CN106228551A - The adaptive decomposition method of wave filter is produced based on image Segmentation Technology - Google Patents

The adaptive decomposition method of wave filter is produced based on image Segmentation Technology Download PDF

Info

Publication number
CN106228551A
CN106228551A CN201610569963.2A CN201610569963A CN106228551A CN 106228551 A CN106228551 A CN 106228551A CN 201610569963 A CN201610569963 A CN 201610569963A CN 106228551 A CN106228551 A CN 106228551A
Authority
CN
China
Prior art keywords
time
frequency
group
image segmentation
wave filter
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
Application number
CN201610569963.2A
Other languages
Chinese (zh)
Other versions
CN106228551B (en
Inventor
阎绍泽
刘涛
果晓东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN201610569963.2A priority Critical patent/CN106228551B/en
Publication of CN106228551A publication Critical patent/CN106228551A/en
Application granted granted Critical
Publication of CN106228551B publication Critical patent/CN106228551B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20048—Transform domain processing
    • G06T2207/20056—Discrete and fast Fourier transform, [DFT, FFT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20048—Transform domain processing
    • G06T2207/20064—Wavelet transform [DWT]

Landscapes

  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

The present invention proposes a kind of adaptive decomposition method producing wave filter based on image Segmentation Technology, comprises the following steps: the first step, signal to be analyzed is carried out time-frequency conversion, obtains time-frequency coefficients and the time-frequency plane of correspondence；Second step, carries out Threshold segmentation to time-frequency plane, to obtain binary picture；3rd step, the connected domain in labelling binary picture；4th step, generates one group of Time frequency Filter according to connected domain, and is filtered the time frequency system according to Time frequency Filter, exports filter result；5th step, carries out time-frequency inverse transformation to filter result, to obtain decomposition result.The method of the present invention is when signal to be decomposed each frequency content sampling time long enough, can possess optional frequency resolution, compared with conventional empirical mode decomposition algorithm, substantially increase frequency resolution, the signal of closely spaced frequencies distribution can be carried out adaptive decomposition, and the instantaneous frequency that calculates each composition with this and instantaneous amplitude, thus widen the range of application of adaptive decomposition algorithm.

Description

The adaptive decomposition method of wave filter is produced based on image Segmentation Technology
Technical field
The present invention relates to signal processing technology field, particularly to a kind of based on image Segmentation Technology produce wave filter from Adapt to decomposition method.
Background technology
At present, adaptive decomposition algorithm obtains application widely in signal processing field.Its Typical Representative Empirical Mode State decomposes the issue paper of (Empirical Mode Decomposition, EMD) the most have been had super in Google's science Cross 12000 times quote, it can be seen that one spot.Trace it to its cause, be primarily due to EMD method and combine with Hilbert conversion, it is provided that A kind of each frequency content instantaneous frequency of signal calculated easily and the approach of instantaneous amplitude.Normal as quantitative assessment signal characteristic Parameter, instantaneous frequency and instantaneous amplitude are widely used.Instantaneous frequency is examined at such as radar, sonar, Mechanical System Trouble Disconnected engineering field is extensively applied.Same instantaneous amplitude is applied to modal parameter and determines and Non-Destructive Testing detection field.
The current method calculating instantaneous frequency and instantaneous amplitude can be divided mainly into two classes: a class is based on inner product meter The time-frequency conversion method calculated, such as, Short Time Fourier Transform, wavelet transformation, framework decomposition etc.；Another kind of is that combining adaptive divides (such as, EMD, local mean value decomposes (Local Mean Decompositio, LMD) to resolving Algorithm and experience wavelet transformation decomposes (Empirical Wavelet transform, EWT)) and instantaneous algorithm (such as, the Hilbert of instantaneous frequency and instantaneous amplitude Conversion and energy operator method).
Owing to decomposing the multiformity of waveform, the former different components are easier to the corresponding region accumulating in time-frequency plane, It is easily separated.But carry out between certain section of signal to be analyzed and decomposition waveform owing to inner product calculates, so this is not a kind of Real instantaneous computational methods, result of calculation can not accurately reflect the transient change of signal frequency and amplitude.For the latter, it is excellent Point is that calculating process includes two stages: signal decomposition is different independent components by the first stage, such as, in EMD side Intrinsic mode functions (Intrinsic Mode Function, IMF) in method；Second stage uses Transient Technique to calculate each point The instantaneous frequency of amount and instantaneous amplitude, such as Hilbert conversion and energy operator method.So using the method more accurately to calculate The instantaneous frequency of signal and instantaneous amplitude.But the frequency resolution of the method is not ideal enough.So far, for LMD and EWT method still It is not related to the report of frequency discrimination.For classical adaptive decomposition algorithm EMD method, document has been had to report, but when two (the f when frequency ratio of individual component is more than 0.75Low/fHigh), both just can not be separated by the method, thus limiting adaptive decomposes calculation The range of application of method.
Summary of the invention
It is contemplated that at least solve one of above-mentioned technical problem.
To this end, it is an object of the invention to propose a kind of adaptive decomposition side producing wave filter based on image Segmentation Technology Method, the method, when signal to be decomposed each frequency content sampling time long enough, can possess optional frequency resolution, with conventional Empirical mode decomposition algorithm compare, substantially increase frequency resolution, can to closely spaced frequencies distribution signal carry out adaptive Should decompose, and the instantaneous frequency that calculates each composition with this and instantaneous amplitude, thus widen the application model of adaptive decomposition algorithm Enclose.
To achieve these goals, embodiments of the invention propose a kind of based on image Segmentation Technology generation wave filter Adaptive decomposition method, comprises the following steps: S1: signal to be analyzed is carried out time-frequency conversion, obtain correspondence time-frequency coefficients and Time-frequency plane；S2: described time-frequency plane is carried out Threshold segmentation, to obtain binary picture；S3: in binary picture described in labelling Connected domain；S4: generate one group of Time frequency Filter according to described connected domain, and according to described Time frequency Filter to described the time frequency system It is filtered, exports filter result；And S5: described filter result is carried out time-frequency inverse transformation, to obtain decomposition result.
It addition, the adaptive decomposition method producing wave filter based on image Segmentation Technology according to the above embodiment of the present invention Can also have a following additional technical characteristic:
In some instances, in described S1, the mode of described time-frequency conversion includes: Short Time Fourier Transform or small echo become Change.
In some instances, the transformation for mula of described Short Time Fourier Transform is:
STFT w ( s ) [ m , n ] = Σ k = 0 K - 1 s [ k ] w [ k - l ] e - j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is to become the time of time-frequency plane to point Amount, n is frequency variable, and s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-l] is window function, and | | w | |=1.
In some instances, in described S2, described in obtain the transformation for mula of binary picture and be:
c ^ ( m , n ) = 1 c ( m , n ) > T 0 0 c ( m , n ) ≤ T 0 ,
Wherein,For described binary picture, (m n) is described time-frequency coefficients, T to c0For segmentation threshold.
In some instances, described S3 farther includes: S31: progressively scan described binary picture, by described binary picture Every a line in continuous print value be the sequence of pixel composition of 1 as a group, and write down starting point and the end of described group Point and the line number at described place；S32: be scanned the group in all row in addition to the first row, if currently rolled into a ball with front All groups in a line all do not have an overlapping region, then give one new label of described current group, if current group only with lastrow In a group have overlapping region, then be marked with the label of described current group, if more than 2 of current group and lastrow There is overlapping region in group, then give the minimum label in group of UNICOM to described current group, and by the labelling of the several groups in lastrow Write is of equal value right；S33: by described equivalence to being converted to equivalent sequence, and give an identical label to each equivalent sequence； S34: traversal starts the labelling of group, searches equivalent sequence, and gives the labelling that described group is new；S35: the label of each group is inserted In described binary image.
In some instances, in described S4, described filter result is:
C ^ i - F i ( C ) ,
Wherein,(i=1,2,3 ..., P) it is described filter result, P is Time frequency Filter number, FiFor i-th time-frequency Wave filter, C is the matrix of described time-frequency coefficients.
In some instances, in described S5, the transformation for mula of described time-frequency inverse transformation is:
s [ k ] = Σ m = 1 M Σ n = 1 N STFT w ( s ) [ m , n ] w [ k - M ] e j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is to become the time of time-frequency plane to point Amount, n is frequency variable, and s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-M] is window function, and | | w | |=1.
The adaptive decomposition method producing wave filter based on image Segmentation Technology according to embodiments of the present invention, in conjunction with time-frequency Conversion and inverse transformation method, the method that divides the image into introduces, and is used for constructing Time frequency Filter, realizes the decomposition of signal component, And when signal to be decomposed each frequency content sampling time long enough, can possess optional frequency resolution, with conventional experience Mode decomposition algorithm is compared, and substantially increases frequency resolution, the signal of closely spaced frequencies distribution can be carried out adaptive decomposition, And the instantaneous frequency that calculates each composition with this and instantaneous amplitude, thus widen the range of application of adaptive decomposition algorithm.
The additional aspect of the present invention and advantage will part be given in the following description, and part will become from the following description Obtain substantially, or recognized by the practice of the present invention.
Accompanying drawing explanation
Above-mentioned and/or the additional aspect of the present invention and advantage are from combining the accompanying drawings below description to embodiment and will become Substantially with easy to understand, wherein:
Fig. 1 is the stream of the adaptive decomposition method producing wave filter based on image Segmentation Technology according to embodiments of the present invention Cheng Tu；
Fig. 2 is the Fourier transformation time-frequency figure of the example signal according to one specific embodiment of the present invention；
Fig. 3 is the image of the binary picture obtained after the Threshold segmentation according to one specific embodiment of the present invention；And
Fig. 4 is the adaptive decomposition producing wave filter based on image Segmentation Technology according to one specific embodiment of the present invention Method converts, after decomposing, the instantaneous frequency and the schematic diagram of instantaneous amplitude obtained based on Hilbert.
Detailed description of the invention
Embodiments of the invention are described below in detail, and the example of described embodiment is shown in the drawings, the most from start to finish Same or similar label represents same or similar element or has the element of same or like function.Below with reference to attached The embodiment that figure describes is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.
In describing the invention, it is to be understood that term " " center ", " longitudinally ", " laterally ", " on ", D score, Orientation or the position relationship of the instruction such as "front", "rear", "left", "right", " vertically ", " level ", " top ", " end ", " interior ", " outward " are Based on orientation shown in the drawings or position relationship, it is for only for ease of the description present invention and simplifies description rather than instruction or dark The device or the element that show indication must have specific orientation, with specific azimuth configuration and operation, therefore it is not intended that right The restriction of the present invention.Additionally, term " first ", " second " are only used for describing purpose, and it is not intended that instruction or hint relatively Importance.
In describing the invention, it should be noted that unless otherwise clearly defined and limited, term " is installed ", " phase Even ", " connection " should be interpreted broadly, for example, it may be fixing connection, it is also possible to be to removably connect, or be integrally connected；Can To be mechanical connection, it is also possible to be electrical connection；Can be to be joined directly together, it is also possible to be indirectly connected to by intermediary, Ke Yishi The connection of two element internals.For the ordinary skill in the art, can understand that above-mentioned term is at this with concrete condition Concrete meaning in invention.
Below in conjunction with accompanying drawing, the self adaptation producing wave filter based on image Segmentation Technology according to embodiments of the present invention is described Decomposition method.
Fig. 1 is the adaptive decomposition method producing wave filter according to an embodiment of the invention based on image Segmentation Technology Flow chart.As it is shown in figure 1, the method comprises the following steps:
Step S1: signal to be analyzed is carried out time-frequency conversion, obtains time-frequency coefficients and the time-frequency plane of correspondence.
Wherein, the mode of time-frequency conversion for example, Short Time Fourier Transform or wavelet transformation.In the present embodiment used time Frequently alternative approach is Short Time Fourier Transform, and its transformation for mula is:
STFT w ( s ) [ m , n ] = Σ k = 0 K - 1 s [ k ] w [ k - l ] e - j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is to become the time of time-frequency plane to point Amount, n is frequency variable, and s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-l] is window function, and | | w | |=1.
Certainly, the time-frequency conversion method of the present invention is not limited only to this, can be any transformation results for two dimension time-frequency plane Time-frequency conversion algorithm, be the most only the description for the purpose of exemplary.
Step S2: time-frequency plane is carried out Threshold segmentation, to obtain binary picture.
In one embodiment of the invention, the expression formula of binary picture is for example:
c ^ ( m , n ) = 1 c ( m , n ) > T 0 0 c ( m , n ) ≤ T 0 ,
Wherein,For binary picture, (m n) is time-frequency coefficients, T to c0For segmentation threshold.
Step S3: the connected domain in labelling binary picture.
Specifically, step S3 farther includes:
S31: progressive scan binary picture, that the pixel that continuous print value in every a line of binary picture is 1 is formed Individual sequence is as a group, and the beginning and end of the group of writing down and the line number at group place.
S32: be scanned the group in all row in addition to the first row, if current group is with all groups in previous row all There is no overlapping region, then give current group one new label, if only there is overlapping region in current group, then with a group in lastrow It is marked with the label of current group, if there is overlapping region in the group that current group is with more than 2 of lastrow, then gives and currently roll into a ball tax Give the minimum label in group of UNICOM, and it is right that the labelling of the several groups in lastrow is write equivalence.
S33: by equivalence to being converted to equivalent sequence, and give an identical label to each equivalent sequence.
S34: traversal starts the labelling of group, searches equivalent sequence, and the labelling that the group of giving is new.
S35: the label of each group is inserted in binary image.
Step S4: generate one group of Time frequency Filter according to connected domain, and according to Time frequency Filter, the time frequency system is filtered Ripple, exports filter result.
In one embodiment of the invention, filter result is for example:
C ^ i = F i ( C ) ,
Wherein,(i=1,2,3 ..., P) it is filter result, namely time-frequency coefficients matrix after filtering, P is Time frequency Filter Number, FiFor i-th Time frequency Filter, C is the matrix of time-frequency coefficients.
Step S5: filter result is carried out time-frequency inverse transformation, to obtain decomposition result.
In one embodiment of the invention, the transformation for mula that filter result carries out time-frequency inverse transformation is:
s [ k ] = Σ m = 1 M Σ n = 1 N STFT w ( s ) [ m , n ] w [ k - M ] e j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is to become the time of time-frequency plane to point Amount, n is frequency variable, and s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-M] is window function, and | | w | |=1.
For the ease of being more fully understood that the self adaptation based on image Segmentation Technology generation wave filter of the embodiment of the present invention is divided Solution method, specifically describes the method for the present invention further below in conjunction with accompanying drawing and specific embodiment.
In the present embodiment, the method realizes under matlab software platform.With a concrete example signal it is The method is illustrated by example.The expression formula of this example signal is:
S (t)=s1(t)+s2(t)+s3(t),0≤t≤1
s1=(t-t2)sin(1000·2π·t)
s2=(t-t2)sin(750·2π·t)
s3=(t-t2) sin (400 2 π t+200sin (2 π t)),
This example signal i.e. includes 3 components, its sample frequency fsFor 100kHz.
Based on this, in the present embodiment, the method comprises the following steps:
Step 1: example signal is carried out time-frequency conversion.This example use the most common Short Time Fourier Transform to showing Example signal carries out time-frequency conversion, and concrete transformation for mula is:
STFT w ( s ) [ m , n ] = Σ k = 0 K - 1 s [ k ] w [ k - l ] e - j 2 π n ( k - m ) / K ,
Wherein, k represents the time, m (m=0,1,2 ..., M-1) time of representing in time-frequency conversion moves, n (n=0,1, 2 ..., N-1) represent the frequency change in time-frequency conversion.Further, transformation results (i.e. time-frequency coefficients) c (m, n) example obtained As shown in Figure 2.As shown in Figure 2, the time-frequency coefficients that 3 different frequency compositions are corresponding accumulates in not same district in time-frequency plane respectively Territory.
Step 2: the time-frequency plane obtained in Threshold segmentation step 1 produces binary picture.Specific formula for calculation is:
c ^ ( m , n ) = 1 c ( m , n ) > T 0 0 c ( m , n ) ≤ T 0 ,
Wherein,For the pixel number of the binary image after calculating, T0For segmentation threshold.Herein, segmentation threshold Be set as 0.01, then transformation results is as it is shown on figure 3, from the figure 3, it may be seen that after Threshold segmentation, define 3 independent connected domains, Wherein contain 3 time-frequency coefficients corresponding to frequency content.
Step 3: obtaining the connected domain in binary picture in markers step 2,3 connected regions are by labelling respectively, it is simple to journey Sequence is automatically performed catabolic process.Concrete labelling connected domain algorithmic procedure is as follows:
(1) binary picture obtained in progressive scan step 2, the pixel composition that continuous print value in every a line is 1 One sequence is referred to as a group, and writes down its beginning and end and the line number at its place.
(2) group in all row in addition to the first row is scanned, if it is with all groups in previous row all There is no overlapping region, then give its new label；If only having overlapping region with a group in lastrow, then by with this group Label is marked；If there is overlapping region in group of its with lastrow more than 2, then give during UNICOM rolls into a ball to current group most Little label, and by of equal value right for the labelling write of these groups of lastrow.
(3) by equivalence to being converted to equivalent sequence, each sequence gives an identical label.Give each equivalent sequence One label, from the beginning of 1.
(4) traversal starts the labelling of group, searches equivalent sequence, gives the labelling that they are new.
(5) label of each group is inserted in labelling image.
Step 4: produce one group of Time frequency Filter according to connected domain, and the time-frequency coefficients obtained in step 1 is filtered, Concrete Filtering Formula is:
C ^ i = F i ( C ) ,
Wherein,(i=1,2,3 ..., P, P are Time frequency Filter number) it is filter result, namely time-frequency coefficients after filtering Matrix, FiFor i-th Time frequency Filter, C is the time-frequency coefficients matrix obtained in step 1.
In the present embodiment, create 3 groups of wave filter, after filtering, the most just obtain 3 groups of different time-frequency systems Number.
Step 5: the filter result obtained in step 4 is carried out the time-frequency inverse transformation corresponding with step 1, to be decomposed As a result, concrete computing formula is:
s [ k ] = Σ m = 1 M Σ n = 1 N STFT w ( s ) [ m , n ] w [ k - M ] e j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is to become the time of time-frequency plane to point Amount, n is frequency variable, and s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-M] is window function, and | | w | |=1.
Further, after obtaining decomposition result, instantaneous frequency based on Hilbert transformation calculations decomposition result and wink Time frequency, its result is as shown in Figure 4.From Fig. 4 (a), for instantaneous frequency, except endpoint location, value of calculation and setting value It coincide；From Fig. 4 (b), for instantaneous amplitude, together with value of calculation essentially coincides with setting value curve.Instantaneous frequency and wink Time amplitude result of calculation indicate this algorithm and effectively 3 compositions separated.
To sum up, the adaptive decomposition method producing wave filter based on image Segmentation Technology according to embodiments of the present invention, knot Closing time-frequency conversion and inverse transformation method, the method that divides the image into introduces, and is used for constructing Time frequency Filter, realizes signal component Decompose, and when signal to be decomposed each frequency content sampling time long enough, can possess optional frequency resolution, with conventional Empirical mode decomposition algorithm is compared, and substantially increases frequency resolution, the signal of closely spaced frequencies distribution can be carried out self adaptation Decompose, and the instantaneous frequency that calculates each composition with this and instantaneous amplitude, thus widen the range of application of adaptive decomposition algorithm.
In the description of this specification, reference term " embodiment ", " some embodiments ", " example ", " specifically show Example " or the description of " some examples " etc. means to combine this embodiment or example describes specific features, structure, material or spy Point is contained at least one embodiment or the example of the present invention.In this manual, to the schematic representation of above-mentioned term not Necessarily refer to identical embodiment or example.And, the specific features of description, structure, material or feature can be any One or more embodiments or example in combine in an appropriate manner.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that: not These embodiments can be carried out multiple change in the case of departing from the principle of the present invention and objective, revise, replace and modification, this The scope of invention is limited by claim and equivalent thereof.

Claims (7)

1. the adaptive decomposition method producing wave filter based on image Segmentation Technology, it is characterised in that comprise the following steps:
S1: signal to be analyzed is carried out time-frequency conversion, obtains time-frequency coefficients and the time-frequency plane of correspondence；
S2: described time-frequency plane is carried out Threshold segmentation, to obtain binary picture；
S3: the connected domain in binary picture described in labelling；
S4: generate one group of Time frequency Filter according to described connected domain, and according to described Time frequency Filter, described the time frequency system is entered Row filtering, exports filter result；And
S5: described filter result is carried out time-frequency inverse transformation, to obtain decomposition result.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 1, its feature exists In, in described S1, the mode of described time-frequency conversion includes: Short Time Fourier Transform or wavelet transformation.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 2, its feature exists In, the transformation for mula of described Short Time Fourier Transform is:
STFT w ( s ) [ m , n ] = Σ k = 0 K - 1 s [ k ] w [ k - l ] e - j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is the time variable of time-frequency plane to point, n For frequency variable, s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-l] For window function, and | | w | |=1.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 1, its feature exists In, in described S2, described in obtain the transformation for mula of binary picture and be:
c ^ ( m , n ) = 1 c ( m , n ) > T 0 0 c ( m , n ) ≤ T 0 ,
Wherein,For described binary picture, (m n) is described time-frequency coefficients, T to c0For segmentation threshold.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 1, its feature exists In, described S3 farther includes:
S31: progressively scan described binary picture, by the pixel composition that continuous print value in every a line of described binary picture is 1 A sequence as a group, and write down beginning and end and the line number at described place of described group；
S32: be scanned the group in all row in addition to the first row, if current group does not all have with all groups in previous row Overlapping region, then give one new label of described current group, if only there is overlapping region in current group, then with a group in lastrow It is marked with the label of described current group, if there is overlapping region in the group that current group is with more than 2 of lastrow, then gives described Current group gives the minimum label in group of UNICOM, and by of equal value right for the labelling write of the several groups in lastrow；
S33: by described equivalence to being converted to equivalent sequence, and give an identical label to each equivalent sequence；
S34: traversal starts the labelling of group, searches equivalent sequence, and gives the labelling that described group is new；
S35: the label of each group is inserted in described binary image.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 1, its feature exists In, in described S4, described filter result is:
C ^ i = F i ( C ) ,
Wherein,For described filter result, P is Time frequency Filter number, FiFor i-th Time frequency Filter, C is the matrix of described time-frequency coefficients.
The adaptive decomposition method producing wave filter based on image Segmentation Technology the most according to claim 3, its feature exists In, in described S5, the transformation for mula of described time-frequency inverse transformation is:
s [ k ] = Σ m = 1 M Σ n = 1 N STFT w ( s ) [ m , n ] w [ k - M ] e j 2 π n ( k - m ) / K ,
Wherein, STFTwS () [m, n] is that (m, n) the Short Time Fourier Transform coefficient at place, m is the time variable of time-frequency plane to point, n For frequency variable, s [k] is discrete signal to be analyzed, and k is the time variable of sample point, and K is sampling time length, w [k-M] For window function, and | | w | |=1.
CN201610569963.2A 2016-07-19 2016-07-19 The adaptive decomposition method of filter is generated based on image Segmentation Technology Expired - Fee Related CN106228551B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610569963.2A CN106228551B (en) 2016-07-19 2016-07-19 The adaptive decomposition method of filter is generated based on image Segmentation Technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610569963.2A CN106228551B (en) 2016-07-19 2016-07-19 The adaptive decomposition method of filter is generated based on image Segmentation Technology

Publications (2)

Publication Number Publication Date
CN106228551A true CN106228551A (en) 2016-12-14
CN106228551B CN106228551B (en) 2019-04-23

Family

ID=57532530

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610569963.2A Expired - Fee Related CN106228551B (en) 2016-07-19 2016-07-19 The adaptive decomposition method of filter is generated based on image Segmentation Technology

Country Status (1)

Country Link
CN (1) CN106228551B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109470647A (en) * 2019-01-20 2019-03-15 南京林业大学 A kind of measurement method of vapor Terahertz absorption spectra
CN109741330A (en) * 2018-12-21 2019-05-10 东华大学 A kind of medical image cutting method of mixed filtering strategy and fuzzy C-mean algorithm
CN109800777A (en) * 2018-09-21 2019-05-24 上海营阅企业管理服务中心(有限合伙) A kind of urine test paper physical signs automatic identifying method
CN111862007A (en) * 2020-07-02 2020-10-30 哈尔滨市科佳通用机电股份有限公司 Freight car brake cylinder inclination detection method based on variance region segmentation
CN114759951A (en) * 2022-06-15 2022-07-15 成都中星世通电子科技有限公司 Frequency hopping signal real-time blind detection method, parameter estimation method, system and terminal

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103995289A (en) * 2014-05-19 2014-08-20 中国石油大学(华东) Time-varying mixed-phase seismic wavelet extraction method based on time-frequency spectrum simulation
CN105678781A (en) * 2016-01-19 2016-06-15 中国人民解放军电子工程学院 Object micro Doppler feature separation and extraction method based on edge detection

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103995289A (en) * 2014-05-19 2014-08-20 中国石油大学(华东) Time-varying mixed-phase seismic wavelet extraction method based on time-frequency spectrum simulation
CN105678781A (en) * 2016-01-19 2016-06-15 中国人民解放军电子工程学院 Object micro Doppler feature separation and extraction method based on edge detection

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
SOUNDARARAJAN SRINIVASAN等: "Binary and ratio time-frequency masks for robust speech recognition", 《SPEECH COMMUNICATION》 *
刘奇琦等: "一种二值图像连通区域标记的新方法", 《计算机工程与应用》 *
刘歌等: "基于时频图像处理的多分量ＬＦＭ信号分离", 《航天电子对抗》 *
吴正国等: "《高等数字信号处理》", 30 April 2009 *
王雪飞等: "基于时频子空间的多分量宽带调频信号DOA 估计", 《电讯技术》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109800777A (en) * 2018-09-21 2019-05-24 上海营阅企业管理服务中心(有限合伙) A kind of urine test paper physical signs automatic identifying method
CN109741330A (en) * 2018-12-21 2019-05-10 东华大学 A kind of medical image cutting method of mixed filtering strategy and fuzzy C-mean algorithm
CN109470647A (en) * 2019-01-20 2019-03-15 南京林业大学 A kind of measurement method of vapor Terahertz absorption spectra
CN111862007A (en) * 2020-07-02 2020-10-30 哈尔滨市科佳通用机电股份有限公司 Freight car brake cylinder inclination detection method based on variance region segmentation
CN111862007B (en) * 2020-07-02 2021-01-08 哈尔滨市科佳通用机电股份有限公司 Freight car brake cylinder inclination detection method based on variance region segmentation
CN114759951A (en) * 2022-06-15 2022-07-15 成都中星世通电子科技有限公司 Frequency hopping signal real-time blind detection method, parameter estimation method, system and terminal
CN114759951B (en) * 2022-06-15 2022-09-02 成都中星世通电子科技有限公司 Frequency hopping signal real-time blind detection method, parameter estimation method, system and terminal

Also Published As

Publication number Publication date
CN106228551B (en) 2019-04-23

Similar Documents

Publication Publication Date Title
CN106228551A (en) The adaptive decomposition method of wave filter is produced based on image Segmentation Technology
Wang et al. Complete ensemble local mean decomposition with adaptive noise and its application to fault diagnosis for rolling bearings
Li et al. Diagonal slice spectrum assisted optimal scale morphological filter for rolling element bearing fault diagnosis
Liu et al. Bearing faults diagnostics based on hybrid LS-SVM and EMD method
Gao et al. Spare optimistic based on improved ADMM and the minimum entropy de-convolution for the early weak fault diagnosis of bearings in marine systems
Miaofen et al. Adaptive synchronous demodulation transform with application to analyzing multicomponent signals for machinery fault diagnostics
CN112446323B (en) HHT harmonic analysis method based on improved EMD modal aliasing and endpoint effect
Han et al. Second-order synchroextracting wavelet transform for nonstationary signal analysis of rotating machinery
Wang et al. An improved local characteristic-scale decomposition to restrict end effects, mode mixing and its application to extract incipient bearing fault signal
Cheng et al. Adaptive sparsest narrow-band decomposition method and its applications to rolling element bearing fault diagnosis
He et al. GMC sparse enhancement diagnostic method based on the tunable Q-factor wavelet transform for detecting faults in rotating machines
Yang et al. Study on ways to restrain end effect of Hilbert-Huang transform
Shah et al. Fault identification for IC engines using artificial neural network
CN108061821A (en) A kind of improved dual feedback wind power generation system harmonic detecting method
Yu et al. A non-linear time–frequency tool for machinery fault diagnosis under varying speed condition
CN108761202B (en) Harmonic detection method combining pole symmetric modal decomposition and Hilbert transform
Lv et al. Longitudinal synchroextracting transform: A useful tool for characterizing signals with strong frequency modulation and application to machine fault diagnosis
CN103455470A (en) Time frequency decomposition method for signals including intersections in instantaneous frequencies
CN108680782B (en) Voltage flicker parameter detection method based on extreme point symmetric mode decomposition
Hafiz et al. An approach for classification of power quality disturbances based on Hilbert Huang transform and Relevance vector machine
Li et al. Power quality disturbance detection based on IEWT
Wang et al. Single-phase grounding fault line selection method based on dynamic time warping algorithm
CN105156901A (en) Optical fiber early warning system noise reduction method and device based on wavelet analysis
Wang et al. An underdetermined environmental sound source separation algorithm based on improved complete ensemble EMD with adaptive noise and ICA
Goh et al. An integrated approach improved fast S-transform and SVD noise reduction for classification of power quality disruptions in noisy environments

Legal Events

Date Code Title Description
C06 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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20190423

Termination date: 20190719

CF01 Termination of patent right due to non-payment of annual fee