CN104794894A - Automobile whistle noise monitoring device, system and method - Google Patents

Abstract

The invention discloses an automobile whistle noise monitoring device, system and method. The automobile whistle noise monitoring system comprises a whistle noise identification module, a noise source positioning module and a camera shooting module; the whistle noise identification module filters sound waves with the frequency being smaller than or larger than that of the whistle noise through a band-pass filter which is arranged at the front end of a sound pickup according to frequency characteristics of the automobile whistle noise and then detects to determine whistle noise signals; the noise source positioning module positions based on multiple sound sources of ESPRIT algorithm; a specific matrix pencil is constructed and performed on generalized characteristic decomposition to obtain a translation operator which is corresponding to the direction of the sound source according to the translation invariance of signal subspace which is brought by translation invariance of a category of array through the ESPRIT algorithm and accordingly directional parameter estimation is obtained; the camera shooting shoots confirmed noise source searching points, marks violations and stores or transmits photos of marked violation vehicles to a related traffic management department.

Description

A kind of vehicle whistle noise monitoring arrangement, system and method

Technical field

The invention belongs to Audio Signal Processing, pattern-recognition and array signal process technique field, relate to a kind of vehicle whistle monitoring arrangement, system and method.

Background technology

Judge that the coordinate position of moving sound is the core objective of sound localization technology quickly and accurately.At present, with microphone array auditory localization technology for mainstream technology, it is that a kind of microphone pickup array that utilizes picks up voice signal and utilizes the sound localization method of Digital Signal Processing, has good spatial choice characteristic.Following three kinds of methods can be had to carry out auditory localization based on this principle:

1. based on the directional technology of High-Resolution Spectral Estimation.It is the method by solving the correlation matrix determination Sounnd source direction angle between microphone signal, solves sound source position further.The method comes from High-Resolution Spectral Estimation technology, as minimum variance Power estimation method and Eigenvalues Decomposition method etc.It requires that sound-source signal has stationarity, and, in order to reduce the impact of external interference factor and meet the specific condition of this technology application, the operand of system need be improved exponentially, thus the hardware device of system is had higher requirement.

2. the steerable beam based on peak power output forms technology.This technology is processed by the voice signal received microphone array, directly controls the direction that microphone points to sound-source signal peak power wave beam.

3. based on the auditory localization technology of step-out time (TDOA).The method is a kind of wireless location technology, and it arrives the method for the mistiming of each acoustic pickup by the sound-source signal that measurement moving sound sends, and realizes positioning function.Accurately and neatly can estimate that delay length is the key factor affecting auditory localization.

Summary of the invention

Blow a whistle based on urban population compact district motor vehicle and remain incessant after repeated prohibition and the phenomenon such as direct surveillance's efficiency is low.Object of the present invention, for above-mentioned phenomenon, proposes a kind of vehicle whistle noise monitoring system and method thereof.Detect the horning against traffic regulation vehicle on no tooting road, and seek a shooting with highway video camera.

The invention provides a kind of vehicle whistle noise monitoring system, it comprises:

Noise identification of blowing a whistle module, it, for identifying the sound wave of pickup, finally determines vehicle whistle noise signal;

Noise source locating module, noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out of blowing a whistle described in its basis;

Photographing module, it is for taking determined vehicle whistle noise source.

Present invention also offers a kind of vehicle whistle noise supervision method, it comprises:

The sound wave of pickup is identified, finally determines vehicle whistle noise signal;

According to described noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out of blowing a whistle;

Determined vehicle whistle noise source is taken.

Present invention also offers a kind of vehicle whistle noise monitoring system, it comprises:

Noise identification of blowing a whistle module, it comprises multiple pickup unit, is respectively used to the sound wave picking up surrounding environment, and identifies the sound wave of pickup, finally determine vehicle whistle noise signal;

Noise source locating module, it comprises the first microphone array and second microphone array, described microphone array comprises and is distributed in the first microphone array in X-axis and Y-axis and second microphone array respectively, and the first microphone array and second microphone array comprise interval uniform equivalent microphone array element; Described first microphone array and second microphone array according to described in blow a whistle noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out;

Photographing module, it comprises multiple camera, is respectively used to take determined vehicle whistle noise source.

A taking module of seeking of highway video camera carries out video camera according to the three-dimensional point that auditory localization module is determined and to seek some shooting.Then identify the picture of shooting, the violation vehicle pictures of mark is preserved or real-time sends vehicle supervision department to.

Accompanying drawing explanation

Fig. 1 is vehicle whistle noise monitoring system theory diagram in the present invention;

Fig. 2 is the system diagram of vehicle whistle sound recognition system in the present invention;

Array battle array source distribution structural drawing in Fig. 3 the present invention;

System and device installation diagram in Fig. 4 the present invention.

Embodiment

For making the object, technical solutions and advantages of the present invention clearly understand, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in further detail.

Below in conjunction with accompanying drawing, the present invention is described further.

Fig. 1 is the structured flowchart of the vehicle whistle noise monitoring system that the present invention proposes.As shown in Figure 1, this system comprises three modules, is respectively noise identification module of blowing a whistle, noise source locating module, photographing module.The vehicle whistle noise monitoring system that the present invention proposes utilizes microphone array pick-up to pick up sound-source signal, is then processed by the sound-source signal of central processing unit to pickup.

Described identification module comprises multiple pickup unit, for carrying out bandpass filtering to the sound wave of pickup, filters out the highway noise that frequency is less than and is greater than vehicle whistle sound, only retains the sound signal near sound of blowing a whistle.Then, described identification module carries out short-time energy to the sound signal near retained sound of blowing a whistle and crosses limit detection, detects doubtful vehicle whistle acoustical signal.Finally carry out frequency right, determine acoustical signal of blowing a whistle.

Alternatively, described multiple pickup unit carries out bandpass filtering according to the sound wave of upper-lower door limit value to pickup, and described upper-lower door limit value can be determined according to the acoustic frequency of blowing a whistle of the general motor vehicles of statistics.Alternatively, the motor vehicles experimentally recorded are blown a whistle acoustic frequency characteristic, adding a cutoff frequency in microphone sensor front end is the bandpass filter of 200Hz and 5KHz, can filtering highway noise to greatest extent, retains the vehicle whistle acoustical signal composition of intermediate-frequency section simultaneously.

Wherein, analog to digital conversion is carried out to the bandpass signal after process, determines whether further as noise of blowing a whistle.

Identification module of the present invention also carries out short-time energy to the sound signal near the sound of blowing a whistle obtained after carrying out bandpass filtering and crosses limit and detect.This is because, due to the time-varying characteristics of signal, utilize window function signal to be divided into the frame of certain length, and think that the feature of signal is substantially constant within these short time.Comparatively level and smooth frequency spectrum is obtained in the present invention with the Hamming window that main lobe is wider.

The present invention adopts the end-point detection method based on time domain short-time energy, and the object of end-point detection is the position of the voice signal detecting doubtful vehicle whistle sound.

The short-time energy of one frame voice signal x (n) is defined as:

$E_n=\underset{m=-\∞}{\overset{\∞}{\mathrm{\Σ}}}{\left|x\left(m\right)w(n-m)\right|}^2=\underset{m=n}{\overset{n+N-1}{\mathrm{\Σ}}}{\left|x\left(m\right)w(n-m)\right|}^2$

Wherein, w (n) is window function, and N is that window is long, is also the length of one section of acoustic processing frame, the frame signal that x (n) and x (m) is n, the m moment.Voice signal after first step bandpass filtering, eliminates most highway noise, and can set detection threshold is herein E _l, work as E _n> E _ltime judge have doubtful vehicle whistle acoustical signal to occur in current sound framing.

Wherein, E _l=min [0.03 (E _max-E _min)+E _r, 4E _r]

Wherein, E _maxfor Energy maximum value, E _minfor energy-minimum, E _rfor the mean value of all framing energy.

After detecting doubtful vehicle whistle acoustical signal frame, described identification module also carries out frequency comparison, to determine acoustical signal of blowing a whistle from doubtful vehicle whistle acoustical signal frame.Particularly, the doubtful vehicle whistle acoustical signal that short-time energy method detects by identification module transforms to frequency field, utilizes the statistical property of blow a whistle sound and noise spectrum to identify sound of blowing a whistle.Namely utilize the noise frequency characteristic of blowing a whistle that statistics is determined in advance to mate it, if coupling, think that doubtful vehicle whistle acoustical signal is the acoustical signal of blowing a whistle determined.

In the present invention, described many noise sources locating module, it is according to the determined voice signal localization of sound source of blowing a whistle of identification module.Alternatively, the translation invariance of the signal subspace that described many noise sources locating module utilizes the translation invariance of a class array to bring based on ESPRIT algorithm, try to achieve to come to corresponding translation operator with sound source by constructing special pencil of matrix and carrying out generalized eigen decomposition to it, thus obtain direction parameter estimation.In this step, the three-dimensional coordinate of signal source can be calculated by the signal source direction determined each pickup unit.The implementation procedure of this module will be specifically introduced below according to details.

In the present invention, described many noise sources locating module comprises the first microphone array and second microphone array that are installed on sound pick up equipment inside.Fig. 3 shows the array distribution schematic diagram of the first microphone array and second microphone array in the present invention.As shown in Figure 3, first microphone array and second microphone array distribution are in x-axis and y-axis, be dispersed with the even linear array X that array number is N respectively, Y, the array element at initial point place is two linear array X, Y has, and the first microphone array is made up of identical array element with second microphone array, and the first microphone array and second microphone array are about plane symmetry, and the center of two arrays sets according to actual needs at a distance of L, L.

Alternatively, in the present invention when carrying out three-dimensional localization to the acoustical signal of blowing a whistle determined, the position angle of noise source relative to array center's point can be obtained based on ESPRIT algorithm, suppose that the center of the first microphone array is three-dimensional coordinate initial point (0,0,0), the array element distance of linear array is d, and the output noise of array is zero-mean, variance is σ ²the white Gaussian noise of statistical iteration, and uncorrelated with noise source.

If there is the noise source of M statistical iteration in space, described noise source is respectively f relative to the frequency of reference array element and incident angle _i, θ _i, the signal phasor received for linear array X and linear array Y is respectively

$\left\{\begin{array}{c}X\left(t\right)=A_{x}S\left(t\right)+N_x\left(t\right)\\ Y\left(t\right)=A_{y}S\left(t\right)+N_y\left(t\right)\end{array}\right.$

Wherein: S (t)=[s _l(t) ..., s _k(t) ... s _m(t)] ^tfor noise vector of blowing a whistle;

represent the direction matrix of linear array X;

represent the direction matrix of linear array Y;

A _x(f _i, θ _i, φ _i) receiver function of expression to i noise source, the direction matrix of different reception array element is different;

N _x(t) and N _yt () represents the noise signal of the non-noise of blowing a whistle that linear array receives respectively.

For the Combined estimator of the two dimensional arrival angles and frequency of asking noise signal of blowing a whistle, the present invention is based on a kind of Combined estimator method of ESPRIT algorithm, but the present invention does not limit and only has the method to can be applicable to native system more.In this microphone array, front N-1 the array element of submatrix X and the Received signal strength vector of a rear N-1 array element are designated as X respectively ₁(t), X ₂t (), front N-1 the array element of submatrix Y and the Received signal strength vector of a rear N-1 array element are designated as Y respectively ₁(t), Y ₂t (), then have

Wherein:

A _xfor the direction matrix of X array, A _yfor the direction matrix of Y array.

For the frequency of estimated signal, the Received signal strength of array adds and postpones τ.

Only N-1 array element before in X submatrix need be carried out a joint delay, the data obtained after delay are designated as X ₃, according to space two-dimensional Power estimation:

wherein ${\mathrm{\Φ}}_{\mathrm{\τ}}=\mathrm{diag}\{e^{-j2{\mathrm{\πf}}_1\mathrm{\τ}},...,e^{-j2{\mathrm{\πf}}_{N-1}\mathrm{\τ}}\}$

From the principle of invariable rotary subspace, if spatial noise is white noise, X can be constructed ₁the auto-covariance matrix of (t) x ₁(t) and X ₂the Cross-covariance of (t) y ₁the auto-covariance matrix of (t) y ₁(t) and Y ₂the Cross-covariance of (t) x ₁(t) and X ₃the Cross-covariance of (t) hypothesis space noise is white Gaussian noise, to above-mentioned 5 covariance matrix denoisings, then has

Try to achieve respectively the generalized eigenvalue that 3 matrixes are right:

vf _i＝exp(-j2πf _iτ)，

Wherein, i=1,2 ..., M, M are noise source number of blowing a whistle.

More than simultaneous three formulas, can solve the frequency of noise source, position angle and depression angle:

$f_i=\left|\frac{\mathrm{angle}\left({\mathrm{vf}}_i\right)}{2\mathrm{\π\τ}}\right|$

vf _i＝exp(-j2πf _iτ)

Wherein, θ _iand φ _ibe respectively position angle and the depression angle of i-th noise source; f _ibe the frequency of i-th noise source, τ is for receiving linear array X ₃relative to reception linear array X ₁receiving signal delayed; D is the spacing in described first microphone array and second microphone array between array element, and c is the light velocity; Vf _i, vX _i, vY _ibe respectively matrix pair generalized eigenvalue.In solution procedure to above-mentioned equation, by constructing special pencil of matrix and carrying out generalized eigen decomposition to it, suppose vf _i, vX _i, vY _ibe respectively matrix generalized eigenvalue, then matrix following expression:

$\left\{\begin{array}{c}{C}_{X_{11}}=R_{X_{11}}-{\mathrm{\σ}}^{2}I\\ C_{X_{12}}=R_{X_{12}}-{\mathrm{\σ}}^{2}I\\ C_{Y_{11}}=R_{Y_{11}}-{\mathrm{\σ}}^{2}I\\ C_{Y_{12}}=R_{Y_{12}}-{\mathrm{\σ}}^{2}I\\ C_{X_{13}}=R_{X_{13}}-{\mathrm{\σ}}^{2}I\end{array}\right.$

Wherein, for X ₁the auto-covariance matrix of (t), for X ₁(t) and X ₂the Cross-covariance of (t), for Y ₁the auto-covariance matrix of (t), for Y ₁(t) and Y ₂the Cross-covariance of (t), for X ₁(t) and X ₃t the Cross-covariance of (), I is unit matrix, σ ²for the variance of white Gaussian noise; X ₁(t) and X ₂t () is respectively the front N-1 of a submatrix X array element of the first microphone array and the Received signal strength vector of a rear N-1 array element; Y ₁(t) and Y ₂t () is respectively the front N-1 of a submatrix Y array element of the first microphone array and the Received signal strength vector of a rear N-1 array element; X ₃t () is that front N-1 array element in the submatrix X of the first microphone array is to the signal phasor after the signal delay τ received.

According to the Received signal strength vector of the first microphone array array element corresponding to second microphone array, the volume coordinate of i-th noise source can be calculated

$\left\{\begin{array}{c}{x}_i=\frac{L\tan {\mathrm{\θ}}_{i2}}{\tan {\mathrm{\θ}}_{i1}+\tan {\mathrm{\θ}}_{i2}}\\ y_i=x_i\tan {\mathrm{\θ}}_{i1}\end{array}\right.$

z _i＝(z _i1+z _i2)/2

Wherein, x _i, y _iand z _ibe respectively the volume coordinate of i-th noise source; θ _i1with be respectively i-th noise source relative to the deflection of the first microphone array central point and the angle of pitch; θ _i2with be respectively i-th noise source relative to the deflection of second microphone array center's point and the angle of pitch.

Photographing module described in the present invention comprises multiple camera, between its first microphone array being evenly distributed on described many noise sources locating module and second microphone array, as shown in Figure 4; When many noise sources locating module obtain blowing a whistle acoustical signal volume coordinate after, described photographing module regulation and control camera rotates focusing, and takes noise source.Specifically, all cameras can be selected all to aim at noise source and to take, also can chosen distance nearest at least one camera be taken by noise source.

Wherein, violation is carried out to the photo of video camera shooting and mark process.

Wherein, the photo of mark is stored or sends relevant traffic department in real time to.

Wherein, regulate and control camera and a shooting is sought to closest approach noise source.

Wherein, the vehicle whistle frequency that can add up according to this section of the number of camera and determining.

The automobile of blowing a whistle in violation of rules and regulations is taken, and by photo storage in video camera internal memory, or send photo to relevant traffic administrative authority in real time.

Wherein, the noise source photo of shooting is blown a whistle in violation of rules and regulations mark.

Wherein, regular process is carried out to the internal memory in video camera, in case internal memory overfill, can not take.

Above-described specific embodiment; object of the present invention, technical scheme and beneficial effect are further described; be understood that; the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. a vehicle whistle noise monitoring system, it comprises:

Noise identification of blowing a whistle module, it, for identifying the sound wave of pickup, finally determines vehicle whistle noise signal;

Noise source locating module, noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out of blowing a whistle described in its basis;

Photographing module, it is for taking determined vehicle whistle noise source.

2. the system as claimed in claim 1, wherein, described in noise identification module of blowing a whistle utilize multiple sound pick up equipment to carry out the pickup of sound wave; Described noise source locating module utilizes the microphone array being installed on sound pick up equipment inside to position noise source; Described photographing module utilizes multiple camera to take described vehicle whistle noise source.

3. the system as claimed in claim 1, wherein, described in blow a whistle noise identification module as follows to pickup sound wave identify:

Bandpass filtering is carried out to picked up sound wave, filters out frequency other noise signals not outside vehicle whistle noise signal range;

Carry out short-time energy to the sound wave after carrying out bandpass filtering to cross limit and detect, detect doubtful vehicle whistle noise signal;

The frequency of the frequency of doubtful vehicle whistle noise signal with the vehicle whistle noise signal preset is mated, to determine vehicle whistle noise signal.

4. system as claimed in claim 3, wherein, describedly short-time energy is carried out to the sound wave after carrying out bandpass filtering cross limit and detect and specifically comprise:

Utilize window function that the sound wave after described bandpass filtering is divided into multiple frame signal, wherein the feature invariant of signal in each frame signal;

Calculate the short-time energy of each frame signal, and itself and energy measuring threshold value are compared;

When the short-time energy of described frame signal is greater than described energy measuring threshold value, be defined as doubtful vehicle whistle noise signal.

5. the system as claimed in claim 1, wherein, noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out of blowing a whistle described in the following basis of described noise source locating module:

Obtain the signal phasor of each vehicle whistle noise source that microphone array receives; Wherein, described microphone array comprises and is distributed in the first microphone array in X-axis and Y-axis and second microphone array respectively, and the first microphone array and second microphone array comprise interval uniform equivalent microphone array element;

According to the signal of obtained each vehicle whistle noise source, estimate to obtain the frequency of each vehicle whistle noise source, position angle and depression angle based on ESPRIT algorithm;

The volume coordinate of described each vehicle whistle noise source is calculated according to the frequency of described each vehicle whistle noise source, position angle and depression angle.

6. system as claimed in claim 5, wherein, the frequency of described each vehicle whistle noise source, position angle and depression angle calculate as follows:

${\mathrm{\θ}}_i=\arctan \left\{\frac{\mathrm{angle}\left({\mathrm{vY}}_i\right)}{\mathrm{angle}\left({\mathrm{vX}}_i\right)}\right\},$

$f_i=\left|\frac{\mathrm{angle}\left({\mathrm{vf}}_i\right)}{2\mathrm{\π\τ}}\right|$

vf _i＝exp(-j2πf _iτ)，

Wherein, θ _iand φ _ibe respectively position angle and the depression angle of i-th noise source; f _ibe the frequency of i-th noise source, τ is time delay; D is the spacing in described first microphone array and second microphone array between array element, and c is the light velocity; Vf _i, vX _iand vY _icalculate intermediate value respectively.

7. system as claimed in claim 5, wherein, the volume coordinate of described each vehicle whistle noise source calculates as follows:

$\left\{\begin{array}{c}{x}_i=\frac{L\tan {\mathrm{\θ}}_{i2}}{\tan {\mathrm{\θ}}_{i1}+\tan {\mathrm{\θ}}_{i2}}\\ y_i=x_i\tan {\mathrm{\θ}}_{i1}\end{array}\right.$

z _i＝(z _i1+z _i2)/2

Wherein, x _i, y _iand z _ibe respectively the volume coordinate of i-th noise source; θ _i1with be respectively i-th noise source relative to the deflection of the first microphone array central point and the angle of pitch; θ _i2with be respectively i-th noise source relative to the deflection of second microphone array center's point and the angle of pitch; L is the center spacing of the first microphone array and second microphone array.

8. the system as claimed in claim 1, wherein, the short-time energy of described each frame signal calculates as follows:

$E_n=\underset{m=-\∞}{\overset{\∞}{\mathrm{\Σ}}}{\left|x\left(m\right)w(n-m)\right|}^2=\underset{m=n}{\overset{n+N+1}{\mathrm{\Σ}}}{\left|x\left(m\right)w(n-m)\right|}^2$

Wherein, E _nbe the short-time energy of the n-th frame signal, w (n) is window function, and N is that window is long, and x (n) is the n-th frame signal;

Described energy monitoring threshold value calculates as follows:

E _l＝min[0.03(E _max-E _min)+E _r，4E _r]

Wherein, E _maxfor the Energy maximum value in all frame signals, E _minfor the energy-minimum in all frame signals, E _rfor the mean value of all frame signal energy.

9. a vehicle whistle noise supervision method, it comprises:

The sound wave of pickup is identified, finally determines vehicle whistle noise signal;

According to described noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out of blowing a whistle;

Determined vehicle whistle noise source is taken.

10. a vehicle whistle noise monitoring system, it comprises:

Noise identification of blowing a whistle module, it comprises multiple pickup unit, is respectively used to the sound wave picking up surrounding environment, and identifies the sound wave of pickup, finally determine vehicle whistle noise signal;

Noise source locating module, it comprises the first microphone array and second microphone array, described microphone array comprises and is distributed in the first microphone array in X-axis and Y-axis and second microphone array respectively, and the first microphone array and second microphone array comprise interval uniform equivalent microphone array element; Described first microphone array and second microphone array according to described in blow a whistle noise identification Module recognition vehicle whistle noise signal determination vehicle whistle noise source out;

Photographing module, it comprises multiple camera, is respectively used to take determined vehicle whistle noise source.