 2009-06-03 | 北京大学深圳研究生院 | Voice enhancement method employing combination of nesting-subarray-based post filtering and spectrum-subtraction |
| CN102800322A (en) * | 2011-05-27 | 2012-11-28 | 中国科学院声学研究所 | Method for estimating noise power spectrum and voice activity |
| CN102708860A (en) * | 2012-06-27 | 2012-10-03 | 昆明信诺莱伯科技有限公司 | Method for establishing judgment standard for identifying bird type based on sound signal |
| US20140226838A1 (en) * | 2013-02-13 | 2014-08-14 | Analog Devices, Inc. | Signal source separation |
| CN106504762A (en) * | 2016-11-04 | 2017-03-15 | 中南民族大学 | Bird community quantity survey system and method |
| CN108694953A (en) * | 2017-04-07 | 2018-10-23 | 南京理工大学 | A kind of chirping of birds automatic identifying method based on Mel sub-band parameter features |
| CN107393542A (en) * | 2017-06-28 | 2017-11-24 | 北京林业大学 | A kind of birds species identification method based on binary channels neutral net |
| CN107393549A (en) * | 2017-07-21 | 2017-11-24 | 北京华捷艾米科技有限公司 | Delay time estimation method and device |
| CN107644650A (en) * | 2017-09-29 | 2018-01-30 | 山东大学 | A kind of improvement sound localization method based on progressive serial orthogonalization blind source separation algorithm and its realize system |
Non-Patent Citations (2)
| Title |
|---|
| ZHAO ZHAO ETC: "Automated bird acoustic event detection and robust species classification", 《ECOLOGICAL INFORMATICS》 * |
| 吴烨: "面向鸟声传感网数据中心的边缘虚拟网桥技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110335613A (en) * | 2019-05-28 | 2019-10-15 | 广东工业大学 | A kind of birds recognition methods using sound pick-up real-time detection |
| CN110335613B (en) * | 2019-05-28 | 2021-07-09 | 广东工业大学 | Bird identification method adopting pickup for real-time detection |
| CN111276151A (en) * | 2020-01-20 | 2020-06-12 | 北京正和恒基滨水生态环境治理股份有限公司 | Bird sound identification system and identification method |
| CN111276151B (en) * | 2020-01-20 | 2023-04-07 | 北京正和恒基滨水生态环境治理股份有限公司 | Bird sound identification system and identification method |
| CN113314127A (en) * | 2021-04-23 | 2021-08-27 | 广州大学 | Space orientation-based bird song recognition method, system, computer device and medium |
| CN113314127B (en) * | 2021-04-23 | 2023-10-10 | 广州大学 | Bird song identification method, system, computer equipment and medium based on space orientation |
| CN117724042A (en) * | 2024-02-18 | 2024-03-19 | 百鸟数据科技(北京)有限责任公司 | Method and system for positioning bird song sound source based on acoustic bispectrum |
| CN117724042B (en) * | 2024-02-18 | 2024-04-19 | 百鸟数据科技(北京)有限责任公司 | Method and system for positioning bird song sound source based on acoustic bispectrum |
| CN118173102A (en) * | 2024-05-15 | 2024-06-11 | 百鸟数据科技(北京)有限责任公司 | Bird voiceprint recognition method in complex scene |
| CN118173102B (en) * | 2024-05-15 | 2024-07-16 | 百鸟数据科技(北京)有限责任公司 | Bird voiceprint recognition method in complex scene |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109741759B (en) | 2020-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109741759B (en) | Acoustic automatic detection method for specific bird species | |
| CN103811020B (en) | A kind of intelligent sound processing method | |
| Potamitis et al. | On automatic bioacoustic detection of pests: the cases of Rhynchophorus ferrugineus and Sitophilus oryzae | |
| CN109658948B (en) | Migratory bird migration activity-oriented acoustic monitoring method | |
| CN109949823A (en) | A kind of interior abnormal sound recognition methods based on DWPT-MFCC and GMM | |
| CN111261189B (en) | Vehicle sound signal feature extraction method | |
| CN103854660B (en) | A kind of four Mike's sound enhancement methods based on independent component analysis | |
| CN102097095A (en) | Speech endpoint detecting method and device | |
| CN107731235B (en) | Method and device for extracting and classifying sound pulse characteristics of sperm whales and long fin pilot whales | |
| CN109884591B (en) | Microphone array-based multi-rotor unmanned aerial vehicle acoustic signal enhancement method | |
| Himawan et al. | Deep Learning Techniques for Koala Activity Detection. | |
| CN113314127B (en) | Bird song identification method, system, computer equipment and medium based on space orientation | |
| Roch et al. | Compensating for the effects of site and equipment variation on delphinid species identification from their echolocation clicks | |
| CN107886050A (en) | Utilize time-frequency characteristics and the Underwater targets recognition of random forest | |
| CN112394324A (en) | Microphone array-based remote sound source positioning method and system | |
| Boulmaiz et al. | The use of WSN (wireless sensor network) in the surveillance of endangered bird species | |
| Castro et al. | Automatic manatee count using passive acoustics | |
| Wang et al. | Multi-speaker Speech Separation under Reverberation Conditions Using Conv-Tasnet | |
| CN101125100A (en) | Method for measuring and identifying bat kind using with bat echo location acoustic | |
| Yin et al. | Simultaneous multiple bands time delay and frequency component estimation based on expectation–maximization | |
| Shkury et al. | Robust automatic detector and feature extractor for dolphin whistles | |
| CN112086105A (en) | Target identification method based on Gamma atom sub-band continuous spectrum characteristics | |
| Bougher et al. | Generalized marine mammal detection based on improved band-limited processing | |
| Lu et al. | Joint classification of whistles and echolocation clicks from odontocetes | |
| Sánchez-García et al. | An efficient statistics-based method for the automated detection of sperm whale clicks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |