CN117009746A - Noise separation method based on probability deconvolution method - Google Patents
Noise separation method based on probability deconvolution method Download PDFInfo
- Publication number
- CN117009746A CN117009746A CN202310920894.5A CN202310920894A CN117009746A CN 117009746 A CN117009746 A CN 117009746A CN 202310920894 A CN202310920894 A CN 202310920894A CN 117009746 A CN117009746 A CN 117009746A
- Authority
- CN
- China
- Prior art keywords
- signal
- noise
- tot
- discrete
- probability
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000000926 separation method Methods 0.000 title claims abstract description 19
- 238000005259 measurement Methods 0.000 claims abstract description 24
- 238000005315 distribution function Methods 0.000 claims abstract description 18
- 238000012935 Averaging Methods 0.000 claims description 4
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000010183 spectrum analysis Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Algebra (AREA)
- Operations Research (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to a noise separation method based on a probability deconvolution method, which comprises the following steps of step 1, recording a discrete signal n of a noise floor signal 1 [x]And a discrete signal n of the measurement signal tot [x]The method comprises the steps of carrying out a first treatment on the surface of the Step 2, the signal array n 1 [x]And n tot [x]Classifying each element according to the size of the element, and recording the number of discrete signal array elements in each value interval; step 3, dividing the number of elements in the value interval by the total number of elements in the discrete signal array, and dividing by the length of the value interval to obtain a distribution function R of the discrete array 1 (n) and R tot (n). The invention carries out deconvolution processing on the known probability density functions of the noise floor, the target noise superposition signal and the noise floor signal, and can obtain the probability density function corresponding to the target noiseA number.
Description
Technical field:
the invention belongs to the technical field of metering test, and particularly relates to a noise separation method based on a probability deconvolution method.
The background technology is as follows:
noise floor is an important technical parameter for determining the minimum value of physical quantity measurement in various disciplines, and is generally composed of thermal noise of electronic elements in instruments and equipment and noise generated by the influence of physical fields such as pressure, temperature, electromagnetic waves, vibration and the like in the environment.
The current noise measurement method mostly adopts a spectrum analysis method: and the modulus distribution of the noise at each frequency is obtained by averaging the noise spectrum signals acquired by the acquisition system for multiple times under the steady state to define the noise level of the acquisition system after the noise floor and the environmental noise are overlapped. Because of randomness and irrelevance of noise in time domain and frequency domain, the negative influence of noise floor on the accuracy of measurement results can not be thoroughly eliminated, and the influence of noise floor can be reduced only by constructing a low noise floor test domain, increasing the number of sensors, improving the signal-to-noise ratio of measured data in the order of magnitude of measured and other modes.
However, in the field of underwater noise measurement, the noise level of the measuring instrument is quite equal to the target noise signal level, so that the background noise cannot be ignored in the measurement result, and the measurement accuracy is greatly affected. In particular, in the research process of the method for measuring the radiation noise of the underwater unmanned aircraft, background noise of a test water area is measured by using hydrophones of different types, and the consistency of measurement results is poor. It is therefore desirable to use noise separation techniques to minimize the effect of noise floor on the measurement results.
The invention comprises the following steps:
the invention aims to solve the technical problem of providing a noise separation method based on a probability deconvolution method, which is known by analyzing probability density characteristics of random noise superposition, and if one random noise signal is formed by time domain superposition of a plurality of random noises, a probability density function of the noise signal is convolution of probability density functions of all source functions. By utilizing the characteristic, the probability density functions of the known noise floor, the target noise superposition signal and the noise floor signal are subjected to deconvolution processing, so that the probability density function corresponding to the target noise is obtained.
The technical proposal of the invention is to provide a noise separation method based on a probability deconvolution method, which comprises the following steps,
step 1, recording a discrete signal n of a noise floor signal 1 [x]And a discrete signal n of the measurement signal tot [x]；
Step 2, the signal array n 1 [x]And n tot [x]Classifying each element according to the size of the element, and recording the number of discrete signal array elements in each value interval;
step 3, dividing the number of elements in the value interval by the total number of elements in the discrete signal array, and dividing by the length of the value interval to obtain a distribution function R of the discrete array 1 (n) and R tot (n)；
Step 4, measuring for multiple times, adding the distribution functions R (n) obtained by each measurement, and averaging to obtain the final probability distribution function P of the noise floor and the measurement signal 1 (n) and P tot (n)；
Step 5, probability distribution function P 1 (n) and P tot (n) substitution into formula
Obtaining probability distribution function P of target noise signal 2 (n) completing noise separation.
Probability density convolution characteristics of random noise superposition
In noise measurement based on spectral analysis, superposition between noise lacks a quantization means due to uncertainty and uncorrelation of the corresponding spectrum of the measured signals at different times. In a noise evaluation system based on a distribution function, this problem can be solved.
Set white noise N 1 And white noise N 2 The probability distribution functions of (a) are the same and are P respectively 1 (x) And P 2 (x) Then for white noise N 1 Determining a value x 1 Superimposed white noise N 2 X of the determined value 2 X obtained after 1 +x 2 Probability of occurrence P tot (x 1+ x 2 ) The method comprises the following steps:
P tot (x 1 +x 2 )＝P 1 (x 1 )P 2 (x 2 ) (1)
let x be 0 ＝x 1+ x 2 The following steps are:
P tot (x 0 )＝P 1 (x 1 )P 2 (x 0 -x 1 ) (2)
then for any x 1 ，P tot (x 0 ) The value of (2) should be all x 1 The corresponding sum of probabilities:
obviously, P tot (x) Is a function P 1 (x) And P 2 (x) Is a convolution of (a) and (b). It can be concluded that: if a random noise signal is a time domain superposition of random noise, then the probability density function of this noise signal is a convolution of all the source function probability density functions.
Noise separation algorithm based on probability deconvolution
As can be seen from the formula (3), the probability density function of the random noise signal can be separated by convolution operation which is linearly reversible in the frequency domain, so that the probability density function P of the measurement signal can be first calculated tot (x) And noise floor signal P of instrument 1 (x) Fourier transform is performed to obtain fft (P) tot (x) And fft (P) 1 (x) Dividing the Fourier transform result, performing inverse Fourier transform to obtain probability density distribution of the target signal, as shown in formula (4),
compared with the prior art, the invention has the following advantages:
the probability density functions corresponding to the target noise can be obtained by deconvoluting the known probability density functions of the noise floor, the target noise superposition signal and the noise floor signal, and the consistency of test results is improved by using a noise separation method based on a probability deconvolution method. Test results show that the method can effectively separate different types of noise so as to analyze the characteristics of various types of noise.
Description of the drawings:
fig. 1 is a graph of a two-segment white noise time domain signal and a corresponding probability density function.
Fig. 2 is a diagram of a noise image and a corresponding probability density function obtained by superimposing source noise at two ends.
Fig. 3 is a schematic diagram of probability density distribution of source noise signals separated after being processed by a probability deconvolution method.
The specific embodiment is as follows:
the invention is further described in terms of specific embodiments in conjunction with the following drawings:
before underwater noise source radiation noise power measurement is carried out, a hydrophone is firstly installed at a specified position, and equivalent noise time domain discrete signals of the hydrophone in the current water area are measured; and starting an underwater noise source, recording discrete signals obtained by measurement of the hydrophone at the moment, and then performing the following operations:
step 1, recording a discrete signal n of a noise floor signal 1 [x]And a discrete signal n of the measurement signal tot [x]；
Step 2, the signal array n 1 [x]And n tot [x]Classifying each element according to the size of the element, and recording the number of discrete signal array elements in each value interval;
step 3, dividing the number of elements in the value interval by the total number of elements in the discrete signal array, and dividing by the length of the value interval to obtain a distribution function R of the discrete array 1 (n) and R tot (n)；
Step 4, measuring for multiple times, adding the distribution functions R (n) obtained by each measurement, and averaging to obtain the noise floorAnd measuring the final probability distribution function P of the signal 1 (n) and P tot (n)；
Step 5, probability distribution function P 1 (n) and P tot (n) substitution into equation (4)
Obtaining probability distribution function P of target noise signal 2 (n) completing noise separation.
By taking superposition and separation of two white noise segments as an example, as shown in fig. 1, the time domain signals of the two white noise segments and the corresponding probability density function diagrams are shown in the figure. The noise image and the corresponding probability density function diagram obtained after superposition are shown in fig. 2. It can be seen that the time domain signal of the noise in fig. 2 is also a time domain signal, and the nature of the source noise cannot be judged from the time domain image; the probability density function is an isosceles triangle, so that the characteristic that the probability density of the source noise is uniformly distributed is clearly reflected.
The probability density distribution of the signal on the right side in fig. 1 is completely restored by substituting the probability density functions of the signal on the left side in fig. 1 and the signal on the right side in fig. 2 into the formula (4) respectively, which shows that the noise separation technology based on the probability deconvolution method is effective.
At present, a frequency spectrum analysis method is mostly adopted in the method for measuring noise, and the negative influence of noise floor on the accuracy of a measurement result cannot be thoroughly eliminated due to the randomness and the irrelevance of noise in a time domain and a frequency domain. The invention uses noise separation technology to reduce the influence of noise floor on the measurement result as much as possible, and obtains the probability density function corresponding to the target noise by carrying out deconvolution processing on the known probability density functions of noise floor, target noise superposition signal and noise floor signal.
The foregoing is illustrative of the preferred embodiments of the present invention, and is not to be construed as limiting the claims. All equivalent flow changes made by the specification of the invention are included in the protection scope of the invention.
Claims (4)
1. A noise separation method based on a probability deconvolution method is characterized by comprising the following steps of: comprises the steps of,
step 1, recording a discrete signal n of a noise floor signal 1 [x]And a discrete signal n of the measurement signal tot [x]；
Step 2, the signal array n 1 [x]And n tot [x]Classifying each element according to the size of the element, and recording the number of discrete signal array elements in each value interval;
step 3, dividing the number of elements in the value interval by the total number of elements in the discrete signal array, and dividing by the length of the value interval to obtain a distribution function R of the discrete array 1 (n) and R tot (n)；
Step 4, measuring for multiple times, adding the distribution functions R (n) obtained by each measurement, and averaging to obtain the final probability distribution function P of the noise floor and the measurement signal 1 (n) and P tot (n)；
Step 5, probability distribution function P 1 (n) and P tot (n) substitution into formula
Obtaining probability distribution function P of target noise signal 2 (n) completing noise separation.
2. The noise separation method based on the probabilistic deconvolution method of claim 1, wherein: recording noise floor signal n with acquisition system 1 [x]And a discrete signal n of the measurement signal tot [x]。
3. The noise separation method based on the probabilistic deconvolution method of claim 1, wherein: discrete signal n using hydrophone noise floor signal 1 [x]And a discrete signal n of the measurement signal tot [x]。
4. The noise separation method based on the probabilistic deconvolution method of claim 1, wherein: formula (VI)
In P tot (x) To measure probability density function of signal, P 1 (x) For the noise floor signal of the hydrophone, the probability density function P of the signal is measured tot (x) And noise floor signal P of hydrophone 1 (x) Fourier transform is performed to obtain fft (P) tot (x) And fft (P) 1 (x) And then dividing the Fourier transform result, and then carrying out inverse Fourier transform to obtain the probability density distribution of the target signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310920894.5A CN117009746A (en) | 2023-07-26 | 2023-07-26 | Noise separation method based on probability deconvolution method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310920894.5A CN117009746A (en) | 2023-07-26 | 2023-07-26 | Noise separation method based on probability deconvolution method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117009746A true CN117009746A (en) | 2023-11-07 |
Family
ID=88562965
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310920894.5A Pending CN117009746A (en) | 2023-07-26 | 2023-07-26 | Noise separation method based on probability deconvolution method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117009746A (en) |
-
2023
- 2023-07-26 CN CN202310920894.5A patent/CN117009746A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5496192B2 (en) | Method for reconstructing an acoustic field | |
| CN107247251B (en) | Three-dimensional sound source positioning method based on compressed sensing | |
| Levine et al. | Model-based imaging of damage with Lamb waves via sparse reconstruction | |
| Wear et al. | Application of autoregressive spectral analysis to cepstral estimation of mean scatterer spacing | |
| US8036836B2 (en) | Dynamic environmental change compensation of sensor data in structural health monitoring systems | |
| CN111487046B (en) | Fault diagnosis method for circuit breaker voiceprint and vibration entropy feature fusion | |
| CN109375171B (en) | Sound source positioning method based on orthogonal matching pursuit algorithm | |
| JP6386556B2 (en) | Wide frequency band acoustic holography | |
| Austin et al. | Special problems in the estimation of power-law spectra as applied to topographical modeling | |
| CN110717478A (en) | Object detection system and method based on microwaves | |
| CN110702042B (en) | Thickness estimation method under signal aliasing condition in ultrasonic pulse echo thickness measurement | |
| CN116304549A (en) | Wavelet threshold denoising method for tunnel health monitoring data | |
| CN110119690B (en) | Parallel reactor vibration sensitive area selection method based on CRP and RQA | |
| CN117009746A (en) | Noise separation method based on probability deconvolution method | |
| CN109752633B (en) | Method and system for positioning partial discharge signal of transformer substation | |
| CN115435895A (en) | Background noise measuring method based on probability distribution function | |
| Sodagar et al. | An alternative approach for measuring the scattered acoustic pressure field of immersed single and multiple cylinders | |
| CN115201821B (en) | Small target detection method based on strong target imaging cancellation | |
| CN115614309A (en) | Circumferential acoustic modal measurement method of aero-engine based on Lp norm regularization | |
| CN112818762A (en) | Large-size composite material and rapid nondestructive testing method for sandwich structure thereof | |
| CN114047260B (en) | Damage identification method and system based on multi-scale fuzzy entropy | |
| CN113381793A (en) | Coherent information source estimation-oriented non-grid direction-of-arrival estimation method | |
| CN111722026A (en) | Insulating medium space charge measuring method and system based on magnetoacoustic system | |
| CN112033518A (en) | Hydrophone sensitivity calibration method based on sine function fitting | |
| CN115307774B (en) | Object surface temperature field detection method and system based on ultrasonic guided waves |
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 |