CN107179310A - Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation - Google Patents
Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation Download PDFInfo
- Publication number
- CN107179310A CN107179310A CN201710403411.9A CN201710403411A CN107179310A CN 107179310 A CN107179310 A CN 107179310A CN 201710403411 A CN201710403411 A CN 201710403411A CN 107179310 A CN107179310 A CN 107179310A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msup
- mfrac
- raman spectrum
- noise
- 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
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
Landscapes
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
The invention discloses a kind of Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation, step is as follows：(1) it is reference by the Gaussian noise that 0, variance is 1 of average, it is determined that with reference to hundredths and its percentile；Difference normalization is carried out to Raman spectrum data, data after normalization are sorted from small to large, calculate the hundredths of each data, then the percentile with reference to corresponding to hundredths is tried to achieve by linear interpolation, and by it with being divided by with reference to percentile, series of noise standard deviation is obtained, the median of standard deviation is taken as the noise estimated standard deviation σ of the spectroscopic data；(2) peak value and valley of Raman spectrum data are asked for, each peak value and the minimum valley of its left and right sides are compared, if greater than r times of noise criteria difference σ, then it is assumed that be the characteristic peak of Raman spectrum.This method need not go background process to Raman spectrum in advance, and without artificially setting any parameter, it is possible to achieve the automation of spectral peak identification.
Description
Technical field
The present invention relates to a kind of recognition methods of spectral signature peak, more particularly to a kind of drawing based on robust noise variance evaluation
The recognition methods of graceful spectral signature peak.
Background technology
Raman spectrum analysis due to have the advantages that it is quick, simple, repeatable, sample is not destroyed, in agricultural, medical science, food
The fields such as product, petrochemical industry, which are obtained for, to be widely applied.Raman spectrum analysis includes pretreatment, feature extraction and the spy of spectrum
The steps such as classification are levied, wherein feature extraction is the core link in Raman spectrum analysis system design.Due to Raman line displacement
Frequency with incident light is unrelated, only relevant with the vibration-rotation energy level of molecule, therefore each material has corresponding Raman light
Can the distribution of spectrum signature peak, the position of Raman spectrum characteristic peak and size directly reflect the structure and content information of material, very
Differentiate that spectral peak will directly influence the accuracy of sample characteristic classification well.
Conventional spectral peak recognition methods has amplitude method, continuous wavelet transform method, derivative method etc. at present.Amplitude method passes through setting
One threshold value, using first than the big point of threshold value as the starting point of spectral peak, and follow-up maximum point then regards as the high point of spectral peak,
Then using point ensuing first smaller than threshold value as the terminal of spectral peak, this Method And Principle is fairly simple, calculating speed
Hurry up, but be vulnerable to the influence of baseline drift, and compare the selection of threshold value influences larger to spectral peak accuracy in detection；Continuous wavelet
Converter technique seeks the crestal line that time domain peak-seeking is changed into matrix of wavelet coefficients a series of superposition that signal decomposition is wavelet functions
Peak, this method peak-seeking accuracy rate is higher, and has stronger rejection ability for noise and background, but amount of calculation is larger, no
It is adapted to real-time operation, and it needs to be determined that ridge line length threshold value and crestal line snr threshold, but ridge line length and wavelet scale
Selection is closely related, and the meaning of crestal line signal to noise ratio is nor very clearly, this causes this method not sane enough, is difficult to use；Lead
The basic thought of number method is that spectral line is regarded as a continuous curve, by being differentiated to each point on spectral line, then foundation
The property of derivative determines the position of spectral peak, and this method has higher searching accuracy, and calculating speed base for smooth curve
Originally real-time requirement can be met, but false peak is also easy to produce, it is necessary to set threshold parameter mistake for the larger complicated spectral line of noise
False peak is filtered, the selection of threshold value influences larger to analysis result.Local noise at the definition of foundation local SNR, spectral peak
It is poorer than the noise criteria that lower limit should be 6 times.But actual spectroscopic data also exists in addition to comprising characteristic peak, baseline noise
Baseline drift, how therefrom to estimate noise criteria difference is a good problem to study.
Traditional noise estimation method is manual or semi-manual mostly, it is necessary to which advance find out one section from spectrum and do not wrap
Data containing characteristic peak, outlier and obvious baseline tilt, then estimate that noise criteria is poor.If the spectroscopic data gathered every time
All handle in this way, due to the influence of artificial subjective factor, estimated result can be caused unreliable, and be unfavorable for Raman light
The automation mechanized operation of spectrometer.
The content of the invention
In order to overcome the shortcomings of that existing Raman spectrum characteristic peak searching algorithm needs to set threshold value manually, the present invention is proposed
A kind of Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation, this method need not be done to Raman spectrum in advance
Background process is gone, and without artificially setting any parameter, it is possible to achieve the automation of spectral peak identification.
The technical proposal for solving the technical problem of the invention is as follows：A kind of Raman based on robust noise variance evaluation
The recognition methods of spectral signature peak, comprises the following steps：
(1) it is reference by the Gaussian noise that 0, variance is 1 of average, it is determined that with reference to hundredths and its percentile；To drawing
Graceful spectroscopic data carries out forward difference computing and normalized, and by the data after normalization according to sorting from small to large, calculates each number
According to hundredths, then tried to achieve by linear interpolation with reference to the percentile corresponding to hundredths, and by it with referring to hundredths
Number is divided by, and obtains series of noise standard deviation, takes the median of standard deviation as the noise estimated standard deviation σ of the spectroscopic data；
(2) its peak value and valley are obtained by carrying out first derivation to Raman spectrum data, by each peak value and its left side
The minimum valley of right both sides is compared, if greater than the noise criteria difference σ of r times (generally taking r >=6), then it is assumed that it is Raman
The characteristic peak of spectrum.
Further, the step (1) is specially：
(1.1) it is 0 to assume that stochastic variable ε obeys average, and variance is σ2Gaussian Profile, then its probability density function be：
Known hundredths p (x) refers to that stochastic variable ε is less than percentile x cumulative probability, i.e.,：
According to the definition of error function：
It can obtain：
It can further obtain：
Using average as 0, the Gaussian noise that variance is 1 is reference, and m point is taken in the range of hundredths p (x) ∈ [0,0.5]
Constitute percentage bit vector p0, its percentile vector can be obtained
(1.2) 1,2 is carried out respectively to Raman spectrum data vector x ..., n rank forward difference computings are simultaneously normalized, can obtain
Differential data vector x1,x2,…,xn；
(1.3) to i=1:N is proceeded as follows：
(1.3.1) is to xiAccording to being ranked up from small to large；
X known to (1.3.2)0Length be m, to j=1:M is proceeded as follows：
(1.3.2.1) assumes xiLength be q, then sort after xiIn each element corresponding to percentage bit vectorWherein τ=1,2 ..., q, 0<κ<1.Remember p0(j) it is p0J-th of element, by piIt is linear between adjacent element
Interpolation can try to achieve pi=p0And p (j)i=1-p0(j) corresponding xiRespectively percentile xi,jWith x 'i,j；
(1.3.2.2) calculates xiIn pi=p0(j) noise estimated standard deviation when
(1.3.3) takes σi,j(j=1,2 ..., median m) is used as xiNoise estimated standard deviation σi；
(1.4) σ is takeni(i=1,2 ..., median n) as Raman spectrum data vector x noise estimated standard deviation σ.
Further, the step (2) is specially：
(2.1) first derivation is carried out to Raman spectrum data and obtains its peak value and valley, peak value and valley are constituted one
New vectorial z；
(2.2) judge that first element is peak value or valley in z, i.e. whether z (1) is more than z (2)：If so, then k=0；
If it is not, then k=1, so as to ensure following circulations since a peak value；
(2.3) defined variable tempmax=min (z) and leftmin=x (1), tempmax represent temporary spike and
Leftmin represents left side valley.If z length is l, work as k<During l, circulated as follows：
(2.3.1) k=k+1, and judge whether z (k) is more than tempmax and leftmin+threshold：If so, then
Tempmax=z (k), wherein threshold=r σ represent spectral peak judgment threshold (generally taking r >=6)；
(2.3.2) k=k+1, and judge whether tempmax is more than z (k)+threshold：If so, then tempmax is light
Spectrum signature peak, the position at record this feature peak and size, then reset to tempmax=min (z), leftmin=z by variable
(k)；If it is not, then judging whether z (k) is less than leftmin, if then leftmin=z (k).
By step (2.3) process, position and the size letter of each characteristic peak can be extracted from Raman spectrum data
Breath.
The beneficial effects of the invention are as follows：
(1) it is insensitive to baseline drift and outlier, the noise system of spectrum can be accurately estimated without artificial selection region
Characteristic is counted, peak identification is characterized and provides an objective, rational judgment threshold；
(2) without artificially setting any parameter, the automation of spectral peak identification can be achieved, and need not be in advance to Raman light
Spectrum goes background process, it is to avoid spectral information loses caused by removing background operation, and spectral peak recognition accuracy is high；
(3) calculate simple, the memory space and computing capability to processor require relatively low, can be including embedded system
Used on various types of hardware platform inside.
Brief description of the drawings
Fig. 1 is the data processing the general frame of the present invention；
Fig. 2 is the data processing detail flowchart of the present invention；
Fig. 3 is the Raman spectrogram and its spectral peak recognition result figure of PS plastic samples；
Fig. 4 is the Raman spectrogram and its spectral peak recognition result figure of PC acrylic samples.
Embodiment
The invention will be further described with reference to the accompanying drawings and examples.
Traditional spectral peak recognition methods is generally comprised：Baseline correction and spectrum peak search, wherein spectrum peak search need operator
Member sets judgment threshold manually.It is demonstrated experimentally that different baseline correction algorithms and judgment threshold are chosen at and are largely fixed
The result of spectral peak identification.As shown in figure 1, the spectroscopic data in the present invention needs not move through baseline correction, spectral peak knowledge is sent directly into
Other unit, this avoid the spectral information loss that baseline correction is brought；Noise of the spectral peak recognition unit first to spectroscopic data
Variance is estimated that the noise variance acts not only as the Appreciation gist of the spectrometer collection quality of data, and can conduct
The judgment threshold of rear end spectrum peak search algorithm, this method is very directly perceived, objective, significantly reduces the burden of operating personnel.
As shown in Fig. 2 the present invention provides a kind of Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation,
Specifically include following steps：
1. it is 0 to assume that stochastic variable ε obeys average, variance is σ2Gaussian Profile, then its probability density function be：
Known hundredths p (x) refers to that stochastic variable ε is less than percentile x cumulative probability, i.e.,：
According to the definition of error function：
It can obtain：
It can further obtain：
Due to inverse error function erf-1(2p-1) is an odd function, and 2p-1 span is [- 1,1], so only
Values of the percentile x in the range of hundredths 2p-1 ∈ [- 1,0] (i.e. p ∈ [0,0.5]) need to be provided and can determine that its complete point
Cloth curve.
The present invention is using average as 0, and the Gaussian noise that variance is 1 is reference, is taken in the range of hundredths p (x) ∈ [0,0.5]
M point composition percentage bit vector p0, its percentile vector can be obtained
2. a pair Raman spectrum data vector x carries out 1,2 respectively ..., n rank forward difference computings are simultaneously normalized, and can obtain difference
Divided data vector x1,x2,…,xn；
3. couple i=1:N is proceeded as follows：
3.1 couples of xiAccording to being ranked up from small to large；
X known to 3.20Length be m, to j=1:M is proceeded as follows：
3.2.1 x is assumediLength be q, then sort after xiIn each element corresponding to percentage bit vector Wherein τ=1,2 ..., q, 0<κ<1.Remember p0(j) it is p0J-th of element, by piIt is linear interior between adjacent element
P can be tried to achieve by insertingi=p0And p (j)i=1-p0(j) corresponding xiRespectively percentile xi,jWith x 'i,j；
3.2.2 x is calculatediIn pi=p0(j) noise estimated standard deviation when
3.3 take σi,j(j=1,2 ..., median m) is used as xiNoise estimated standard deviation σi；
4. take σi(i=1,2 ..., median n) as Raman spectrum data vector x noise estimated standard deviation σ.
5. in the first-order difference data vector x of above-mentioned steps 21In find out the data that saltus step occurs for front and rear sign symbol, this
Corresponding a little data are exactly Raman spectrum peak value or valley, and these data are constituted into a new vectorial z；
6. judging that first element is peak value or valley in z, i.e. whether z (1) is more than z (2)：If so, then k=0；If
It is no, then k=1, so as to ensure following circulations since a peak value；
7. defined variable tempmax=min (z) and leftmin=x (1), is respectively used to storage temporary spike and left side paddy
Value.If z length is l, work as k<During l, circulated as follows：
7.1 k=k+1, and judge whether z (k) is more than tempmax and leftmin+threshold：If so, then
Tempmax=z (k), wherein threshold=r σ represent spectral peak judgment threshold (generally taking r >=6)；
7.2 k=k+1, and judge whether tempmax is more than z (k)+threshold：If so, then tempmax is Spectral Properties
Peak, the position at record this feature peak and size are levied, variable is then reset into tempmax=min (z), leftmin=z (k)；
If it is not, then judging whether z (k) is less than leftmin：If then leftmin=z (k).
By said process, position and the size information of each characteristic peak can be extracted from Raman spectrum data, so
Afterwards by setting up standard database and searching algorithm that material differentiates, it is possible to realize the judgement to unknown sample and identify.
Embodiment：
By taking PS plastics and PC acrylic sample datas that Raman spectrometer is gathered as an example, its original spectrum is respectively such as Fig. 3, figure
Shown in 4.As seen from the figure, there are different characteristic spectral lines in both materials, but due to being influenceed by fluorescence background, cause baseline
Produce certain skew.
First, spectroscopic data is handled using robust noise variance evaluation method proposed by the present invention, takes hundredths
Vectorial p0={ 0.05,0.10,0,15,0.20 ..., 0.40 }, κ=0.5 can obtain the Raman spectrum noise estimation standard of PS plastics
Poor σ1The Raman spectrum noise estimated standard deviation σ of=38.9859, PC acrylic2=35.8535.In order to verify these noise criterias
Whether poor estimated result is reasonable, manually selects respectively in Fig. 3 1700 in 1600-2500 scopes and Fig. 4:Data in the range of 2600
The data carried out in noise variance statistics, these regions are not influenceed by characteristic spectral line and needle position misalignment substantially, can be relatively defined
The noise statisticses of the true respective spectrum of reflection, the poor result of calculation of noise criteria for obtaining PS plastics is σ '1=38.1571,
The poor result of calculation of the noise criteria of PC acrylics is σ '2=34.8209, this and robust noise variance evaluation knot proposed by the present invention
It is really basically identical.Then, r=6 is taken, i.e., using 6 times of noise estimated standard deviation as the judgment threshold of spectral peak, further utilizes this
The characteristic peak searching algorithm that invention is proposed is handled spectrum, marks the characteristic peak searched in spectrogram with small circle,
Final result is as shown in Figure 3, Figure 4.As seen from the figure, the Raman spectral characteristics proposed by the present invention based on robust noise variance evaluation
Peak recognition methods can efficiently identify out the characteristic peak of respective spectrum.
Claims (3)
1. a kind of Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation, it is characterised in that including following step
Suddenly：
(1) it is reference by the Gaussian noise that 0, variance is 1 of average, it is determined that with reference to hundredths and its percentile；To Raman light
Modal data carries out forward difference computing and normalized, and by the data after normalization according to sorting from small to large, calculates each data
Hundredths, then tries to achieve the percentile with reference to corresponding to hundredths by linear interpolation, and by it with referring to percentile phase
Remove, obtain series of noise standard deviation, take the median of standard deviation as the noise estimated standard deviation σ of the spectroscopic data.
(2) its peak value and valley are obtained by carrying out first derivation to Raman spectrum data, by each peak value and its left and right two
The minimum valley of side is compared, if greater than the noise criteria difference σ of r times (generally taking r >=6), then it is assumed that it is Raman spectrum
Characteristic peak.
2. the Raman spectrum characteristic peak recognition methods according to claim 1 based on robust noise variance evaluation, its feature
It is, the step (1) is specially：
(1.1) it is 0 to assume that stochastic variable ε obeys average, and variance is σ2Gaussian Profile, then its probability density function be：
<mrow>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<msqrt>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
</mrow>
</msqrt>
<mi>&sigma;</mi>
</mrow>
</mfrac>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mfrac>
<msup>
<mi>x</mi>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
</mrow>
</msup>
</mrow>
Known hundredths p (x) refers to that stochastic variable ε is less than percentile x cumulative probability, i.e.,：
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<msqrt>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
</mrow>
</msqrt>
<mi>&sigma;</mi>
</mrow>
</mfrac>
<msubsup>
<mo>&Integral;</mo>
<mrow>
<mo>-</mo>
<mi>&infin;</mi>
</mrow>
<mi>x</mi>
</msubsup>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mfrac>
<msup>
<mi>t</mi>
<mn>2</mn>
</msup>
<mrow>
<mn>2</mn>
<msup>
<mi>&sigma;</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
</mrow>
</msup>
<mi>d</mi>
<mi>t</mi>
</mrow>
According to the definition of error function：
<mrow>
<mi>e</mi>
<mi>r</mi>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msqrt>
<mi>&pi;</mi>
</msqrt>
</mfrac>
<msubsup>
<mo>&Integral;</mo>
<mrow>
<mo>-</mo>
<mi>x</mi>
</mrow>
<mi>x</mi>
</msubsup>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<msup>
<mi>t</mi>
<mn>2</mn>
</msup>
</mrow>
</msup>
<mi>d</mi>
<mi>t</mi>
<mo>=</mo>
<mfrac>
<mn>2</mn>
<msqrt>
<mi>&pi;</mi>
</msqrt>
</mfrac>
<msubsup>
<mo>&Integral;</mo>
<mn>0</mn>
<mi>x</mi>
</msubsup>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<msup>
<mi>t</mi>
<mn>2</mn>
</msup>
</mrow>
</msup>
<mi>d</mi>
<mi>t</mi>
</mrow>
It can obtain：
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<mo>&lsqb;</mo>
<mn>1</mn>
<mo>+</mo>
<mi>e</mi>
<mi>r</mi>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mi>x</mi>
<mrow>
<msqrt>
<mn>2</mn>
</msqrt>
<mi>&sigma;</mi>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
</mrow>
It can further obtain：
<mrow>
<mi>x</mi>
<mo>=</mo>
<msqrt>
<mn>2</mn>
</msqrt>
<mi>&sigma;</mi>
<mo>&CenterDot;</mo>
<msup>
<mi>erf</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>&lsqb;</mo>
<mn>2</mn>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mn>1</mn>
<mo>&rsqb;</mo>
</mrow>
Using average as 0, the Gaussian noise that variance is 1 is reference, and m point composition is taken in the range of hundredths p (x) ∈ [0,0.5]
Percentage bit vector p0, its percentile vector can be obtained
(1.2) 1,2 is carried out respectively to Raman spectrum data vector x ..., n rank forward difference computings are simultaneously normalized, and can obtain difference
Data vector x1,x2,…,xn；
(1.3) to i=1:N is proceeded as follows：
(1.3.1) is to xiAccording to being ranked up from small to large；
X known to (1.3.2)0Length be m, to j=1:M is proceeded as follows：
(1.3.2.1) assumes xiLength be q, then sort after xiIn each element corresponding to percentage bit vectorWherein τ=1,2 ..., q, 0<κ<1.Remember p0(j) it is p0J-th of element, by piIt is linear between adjacent element
Interpolation can try to achieve pi=p0And p (j)i=1-p0(j) corresponding xiRespectively percentile xi,jWith x 'i,j；
(1.3.2.2) calculates xiIn pi=p0(j) noise estimated standard deviation when
(1.3.3) takes σi,j(j=1,2 ..., median m) is used as xiNoise estimated standard deviation σi；
(1.4) σ is takeni(i=1,2 ..., median n) as Raman spectrum data vector x noise estimated standard deviation σ.
3. the Raman spectrum characteristic peak recognition methods according to claim 1 based on robust noise variance evaluation, its feature
It is, the step (2) is specially：
(2.1) first derivation is carried out to Raman spectrum data and obtains its peak value and valley, peak value and valley composition one is new
Vectorial z；
(2.2) judge that first element is peak value or valley in z, i.e. whether z (1) is more than z (2)：If so, then k=0；If it is not,
Then k=1, so as to ensure following circulations since a peak value；
(2.3) defined variable tempmax=min (z) and leftmin=x (1), tempmax represents temporary spike and leftmin
Represent left side valley.If z length is l, work as k<During l, circulated as follows：
(2.3.1) k=k+1, and judge whether z (k) is more than tempmax and leftmin+threshold：If so, then tempmax
=z (k), wherein threshold=r σ represent spectral peak judgment threshold (generally taking r >=6)；
(2.3.2) k=k+1, and judge whether tempmax is more than z (k)+threshold：If so, then tempmax is Spectral Properties
Peak, the position at record this feature peak and size are levied, variable is then reset into tempmax=min (z), leftmin=z (k)；
If it is not, then judging whether z (k) is less than leftmin, if then leftmin=z (k).
By step (2.3) process, position and the size information of each characteristic peak can be extracted from Raman spectrum data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710403411.9A CN107179310B (en) | 2017-06-01 | 2017-06-01 | Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710403411.9A CN107179310B (en) | 2017-06-01 | 2017-06-01 | Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107179310A true CN107179310A (en) | 2017-09-19 |
| CN107179310B CN107179310B (en) | 2018-02-27 |
Family
ID=59836785
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710403411.9A Active CN107179310B (en) | 2017-06-01 | 2017-06-01 | Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107179310B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109060771A (en) * | 2018-07-26 | 2018-12-21 | 温州大学 | A kind of common recognition model building method based on spectrum different characteristic collection |
| CN109283169A (en) * | 2018-11-22 | 2019-01-29 | 深圳市雷泛科技有限公司 | A kind of Raman spectral peaks recognition methods of robust |
| CN109993155A (en) * | 2019-04-23 | 2019-07-09 | 北京理工大学 | For the characteristic peak extracting method of low signal-to-noise ratio uv raman spectroscopy |
| CN111257506A (en) * | 2020-01-15 | 2020-06-09 | 交通运输部天津水运工程科学研究所 | Automatic recognition algorithm for fuel oil sulfur content |
| CN112418051A (en) * | 2020-11-18 | 2021-02-26 | 温州大学 | State estimation method for nonlinear dynamic system under non-Gaussian noise |
| CN112697774A (en) * | 2020-12-24 | 2021-04-23 | 江南大学 | Method for detecting flunixin meglumine residue based on surface enhanced Raman spectroscopy |
| CN114166814A (en) * | 2021-05-25 | 2022-03-11 | 北京理工大学 | Raman spectrum peak identification method based on dual-scale correlation operation |
| CN114354571A (en) * | 2021-12-22 | 2022-04-15 | 中国科学院合肥物质科学研究院 | Easy-to-prepare chemical Raman characteristic peak identification method based on half-peak width and peak height |
| CN115980744A (en) * | 2022-11-10 | 2023-04-18 | 国家卫星海洋应用中心 | Method for separating spectrum of multi-peak sea wave image of satellite-borne SAR image data without overlapping |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5819208A (en) * | 1996-10-29 | 1998-10-06 | Northern Telecom Limited | Quantifying circuit performance |
| CN101465122A (en) * | 2007-12-20 | 2009-06-24 | 株式会社东芝 | Method and system for detecting phonetic frequency spectrum wave crest and phonetic identification |
| CN102274016A (en) * | 2011-07-08 | 2011-12-14 | 重庆大学 | Method for recognizing intracranial pressure signal characteristic peaks |
| CN103457890A (en) * | 2013-09-03 | 2013-12-18 | 西安电子科技大学 | Method for effectively recognizing digital modulating signals in non-Gaussian noise |
| CN103913765A (en) * | 2014-03-24 | 2014-07-09 | 中国船舶重工集团公司第七一九研究所 | Nuclide energy spectrum peak searching method |
| CN103983588A (en) * | 2014-05-20 | 2014-08-13 | 核工业北京地质研究院 | Rock and mineral spectral feature absorption peak position identification method |
| CN105518455A (en) * | 2013-09-09 | 2016-04-20 | 株式会社岛津制作所 | Peak detection method |
-
2017
- 2017-06-01 CN CN201710403411.9A patent/CN107179310B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5819208A (en) * | 1996-10-29 | 1998-10-06 | Northern Telecom Limited | Quantifying circuit performance |
| CN101465122A (en) * | 2007-12-20 | 2009-06-24 | 株式会社东芝 | Method and system for detecting phonetic frequency spectrum wave crest and phonetic identification |
| CN102274016A (en) * | 2011-07-08 | 2011-12-14 | 重庆大学 | Method for recognizing intracranial pressure signal characteristic peaks |
| CN103457890A (en) * | 2013-09-03 | 2013-12-18 | 西安电子科技大学 | Method for effectively recognizing digital modulating signals in non-Gaussian noise |
| CN105518455A (en) * | 2013-09-09 | 2016-04-20 | 株式会社岛津制作所 | Peak detection method |
| CN103913765A (en) * | 2014-03-24 | 2014-07-09 | 中国船舶重工集团公司第七一九研究所 | Nuclide energy spectrum peak searching method |
| CN103983588A (en) * | 2014-05-20 | 2014-08-13 | 核工业北京地质研究院 | Rock and mineral spectral feature absorption peak position identification method |
Non-Patent Citations (2)
| Title |
|---|
| PENG DAI ET AL.: "Robust speech recognition by using spectral subtraction with noise peak shifting", 《THE INSTITUTION OF ENGINEERING AND TECHNOLOGY SIGNAL PROCESSING》 * |
| 林冬秀 等: "基于近红外光谱的重金属污染泥蚶的快速检测", 《中国食品学报》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109060771B (en) * | 2018-07-26 | 2020-12-29 | 温州大学 | Consensus model construction method based on different characteristic sets of spectrum |
| CN109060771A (en) * | 2018-07-26 | 2018-12-21 | 温州大学 | A kind of common recognition model building method based on spectrum different characteristic collection |
| CN109283169A (en) * | 2018-11-22 | 2019-01-29 | 深圳市雷泛科技有限公司 | A kind of Raman spectral peaks recognition methods of robust |
| CN109993155A (en) * | 2019-04-23 | 2019-07-09 | 北京理工大学 | For the characteristic peak extracting method of low signal-to-noise ratio uv raman spectroscopy |
| CN111257506A (en) * | 2020-01-15 | 2020-06-09 | 交通运输部天津水运工程科学研究所 | Automatic recognition algorithm for fuel oil sulfur content |
| CN112418051B (en) * | 2020-11-18 | 2024-05-03 | 温州大学 | State estimation method for nonlinear dynamic system under non-Gaussian noise |
| CN112418051A (en) * | 2020-11-18 | 2021-02-26 | 温州大学 | State estimation method for nonlinear dynamic system under non-Gaussian noise |
| CN112697774A (en) * | 2020-12-24 | 2021-04-23 | 江南大学 | Method for detecting flunixin meglumine residue based on surface enhanced Raman spectroscopy |
| CN112697774B (en) * | 2020-12-24 | 2022-05-17 | 江南大学 | Method for detecting flunixin meglumine residue based on surface enhanced Raman spectroscopy |
| CN114166814A (en) * | 2021-05-25 | 2022-03-11 | 北京理工大学 | Raman spectrum peak identification method based on dual-scale correlation operation |
| CN114354571A (en) * | 2021-12-22 | 2022-04-15 | 中国科学院合肥物质科学研究院 | Easy-to-prepare chemical Raman characteristic peak identification method based on half-peak width and peak height |
| CN114354571B (en) * | 2021-12-22 | 2024-05-28 | 中国科学院合肥物质科学研究院 | Method for identifying Raman characteristic peak of easily-toxic chemical based on half-peak width and peak height |
| CN115980744A (en) * | 2022-11-10 | 2023-04-18 | 国家卫星海洋应用中心 | Method for separating spectrum of multi-peak sea wave image of satellite-borne SAR image data without overlapping |
| CN115980744B (en) * | 2022-11-10 | 2024-03-22 | 国家卫星海洋应用中心 | Method for separating satellite-borne SAR image data from non-overlapping masking peak sea wave image spectrum |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107179310B (en) | 2018-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107179310A (en) | Raman spectrum characteristic peak recognition methods based on robust noise variance evaluation | |
| CN107092829B (en) | Malicious code detection method based on image matching | |
| CN105334185B (en) | The near-infrared model maintaining method differentiated based on spectrum projection | |
| CN110243806B (en) | Mixture component identification method based on similarity under Raman spectrum | |
| CN107515895A (en) | A kind of sensation target search method and system based on target detection | |
| CN108765470A (en) | One kind being directed to the improved KCF track algorithms of target occlusion | |
| CN101534305A (en) | Method and system for detecting network flow exception | |
| CN103020265B (en) | The method and system of image retrieval | |
| CN104713835B (en) | A kind of tobacco leaf color quantizes recognition methods online | |
| CN101216436A (en) | Fabric flaw automatic detection method based on Support Vector data description theory | |
| CN111506637B (en) | Multi-dimensional anomaly detection method and device based on KPI (Key Performance indicator) and storage medium | |
| CN101833664A (en) | Video image character detecting method based on sparse expression | |
| CN103325122A (en) | Pedestrian retrieval method based on bidirectional sequencing | |
| CN105023027A (en) | Sole trace pattern image retrieval method based on multi-feedback mechanism | |
| CN104881795A (en) | E-commerce false comment judging and recognizing method | |
| CN107292302A (en) | Detect the method and system of point of interest in picture | |
| CN110288013A (en) | A kind of defective labels recognition methods based on block segmentation and the multiple twin convolutional neural networks of input | |
| CN104038792A (en) | Video content analysis method and device for IPTV (Internet Protocol Television) supervision | |
| CN105528580A (en) | Hyperspectral curve matching method based on absorption peak characteristic | |
| Park et al. | Evaluation and comparison of methods for forensic glass source conclusions | |
| CN104155245B (en) | A kind of detection method of the oil product multiclass character mapped with spectrogram based on pattern recognition | |
| CN109829902A (en) | A kind of lung CT image tubercle screening technique based on generalized S-transform and Teager attribute | |
| CN109146003A (en) | A kind of classifying identification method based on high light spectrum image-forming, apparatus and system | |
| CN104980442B (en) | A kind of network inbreak detection method based on first sample rarefaction representation | |
| CN110990383A (en) | Similarity calculation method based on industrial big data set |
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 | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170919 Assignee: Yueqing Institute of industry, Wenzhou University Assignor: Wenzhou University Contract record no.: X2020330000105 Denomination of invention: Raman spectrum characteristic peak recognition method based on robust noise variance estimation Granted publication date: 20180227 License type: Common License Record date: 20201125 |
|
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20170919 Assignee: Zhejiang Zhongshi wanwei Zhilian Co.,Ltd. Assignor: Wenzhou University Contract record no.: X2021330000830 Denomination of invention: Raman spectrum characteristic peak recognition method based on robust noise variance estimation Granted publication date: 20180227 License type: Common License Record date: 20211222 |
|
| EE01 | Entry into force of recordation of patent licensing contract |