CN105827218B - A kind of grass implemented to FFT data limits filtering method - Google Patents
A kind of grass implemented to FFT data limits filtering method Download PDFInfo
- Publication number
- CN105827218B CN105827218B CN201610323875.4A CN201610323875A CN105827218B CN 105827218 B CN105827218 B CN 105827218B CN 201610323875 A CN201610323875 A CN 201610323875A CN 105827218 B CN105827218 B CN 105827218B
- Authority
- CN
- China
- Prior art keywords
- filtering
- value
- average noise
- data value
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0201—Wave digital filters
Abstract
The invention discloses a kind of grass implemented to FFT data to limit filtering method, different filtering process is carried out according to the difference of filtering process band, including being filtered between limit value maximum, to in FFT signal datas to be dealt be higher than reference value dRef's and less than maximum decline δ amplitude section in signal carry out analysis filtering, with average noise replace the section in data value handled；Band filtering on average noise, analyzed by the signal multiplied in factor θ setting multiple section being higher than average noise in FFT signal datas to be dealt with, replace data value in the section to be handled with average noise；The average noise upper limit filters, and is analyzed less than the signal in reference value dRef sections being counted to again higher than the setting of average noise lower limit factor-beta in FFT signal datas to be dealt with, replaces data value in the section to be handled with average noise.Effectively filter out the garbage signal in known pre-selection band or be indifferent to signal.
Description
Technical field
The present invention relates to digital signal processing technique field, more particularly to a kind of grass implemented to FFT data to limit filter
Wave method.
Background technology
In high speed acquisition process field, because high-speed a/d sampling rate is high, at back-end hardware design and real-time data analysis
Reason brings very big difficulty.Sampled data is after FFT processing, many times because big bandwidth problem, signal leakage are asked
Topic, signal aliasing problem, with interior self-excitation problem etc. so that effect is unsatisfactory after FFT, some undesired signals or and is not related to
The signal of system can occur in useful signal band, and high-speed case lower front end hardware and FPGA process part have been difficult solution
Certainly, data process effects are influenceed.
Under normal circumstances, when hardware environment is undesirable, especially in the case of high-speed applications, general is difficult excellent by hardware
Change means accomplish the thorough improvement of signal effect.Meanwhile the improvement of hardware environment need to a certain extent increase manpower costs, when
Between extension, hardware resource expend etc., and in many cases, because of the difference of environment, debugging result once does not possess multiple environment
Under versatility demand.
The content of the invention
The purpose of the present invention is exactly to solve the above problems, there is provided a kind of noise bandlimiting filtering implemented to FFT data
Method, effectively filter out the garbage signal in known pre-selection band or be indifferent to signal, purified treatment effect, at high-speed data signal
Reason is issued to preferable processing requirement.
To achieve these goals, the present invention adopts the following technical scheme that：
A kind of grass implemented to FFT data limits filtering method, to entering less than the big signal of the apex zone of peak signal
Filtered between row limit value maximum, band filtering on average noise is carried out to bottom zone small-signal above grass, in signal band
Between section region carry out upper noise limit filtering；
Filtered between limit value maximum, to be higher than in FFT signal datas to be dealt with reference value dRef's and less than maximum
Value declines signal in δ amplitude section and carries out analysis filtering, replaces data value in the section to be handled with average noise；
Band filtering on average noise, to being multiplied factor θ setting times higher than average noise in FFT signal datas to be dealt with
Signal in number section carries out analysis filtering, replaces data value in the section to be handled with average noise；
The average noise upper limit filters, to being higher than the setting times of average noise lower limit factor-beta in FFT signal datas to be dealt with
Count to and carry out analysis filtering less than the signal in reference value dRef sections, replaced with average noise in the section at data value
Reason.
For the given data value group X being made up of N number of FFT data, and reference value dRef, data value group X counting rope
L is cited as, average noise baseline value dAvgN is determined by being averaging noisy base line method first, is determined by maximizing method
This data value group X maximum dMax；Meet that the different of condition filtered, on average noise between limit value maximum according to numerical value
Band filtering or the filtering of the average noise upper limit.
The acquiring method of average noise baseline value includes,
Step 11：If the counting index that data value group X asks for average noise span is initiated with Lmin, count index eventually
It is only Lmax；If evaluation summation is XT, count value LT；It is L to make Lmin, XTFor 0, LTFor 0；Subsequently into step 12；
Step 12：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether reference value dRef is less than,
It is then to enter step 13, otherwise into step 14；
Step 13：Make XT=XT+XL, LT=LT+1；Into step 14；
Step 14：Count index L and add 1；Into step 15；
Step 15：Judge whether L is less than or equal to Lmax, it is then return to step 12；Otherwise step 16 is entered；
Step 16：UtilizeTry to achieve average noise baseline value dAvgN.
Counting index starting L described in the step 11minL is terminated with counting to indexmaxDetermination method be：
The specific method filtered between limit value maximum includes,
Step 21：It is 0 to initialize L, is calculated by descending factors δ and filters higher limit determined by maximum dMax
DMaxU, into step 22；
Step 22：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether reference value is more than or equal to
DRef, and be then to enter step 23, otherwise into step 24 less than filtering higher limit dMaxU；
Step 23：Replace indexing value X corresponding to L in X with dAvgNL；Into step 24；
Step 24：Count index L and add 1；Into step 25；
Step 25：Judge whether L is less than N, be then return to step 22；Otherwise whole handling process terminates.
Specific method with filtering on average noise includes,
Step 31：It is 0 to initialize L, is multiplied factor θ by bandwidth and determines the filtering upper limit with filtering method on average noise
Value dMaxU2 and filtering lower limit dMaxD2, into step 32；
Step 32：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
Value dMaxD2, and be then to enter step 33, otherwise into step 34 less than filtering higher limit dMaxU2；
Step 33：Replace indexing value X corresponding to L in X with dAvgNL；Into step 34；
Step 34：Count index L and add 1；Into step 35；
Step 35：Judge whether L is less than N, be then return to step 32；Otherwise whole handling process terminates.
The specific method of average noise upper limit filtering includes,
Step 41：It is 0 to initialize L, and the filtering lower limit of average noise upper limit filtering method is determined by lower limit factor-beta
DMaxD, into step 42；
Step 42：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
Value dMaxD, and it is less than reference value dRef, it is then to enter step 43, otherwise into step 44；
Step 43：Replace indexing value X corresponding to L in X with dAvgNL；Into step 44；
Step 44：Count index L and add 1；Into step 45；
Step 45：Judge whether L is less than N, be then return to step 42；Otherwise whole handling process terminates.
Pass through descending factors δ and maximum dMax the filtering higher limit dMaxU implemented determination side in the step 21
Method is：DMaxU=dMax- | dMax | × δ；Wherein δ value is the number more than 0, and dMaxU determined by its value simultaneously will
Seek the condition of satisfaction：dMaxU>dRef.
Factor θ is multiplied by bandwidth in the step 31 and determines the filtering higher limit with filtering method on average noise
DMaxU2 and filtering lower limit dMaxD2 determination method is：DMaxD2=dAvgN+ θ, dMaxU2=dAvgN+ θ × 2；
Wherein θ value is the number more than 0, and dMaxU2 determined by its value is required to meet condition simultaneously：dMaxU2<
dMax。
Determine the filtering lower limit dMaxD's of average noise upper limit filtering method in the step 41 by lower limit factor-beta
The method of determination is：DMaxD=dAvgN+ β；Wherein β value is the number more than 0, and dMaxD determined by its value simultaneously will
Seek the condition of satisfaction：dMaxD<dRef.
Beneficial effects of the present invention：
The present invention is for the undesirable feelings of effect after sampled data FFT processing on the filter processing method of FFT data
Garbage signal is carried out under condition or is indifferent to filtering out for signal handling operation, in the case of precognition is selected, FFT treatment effects can be made
It is more preferably preferable, provide more optional condition filtering modes for signal transacting.Under certain condition, high speed acquisition can effectively be alleviated
The difficulty of hardware design difficulty and high speed FPGA signal transactings, easy to implement as software processing method, optional to match somebody with somebody, application
Flexibly, signal analysis effect is targetedly improved.And debug time is short, human and material resources cost is saved, improves benefit.
Brief description of the drawings
Filtering Analysis schematic illustration when Fig. 1 is filtered between limit value maximum；
Fig. 2 is filter effect between the limit value maximum that -50dB is referred to；
Fig. 3 is filter effect between the limit value maximum that -70dB is referred to；
Fig. 4 is filter effect between the limit value maximum that -75dB is referred to；
Filtering Analysis schematic illustration when Fig. 5 is band filtering on average noise；
Fig. 6 is band filter effect on average noise；
Fig. 7 is Filtering Analysis schematic illustration when the average noise upper limit filters；
Fig. 8 is the average noise upper limit filter effect of -60dB references；
Fig. 9 is the average noise upper limit filter effect of -40dB references.
Embodiment
The invention will be further described with embodiment below in conjunction with the accompanying drawings.
A kind of grass implemented to FFT data limits filtering method, and the stronger filtering of specific aim is specially carried out to FFT data
Processing, made a distinction, including filtered between limit value maximum, band filter on average noise with different filtering process band dividing modes
Ripple, the average noise upper limit filter three kinds of filtering methods, effectively filter out the garbage signal in known pre-selection band or are indifferent to signal, only
Change treatment effect, preferable processing requirement is issued in high-speed data signal processing.
The present invention carries out different filtering process according to the difference of filtering process band, including is filtered between limit value maximum, be flat
Band filtering, the filtering of the average noise upper limit on equal noise, wherein between limit value maximum filtering be mainly used in attached less than peak signal
The big signal filtering process of near apex zone；It is small to be mainly used in bottom zone near grass top for band filtering on average noise
The filtering process of signal；Upper noise limit filtering is mainly used in signal interlude region filtering process.
(1) filtered between limit value maximum
The filtering method is to being higher than reference value dRef in FFT signal datas to be dealt with and declining less than maximum
Signal carries out analysis filtering in δ amplitude section, replaces data value in the section to be handled with average noise.
For the handling process by amplitude section partition, it is considered as interference or mistake to filter out in set amplitude section, or
It is not concerned with or undesired signal, so as to realize the purifying of information.If Fig. 1 is that Filtering Analysis of the descending factors δ equal to 20% shows
Meaning process.
As shown in Fig. 2 filtering situation when being equal to -50dB for dRef, zero-frequency nearby the dc noise less than -50dB and
Interference signal at 40MHz is narrower due to filtering section, is not filtered out.
As shown in figure 3, filtering situation when being equal to -70dB for dRef, the dc noise near zero-frequency be effectively controlled in -
Interference signal at below 70dB and 40MHz is also effectively controlled below -70dB, and noise and interference signal are effectively pressed down
System, effect are obvious.
It is dry at the dc noise and 40MHz near zero-frequency as shown in figure 4, filtering situation when being equal to -75dB for dRef
Disturb signal to be also effectively controlled below -75dB, noise and interference signal are more effectively suppressed, substantially equal with bottom of making an uproar.
(2) band filtering on average noise
The filtering method is to being multiplied factor θ one again to twice higher than average noise in FFT signal datas to be dealt with
Signal in section is analyzed, and replaces data value in the section to be handled with average noise.The handling process passes through amplitude area
Section division, it is considered as interference or mistake to filter out in set amplitude section, or is not concerned with or undesired signal, so as to realize
The purifying of information.If Fig. 5 is to be multiplied Filtering Analysis of the factor θ equal to 10 to illustrate.
When reference value with filtering on average noise is set as -50dB, its filter effect is as shown in fig. 6, for falling into filter
40MHz interference signals in ripple section have obtained effective inhibitory action.
(3) the average noise upper limit filters
The filtering method is arrived less than ginseng to being higher than one times of average noise lower limit factor-beta in FFT signal datas to be dealt with
The signal examined in value dRef sections is analyzed, and replaces data value in the section to be handled with average noise.The handling process is led to
Amplitude section partition is crossed, it is considered as interference or mistake to filter out in set amplitude section, or is not concerned with or undesired signal,
So as to realize the purifying of information.If Fig. 7 is that Filtering Analysis of the lower limit factor-beta equal to 10 is illustrated.
Filter effect when reference value dRef is -60dB is as shown in figure 8, interference signal is effectively suppressed at 40MHz
With filter out, the DC influence signal near zero-frequency is effectively controlled, illustrate that higher limit is too low, filter section it is too narrow,
Filtration result is limited.
Filter effect when reference value dRef is -40dB as shown in figure 9, DC influence signal near zero-frequency and
Interference signal is all effectively suppressed and filtered out at 40MHz, and filter effect reaches preferable requirement.
Wherein average noise baseline ask for process be to FFT signal datas 10%~90% to be dealt with the range of
Data and less than reference value dRef data carry out average value ask for, why take data in the range of 10%~90%, be in order to
The abnormal conditions that signal data stem and afterbody are likely to occur in resampling process are avoided, and influence the degree of accuracy of baseline value.
A kind of grass implemented to FFT data limits filtering method, can be according to concrete application demand 3 kinds of filtering methods of progress
Selection uses, and when different methods is applied, is filtered for band on the descending factors δ in being filtered between limit value maximum, average noise
In multiplied factor θ, the lower limit factor-beta in the filtering of the average noise upper limit and different reference value dRef value can be according to tool
Body hardware effort environment, software debugging condition etc. are changed in the case where meeting value requirement, to adapt under different condition
The optimal selection of the inventive method filtering bandwidth, so as to be optimal filter effect.
For the given data value group X being made up of N number of FFT data, and reference value dRef, data value group X counting rope
L is cited as, average noise baseline value dAvgN is determined by being averaging noisy base line method first, is determined by maximizing method
This data value group X maximum dMax；Signal filtering is carried out according to selected filtering method.
The acquiring method of average noise baseline value includes,
Step 11：If the counting index that data value group X asks for average noise span is initiated with Lmin, count index eventually
It is only Lmax；If evaluation summation is XT, count value LT；It is L to make Lmin, XTFor 0, LTFor 0；Subsequently into step 12；
Step 12：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether reference value dRef is less than,
It is then to enter step 13, otherwise into step 14；
Step 13：Make XT=XT+XL, LT=LT+1；Into step 14；
Step 14：Count index L and add 1；Into step 15；
Step 15：Judge whether L is less than or equal to Lmax, it is then return to step 12；Otherwise step 16 is entered；
Step 16：UtilizeAverage noise baseline value dAvgN is tried to achieve, carries out next data processing link.
Counting index starting L described in the step 11minL is terminated with counting to indexmaxDetermination method be：
The specific method filtered between limit value maximum includes,
Step 21：It is 0 to initialize L, is calculated by descending factors δ and filters higher limit determined by maximum dMax
DMaxU, into step 22；
Step 22：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether reference value is more than or equal to
DRef, and be then to enter step 23, otherwise into step 24 less than filtering higher limit dMaxU；
Step 23：Replace indexing value X corresponding to L in X with dAvgNL；Into step 24；
Step 24：Count index L and add 1；Into step 25；
Step 25：Judge whether L is less than N, be then return to step 22；Otherwise whole handling process terminates, by data value group X
It is sent to next processing links or carries out display processing, carries out next round data processing.
Specific method with filtering on average noise includes,
Step 31：It is 0 to initialize L, is multiplied factor θ by bandwidth and determines the filtering upper limit with filtering method on average noise
Value dMaxU2 and filtering lower limit dMaxD2, into step 32；
Step 32：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
Value dMaxD2, and be then to enter step 33, otherwise into step 34 less than filtering higher limit dMaxU2；
Step 33：Replace indexing value X corresponding to L in X with dAvgNL；Into step 34；
Step 34：Count index L and add 1；Into step 35；
Step 35：Judge whether L is less than N, be then return to step 32；Otherwise whole handling process terminates, by data value group X
It is sent to next processing links or carries out display processing, carries out next round data processing.
The specific method of average noise upper limit filtering includes,
Step 41：It is 0 to initialize L, and the filtering lower limit of average noise upper limit filtering method is determined by lower limit factor-beta
DMaxD, into step 42；
Step 42：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
Value dMaxD, and it is less than reference value dRef, it is then to enter step 43, otherwise into step 44；
Step 43：Replace indexing value X corresponding to L in X with dAvgNL；Into step 44；
Step 44：Count index L and add 1；Into step 45；
Step 45：Judge whether L is less than N, be then return to step 42；Otherwise whole handling process terminates, by data value group X
It is sent to next processing links or carries out display processing, carries out next round data processing.
Pass through descending factors δ and maximum dMax the filtering higher limit dMaxU implemented determination side in the step 21
Method is：DMaxU=dMax- | dMax | × δ；Wherein δ value is the number more than 0, and dMaxU determined by its value simultaneously will
Seek the condition of satisfaction：dMaxU>dRef.
Factor θ is multiplied by bandwidth in the step 31 and determines the filtering higher limit with filtering method on average noise
DMaxU2 and filtering lower limit dMaxD2 determination method is：DMaxD2=dAvgN+ θ, dMaxU2=dAvgN+ θ × 2；
Wherein θ value is the number more than 0, and dMaxU2 determined by its value is required to meet condition simultaneously：dMaxU2<
dMax。
Determine the filtering lower limit dMaxD's of average noise upper limit filtering method in the step 41 by lower limit factor-beta
The method of determination is：DMaxD=dAvgN+ β；Wherein β value is the number more than 0, and dMaxD determined by its value simultaneously will
Seek the condition of satisfaction：dMaxD<dRef.
The step of determining data value group X maximum dMax by maximizing method be：
Step 51：It is 0 to make L, obtains the data value X for counting index L corresponding data value groups XL；By XLAssignment is in dMax, then
Into step 54；
Step 52：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether it is more than dMax, is to enter
Enter step 53, otherwise into step 54；
Step 53：By XLAssignment is in dMax, into step 54；
Step 54：Count index L and add 1；Into step 55；
Step 55：Judge whether L is less than N, be then return to step 52；Otherwise maximum dMax is obtained.
The present invention carries out bandlimiting filtering on the basis of cake resistancet, is made a distinction with different filtering process band dividing modes,
It includes filtering between limit value maximum, filters three kinds of filtering methods with filtering, the average noise upper limit on average noise.
The present invention is for the undesirable feelings of effect after sampled data FFT processing on the filter processing method of FFT data
Garbage signal is carried out under condition or is indifferent to filtering out for signal handling operation, in the case of precognition is selected, FFT treatment effects can be made
It is more preferably preferable, more optional condition filtering modes are provided for signal transacting, effectively filter out the garbage signal in known pre-selection band
Or it is indifferent to signal, purified treatment effect, preferable processing requirement is issued in high-speed data signal processing.
Especially for high speed acquisition signal analysis and processing high speed sampled data after FFT processing, because of high sampling bar
Hardware design and the limitation of real-time data analysis intractability under part, because big bandwidth problem, signal leakage problem, signal aliasing are asked
Topic, with interior self-excitation problem etc. so that effect is unsatisfactory after FFT, and some undesired signals or the not signal of relation can be
Occur in useful signal band, by improving hardware and in the case that optimization FPGA sequential has been difficult solution under high-speed case
Use, in the case of precognition is selected, help to filter and bottom purified treatment of making an uproar, improve FFT treatment effects, optimize data processing knot
Fruit.Due to by the way of software processing, with can burdening mode be further processed, using flexible, to a certain degree
On can also simplify requirement to hardware, save hardware resource cost, reduce hardware debug time, save human and material resources cost, carry
High benefit.
Although above-mentioned the embodiment of the present invention is described with reference to accompanying drawing, model not is protected to the present invention
The limitation enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme those skilled in the art are not
Need to pay various modifications or deformation that creative work can make still within protection scope of the present invention.
Claims (1)
1. a kind of grass implemented to FFT data limits filtering method, it is characterized in that, to big less than the apex zone of peak signal
Signal filtered between limit value maximum, band filtering on average noise is carried out to bottom zone small-signal above grass, to letter
Number band interlude region carries out upper noise limit filtering；
Filtered between limit value maximum, to be higher than in FFT signal datas to be dealt with reference value dRef's and less than under maximum
Drop signal in δ amplitude section and carry out analysis filtering, replace data value in the section to be handled with average noise；
Band filtering on average noise, to being multiplied factor θ setting multiple area higher than average noise in FFT signal datas to be dealt with
Signal in section carries out analysis filtering, replaces data value in the section to be handled with average noise；
The average noise upper limit filters, and is counted to again to being higher than the setting of average noise lower limit factor-beta in FFT signal datas to be dealt with
Analysis filtering is carried out less than the signal in reference value dRef sections, replaces data value in the section to be handled with average noise；
Counting for the given data value group X being made up of N number of FFT data, and reference value dRef, data value group X, which indexes, is
L, determines average noise baseline value dAvgN by being averaging noisy base line method first, and this number is determined by maximizing method
According to value group X maximum dMax；According to filtering process band it is different carry out filtering between limit value maximums, band filtering on average noise
Or average noise upper limit filtering；
The acquiring method of average noise baseline value includes,
Step 11：If the counting index that data value group X asks for average noise span is initiated with Lmin, count index and terminate as
Lmax；If evaluation summation is XT, count value LT；It is L to make Lmin, XTFor 0, LTFor 0；Subsequently into step 12；
Step 12：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether it is less than reference value dRef, is then
Into step 13, otherwise into step 14；
Step 13：Make XT=XT+XL, LT=LT+1；Into step 14；
Step 14：Count index L and add 1；Into step 15；
Step 15：Judge whether L is less than or equal to Lmax, it is then return to step 12；Otherwise step 16 is entered；
Step 16：UtilizeTry to achieve average noise baseline value dAvgN；
Counting index starting L described in the step 11minL is terminated with counting to indexmaxDetermination method be：
To round up；
The specific method filtered between limit value maximum includes,
Step 21：It is 0 to initialize L, is calculated by descending factors δ and higher limit dMaxU is filtered determined by maximum dMax, entered
Enter step 22；
Step 22：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether reference value dRef is more than or equal to,
And it is then to enter step 23, otherwise into step 24 less than filtering higher limit dMaxU；
Step 23：Replace indexing value X corresponding to L in X with dAvgNL；Into step 24；
Step 24：Count index L and add 1；Into step 25；
Step 25：Judge whether L is less than N, be then return to step 22；Otherwise whole handling process terminates；
The filtering higher limit dMaxU implemented in the step 21 by descending factors δ and maximum dMax determination method is：
DMaxU=dMax- | dMax | × δ；Wherein δ value is the number more than 0, and dMaxU determined by its value requires full simultaneously
Sufficient condition：dMaxU>dRef；
Specific method with filtering on average noise includes,
Step 31：It is 0 to initialize L, is multiplied factor θ by bandwidth and determines the filtering higher limit with filtering method on average noise
DMaxU2 and filtering lower limit dMaxD2, into step 32；
Step 32：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
DMaxD2, and be then to enter step 33, otherwise into step 34 less than filtering higher limit dMaxU2；
Step 33：Replace indexing value X corresponding to L in X with dAvgNL；Into step 34；
Step 34：Count index L and add 1；Into step 35；
Step 35：Judge whether L is less than N, be then return to step 32；Otherwise whole handling process terminates；
In the step 31 by bandwidth multiplied factor θ determine on average noise the filtering higher limit dMaxU2 with filtering method and
Filtering lower limit dMaxD2 determination method is：DMaxD2=dAvgN+ θ, dMaxU2=dAvgN+ θ × 2；
Wherein θ value is the number more than 0, and dMaxU2 determined by its value is required to meet condition simultaneously：dMaxU2<
dMax；
The specific method of average noise upper limit filtering includes,
Step 41：It is 0 to initialize L, and the filtering lower limit dMaxD of average noise upper limit filtering method is determined by lower limit factor-beta,
Into step 42；
Step 42：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether filtering lower limit is more than or equal to
DMaxD, and it is less than reference value dRef, it is then to enter step 43, otherwise into step 44；
Step 43：Replace indexing value X corresponding to L in X with dAvgNL；Into step 44；
Step 44：Count index L and add 1；Into step 45；
Step 45：Judge whether L is less than N, be then return to step 42；Otherwise whole handling process terminates；
The filtering lower limit dMaxD of average noise upper limit filtering method determination is determined in the step 41 by lower limit factor-beta
Method is：DMaxD=dAvgN+ β；Wherein β value is the number more than 0, and dMaxD determined by its value requires full simultaneously
Sufficient condition：dMaxD<dRef；
The step of determining data value group X maximum dMax by maximizing method be：
Step 51：It is 0 to make L, obtains the data value X for counting index L corresponding data value groups XL；By XLAssignment in dMax, subsequently into
Step 54；
Step 52：Obtain the data value X for counting index L corresponding data value groups XL, judge XLWhether it is more than dMax, is then to enter step
Rapid 53, otherwise into step 54；
Step 53：By XLAssignment is in dMax, into step 54；
Step 54：Count index L and add 1；Into step 55；
Step 55：Judge whether L is less than N, be then return to step 52；Otherwise maximum dMax is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610323875.4A CN105827218B (en) | 2016-05-17 | 2016-05-17 | A kind of grass implemented to FFT data limits filtering method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610323875.4A CN105827218B (en) | 2016-05-17 | 2016-05-17 | A kind of grass implemented to FFT data limits filtering method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105827218A CN105827218A (en) | 2016-08-03 |
| CN105827218B true CN105827218B (en) | 2018-04-06 |
Family
ID=56529731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610323875.4A Active CN105827218B (en) | 2016-05-17 | 2016-05-17 | A kind of grass implemented to FFT data limits filtering method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105827218B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6564184B1 (en) * | 1999-09-07 | 2003-05-13 | Telefonaktiebolaget Lm Ericsson (Publ) | Digital filter design method and apparatus |
| CN101005295A (en) * | 2007-01-29 | 2007-07-25 | 华为技术有限公司 | Signal processing method, signal processor and signal processing module |
| CN102984634A (en) * | 2011-11-22 | 2013-03-20 | 南京工程学院 | Digital hearing-aid unequal-width sub-band automatic gain control method |
| CN103236825A (en) * | 2013-03-22 | 2013-08-07 | 中国科学院光电技术研究所 | Data correcting method for high-precision data acquiring system |
-
2016
- 2016-05-17 CN CN201610323875.4A patent/CN105827218B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6564184B1 (en) * | 1999-09-07 | 2003-05-13 | Telefonaktiebolaget Lm Ericsson (Publ) | Digital filter design method and apparatus |
| CN101005295A (en) * | 2007-01-29 | 2007-07-25 | 华为技术有限公司 | Signal processing method, signal processor and signal processing module |
| CN102984634A (en) * | 2011-11-22 | 2013-03-20 | 南京工程学院 | Digital hearing-aid unequal-width sub-band automatic gain control method |
| CN103236825A (en) * | 2013-03-22 | 2013-08-07 | 中国科学院光电技术研究所 | Data correcting method for high-precision data acquiring system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105827218A (en) | 2016-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105411567B (en) | A kind of ECG data analysis method and system | |
| Benmalek et al. | Digital fractional order operators for R-wave detection in electrocardiogram signal | |
| CN102178522B (en) | Method for detecting and locating R wave in QRS (Quantum Resonance Spectrometer) waves of electrocardiographic signals of mother and fetus | |
| Afonso | Ecg qrs detection | |
| CN104939820B (en) | A kind of pace-making signal detection method and device | |
| KR20100008239A (en) | Eliminating method of motion artifact from ppg signal | |
| CN100589759C (en) | Saturation oxygen detection method | |
| CN106264505A (en) | A kind of heart rate spectral peak system of selection based on support vector machine | |
| WO2012087702A1 (en) | Adaptive ecg wandering correction | |
| CN105509763A (en) | Step-counting interference removal method and device | |
| CN105827218B (en) | A kind of grass implemented to FFT data limits filtering method | |
| CN109009087B (en) | Rapid detection method for electrocardiosignal R wave | |
| JPH0322770B2 (en) | ||
| CN105030232A (en) | Baseline drift correction method for electrocardiosignal | |
| CN107361763B (en) | Electrocardiogram data R wave detection method and device | |
| CN110505839A (en) | For handling the method and system of EMG signal | |
| DE19805530C1 (en) | Device for improving signals in an electrocardiogram (ECG) subject to frequent artefacts | |
| CN201631188U (en) | Multichannel biomedical electric signal preprocessing module | |
| CN107569227A (en) | The processing method and monitoring device of heart rate under a kind of motion state | |
| WO1992015242A1 (en) | Method and device for filtering out baseline fluctuations from an electrocardiogram | |
| CN108694355B (en) | Signal processing method and device | |
| CN106026973B (en) | The average noise noise-reduction method that a kind of pair of FFT data is implemented | |
| CN104617914A (en) | Electrical equipment signal filtering method and system | |
| CN105915198B (en) | A kind of analog-digital hybrid filtering method | |
| CN103438983A (en) | Data processing method of signal random average spectrums |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |