CN110376436A - A kind of multiple dimensioned line-spectrum detection method for noise power spectra - Google Patents
A kind of multiple dimensioned line-spectrum detection method for noise power spectra Download PDFInfo
- Publication number
- CN110376436A CN110376436A CN201910565539.4A CN201910565539A CN110376436A CN 110376436 A CN110376436 A CN 110376436A CN 201910565539 A CN201910565539 A CN 201910565539A CN 110376436 A CN110376436 A CN 110376436A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- line
- data
- group
- parameter
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
- G01R23/16—Spectrum analysis; Fourier analysis
Landscapes
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Monitoring And Testing Of Transmission In General (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a kind of multiple dimensioned line-spectrum detection method for noise power spectra, and this method comprises the following steps: step 1: obtaining the power spectrum data of noise signal to be detected;Step 2: parameter initialization, sets multi-group data window and threshold parameter;Step 3: the maximum in search power spectrum data;Step 4: carrying out normalized set, preliminary ruling line spectrum to the corresponding window data of each maximum point under each group Parameter Conditions;Adjudicate whether each maximum point is line spectrum point step 5: comprehensive.The partial statistics characteristic of power spectrum line spectrum is utilized in detection method of the invention, comprehensive judgement is carried out under the conditions of multiple data windows and threshold parameter, have the characteristics that algorithm is simple, detection probability is high, strong interference immunity, is suitble to carry out noise power spectrum extraction of line spectrum rapidly and efficiently.
Description
Technical field
The invention belongs to signal processing technology field more particularly to a kind of multiple dimensioned line-spectrum detection method for noise power spectra.
Background technique
Power spectrum line spectrum is the form of expression of the single-frequency narrowband ingredient in power spectrum in signal.Power spectrum extraction of line spectrum is to obtain
Take one of the important means of noise signal feature.There are two types of conventional extraction of line spectrum methods, one is by power spectrum signal into
Row background equalization, then the detection threshold of line spectrum is estimated by the fluctuating statistical property in power spectrum signal, and then carry out line spectrum
Global data are utilized in detection, this method, have preferable robustness, but the signal of statistical nature big rise and fall is examined
It is poor to survey performance;Another method of line-spectrum detection be according to local data's point of power spectrum carry out exceptional value judgement, thus
Line spectrum is detected, this mode chooses local data to non-uniform the well adapting to property of power spectrum data of statistical property
Window it is long and thresholding directly affects detection performance, and be difficult to seek unity of standard, there are larger fluctuation or be locally present in local data
When more line spectrums, it is easy to bring the missing inspection of line spectrum and misjudgement.
Summary of the invention
Goal of the invention: for problem and shortage existing for above-mentioned existing method, the invention proposes a kind of multiple dimensioned noises
Power spectrum line-spectrum detection method, this method by the independent window of setting multiple groups and threshold parameter, using power spectrum compose in line spectrum
Value and the otherness of local background's value, successively detect each maximum point with each group parameter, by maximum point not
The equilibrium treatment and normalized set for the data in window of studying in the same school under each group Parameter Conditions of long scale, and it is adjudicated in this group of parameter
Under whether meet line spectrum condition, finally integrate the discriminative information obtained under all groups of Parameter Conditions provide the maximum point whether be
The judgement of line spectrum, this method have the characteristics that algorithm is simple, detection probability is high, strong interference immunity, are suitble to noise power
Spectrum carries out extraction of line spectrum rapidly and efficiently.
Technical solution: in order to achieve the above-mentioned object of the invention, the invention adopts the following technical scheme: a kind of multiple dimensioned noise function
Rate spectral line spectrum detection method, this method comprises the following steps:
(1) the power spectrum data s (n), n=0,1 ..., N-1 of noise signal to be detected are obtained, the n is power spectrum
Data sequence number, N are power spectrum data number;
(2) parameter initialization sets M group data window and threshold parameter, and wherein M is the natural number greater than 0, and power spectrum is arranged
Serial number indexes i=0, and setting line spectrum collection is combined into E;
(3) i=i+1 operation is executed, if i is equal to N-1, line-spectrum detection work terminates, each in line spectrum set E at this time
Element as detects obtained line spectrum parameter;If i < N-1, further judge whether power spectrum data s (i) is maximum point,
If s (i) is that maximum point thens follow the steps (4), if s (i) is not maximum point, this step is repeated;
(4) under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, the preliminary ruling maximum point exists
Whether line spectrum point condition is met under each group Parameter Conditions;
(5) whether comprehensive judgement s (i) is line spectrum point, if it is line spectrum point, is then added in set E;
(6) if i < N-1, return step (3) is continued to execute, and otherwise, this method terminates.
Further, in step (1), the power spectrum data s (n) of noise signal to be detected is obtained with the following method:
From the received acquisition data of sensor and carries out power spectrumanalysis and obtain s (n);Or noise signal to be detected is read from memory
Power spectrum data s (n).
Further, in step (2), parameter initialization is carried out, M group independence and the different parameter of the long scale of window are set,
M group parameter includes the long L of left windowm, the right long R of windowm, big value exclusion number Um, small value exclusion number Dm, amplitude threshold Gm, wherein
M=1,2 ..., M, Lm, Rm, Um, DmIt is the natural number greater than 0, GmFor the real number greater than 0, comprehensive thresholding piece is set, and H is big
In 0 natural number, power spectrum serial number is set and indexes i=0, setting line spectrum set E is empty set.
Further, in step (3), the specific steps are as follows:
(3-1) executes i=i+1 operation;
(3-2) if i is equal to N-1, line-spectrum detection work terminates, and each element in set E as detects obtained line
Compose parameter;If i < N-1, further judge whether power spectrum data s (i) is maximum point, i.e., whether meet condition s (i) >=
S (i-1) and s (i) > s (i+1), if s (i) is that maximum point thens follow the steps (4), if s (i) is not maximum point, weight
Execute step (3) again.
Further, in step (4), under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, just
Step adjudicates whether the maximum point meets line spectrum condition under each group Parameter Conditions, and when selected m group parameter, step is such as
Under:
(4-1) is directed to maximum point s (i), according to the m group parameter that step (2) are set, intercepts out from s (n) corresponding
Data take data length Km=min (N-1, i+Rm)-max (0, i-Lm)+1, max () and min () respectively indicate be maximized and
It is minimized, the data using the interception of m group parameter are gm(j)=s (j+max (0, i-Lm)), j=0,1 ..., Km-1;
(4-2) calculates gm(j) mean value a0;
(4-3) sequence of calculation bm(j), formula is as follows:
(4-4) is to sequence gm(j) background equalization is carried out to obtainFormula is as follows:
Wherein, operation is inner product operation;
(4-6) uses data sequenceElement constitute a data acquisition system
(4-7) is in data acquisition system PmIn, find out preceding UmA maximum value element and from data acquisition system PmMiddle rejecting;
(4-8) is in data acquisition system PmIn, find out preceding DmA minimum value element and from data acquisition system PmMiddle rejecting;
(4-9) calculates the data acquisition system P after (4-7) and (4-8) processingmMean value emAnd standard deviation sigmam;
(4-10) is when meeting s (i) > ao+em+Gm×σmWhen, the line spectrum mark L of m group is setm=1, if conditions are not met, then
The line spectrum mark L of m group is setm=0;
(4-11) is repeated the above process, and is finished until all M group parameters all calculate.
Further, in step (5), comprehensive to adjudicate whether each maximum point is line spectrum point, judgment method is as follows: each group
Line spectrum mark meetsWhen, s (i) the point conclusive judgement in power spectrum data is line spectrum, by (i, s at this time
(i)) it is used as element, is added in set E.
The utility model has the advantages that compared with prior art, technical solution of the present invention has following advantageous effects:
The partial statistics characteristic of power spectrum line spectrum is utilized in detection method of the invention, in the data window and door of multiple scales
Comprehensive judgement is carried out under limit Parameter Conditions, has the characteristics that algorithm is simple, detection probability is high, strong interference immunity, is suitble to noise
Power spectrum signal carries out extraction of line spectrum rapidly and efficiently.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention;
Fig. 2 is the power spectrum sequence figure of embodiment 1;
Fig. 3 is the line-spectrum detection result of embodiment 1.
Specific embodiment
The present invention is described further with reference to the accompanying drawings and examples:
As shown in Figure 1, the present invention proposes a kind of multiple dimensioned line-spectrum detection method for noise power spectra, this method includes following step
It is rapid:
(1) the power spectrum data s (n), n=0,1 ..., N-1 of noise signal to be detected are obtained, the n is power spectrum
Data sequence number, N are power spectrum data number;
(2) parameter initialization sets M group data window and threshold parameter, and wherein M is the natural number greater than 0, and power spectrum is arranged
Serial number indexes i=0, and setting line spectrum collection is combined into E;
(3) i=i+1 operation is executed, if i is equal to N-1, line-spectrum detection work terminates, each in line spectrum set E at this time
Element as detects obtained line spectrum parameter;If i < N-1, further judge whether power spectrum data s (i) is maximum point,
If s (i) is that maximum point thens follow the steps (4), if s (i) is not maximum point, this step is repeated;
(4) under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, the preliminary ruling maximum point exists
Whether line spectrum point condition is met under each group Parameter Conditions;
(5) whether comprehensive judgement s (i) is line spectrum point, if it is line spectrum point, is then added in set E;
(6) if i < N-1, return step (3) is continued to execute;Otherwise, this method terminates.
Further, in step (1), the power spectrum data s (n) of noise signal to be detected is obtained with the following method:
From the received acquisition data of sensor and carries out power spectrumanalysis and obtain s (n);Or noise signal to be detected is read from memory
Power spectrum data s (n).
Further, in step (2), parameter initialization is carried out, M group independence and the different parameter of the long scale of window are set,
M group parameter includes the long L of left windowm, the right long R of windowm, big value exclusion number Um, small value exclusion number Dm, amplitude threshold Gm, wherein
M=1,2 ..., M, Lm, Rm, Um, DmIt is the natural number greater than 0, GmFor the real number greater than 0, comprehensive thresholding H is set, H be greater than
0 natural number, setting power spectrum serial number index i=0, and setting line spectrum set E is empty set.
Further, in step (3), the specific steps are as follows:
(3-1) executes i=i+1 operation;
(3-2) if i=N-1, line-spectrum detection work terminates, and each element in set E as detects obtained line spectrum
Parameter;If i < N-1, further judges whether power spectrum data s (i) is maximum point, i.e., whether meet condition s (i) >=s
(i-1) and s (i) > s (i+1) if s (i) is not maximum point, is repeated if s (i) is that maximum point thens follow the steps (4)
It executes step (3).
Further, in step (4), under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, just
Step adjudicates the maximum point is whether to meet line spectrum condition under each group Parameter Conditions, and when selected m group parameter, step is such as
Under:
(4-1) is directed to maximum point s (i), according to the m group parameter that step (2) are set, intercepts out from s (n) corresponding
Data take data length Km=min (N-1, i+Rm)-max (0, i-Lm)+1, max () and min () respectively indicate be maximized and
It is minimized, the data using the interception of m group parameter are gm(j)=s (j+max (0, i-Lm)), j=0,1 ..., Km-1;
(4-2) calculates gm(j) mean value a0;
(4-3) sequence of calculation bm(j), formula is as follows:
(4-4) is to sequence gm(j) background equalization is carried out to obtainFormula is as follows:
Wherein, operation is inner product operation;
(4-6) uses data sequenceElement constitute a data acquisition system
(4-7) is in data acquisition system PmIn, find out preceding UmA maximum value element and from data acquisition system PmMiddle rejecting;
(4-8) is in data acquisition system PmIn, find out preceding DmA minimum value element and from data acquisition system PmMiddle rejecting;
(4-9) calculates the data acquisition system P after (4-7) and (4-8) processingmMean value emAnd standard deviation sigmam;
(4-10) is when meeting s (i) > a0+em+Gm×σmWhen, the line spectrum mark L of m group is setm=1, if conditions are not met, then
The line spectrum mark L of m group is setm=0;
(4-11) is repeated the above process, and is finished until all M group parameters all calculate.
Further, in step (5), comprehensive to adjudicate whether each maximum point is line spectrum point, judgment method is as follows: each group
Line spectrum mark meetsWhen, s (i) the point conclusive judgement in power spectrum data is line spectrum, by (i, s at this time
(i)) it is used as element, is added in set E.
Embodiment 1
Emulate coloured a power spectrum signal s (n), n=0,1 ..., 1000, power spectrum data number 1001, power
A data point all contains random noise in spectrum signal, (100) s in power spectrum signal, s (150), s (500), s (508), s
(700) line spectrum ingredient is contained, amplitude is respectively 6,9,10,5,4, and power spectrum signal s (n) is as shown in Figure 2;
According to (2) step, parameter initialization sets 4 groups of data windows and threshold parameter, each group parameter setting is as follows: 1 ginseng of group
Number has the long L of left window1=10, the right long R of window1=10, big value excludes number U1=3, small value excludes number D1=3, amplitude threshold G1=
8;2 parameters of group have the long L of left window2=15, the right long R of window2=15, big value excludes number U2=4, small value excludes number D2=4, amplitude
Thresholding G2=8;3 parameters of group have the long L of left window3=20, the right long R of window3=20, big value excludes number U3=5, small value excludes number D3=
5, amplitude threshold G3=7;4 parameters of group have the long L of left window4=30, the right long R of window4=30, big value excludes number U4=8, small value excludes
Number D4=8, amplitude threshold G4=7;Comprehensive thresholding H=2, setting power spectrum serial number index i=0, and setting line spectrum set E is sky
Collection.
According to (3) step, i=i+1 operation is executed, if i is equal to N-1, line-spectrum detection work terminates, at this time line spectrum collection
The each element closed in E as detects obtained line spectrum parameter, E=(100,186.70), (150,187.97), (500,
181.21), (508,177.96), (700,161.95) }, as shown in figure 3, wherein line spectrum is marked with No. *;If i < N-1, into
One step judges whether power spectrum data s (i) is maximum point, if s (i) is that maximum point thens follow the steps (4), if s (i)
It is not maximum point, repeats this step;
(4) under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, the preliminary ruling maximum point exists
Whether line spectrum point condition is met under each group Parameter Conditions;
(5) whether comprehensive judgement s (i) is line spectrum point, if it is line spectrum point, is then added in set E;
(6) if i < N-1, return step (3) is continued to execute;Otherwise, this method process terminates.
Claims (6)
1. a kind of multiple dimensioned line-spectrum detection method for noise power spectra, which is characterized in that this method comprises the following steps:
(1) the power spectrum data s (n), n=0,1 ..., N-1 of noise signal to be detected are obtained, the n is power spectrum data sequence
Number, N is power spectrum data number;
(2) parameter initialization sets M group data window and threshold parameter, wherein M is the natural number greater than 0, and power spectrum sequence is arranged
Number index i=0, setting line spectrum collection be combined into E;
(3) i=i+1 operation is executed, if i is equal to N-1, line-spectrum detection work terminates, at this time each member in line spectrum set E
Element, the line spectrum parameter as detected;If i < N-1, further judge whether power spectrum data s (i) is maximum point,
If s (i) is that maximum point thens follow the steps (4), if s (i) is not maximum point, step (3) are repeated;
(4) under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, the preliminary ruling maximum point is in each group
Whether line spectrum point condition is met under Parameter Conditions;
(5) whether comprehensive judgement s (i) is line spectrum point, if it is line spectrum point, is then added in set E;
(6) if i < N-1, return step (3) is continued to execute;Otherwise, this method terminates.
2. a kind of multiple dimensioned line-spectrum detection method for noise power spectra according to claim 1, which is characterized in that in step
(1) in, the power spectrum data s (n) of noise signal to be detected is obtained with the following method: being received acquisition data from sensor and is gone forward side by side
Row power spectrumanalysis obtains s (n);Or the power spectrum data s (n) of noise signal to be detected is read from memory.
3. a kind of multiple dimensioned line-spectrum detection method for noise power spectra according to claim 1 or 2, which is characterized in that in step
Suddenly in (2), parameter initialization method is as follows: setting M group independence and the different parameter of the long scale of window, m group parameter includes a left side
The long L of windowm, the right long R of windowm, big value exclusion number Um, small value exclusion number Dm, amplitude threshold Gm, wherein m=1,2 ..., M, Lm,
Rm, Um, DmIt is the natural number greater than 0, GmFor the real number greater than 0, comprehensive thresholding H is set, H is the natural number greater than 0, setting
Power spectrum serial number indexes i=0, and setting line spectrum set E is empty set.
4. a kind of multiple dimensioned line-spectrum detection method for noise power spectra according to claim 3, which is characterized in that in step
(3) in, the specific steps are as follows:
(3-1) executes i=i+1 operation;
(3-2) if i is equal to N-1, line-spectrum detection work terminates, each element in set E, the line spectrum ginseng as detected
Number;If i < N-1, further judges whether power spectrum data s (i) is maximum point, i.e., whether meet condition s (i) >=s (i-
1)] and s (i) > s (i+1) if s (i) is not maximum point, repeats to hold if s (i) is that maximum point thens follow the steps (4)
Row step (3).
5. a kind of multiple dimensioned line-spectrum detection method for noise power spectra according to claim 4, which is characterized in that in step
(4) in, under each group Parameter Conditions, maximum point s (i) corresponding statistic is calculated, the preliminary ruling maximum point is in each group
Whether line spectrum condition is met under Parameter Conditions, and when selected m group parameter, its step are as follows:
(4-1) is directed to maximum point s (i), and according to selected m group parameter, corresponding data are intercepted out from s (n), evidence of fetching
Length Km=min (N-1, i+Rm)-max (0, i-Lm)+1, max () and min () respectively indicate and be maximized and be minimized, benefit
It is g with the data that m group parameter interceptsm(j)=s (j+max (0, i-Lm)), j=0,1 ..., Km-1;
(4-2) calculates gm(j) mean value a0;
(4-3) sequence of calculation bm(j), formula is as follows:
(4-4) is to sequence gm(j) background equalization is carried out to obtainFormula is as follows:
Wherein, operation is inner product operation;
(4-6) uses data sequenceElement constitute a data acquisition system
(4-7) is in data acquisition system PmIn, find out preceding UmA maximum value element and from data acquisition system PmMiddle rejecting;
(4-8) is in data acquisition system PmIn, find out preceding DmA minimum value element and from data acquisition system PmMiddle rejecting;
(4-9) calculates the data acquisition system P after (4-7) and (4-8) processingmMean value emAnd standard deviation sigmam;
(4-10) is when meeting s (i) > a0+em+Gm×σmWhen, the line spectrum mark L of m group is setm=1, if conditions are not met, being then arranged
The line spectrum mark L of m groupm=0;
(4-11) is repeated the above process, and is finished until all M group parameters all calculate.
6. a kind of multiple dimensioned line-spectrum detection method for noise power spectra according to claim 5, which is characterized in that in step
(5) in, comprehensive to adjudicate whether each maximum point is line spectrum point, judgment method is as follows: each group line spectrum mark meetsWhen, s (i) the point conclusive judgement in power spectrum data is line spectrum, regard (i, s (i)) at this time as element, adds
It is added in set E.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910565539.4A CN110376436B (en) | 2019-06-27 | 2019-06-27 | Multi-scale noise power spectral line spectrum detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910565539.4A CN110376436B (en) | 2019-06-27 | 2019-06-27 | Multi-scale noise power spectral line spectrum detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110376436A true CN110376436A (en) | 2019-10-25 |
CN110376436B CN110376436B (en) | 2021-06-01 |
Family
ID=68250717
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910565539.4A Active CN110376436B (en) | 2019-06-27 | 2019-06-27 | Multi-scale noise power spectral line spectrum detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110376436B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112379357A (en) * | 2020-11-16 | 2021-02-19 | 东南大学 | Signal discrimination method based on pulse signal pre-estimation time parameter secondary processing |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002148329A (en) * | 2000-11-10 | 2002-05-22 | Taiyo Musen Co Ltd | Disturbance wave removing method |
CN104217616A (en) * | 2014-08-19 | 2014-12-17 | 中国电子科技集团公司第二十三研究所 | Inland waterway traffic monitoring implementation method based on optical fiber acoustic sensors |
CN109061591A (en) * | 2018-07-23 | 2018-12-21 | 东南大学 | A kind of time-frequency line-spectrum detection method based on sequential cluster |
CN109285561A (en) * | 2018-09-06 | 2019-01-29 | 东南大学 | A kind of ship propeller cavitation noise Modulation Spectral Feature fidelity Enhancement Method based on adaptive window length |
CN109655148A (en) * | 2018-12-19 | 2019-04-19 | 南京世海声学科技有限公司 | A kind of autonomous extracting method of ship noise non-stationary low frequency spectrum lines |
-
2019
- 2019-06-27 CN CN201910565539.4A patent/CN110376436B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002148329A (en) * | 2000-11-10 | 2002-05-22 | Taiyo Musen Co Ltd | Disturbance wave removing method |
CN104217616A (en) * | 2014-08-19 | 2014-12-17 | 中国电子科技集团公司第二十三研究所 | Inland waterway traffic monitoring implementation method based on optical fiber acoustic sensors |
CN109061591A (en) * | 2018-07-23 | 2018-12-21 | 东南大学 | A kind of time-frequency line-spectrum detection method based on sequential cluster |
CN109285561A (en) * | 2018-09-06 | 2019-01-29 | 东南大学 | A kind of ship propeller cavitation noise Modulation Spectral Feature fidelity Enhancement Method based on adaptive window length |
CN109655148A (en) * | 2018-12-19 | 2019-04-19 | 南京世海声学科技有限公司 | A kind of autonomous extracting method of ship noise non-stationary low frequency spectrum lines |
Non-Patent Citations (1)
Title |
---|
周清平: ""舰船噪声包络的高阶统计量特征研究"", 《中国优秀硕士学位论文全文数据库(工程科技II辑)》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112379357A (en) * | 2020-11-16 | 2021-02-19 | 东南大学 | Signal discrimination method based on pulse signal pre-estimation time parameter secondary processing |
CN112379357B (en) * | 2020-11-16 | 2023-10-27 | 东南大学 | Signal discrimination method based on secondary processing of pulse signal estimated time parameter |
Also Published As
Publication number | Publication date |
---|---|
CN110376436B (en) | 2021-06-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111913156B (en) | Radar radiation source individual identification method based on deep learning model and feature combination | |
CN106658590B (en) | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN103065149B (en) | Muskmelon Fruit phenotype is extracted and quantization method | |
CN113255848B (en) | Water turbine cavitation sound signal identification method based on big data learning | |
CN109213121B (en) | Method for diagnosing clamping cylinder fault of fan braking system | |
CN111126471A (en) | Microseism event detection method and system | |
CN112435221A (en) | Image anomaly detection method based on generative confrontation network model | |
CN108073158A (en) | Based on PCA and KNN density algorithm Wind turbines Method for Bearing Fault Diagnosis | |
CN112257530B (en) | Rolling bearing fault diagnosis method based on blind signal separation and support vector machine | |
CN111275288A (en) | XGboost-based multi-dimensional data anomaly detection method and device | |
CN108435819B (en) | Energy consumption abnormity detection method for aluminum profile extruder | |
CN110133714B (en) | Microseismic signal classification and identification method based on deep learning | |
CN104538041A (en) | Method and system for detecting abnormal sounds | |
CN110175541B (en) | Method for extracting sea level change nonlinear trend | |
CN108090515B (en) | Data fusion-based environment grade evaluation method | |
CN111698258B (en) | WiFi-based environmental intrusion detection method and system | |
CN103048273A (en) | Fruit near infrared spectrum sorting method based on fuzzy clustering | |
CN116522268B (en) | Line loss anomaly identification method for power distribution network | |
CN112613443A (en) | Robustness communication radiation source intelligent identification method based on deep learning | |
Cao et al. | Bearing fault diagnosis method based on GMM and coupled hidden Markov model | |
CN115797772A (en) | Classified early warning method and system for icing thickness of overhead transmission line | |
CN112213687A (en) | Gateway electric energy meter data anomaly detection method and system based on pseudo anomaly point identification | |
CN112765550A (en) | Target behavior segmentation method based on Wi-Fi channel state information | |
CN110376436A (en) | A kind of multiple dimensioned line-spectrum detection method for noise power spectra |
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 |