CN108802000A - A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman - Google Patents
A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman Download PDFInfo
- Publication number
- CN108802000A CN108802000A CN201810219592.4A CN201810219592A CN108802000A CN 108802000 A CN108802000 A CN 108802000A CN 201810219592 A CN201810219592 A CN 201810219592A CN 108802000 A CN108802000 A CN 108802000A
- Authority
- CN
- China
- Prior art keywords
- cholecalciferol
- cholesterol
- raman
- analysis
- spectrum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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)
Abstract
The present invention provides a kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman, wherein：Step 1, cholesterol and vitamin mixtures are configured to the sample of different quality degree cholecalciferol-cholesterol with cholecalciferol-cholesterol；Step 2, the spectral signal that above-mentioned different content cholecalciferol-cholesterol sample is measured by Raman spectrometer, obtains Raman spectrum data after pretreatment；Step 3, analyzing processing is carried out to the full modal data of pretreated Raman spectrum using Multielement statistical analysis method and establishes Quantitative Analysis Model；Step 4, cholecalciferol-cholesterol quantitative analysis is carried out to the sample of unknown content using established Quantitative Analysis Model.The present invention is so effective that solve influence of the cholecalciferol-cholesterol sample raman spectra offset to quantitative measurment precision；Forecast analysis model of the present invention realizes prediction to the cholecalciferol-cholesterol sample of different content, as a result demonstrates validity of the full spectrum analysis method based on Raman spectrum in the analysis of cholecalciferol-cholesterol content quantitative.
Description
Technical field
The present invention relates to a kind of cholecalciferol-cholesterol content quantitative methods, and in particular, to a kind of nothing based on the full spectrum analysis of Raman
Damage quick cholecalciferol-cholesterol content quantitative method.
Background technology
Cholecalciferol-cholesterol is the safely and effectively clinical commonly used drug for preventing and treating rickets, has and calcium, phosphorus is promoted to absorb profit
With with the effect that promotes bone normal calcification, curative effect is suitable with intramuscular injection vitamin D.In order to ensure curative effect, it is necessary to be tieed up in courage
Cholecalciferol-cholesterol component content is detected in production process and the final product quality control of fourth drug, studies lossless quick cholecalciferol-cholesterol content quantitative
Method have great importance.
Cholecalciferol-cholesterol raw material is crystal, at present only a small number of document reports its by vitamin D3Pass through Hydrogenbond with cholesterol
Equimolar molecule addition product is formed, but not yet its structure is furtherd investigate.Cholecalciferol-cholesterol is by vitamin D3It is formed with cholesterol,
Its structure is by vitamin D3Equimolar non-covalent bond (hydrogen bond) conjugate, chemical equation are formed on hydroxyl with cholesterol
For：
Hydrogen bond energy is about 200kJ/mol, generally 5-30kJ/mol, than general covalent bond, ionic bond and gold
Belong to key bond energy smaller, but is better than electrostatic attraction.Since hydrogen bond is unstable, cholecalciferol-cholesterol is also decomposed into free vitamin D3It is solid with courage
Alcohol.Therefore, vitamin D3It is both starting material and degradation impurity with cholesterol.
Free vitamin D cannot be distinguished in existing liquid phase process3With free cholesterol, standards of pharmacopoeia and current pharmaceutical factory
From quasi- quality standard not to free cholesterol and vitamin D3It is controlled as impurity, existing content detection side
Method measured result further includes the vitamin D of free state therein3And cholesterol, it can not accurately measure the real content of cholecalciferol-cholesterol.
On the other hand, common structural identification means for example infrared spectrum, ultra-violet absorption spectrum, nuclear magnetic resonance, mass spectrum and
Elemental analysis etc. can not illustrate the material base of cholecalciferol-cholesterol, can not prove the presence of hydrogen bond.
Through retrieval, at present only Shanghai medicine-feeding institute of materia medica of Xinyi Pharmaceutical Factory Co., Ltd by differential scan thermometric analysis,
Raman spectrum and powder x-ray diffraction are found that in cholecalciferol-cholesterol raw material that there are eutectic phenomenas, and are characterized to it, using powder
Last X-ray diffraction establishes eutectic and free state vitamin D3With the quantitative analysis method of cholesterol level, application is corresponded to
Number of patent application is 201510106721.5, and step is：
(1) cholecalciferol-cholesterol of different cholecalciferol-cholesterol contents, the correct mixture of cholesterol and vitamine D3 are provided；
(2) the X difraction spectrums of above-mentioned standard mixture are provided；
(3) 2 θ=4.0 ° or 8.0 ° of diffraction in the cholecalciferol-cholesterol content based on above-mentioned standard mixture and X difraction spectrums
Peak heights or area provide equation of linear regression；
(4) cholecalciferol-cholesterol, cholesterol and vitamin D are provided3Mixture X difraction spectrums；And
(5) based on the equation of linear regression and cholecalciferol-cholesterol, the mixture of cholesterol and vitamine D3 X difraction spectrums in
2 θ=4.0 ° or 8.0 ° of diffraction peak heights or area determine the content of cholecalciferol-cholesterol.
It is existing to study using Raman spectrum come the composition and content of quantitative analysis substance, often with the Raman spectrum of substance
One or several characteristic peaks establish analysis Quantitative Analysis Model as training data, however, the study found that by one or several
The method of characteristic peak is not particularly suited for the quantitative analysis of cholecalciferol-cholesterol content.Cholecalciferol-cholesterol is in 960cm-1、1070cm-1、1225cm-1、
1437cm-1、1638cm-1There are apparent raman characteristic peak, 1638cm therein-1Raman characteristic peak is most strong, cholesterol and vitamin
D3Raman spectrum in 1646cm-1There is stronger raman characteristic peak at place, when the reduction of cholecalciferol-cholesterol concentration, free cholesterol in sample
And vitamin D3Concentration when increasing, original 1638cm in Raman spectrum-1The cholecalciferol-cholesterol raman characteristic peak at place disappears, and
1646cm-1Place shows raman characteristic peak, due to 1638cm-1With 1646cm-1Two characteristic peaks are very close to directly affecting
The quantitative precision of analysis Quantitative Analysis Model is established using one or several characteristic peaks as training data.
Invention content
For the defects in the prior art, the object of the present invention is to provide a kind of based on the lossless of the full spectrum analysis of Raman spectrum
Quick cholecalciferol-cholesterol content quantitative method carries out analyzing processing to the full modal data of pretreated Raman spectrum and establishes quantitative analysis mould
Type can with quick nondestructive detect the spectral signal of different content cholecalciferol-cholesterol.
The present invention provides a kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman spectrum, including such as
Lower step：
Step 1, cholesterol and vitamin mixtures are configured to different quality degree cholecalciferol-cholesterol with cholecalciferol-cholesterol
Sample；
Step 2, the spectral signal that above-mentioned different content cholecalciferol-cholesterol sample is measured by Raman spectrometer, after pretreatment
Obtain Raman spectrum data；
Step 3, analyzing processing is carried out to the full modal data of pretreated Raman spectrum using Multielement statistical analysis method to build
Vertical Quantitative Analysis Model；
Step 4, cholecalciferol-cholesterol quantitative analysis is carried out to the sample of unknown content using Quantitative Analysis Model.
Preferably, described that Raman spectrum data is obtained after pretreatment, wherein pretreatment includes baseline correction, smooth filter
Three processes of wave and area normalization, these three processes are individually performed or execute successively in sequence.
The baseline correction refers to carrying out baseline correction to original Raman spectrum data.
The smothing filtering refers to carrying out smothing filtering to spectroscopic data.Using SGF filter methods (Savitzky-
GolayFilter), polynomial least mean square fitting is done by the data to window interior, obtains pixel at window center position
Value after smooth, the average value at wavelength k after convolution is smooth are：
In formula, hiFor smoothing factor, H is normalization factor,Wherein hiPassed through based on principle of least square method
Fitting of a polynomial acquires.
The area normalization refers to carrying out area normalization to spectroscopic data, and calculation formula is：
In formula, f (x0) it is former spectral intensity, y is to normalize transformed raman spectrum strength, and a, b are to need to normalize
Starting Raman-shifted wavenumbers and cut-off Raman-shifted wavenumbers.
Preferably, described right using Multielement statistical analysis method (Partial Least Squares PLS, PCA-LDA, clustering etc.)
The pretreated full modal data of Raman spectrum carries out analyzing processing and establishes Quantitative Analysis Model, is first divided into the spectroscopic data of sample
Training sample set and test sample collection are then divided into two steps：
(1), Quantitative Analysis Model is built using training sample set；
The correspondence model of concentration and Raman spectrum is established using training sample set, including wavelength band selection is quantified with foundation
Analysis model step.
Above-mentioned wavelength band selects full spectral region.It is quantitative that foundation is trained to the full modal data of the Raman spectrum of input
Analysis model, without extracting Raman spectrum characteristic peak.
(2), Quantitative Analysis Model is verified by test sample；
Test sample is substituted into the Quantitative Analysis Model established and carries out prediction and Fitting Analysis, including wavelength band selection
With Quantitative Analysis Model prediction steps, wavelength band selects full spectral region, without extracting Raman spectrum characteristic peak.
Preferably, the different content cholecalciferol-cholesterol, specially：
It is 1 to take mass ratio:1 cholesterol and vitamin is uniformly mixed, and mixture is made；Cholecalciferol-cholesterol and mixture are prepared
At the mixing sample of different weight percentage content of the cholecalciferol-cholesterol mass percentage content between 0~100%.
Preferably, the influence in order to avoid stray light to result, spectrum data gathering process is in darkroom or shading environment
It carries out.
The present invention carries out cholecalciferol-cholesterol quantitative analysis using Quantitative Analysis Model to the sample of unknown content, adopts first
Collect sample Raman spectrum data, pretreatment operation then carried out to the data that measure, test condition and preprocess method with
It is identical when modeling sample, it is pre- that the full modal data of pretreated Raman is finally substituted into established Quantitative Analysis Model progress concentration
It surveys.
Compared with prior art, the present invention has following advantageous effect：
(1) present invention can with quick nondestructive detect the spectral signal of different content cholecalciferol-cholesterol, Raman after treatment
Spectroscopic data carries out full spectrum analysis in conjunction with Multielement statistical analysis method (Partial Least Squares PLS, PCA-LDA, clustering etc.)
Method establish Quantitative Analysis Model, it is so effective that solve spectral peak offset and the influence of intensity error, cholecalciferol-cholesterol may be implemented and contain
The prediction of amount.
(2) the forecast analysis model in the present invention in one embodiment realizes the cholecalciferol-cholesterol sample of 13 different contents
Prediction, the right R of Linear Quasi of prediction result and actual result2Reach 99.9%；As a result it demonstrates based on the complete of Raman spectrum
Validity of the spectrum analysis method in the analysis of cholecalciferol-cholesterol content quantitative.
Description of the drawings
Upon reading the detailed description of non-limiting embodiments with reference to the following drawings, other feature of the invention,
Objects and advantages will become more apparent upon：
Fig. 1 is the Raman spectrogram of the cholecalciferol-cholesterol sample for four kinds of different contents that one embodiment of the invention measures；
Fig. 2 is one embodiment of the invention cholecalciferol-cholesterol content quantitative analysis modeling flow chart；
Fig. 3 is one embodiment of the invention cholecalciferol-cholesterol Pretreated spectra baseline correction flow chart；
Fig. 4 is that one embodiment of the invention predicts content and actual content fitting result linear correlation figure.
Specific implementation mode
With reference to specific embodiment, the present invention is described in detail.Following embodiment will be helpful to the technology of this field
Personnel further understand the present invention, but the invention is not limited in any way.It should be pointed out that the ordinary skill of this field
For personnel, without departing from the inventive concept of the premise, various modifications and improvements can be made.These belong to the present invention
Protection domain.
The present invention provides the lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman spectrum, pre- by suggesting
Analysis model is surveyed, the spectral signal of different content cholecalciferol-cholesterol can with quick nondestructive be detected.When it is implemented, being referred to as follows
Step carries out：
Step 1, cholesterol and vitamin mixtures are configured to different quality degree cholecalciferol-cholesterol with cholecalciferol-cholesterol
Sample；
Step 2, the spectral signal that above-mentioned different content cholecalciferol-cholesterol sample is measured by Portable Raman spectrometer, by base
Obtain Raman spectrum data after the pretreatments such as line fitting, smooth filter and area normalization, i.e. sample database in Fig. 2；
Baseline fitting, smooth filter and area normalization can be individually performed or execute successively in sequence, execute effect successively
Fruit is more preferable.
The present invention reaches removal noise, background and sample otherness by being pre-processed to Raman spectrum data
The purpose of interference.It is easy by the floating of excitation laser power, detector thermal noise, sample shape in the acquisition process of Raman signal
The spectroscopic data of the influence of many factors such as looks, external environment, acquisition usually has larger noise.In addition, sample or sample
The fluorescence background of product container itself, fluorescence background have larger impact to subsequent qualitative and quantitative analysis, it should from spectrum number
According to middle removal.Meanwhile the concentration of sample is different, the variation of the difference and test condition of sampled point can all lead to different samples
Raman spectrum strength and background difference it is larger, higher analysis precision, is also carried out standard normalization in order to obtain.
Step 3, Raman spectrum data combination Multielement statistical analysis method is (such as Partial Least Squares PLS, PCA-LDA, poly-
Alanysis etc.) method that carries out full spectrum analysis establishes Quantitative Analysis Model；
By taking Partial Least Squares as an example, principal component decomposition first is carried out to the Raman spectrum matrix of input, while to input
Concentration matrix principal component decomposes, and is then returned using spectrum principal component scores matrix and concentration principal component scores matrix.
The formula decomposed to spectrum matrix X and concentration matrix Y is as follows：
T and U is respectively the score matrix of X and Y in formula, and P and Q are respectively the loading matrix of X and Y, tk、pk、uk、qkPoint
Do not refer to the kth row column vector of matrix T, P, U, K, f is above-mentioned matrix column number, EXAnd EYIt is respectively quasi- with Partial Least Squares
Close the error introduced when matrix X and Y.
Characteristic spectrum matrix T and characteristic concentration matrix U are obtained after decomposition, then by characteristic spectrum matrix T and characteristic concentration
Matrix U is returned, and formula is：U=TB；
To obtain incidence matrix B：B=(TTT)-1TTU；
Step 4, cholecalciferol-cholesterol quantitative analysis is carried out to the sample of unknown content using Quantitative Analysis Model.When prediction,
Score matrix T is found out according to unknown sample spectrum matrix X and loading matrix P, concentration prediction value is：Y=TBQT。
Specific sample detection is solved below the implementation detail of the above method to be described, it should be understood that with
Lower sample and step etc. are only the selection in part preferred embodiment of the present invention, are not the limitation for the present invention, early
In other embodiments of the invention, the sample of other different contents can also be used, this is to those skilled in the art
It is easily achieved.
In one embodiment, described that cholesterol and vitamin mixtures are configured to different quality percentage with cholecalciferol-cholesterol
Than the sample of content cholecalciferol-cholesterol, it is referred to following progress：
It is 1 to take mass ratio:1 cholesterol and vitamin is uniformly mixed, and mixture is made；Cholecalciferol-cholesterol and mixture are prepared
Be respectively 0.78125% at cholecalciferol-cholesterol mass percentage content, 1.5625%, 3.125%, 6.25%, 12.5%, 25%,
50.0%, 75.0%, 87.5%, 93.75%, 96.875%, 98.4375%, 99.21875% mixing sample；
Take 10g cholecalciferol-cholesterols and 5g cholesterol and 5g vitamin Ds3It is uniformly mixed, obtains 50% concentration cholecalciferol-cholesterol sample；It takes
50% concentration cholecalciferol-cholesterol samples of 5g are stated, with 2.5g cholesterol and 2.5g vitamin Ds3It is uniformly mixed, obtains 25% concentration cholecalciferol-cholesterol
Sample；25% concentration cholecalciferol-cholesterol samples of above-mentioned 5g are taken again, with 2.5g cholesterol and 2.5g vitamin Ds3It is uniformly mixed, obtains
12.5% concentration cholecalciferol-cholesterol sample；Prepare respectively as stated above concentration be respectively 6.25%, 3.125%, 1.5625%,
0.78125% cholecalciferol-cholesterol sample；
Above-mentioned 5g50% concentration cholecalciferol-cholesterol sample is taken, is uniformly mixed with 5g cholecalciferol-cholesterols, 75% concentration cholecalciferol-cholesterol sample is obtained；
Above-mentioned 5g75% concentration cholecalciferol-cholesterol sample is taken, is uniformly mixed with 5g cholecalciferol-cholesterols, 87.5% concentration cholecalciferol-cholesterol sample is obtained, by above-mentioned
Method prepares a concentration of 93.75%, 96.875%, 98.4375%, 99.21875% cholecalciferol-cholesterol sample respectively；
The determination condition of the Portable Raman spectrometer：Optical maser wavelength 785nm, output power 200mW, the time of integration
5s, pendulous frequency 3 times；Meanwhile the influence in order to avoid stray light to result, whole process carry out in darkroom.
(1) Raman spectrum analysis is carried out to data：
Fig. 1 is the Raman spectrogram of the cholecalciferol-cholesterol sample of the four kinds of different contents measured, and cholecalciferol-cholesterol sample exists as shown in Figure 1
960cm-1、1070cm-1、1225cm-1、1437cm-1、1638cm-1There is apparent raman characteristic peak, and when concentration reduces, most
Strong raman characteristic peak is from 1638cm-1Become 1646cm-1.It is found by inquiring literature survey, cholesterol and vitamin D3Drawing
Graceful spectrum is in 1646cm-1There is stronger raman characteristic peak at place, with the reduction of cholecalciferol-cholesterol concentration in sample, free cholesterol and
Vitamin D3Concentration increase, lead to original 1638cm in low concentration sample Raman spectrum-1The cholecalciferol-cholesterol raman characteristic peak at place
It disappears, in 1646cm-1Place shows raman characteristic peak.It is by above-mentioned Raman spectrum analysis it is found that traditional special with sample Raman
Sign peak is not suitable for the quantitative analysis of cholecalciferol-cholesterol content as the method that training data establishes analysis Quantitative Analysis Model.
In the present embodiment, the Portable Raman spectrometer uses 785nm laser for excitation light source, laser maximum output
Power 300mW, 200~2700cm of spectral region-1, 6~8.5cm of resolution ratio-1。
(2) the offset minimum binary quantitative analysis method based on the full modal data of Raman：
According to above-mentioned, with the reduction of sample concentration, the cholecalciferol-cholesterol sample Raman spectrum measured is in 1638cm-1The Raman at place
Characteristic peak can disappear, and in 1646cm-1Place shows raman characteristic peak.Therefore selection a few features peak progress modeling analysis is difficult
Obtain accurately and effectively conclusion.For this problem, the progress of Raman spectrum combination Partial Least Squares is had chosen in the present embodiment
The method of full spectrum analysis establishes Quantitative Analysis Model, efficiently solves the influence of Raman spectrum spectral peak offset and intensity error.
(3) forecast analysis model modelling approach：
Full spectrum analysis is carried out to the concentration of cholecalciferol-cholesterol using Raman spectroscopy combination partial least squares algorithm, it is specific
Method flow diagram, as shown in Fig. 2, being divided into two steps：
The first step is structure training sample set model：Including wavelength band select with establish Quantitative Analysis Model and etc. and
Process.
It is characterized in that wavelength band selects full spectral region, to the full modal data of collected Raman spectrum rather than Individual features
Peak data is trained, and establishes Quantitative Analysis Model, and there is no need to extract Raman spectrum characteristic peak.
The process for establishing Quantitative Analysis Model is：By taking Partial Least Squares as an example, first to the Raman spectrum matrix of input into
Row principal component decomposition, while the concentration matrix principal component of input is decomposed, then utilize spectrum principal component scores matrix and dense
Degree principal component scores matrix is returned.The formula decomposed to spectrum matrix X and concentration matrix Y is as follows：
T and U is respectively the score matrix of X and Y in formula, and P and Q are respectively the loading matrix of X and Y, EXAnd EYRespectively transport
The error introduced when with Partial Least Squares fit metric X and Y.
Characteristic spectrum matrix T and characteristic concentration matrix U are obtained after decomposition, then by characteristic spectrum matrix T and characteristic concentration
Matrix U is returned, and formula is：U=TB
To obtain incidence matrix B：B=(TTT)-1TTU
Second step is to verify Quantitative Analysis Model by test sample, and test sample is substituted into the model established and is carried out in advance
Survey and Fitting Analysis, including wavelength band selection and Quantitative Analysis Model prediction steps, Pretreated spectra is identical with the first step, special
Sign is that wavelength band selects full spectral region, is carried out to the full modal data of collected Raman spectrum rather than Individual features peak data
Training, establishes Quantitative Analysis Model, there is no need to extract Raman spectrum characteristic peak.
X in figure0xIt is the original Raman spectrum of training set sample, X1sIt is that training set after series of preprocessing is returned
One changes Raman spectrum, XsThe Raman spectrum being to determine after wavelength band, YsIt is the known concentration of training set sample, X0tIt is to survey
The original Raman spectrum of examination collection sample, X1tIt is the test set normalization Raman spectrum pre-processed, XtBe to determine wavelength band it
Test set Raman spectrum afterwards, YtIt is the prediction concentrations of the test set sample provided after Quantitative Analysis Model is analyzed.
In order to which noise and laser power shake for eliminating CCD noises, Acquisition Circuit etc. is to the shadow of Raman signal intensity
It rings, using three kinds of baseline fitting and correction, smothing filtering, spectrum area normalization preconditioning techniques to the original Raman light that measures
Spectrum carries out data prediction.In the present embodiment, Pretreated spectra includes baseline correction, smooth filter and three mistakes of area normalization
Journey：
A) baseline correction is carried out to original Raman spectrum data.
For baseline correction flow as shown in figure 3, R (v) is original Raman signal intensity, P (v) is single order polynomial fitting, RS
(v) it is finally obtained raman spectral signal intensity, U (v) is residual error, and error function E (w) is with root-mean-square error (RMS) come table
Show, then：
U (v)=R (v)-P (v) (1)
Error function E (w) is
N is data point number in spectral information, and v indicates that Raman shift, (2) formula refer to calculating process and traverse entire spectrum, mark
Quasi- difference E (w) is considered as the approximation of noise level.
Raman original signal is assigned to R0(v), it while by this calculating is recorded as first time iteration, obtains P1(v)。
Judge whether result of calculation is eligible
Ri(v)<Pi(v)+E(w) (4)
If eligible
Ri(v)=Ri-1(v) (5)
Conversely, then
Ri(v)=Pi(v)+E(w) (6)
Wherein i indicates iterations, works as Ri(v) result of calculation meets the condition of convergence
Pi(v)=Pi-1(v) (7)
Iterative process terminates, and finally obtains Raman data RS (v).
RS (v)=P0(v)-Pi(v) (8)
B) smothing filtering is carried out to spectroscopic data.Using SGF filter methods (Savitzky-Golay Filter), by right
The data of window interior do polynomial least mean square fitting, obtain the value after pixel is smooth at window center position, are passed through at wavelength k
Average value after convolution is smooth is：
In formula, hiFor smoothing factor, H is normalization factor,Wherein hiPassed through based on principle of least square method
Fitting of a polynomial acquires.
C) area normalization is carried out to spectroscopic data, calculation formula is：
In formula, f (x0) it is former spectral intensity, y is to normalize transformed raman spectrum strength, and a, b are to need to normalize
Starting Raman-shifted wavenumbers and cut-off Raman-shifted wavenumbers.
(4) model result is analyzed：
39 groups of Raman spectrum datas of measure 13 various concentration cholecalciferol-cholesterol samples are divided into training set and forecast set.Often
A concentration samples take two groups of spectroscopic datas, and totally 26 groups of spectroscopic datas are as training set, using partial least squares algorithm to training set
Data establish the correspondence model of concentration and Raman spectrum；Remaining 13 groups of spectroscopic datas are brought into foundation as forecast set sample
Prediction and Fitting Analysis are carried out in model.Obtaining that the results are shown in Figure 4, small circle is with actual content (%) for abscissa,
Prediction content (%) is the dispersion scatterplot that ordinate obtains, and figure bend is the y=x function curves fitted, which intends
Close goodness R2It is 0.999, the concentration of prediction and the linearly dependent coefficient R of actual concentrations are 0.9994.
In conclusion the present invention can with quick nondestructive detect the spectral signal of different content cholecalciferol-cholesterol, after treatment
Raman spectrum data, establish Quantitative Analysis Model in conjunction with the method that Partial Least Squares carries out full spectrum analysis, it is effective to solve
Spectral peak offset and the influence of intensity error, may be implemented the prediction of cholecalciferol-cholesterol content.Forecast analysis mould in above-described embodiment
Type realizes prediction, the right R of Linear Quasi of prediction result and actual result to the cholecalciferol-cholesterol sample of 13 different contents2Reach
99.9%.Demonstrate validity of the full spectrum analysis method based on Raman spectrum in the analysis of cholecalciferol-cholesterol content quantitative.
Specific embodiments of the present invention are described above.It is to be appreciated that the invention is not limited in above-mentioned
Particular implementation, those skilled in the art can make various deformations or amendments within the scope of the claims, this not shadow
Ring the substantive content of the present invention.
Claims (10)
1. a kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman spectrum, which is characterized in that including such as
Lower step：
Step 1, cholesterol and vitamin mixtures are configured to the sample of different quality degree cholecalciferol-cholesterol with cholecalciferol-cholesterol；
Step 2, the spectral signal that above-mentioned different content cholecalciferol-cholesterol sample is measured by Raman spectrometer, obtains after pretreatment
Raman spectrum data；
Step 3, using Multielement statistical analysis method, to treated, the full modal data progress analyzing processing foundation of Raman spectrum is quantitative
Analysis model；
Step 4, cholecalciferol-cholesterol quantitative analysis is carried out to the sample of unknown content using Quantitative Analysis Model.
2. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 1 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, described using Multielement statistical analysis method, to treated, the full modal data progress analyzing processing foundation of Raman spectrum is fixed
Analysis model is measured, refers to：The spectroscopic data of pretreated sample is first divided into training sample set and test sample collection, is then divided
At two steps：
The first step is built model using training sample set, the correspondence model of concentration and Raman spectrum is established using training sample set,
It is selected including wavelength band and establishes Quantitative Analysis Model step；
Second step verifies model by test sample, test sample is substituted into the model established and carries out prediction and Fitting Analysis,
Including wavelength band selection and Quantitative Analysis Model prediction steps.
3. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 2 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the Multielement statistical analysis method is any in Partial Least Squares PLS, PCA-LDA, clustering.
4. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 1 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, in step 2：The pretreatment includes baseline correction, smothing filtering and area normalization, these three processing are individually held
Row executes successively in sequence.
5. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 4 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the baseline correction, refers to that baseline correction is carried out to untreated original Raman spectrum data；
Baseline correction flow is：If R (v) is original Raman signal intensity, P (v) is single order polynomial fitting, and U (v) is residual error,
RS (v) is finally obtained raman spectral signal intensity, and error function E (w) is indicated with root-mean-square error (RMS), then：
U (v)=R (v)-P (v) (1)
Error function E (w) is
N is data point number in spectral information, and v indicates that Raman shift, (2) formula refer to calculating process and traverse entire spectrum, standard deviation E
(w) it is considered as the approximation of noise level；
Raman original signal is assigned to R0(v), it while by this calculating is recorded as first time iteration, obtains P1(v)；
Judge whether result of calculation is eligible
Ri(v)<Pi(v)+E(w) (4)
If eligible
Ri(v)=Ri-1(v) (5)
Conversely, then
Ri(v)=Pi(v)+E(w) (6)
Wherein i indicates iterations, works as Ri(v) result of calculation meets the condition of convergence
Pi(v)=Pi-1(v) (7)
Iterative process terminates, and finally obtains Raman data RS (v)：
RS (v)=P0(v)-Pi(v) (8)。
6. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 4 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the smothing filtering, refers to that smothing filtering is carried out to the spectroscopic data after baseline correction；
Using SGF filter methods (Savitzky-Golay Filter), multinomial least square is done by the data to window interior
Fitting, obtains the value after pixel is smooth at window center position, and the average value at wavelength k after convolution is smooth is：
In formula, hiFor smoothing factor, H is normalization factor,Wherein hiPassed through based on principle of least square method more
Item formula fitting acquires, and 2 ω are window width, xk+iFor spectral intensity at wavelength k+i.
7. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 4 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the area normalization, refers to that area normalization is carried out to the spectroscopic data after smothing filtering, calculation formula is：
In formula, f (x0) it is former spectral intensity, y is to normalize transformed raman spectrum strength, and a, b are to need normalized starting
Raman-shifted wavenumbers and cut-off Raman-shifted wavenumbers.
8. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 2 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the wavelength band selection, is the full spectral region of selection, without extracting Raman spectrum characteristic peak.
9. according to lossless quick cholecalciferol-cholesterol content quantitative of the claim 1-8 any one of them based on the full spectrum analysis of Raman spectrum
Method, which is characterized in that the different content cholecalciferol-cholesterol, specially：
It is 1 to take mass ratio:1 cholesterol and vitamin is uniformly mixed, and mixture is made；Cholecalciferol-cholesterol and mixture are configured to courage
Tie up the mixing sample of fourth mass percentage content different weight percentage content between 0~100%.
10. the lossless quick cholecalciferol-cholesterol content quantitative method according to claim 1 based on the full spectrum analysis of Raman spectrum,
It is characterized in that, the Raman spectrometer is the micro-Raman spectroscopy either Portable Raman spectrometer of large size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810219592.4A CN108802000A (en) | 2018-03-16 | 2018-03-16 | A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810219592.4A CN108802000A (en) | 2018-03-16 | 2018-03-16 | A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108802000A true CN108802000A (en) | 2018-11-13 |
Family
ID=64095235
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810219592.4A Pending CN108802000A (en) | 2018-03-16 | 2018-03-16 | A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108802000A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109557071A (en) * | 2018-11-14 | 2019-04-02 | 公安部第研究所 | A kind of Raman spectra qualitative quantitative identification method of dangerous liquid mixture |
| CN111650184A (en) * | 2020-07-13 | 2020-09-11 | 中国科学院烟台海岸带研究所 | Raman spectrum-based combustible liquid rapid measurement method |
| CN112268889A (en) * | 2020-11-13 | 2021-01-26 | 纽斯葆广赛（广东）生物科技股份有限公司 | Method for measuring content of vitamin D3 based on Raman spectroscopy |
| CN112712108A (en) * | 2020-12-16 | 2021-04-27 | 西北大学 | Raman spectrum multivariate data analysis method |
| CN113109318A (en) * | 2021-03-26 | 2021-07-13 | 中国科学院西安光学精密机械研究所 | Raman spectrum quantitative analysis method and system based on spectral peak height direct extraction |
| CN113218932A (en) * | 2021-04-30 | 2021-08-06 | 赵艳茹 | Discrimination method for detecting content of titanium dioxide in persimmon frost based on Raman spectrum |
| CN113640276A (en) * | 2021-08-06 | 2021-11-12 | 北京科技大学 | Method for representing absorption and desorption process of corrosion inhibitor |
| CN113655050A (en) * | 2021-08-17 | 2021-11-16 | 南京富岛信息工程有限公司 | Method for improving Raman spectrum detection limit of trace crude oil in light oil |
| CN114651218A (en) * | 2019-11-15 | 2022-06-21 | 赛多利斯司特蒂姆数据分析公司 | Method and device assembly for predicting parameters in a biological process based on raman spectroscopy and method and device assembly for controlling a biological process |
| CN114674803A (en) * | 2022-02-16 | 2022-06-28 | 北京化工大学 | Quantitative analysis method for components of marine environment micro-plastic mixture |
| CN115736290A (en) * | 2022-11-07 | 2023-03-07 | 南京邦康生物技术有限公司 | Production process control method of product for increasing bone mineral density |
| CN116959628A (en) * | 2023-07-25 | 2023-10-27 | 安及义实业（上海）有限公司 | Method and device for analyzing substance components in whole cell culture process |
| CN117288739A (en) * | 2023-11-27 | 2023-12-26 | 奥谱天成（厦门）光电有限公司 | Asymmetric Raman spectrum baseline correction method, device and storage medium |
| US12019024B2 (en) | 2019-11-15 | 2024-06-25 | Sartorius Stedim Data Analytics Ab | Method and device assembly for predicting a parameter in a bioprocess based on Raman spectroscopy and method and device assembly for controlling a bioprocess |
| CN114674803B (en) * | 2022-02-16 | 2024-07-12 | 北京化工大学 | Quantitative analysis method for components of marine environment microplastic mixture |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101403696A (en) * | 2008-10-21 | 2009-04-08 | 浙江大学 | Method for measuring gasoline olefin content based on Raman spectrum |
| US20120065899A1 (en) * | 2010-02-25 | 2012-03-15 | Olympus Corporation | Analysis method of vibrational spectra |
| CN104777149A (en) * | 2015-04-17 | 2015-07-15 | 浙江大学 | Method for rapidly measuring content of trace methylbenzene in benzene based on Raman spectrum |
| CN105588828A (en) * | 2015-12-16 | 2016-05-18 | 新希望双喜乳业（苏州）有限公司 | Method for quickly detecting butterfat content through Raman spectra |
-
2018
- 2018-03-16 CN CN201810219592.4A patent/CN108802000A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101403696A (en) * | 2008-10-21 | 2009-04-08 | 浙江大学 | Method for measuring gasoline olefin content based on Raman spectrum |
| US20120065899A1 (en) * | 2010-02-25 | 2012-03-15 | Olympus Corporation | Analysis method of vibrational spectra |
| CN104777149A (en) * | 2015-04-17 | 2015-07-15 | 浙江大学 | Method for rapidly measuring content of trace methylbenzene in benzene based on Raman spectrum |
| CN105588828A (en) * | 2015-12-16 | 2016-05-18 | 新希望双喜乳业（苏州）有限公司 | Method for quickly detecting butterfat content through Raman spectra |
Non-Patent Citations (1)
| Title |
|---|
| 戴健等: "《胆维丁原料中共晶与游离态维生素D3和胆固醇含量的定量分析》", 《上海医药》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109557071B (en) * | 2018-11-14 | 2021-12-17 | 公安部第一研究所 | Raman spectrum qualitative and quantitative identification method for dangerous liquid mixture |
| CN109557071A (en) * | 2018-11-14 | 2019-04-02 | 公安部第研究所 | A kind of Raman spectra qualitative quantitative identification method of dangerous liquid mixture |
| CN114651218B (en) * | 2019-11-15 | 2023-09-15 | 赛多利斯司特蒂姆数据分析公司 | Method and device assembly for predicting parameters in a biological process based on raman spectroscopy, and method and device assembly for controlling a biological process |
| US12019024B2 (en) | 2019-11-15 | 2024-06-25 | Sartorius Stedim Data Analytics Ab | Method and device assembly for predicting a parameter in a bioprocess based on Raman spectroscopy and method and device assembly for controlling a bioprocess |
| CN114651218A (en) * | 2019-11-15 | 2022-06-21 | 赛多利斯司特蒂姆数据分析公司 | Method and device assembly for predicting parameters in a biological process based on raman spectroscopy and method and device assembly for controlling a biological process |
| CN111650184A (en) * | 2020-07-13 | 2020-09-11 | 中国科学院烟台海岸带研究所 | Raman spectrum-based combustible liquid rapid measurement method |
| CN112268889A (en) * | 2020-11-13 | 2021-01-26 | 纽斯葆广赛（广东）生物科技股份有限公司 | Method for measuring content of vitamin D3 based on Raman spectroscopy |
| CN112712108A (en) * | 2020-12-16 | 2021-04-27 | 西北大学 | Raman spectrum multivariate data analysis method |
| CN112712108B (en) * | 2020-12-16 | 2023-08-18 | 西北大学 | Raman spectrum multivariate data analysis method |
| CN113109318A (en) * | 2021-03-26 | 2021-07-13 | 中国科学院西安光学精密机械研究所 | Raman spectrum quantitative analysis method and system based on spectral peak height direct extraction |
| CN113218932A (en) * | 2021-04-30 | 2021-08-06 | 赵艳茹 | Discrimination method for detecting content of titanium dioxide in persimmon frost based on Raman spectrum |
| CN113640276A (en) * | 2021-08-06 | 2021-11-12 | 北京科技大学 | Method for representing absorption and desorption process of corrosion inhibitor |
| CN113655050A (en) * | 2021-08-17 | 2021-11-16 | 南京富岛信息工程有限公司 | Method for improving Raman spectrum detection limit of trace crude oil in light oil |
| CN113655050B (en) * | 2021-08-17 | 2024-04-26 | 南京富岛信息工程有限公司 | Method for improving Raman spectrum detection limit of trace crude oil in light oil |
| CN114674803A (en) * | 2022-02-16 | 2022-06-28 | 北京化工大学 | Quantitative analysis method for components of marine environment micro-plastic mixture |
| CN114674803B (en) * | 2022-02-16 | 2024-07-12 | 北京化工大学 | Quantitative analysis method for components of marine environment microplastic mixture |
| CN115736290A (en) * | 2022-11-07 | 2023-03-07 | 南京邦康生物技术有限公司 | Production process control method of product for increasing bone mineral density |
| CN116959628A (en) * | 2023-07-25 | 2023-10-27 | 安及义实业（上海）有限公司 | Method and device for analyzing substance components in whole cell culture process |
| CN117288739A (en) * | 2023-11-27 | 2023-12-26 | 奥谱天成（厦门）光电有限公司 | Asymmetric Raman spectrum baseline correction method, device and storage medium |
| CN117288739B (en) * | 2023-11-27 | 2024-02-02 | 奥谱天成（厦门）光电有限公司 | Asymmetric Raman spectrum baseline correction method, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108802000A (en) | A kind of lossless quick cholecalciferol-cholesterol content quantitative method based on the full spectrum analysis of Raman | |
| Jiang et al. | Comparison of algorithms for wavelength variables selection from near-infrared (NIR) spectra for quantitative monitoring of yeast (Saccharomyces cerevisiae) cultivations | |
| WO2018018838A1 (en) | Raman spectrum analysis-based method for quickly detecting quality of coal | |
| CN104251839B (en) | Spectrum separation detection method of compositions of south-Xinjiang red date sample for south-Xinjiang red date modeling | |
| US20050149300A1 (en) | Method and apparatus for enhanced estimation of an analyte property through multiple region transformation | |
| JP6196220B2 (en) | Use of nuclear magnetic resonance and near infrared for biological sample analysis | |
| CN101915744A (en) | Near infrared spectrum nondestructive testing method and device for material component content | |
| CN106248610A (en) | Dynamic, the careless cultivar identification of multiple spot based on terahertz time-domain spectroscopy and authentication method | |
| CN104062258B (en) | Method for rapid determination of soluble solids in compound ass-hide glue pulp by near infrared spectroscopy | |
| CN109085136B (en) | Method for measuring content of oxide components in cement raw material by near-infrared diffuse reflection spectrum | |
| CN113008805B (en) | Radix angelicae decoction piece quality prediction method based on hyperspectral imaging depth analysis | |
| Cao et al. | Terahertz spectral unmixing based method for identifying gastric cancer | |
| Huynh et al. | The ATLAS 5.5 GHz survey of the extended Chandra Deep Field South: catalogue, source counts and spectral indices | |
| CN105486663B (en) | A method of detecting the stable carbon isotope ratio of soil using near infrared spectrum | |
| CN105092513A (en) | Method for determining content of drug components in mixture drug | |
| CN104596975A (en) | Method for measuring lignin of reconstituted tobacco by paper-making process by virtue of near infrared reflectance spectroscopy technique | |
| Ma et al. | Raman spectroscopy for pharmaceutical quantitative analysis by low-rank estimation | |
| CN107076664A (en) | Method for correcting infrared absorption spectroscopy | |
| CN107247033B (en) | Identify the method for Huanghua Pear maturity based on rapid decay formula life cycle algorithm and PLSDA | |
| Zhao et al. | Data fusion of Laser-induced breakdown spectroscopy and Near-infrared spectroscopy to quantitatively detect heavy metals in lily | |
| CN103760130A (en) | Method for measuring tween-80 content in compound musk injection by using near-infrared spectrum | |
| US20110213746A1 (en) | Probabilistic scoring for components of a mixture | |
| CN104596976A (en) | Method for determining protein of paper-making reconstituted tobacco through ear infrared reflectance spectroscopy technique | |
| CN105699314B (en) | A method of detecting soil stabilization carbon isotope ratio using middle infrared spectrum | |
| CN109520950A (en) | The insensitive chemical component spectral method of detection of a kind of pair of spectral shift |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |