CN109444066A - Model transfer method based on spectroscopic data - Google Patents
Model transfer method based on spectroscopic data Download PDFInfo
- Publication number
- CN109444066A CN109444066A CN201811271230.6A CN201811271230A CN109444066A CN 109444066 A CN109444066 A CN 109444066A CN 201811271230 A CN201811271230 A CN 201811271230A CN 109444066 A CN109444066 A CN 109444066A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- sample
- instrument
- model
- target
- 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
- 238000012546 transfer Methods 0.000 title claims abstract description 135
- 238000000034 method Methods 0.000 title claims abstract description 89
- 238000004611 spectroscopical analysis Methods 0.000 title claims abstract description 40
- 238000001228 spectrum Methods 0.000 claims abstract description 156
- 230000007704 transition Effects 0.000 claims abstract description 17
- 230000002441 reversible effect Effects 0.000 claims abstract description 6
- 238000002329 infrared spectrum Methods 0.000 claims description 38
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000012937 correction Methods 0.000 claims description 34
- 230000003595 spectral effect Effects 0.000 claims description 32
- 238000004164 analytical calibration Methods 0.000 claims description 18
- 238000005259 measurement Methods 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 238000011156 evaluation Methods 0.000 claims description 5
- 238000004566 IR spectroscopy Methods 0.000 claims description 4
- 230000001419 dependent effect Effects 0.000 claims description 4
- 230000017105 transposition Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000001237 Raman spectrum Methods 0.000 claims description 2
- 238000009499 grossing Methods 0.000 claims description 2
- 238000000655 nuclear magnetic resonance spectrum Methods 0.000 claims description 2
- 238000002371 ultraviolet--visible spectrum Methods 0.000 claims description 2
- 238000007796 conventional method Methods 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 15
- 239000008186 active pharmaceutical agent Substances 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 229940107666 astragalus root Drugs 0.000 description 4
- 238000002835 absorbance Methods 0.000 description 3
- QMNWISYXSJWHRY-YLNUDOOFSA-N astragaloside IV Chemical compound O1[C@H](C(C)(O)C)CC[C@]1(C)[C@@H]1[C@@]2(C)CC[C@]34C[C@]4(CC[C@H](O[C@H]4[C@@H]([C@@H](O)[C@H](O)CO4)O)C4(C)C)[C@H]4[C@@H](O[C@H]4[C@@H]([C@@H](O)[C@H](O)[C@@H](CO)O4)O)C[C@H]3[C@]2(C)C[C@@H]1O QMNWISYXSJWHRY-YLNUDOOFSA-N 0.000 description 3
- QMNWISYXSJWHRY-BCBPIKMJSA-N astragaloside IV Natural products CC(C)(O)[C@@H]1CC[C@@](C)(O1)[C@H]2[C@@H](O)C[C@@]3(C)[C@@H]4C[C@H](O[C@@H]5O[C@H](CO)[C@H](O)[C@@H](O)[C@H]5O)[C@H]6C(C)(C)[C@H](CC[C@@]67C[C@@]47CC[C@]23C)O[C@@H]8OC[C@@H](O)[C@H](O)[C@H]8O QMNWISYXSJWHRY-BCBPIKMJSA-N 0.000 description 3
- PFKIBRPYVNVMRU-UHFFFAOYSA-N cyclosieversioside F Natural products CC(C)(O)C1COC(C)(C1)C2C(O)CC3(C)C4CC(OC5OC(CO)C(O)C(O)C5O)C6C(C)(C)C(CCC67CC47CCC23C)OC8OCC(O)C(O)C8O PFKIBRPYVNVMRU-UHFFFAOYSA-N 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000004497 NIR spectroscopy Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000002401 inhibitory effect Effects 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000010561 standard procedure Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- SMDOOINVMJSDPS-UHFFFAOYSA-N Astragaloside Natural products C1=C(O)C(OC)=CC(C2=C(C(=O)C3=C(O)C=C(O)C=C3O2)OC2C(C(OC3C(C(O)C(O)C(CO)O3)O)C(O)C(CO)O2)O)=C1 SMDOOINVMJSDPS-UHFFFAOYSA-N 0.000 description 1
- 239000009636 Huang Qi Substances 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003556 assay Methods 0.000 description 1
- QMNWISYXSJWHRY-XWJCTJPOSA-N astragaloside Chemical compound O1[C@H](C(C)(O)C)CC[C@]1(C)[C@@H]1[C@@]2(C)CC[C@]34C[C@]4(CC[C@H](O[C@H]4[C@@H]([C@@H](O)[C@H](O)CO4)O)C4(C)C)C4[C@@H](O[C@H]4[C@@H]([C@@H](O)[C@H](O)[C@@H](CO)O4)O)CC3[C@]2(C)C[C@@H]1O QMNWISYXSJWHRY-XWJCTJPOSA-N 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004587 chromatography analysis Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 238000004128 high performance liquid chromatography Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000007791 liquid phase Substances 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010238 partial least squares regression Methods 0.000 description 1
- 239000012071 phase Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012628 principal component regression Methods 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000004454 trace mineral analysis Methods 0.000 description 1
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
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
-
- 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
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (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 Or Analysing Materials By Optical Means (AREA)
Abstract
The embodiment of the present application discloses the Model transfer method based on spectroscopic data, the index determining and spectra collection of sample；The pretreatment of spectrum；Several described calibration samples are divided into calibration samples and forecast sample；Model transfer, i.e. forward transfer are carried out from main instrument to target instrument；Alternatively, carrying out Model transfer, i.e. reverse transition from target instrument to main instrument；Conventional method of the present invention is compared, and is had the advantages that being associated with by sample set spectroscopic data and spectroscopic data structure, can be inhibited the interference of noise and irrelevant information；It can be realized the bi-directional of main instrument and target instrument model, and no limitation whether consistent with the instrument type of target instrument for main instrument, whether unanimously without limitation, this is of great significance for the model sharing of same spectra method different type spectrometer and the applicability of raising model to the spectroscopic data points of acquisition.
Description
Technical field
The invention relates to Infrared Spectrum Technology fields, more particularly to the Model transfer based on spectroscopic data
Method.
Background technique
Modern spectral analysis technique is multiple subjects (spectral detection, Chemical Measurement, computer science and statistics etc.)
The comprehensive analysis technology of mixing together.Spectral analysis technique is answered extensively in petrochemical industry, agricultural, food and pharmaceutical field
With.It is measured using spectra methods for material component content and online process analysis procedure analysis needs to establish multivariate calibration model, that is, adopted
With the index (such as component content or activity value or other physical and chemical index) of standard method measurement calibration samples, then use
Spectrometer measures the spectrum of these samples, and spectroscopic data is associated with achievement data and establishes calibration model, by the light of unknown sample
Spectrum substitutes into calibration model, calculates the predicted value of unknown sample index, realizes the measurement to unknown sample index.
In practical applications, the model for establishing spectrum needs to put into a large amount of human and material resources, costly.If by
The model being set up is applied to other instruments or continues to use, then not only saving a large amount of human and material resources, but also for realizing
Sharing for model, improves the versatility of model, applicability has great importance.With the diversification of spectral instrument and using ring
The complication in border, type, the model such as instrument are identical or different, the aging of instrument and the replacement of component, sample measures environment
The factors such as the variation of temperature, humidity so that acquisition spectrum be usually present the difference of spectral absorbance (or intensity), wavelength (or
Wave number) drift the problems such as, if continuing to use established model, prediction result deviation is larger, and even resulting in model can not
It uses.In order to solve model sharing, this technical bottleneck problem of model commonality is improved, Model transfer or transmitting are a kind of effective
Method.Model transfer is the functional relation by establishing between main instrument and target instrument (or sub- instrument) detection signal, is realized
The consistency and accuracy of main instrument and target instrument model prediction result.According to the affiliated partner of built function, Model transfer
Method is generally divided into three classes: first is that the bearing calibration based on prediction result, such as Tom Fearn, Standardisation and
Calibration Transfer for near Infrared Instruments:A Review,Journal of Near
Infrared Spectroscopy Vol.9, Issue 4, the S/B method that pp.229-244 (2001) is referred to；Second is that being based on light
The bearing calibration of spectrum signal, such as the methods of Shenks, DS, PDS, FIR, OSC；Third is that M.Forina, G.Drava,
C.Armanino, R.Boggia, S.Lanteri, R.Leardi, P.Corti, P.Conti, R.Giangiacomo,
C.Galliena, R.Bigoni, I.Quartari, C.Serra, D.Ferri, O.Leoni, L.Lazzeri, Transfer of
Calibration function in near-infrared spectroscopy, Chemometrics and
Intelligent Laboratory Systems, Volume 27, Issue 2, February 1995, Pages 189-203.
The bearing calibration based on model parameter referred to.Wherein, the DS method PDS method based on spectral signal correction is widely used.
DS method and PDS method have belonged to standard specimen method.Standard specimen, that is, master sample is during Model transfer, to establish
Master sample required for transfer function.It in practical applications, can be using no mark if can not or be difficult to obtain master sample
Quadrat method, such as finite impulse response (FIR) method (FIR).DS method is the association directly by main instrument and target instrument, then by most
Small two multiply principle solving transfer matrix, establish functional relation.Since the variation of spectrum has regionality, under different wave length, spectrum
Variation may be inconsistent；And noise also has region inequality, so that effect of the DS method for Model transfer is not ideal enough.
PDS method assumes that the variation primary limitation of spectrum carries out model turn in certain region, therefore using the local spectrum in window
It moves, Model transfer effect obtains a degree of raising.But this method is complex for the adjusting of window size, it can not
Adaptively, if window size parameter setting is improper, transfer effect is bad, and local noise also has certain influence.
Summary of the invention
In order to solve the deficiencies in the prior art, the embodiment of the present application provides the Model transfer method based on spectroscopic data,
By the spectroscopic data structure connection of the spectroscopic data of main instrument and target instrument (or by the spectroscopic data of target instrument and main instrument
Spectroscopic data structure connection), have inhibiting effect to the interference of noise and redundancy feature, can be realized main instrument and target instrument
The bi-directional of model, and no limitation whether consistent with the instrument type of target instrument for main instrument, the spectrum number of acquisition
Whether strong point number is unanimously without limitation；The spectrum of low resolution can be converted to the light of high-resolution by the method for the invention
Spectrum, has expanded the application range of Model transfer.
A kind of Model transfer method based on spectroscopic data is provided as the embodiment of the present application；
A kind of Model transfer method based on spectroscopic data, comprising:
Step (1): it the index determining and spectra collection of sample: for several calibration samples, is detected using professional standard
Method measures n index；It is utilized respectively the acquisition that main instrument and target instrument carry out spectrum to several described calibration samples, is obtained
The spectrum X acquired to main instrumentMWith the spectrum X of target instrument acquisitionT；
Step (2): the pretreatment of spectrum: to the spectrum X of main instrument acquisitionMWith the spectrum X of target instrument acquisitionTIt adopts respectively
It is pre-processed in the same way；
Step (3): several described calibration samples are divided into calibration samples and forecast sample, according to calibration samples and in advance
This spectrum X for acquiring main instrument of test sampleMIt is divided into calibration set sample and forecast set sample；According to calibration samples and forecast sample
The spectrum X that target instrument is acquiredTIt is divided into calibration set sample and forecast set sample；
Step (4): Model transfer, i.e. forward transfer are carried out from main instrument to target instrument:
Step (401): it selects the sample of setting quantity to collect sample as transfer in main instrumental correction collection sample, then exists
Selected in target instrument calibration set the sample of same sequence number as target instrument shift collection sample, the transfer collection sample of main instrument with
Target instrument transfer collection sample serial number corresponds；Main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage
It is adjusted to shift spectrum X in collection sample with target instrumentTSpectral variables it is consistent with wave-length coverage；
Step (402): main instrument is shifted to the spectrum intensity data of collection sample and the spectrum of target instrument transfer collection sample
Intensity data is associated to obtain relational model to get transition matrix is arrived；
Step (403): substituting into transition matrix for the spectroscopic data of main instrumental correction collection sample, obtains being suitble to target instrument school
The spectroscopic data of positive collection sample；
Step (404): the spectroscopic data based on suitable target instrument calibration set sample establishes target instrument calibration model: will
It is suitble to the spectroscopic data of target instrument calibration set sample as independent variable, using n testing index as dependent variable, using polynary school
The spectroscopic data of suitable target instrument calibration set sample is associated with by correction method with n achievement data, establishes n calibration model；
Step (405): carrying out the acquisition of spectrum using target instrument to sample to be tested, then utilizes target instrument straightening die
The index value of type forecast sample.
After step (404), before step (405), further includes: the spectrum of target instrument forecast set is substituted into correction
Model calculates the predicted value of n index of each forecast set sample by calibration model, by predicted value and actual measured value ratio
Compared with progress model evaluation.
Optionally, step (4) replacement are as follows: carry out Model transfer, i.e. reverse transition from target instrument to main instrument:
Step (411): it selects the sample of setting quantity to collect sample as transfer in main instrumental correction collection sample, then exists
Selected in target instrument calibration set the sample of same sequence number as target instrument shift collection sample, the transfer collection sample of main instrument with
Target instrument transfer collection sample corresponds；Main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage adjustment
To shift spectrum X in collection sample with target instrumentTSpectral variables it is consistent with wave-length coverage；
Step (412): target instrument is shifted into collection sample spectrum intensity data and main instrument transfer collection sample spectrum intensity
Data are associated to obtain relational model to get transition matrix is arrived；
Step (413): carrying out the acquisition of spectrum using target instrument to sample to be tested, is based on relational model for target instrument
The spectrum correction of sample to be tested is the sample to be tested spectrum for being suitble to main instrument；
Step (414): calibration model is established using main instrumental correction collection sample spectrum: by the light of main instrumental correction collection sample
Modal data is as independent variable, using n testing index as dependent variable, establishes n calibration model using multivariate calibration methods；It will fit
The spectrum for closing main instrument sample to be tested substitutes into the calibration model of main instrument and predicts the index value of sample to be predicted.
After step (414), may also include that by the spectrum of target instrument forecast set using relational model be corrected to it is suitable
The forecast sample spectrum of main instrument substitutes into calibration model, calculates the pre- of n index of each forecast set sample by calibration model
Measured value compares predicted value and actual measured value, carries out model evaluation.
Optionally, during forward transfer, target instrument is shifted into collection spectrum XTtPrincipal component decomposition is carried out, is obtained:
XTt=STtVTt'； (1)
Wherein, STtIndicate principal component scores matrix, VTtExpression principal component load matrix, symbol " ' " indicate principal component load
The transposition of matrix, subscript " T " indicate that target instrument, subscript " t " indicate transfer collection；
Set number of principal components lT；Main instrument is shifted into collection spectrum XMtWith STtAssociation:
STt=XMtFMt； (2)
Wherein, FMtIndicate that transition matrix, symbol " M " indicate that main instrument, " t " indicate transfer collection；
According to FMt=X+ MtSTtIt calculates and solves, symbol "+" indicates broad sense inverse operation；X+ MtIndicate XMtGeneralized inverse matrix；
Formula (2) are substituted into formula (1), the spectrum after obtaining model conversion:
Xt T=XMFMtVTt'； (3)
Wherein, Xt TIndicate that the spectrum matrix for being suitable for target instrument sample after converting, symbol " t " indicate conversion, symbol " T "
Indicate target instrument, XMIndicate the spectroscopic data based on main Instrument measuring；
According to formula (3), the spectrum of main instrumental correction collection is corrected as:
Xt Tc=XMcFMtVTt'； (4)
By formula (4) by the spectrum X of main instrumental correction collectionMcIt is corrected to the spectrum X of suitable target instrumentt Tc, wherein symbol
" c " indicates calibration set.
Optionally, during reverse transition, main instrument is shifted into collection spectrum XMtPrincipal component decomposition is carried out, is obtained:
XMt=SMtVMt'； (5)
Wherein, SMtIndicate that the score matrix of principal component, " M " indicate that main instrument, " t " indicate transfer collection；VMtIndicate principal component
Load matrix, symbol " ' " transposition of representing matrix；
Set number of principal components lM, target instrument is shifted into collection spectrum XTtWith SMtAssociation:
SMt=XTtFTt； (6)
Wherein, FTtIndicate transition matrix, symbol " T " indicates target instrument, and " t " indicates transfer collection；
Wherein according to FTt=X+ TtSMtIt calculates and solves；X+ TtIndicate XTtGeneralized inverse matrix；
Formula (6) are substituted into formula (5), the spectrum after model conversion:
Xt M=XTFTtVMt'； (7)
Wherein, Xt MIndicate that the spectrum matrix for being suitable for main instrument sample after converting, symbol " t " indicate conversion, symbol " M " table
Show main instrument；XTIndicate the spectroscopic data measured based on target instrument, symbol " T " indicates target instrument；
According to formula (7), the spectrum correction of the forecast set sample of target instrument are as follows:
Xt Mv=XTvFTtVMt'； (8)
Pass through formula (8), the forecast set spectrum X of target instrumentTvIt is corrected to the forecast set spectrum for being suitble to main instrument requirements
Xt Mv, symbol " v " expression forecast set.
For the sample to be tested of target instrument measurement, converted according to formula (9):
Xt Mu=XTuFTtVMt'； (9)
Symbol " u " indicates sample to be tested, XTuIndicate the spectrum of the sample to be tested of target instrument measurement, Xt MuFor XTuAfter conversion
It is suitble to the spectrum of the sample to be tested of main instrument.
Optionally, n is more than or equal to 1 in the step (1)；Main instrument and target instrument belong to the spectrum of spectrum of the same race
Instrument.
The spectrum includes: infrared spectroscopy, ultraviolet-visible spectrum, Raman spectrum or NMR spectrum；The infrared light
Spectrum, including middle infrared spectrum and near infrared spectrum；
Optionally, the beneficial effect of the step (2) is: for main instrument and target instrument, the pretreatment of sample spectrum
Mode is consistent, and can eliminate the interference of Aimless factors in this way, the spectral profile phase for obtaining main instrument and target instrument
Closely.
Optionally, the pretreatment mode of the step (2), comprising: smoothing processing, first derivative calculating, second dervative meter
Calculation, standardization, baseline drift processing, standard normal variable processing, multiplicative scatter correction are handled and are gone in trend processing
The combination of any one or more；
Optionally, the quantity of calibration set described in the step (3) is greater than or equal to the quantity of forecast set；
Optionally, several described calibration samples are divided into calibration samples and forecast sample in the step (3), divided
Mode includes: any one in KS method, Rank-KS method, SPXY method, Rank-SPXY method and concentration gradients method.
Optionally, the quantity of transfer collection sample should not be very little in the step (4), and the sample information for otherwise including not enough is filled
Point；If the sample size for shifting collection is too many, Model transfer effect may be made undesirable；
Optionally, transfer collection quantity is more than or equal to 10 in the step (4), is less than or equal to calibration set quantity；
Optionally, the selection criteria of transfer collection is according to the minimum original of predicted root mean square error RMSEP in the step (4)
Then, select several samples that the RMSEP of the identical index of target instrument is made to reach minimum from the calibration set of main some index of instrument,
Corresponding sample is to shift collection sample at this time；For main instrument and target instrument, the serial number of transfer collection sample is consistent.
Optionally, described that main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage is adjusted to and target
Spectrum X in instrument transfer collection sampleTSpectral variables and the consistent specific steps of wave-length coverage are as follows:
The spectrum X that main instrument is acquiredMSpectral variables be adjusted to target instrument spectrum XTSpectral variables；
Calculate the spectrum X of main instrument acquisitionMWave-length coverage and target instrument acquisition spectrum XTWave-length coverage friendship
Collection；To the spectrum X of the main instrument acquisition in wave-length coverage intersectionMWith the spectrum X of target instrument acquisitionTRetained, to place
In the spectrum X of the main instrument acquisition outside wave-length coverage intersectionMWith the spectrum X of target instrument acquisitionTIt is deleted.
The calculation method of transition matrix includes: principal component regression, in Partial Least Squares Regression in the formula (2) and (6)
It is a kind of.
In the formula (3) and formula (7), the sample number of the spectrum of main instrument and target instrument is must be equal, but spectrum
Variable number can be it is equal, can also be unequal.
Optionally, target instrument or main instrument transfer collection sample spectrum data are subjected to principal component point in the step (4)
Solution, number of principal components be come according to the RMSEP value minimum of Principal Component Explanation variance ability combining target Instrument measuring some index it is true
It is fixed, maximum value XTtOr XMtOrder.
Beneficial effects of the present invention
1. the method for the invention can be realized the two-way transfer of model, and the spectral variables number of main instrument and target instrument
Unanimously whether there is no limit can mutually be converted the spectrum of different resolution, be realized between different resolution spectrometer
Model transfer, therefore the method for the present invention is more flexible, and the scope of application is more extensive.
2. the spectrum before and after Model transfer is established inner link by data internal structure by the method for the invention, therefore
Influence for noise has inhibiting effect, and can preferably retain the characteristic information of initial data, improves the success rate of Model transfer.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 is the original near infrared spectrum of the main instrument of the embodiment of the present application；
Fig. 2 is the original near infrared spectrum of the target instrument of the embodiment of the present application；
Fig. 3 is the near infrared spectrum of the target instrument calibration set sample of the embodiment of the present application；
Fig. 4 is that the main instrumental correction collection sample near infrared spectrum of the embodiment of the present application is corrected through Model transfer method of the present invention
The calibration set sample near infrared spectrum of suitable target instrument afterwards；
Fig. 5 is spectrum and mesh of the main instrumental correction collection near infrared spectrum of the embodiment of the present application after Model transfer of the present invention
The difference of the near infrared spectrum of nonius instrument calibration set sample；
Fig. 6 is that the embodiment of the present application uses the PDS method to shift target instrument forecast set sample spectrum to be suitble to main instrument
The near infrared spectrum of device forecast set；
Fig. 7 is being shifted the near infrared spectrum of target instrument forecast set sample using the method for the present invention of the embodiment of the present application
To be suitble to main instrument forecast set near infrared spectrum.
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Embodiment 1: with a kind of method for transferring near infrared model CN105842190A returned based on spectrum of Chinese invention patent
Comparison；
1. theory analysis:
Universal model transfer method is all to correct target optical spectrum with the spectrum of high quality (high-resolution, high-precision), according to
Patent (CN105842190A) described principle, host near infrared spectrum XmWith slave near infrared spectrum XsIt is low by respectively corresponding to
The expression Z of dimensionmAnd ZsUsing transformation matrix F1It is associated together.Wherein Zm=Am TXm, AmFor the projection vector collection of host, thus
As it can be seen that ZmIt is XmBy what is centainly converted.Due to being solved and being projected using regular least square method under h feature vector
Vector, and then obtain Am, therefore, ZmIt is under h feature vector to XmApproximate representation, that is, pass through Zm=Am TXmAfterwards, ZmRelatively
XmThere is certain information loss.
And in the present invention, we do not carry out any transformation to main instrument spectral (see the near infrared light of transfer collection sample
Compose XMtWith the near infrared spectrum X of main instrument sampleM), therefore the information of main instrument spectral is fully retained, and is not lost.Cause
This theoretically, relative to patent (CN105842190A) the method, the method for the present invention retains main instrument during Model transfer
The spectral information of device sample is more complete, loses less.
2. profile variation compares before and after Model transfer
Fig. 1 is the original near infrared spectrum of main instrument, the original near infrared spectrum of target instrument shown in Fig. 2.
By Fig. 1 and Fig. 2 as it can be seen that for identical sample, the sample measured using main instrument and target instrument is original close red
The difference of external spectrum is very big, is mainly reflected in abscissa representation method and range, absorbance size, the variable numbers of data this three
A aspect.Wave number and wavelength, which is respectively adopted, in main instrument and target instrument indicates abscissa, if being converted into wavelength, main instrument
The wave-length coverage of measurement is 1000-2500nm, and the wave-length coverage of target instrument is 900-1676nm, it is clear that is differed greatly.If
Model transfer is carried out, is needed first that wave-length coverage is unified；The spectral absorbance range of main Instrument measuring is 0.39~1.04, and
The range of absorbency of target instrument spectrum is 0.09-0.34；The spectral variables number of main Instrument measuring is 1557, and target instrument
Spectral variables number be 125.As it can be seen that the model that main instrument is established will be unable to make in target instrument without Model transfer
With.
Fig. 3 illustrates the near infrared spectrum of target instrument calibration set sample, and Fig. 4 illustrates the close of main instrumental correction collection sample
Infrared spectroscopy is suitble to the calibration set sample near infrared spectrum of target instrument after Model transfer.These spectrum are carrying out model turn
Pretreated spectra (Savitzky-Golay Filtering method, window size 15, polynomial order 2) is carried out before moving.
Fig. 3 and SPECTRAL DIVERSITY shown in Fig. 4 that the method for the present invention shown in Fig. 5 obtains.
By Fig. 3 and Fig. 4 as it can be seen that even if main instrument and the original near infrared spectrum of target instrument differ greatly, but main instrument
Calibration set near infrared spectrum is closely similar with the calibration set near infrared spectrum of target instrument after Model transfer of the present invention.Map is poor
Different very small (see Fig. 5), the order of magnitude 10-6.The number for the profile variation that comparison patent (CN105842190A) the method obtains
Magnitude is 10-2(see shown in patent CN105842190A Fig. 3 c).The above results show: Model transfer method energy provided by the invention
The SPECTRAL DIVERSITY of main instrument and target instrument is enough greatly reduced, effect is very significant.
It is to be appreciated that the prediction of unknown sample is the near infrared spectrum of the forecast set sample based on target instrument measurement, because
This is not necessarily to obtain the near infrared spectrum progress Model transfer of main instrument forecast set sample to be suitable for target instrument forecast set sample
Near infrared spectrum.
2 model forward transfer of embodiment
To measure the content of commercially available Astragloside IV in Extraction of Radix Astragali as example.The assay of Astragaloside IV uses HPLC
Liquid-phase chromatography method, specific method are described with reference to Chinese Pharmacopoeia (2015 editions).
I, the index determining and spectra collection of sample
Commercially available Astragalus Root P.E totally 82 samples are collected, with reference to (2015 editions) measurement Astragaloside contents of official method, are adopted
With main instrument (Thermo Fisher Scientific, USA) and target instrument (Viavi 1700, USA) respectively to these samples
Originally near infrared spectrum, X are measuredMIndicate the near infrared spectrum that main instrument measures, XTIndicate the near infrared spectrum that target instrument measures.
The near infrared spectrum difference that main instrument and target instrument measure is as depicted in figs. 1 and 2.
The pretreatment of II, spectrum
For main instrument and target instrument, the pretreatment of sample spectrum is consistent.Using Savitzky-Golay
Filtering method carries out smooth, window as near infrared light spectrum signal of the preprocessing procedures to main instrument and target instrument
Mouth size is 15, polynomial order 2.
The division of III, sample set
Sample set is divided into calibration set and forecast set using common KS method by the spectroscopic data based on sample；Correction
Collecting sample size is 42, and forecast set sample is 40.The serial number of the calibration set sample and forecast set sample of main instrument and target instrument
It should be consistent.
IV, Model transfer and spectrum correction
In calibration set sample, select a certain number of samples as transfer collection sample for constructing transfer matrix, with reality
Existing Model transfer；The quantity of transfer collection sample should not be very little, and the sample information for otherwise including is not enough；If shifting the sample of collection
This quantity is too many, and Model transfer effect may be made undesirable；Transfer collection sample comes from calibration set sample, and quantity is more than or equal to
10, it is less than or equal to calibration set sample size；The determination of transfer collection sample is according to the minimum original of predicted root mean square error RMSEP
Then, right at this time by selecting a certain number of samples that the RMSEP of some index of target instrument is made to reach minimum in calibration set sample
The sample answered is to shift collection sample；For main instrument and target instrument, the serial number of transfer collection sample should be consistent.
Since the sample near infrared spectrum of main instrument and target instrument differs greatly, before Model transfer, need to carry out
The unification of spectral variables and range.As shown in Fig. 1 and Fig. 2, the spectral variables of main instrument are wave number, and range is 4000cm-1~
10000cm-1, and the spectral variables of target instrument are wavelength, range is 900nm-1676nm.Since the prediction of unknown sample is base
It is carried out in the spectrum of target instrument measurement, so for convenience, spectral variables are used uniformly wavelength, range selects spectral wavelength
Intersection part.Convert wavelength for the spectral variables of main instrument, obtaining wave-length coverage is 1000nm-2500nm, then again with mesh
Nonius instrument range compares, and obtaining intersection range is 1000nm~1676nm.
In calibration set sample, a certain number of transfer collection samples is selected to construct Model transfer matrix.Using Kennard-
The sample of different number is collected sample by Stone (KS) division methods, carries out principal component decomposition according to the following formula:
XTt=STtVTt',
In order to preferably keep the spectral information of target instrument, number of principal components is according to can explain XTtA degree of side
It is determined based on difference.In the present embodiment, if transfer sample size is less than 25, number of principal components XTtOrder；If shifting sample
This quantity is greater than 25, and number of principal components is to be set as 25；By the X of main instrumentMtWith STtBy formula STt=XMtFMtAssociation.FMtCalculating can
To be solved using Partial Least Squares, corresponding latent factor number has leaving-one method (leave-one-out) to determine.It attempts
The transfer collection sample of different number, preferred according to RMSEP value minimum principle, determining transfer collection sample.After transfer collection sample determines,
FMtAlso it thereby determines that.Pass through formula Xt Tc=XMcFMtVTt' near infrared spectrum of main instrumental correction collection sample is corrected to suitable mesh
The near infrared spectrum of nonius instrument calibration set.Based on Xt TcEstablish the calibration model of target instrument, and the prediction to target instrument measurement
Collection sample Astragaloside IV index is predicted, predicted root mean square error and the coefficient of determination are calculated, and evaluates forward model transfer with this
Effect.
Astragalus Root P.E shown in table 1 is through the effect assessment (forward model transmitting) before and after this law Model transfer
As shown in Table 1, the spectral resolution that main instrument measures is high, and precision is high, and model prediction result is optimal.Target instrument is surveyed
The spectral resolution obtained is low, before Model transfer, the worst (R of model prediction resultpMinimum, RMSEP is maximum)；And pass through institute of the present invention
After stating method model transfer, the predictive ability of model before Model transfer than being improved, RpIncrease, and RMSEP reduces.This explanation
The spectrum performance of modeling that main instrument spectral passes through the suitable target instrument obtained after Model transfer is improved, and model prediction obtains
Reinforce.
For the method for the invention compared with traditional PDS method, the model prediction result of this law is better than traditional PDS method, explanation
The Model transfer effect of this law is more preferable, the model that main instrument is established can be successfully transferred in target instrument, implementation model
It is shared.
3 model reverse transition of embodiment
Main instrument is shifted into collection sample near infrared spectrum data and (is denoted as XMt) pass through formula XMt=SMtVMt' carry out principal component
It decomposes, and determines suitable number of principal components lM.It determines that method and above-described embodiment 2 are consistent.Shift collection selection method also and
Embodiment 2 is consistent.
By the X of target instrumentTtWith SMtBy formula SMt=XTtFTtAssociation.Wherein FTtCalculating can use Partial Least Squares
It is solved, corresponding latent factor number has leaving-one method (leave-one-out) to determine.Attempt the transfer collection sample of different number
This, preferred according to RMSEP value minimum principle, determining transfer collection sample.After transfer collection sample determines, FTtAlso it thereby determines that.Due to
Unknown sample is predicted by the near infrared spectrum that target instrument measures, therefore, by the forecast set sample light of target instrument
Spectrum passes through formula Xt Mv=XTvFTtVMt' it is corrected to the near infrared spectrum for being suitble to main instrument forecast set sample.Pass through established master
The content of the Astragaloside IV of instrumental correction collection model prediction forecast set sample, and predicted root mean square error and the coefficient of determination are calculated,
Reversed Model transfer effect is evaluated with this.Trace analysis after reversed Model transfer is shown in Fig. 6 and Fig. 7.
By Fig. 6 and Fig. 7 as it can be seen that this law transfer effect is substantially better than traditional PDS method.Since PDS method is based on mesh
The data that nonius instrument and main instrument measure spectral signal are directly corrected, and the spectral signature data point that target instrument measures is few,
Resolution ratio is low, and the spectral signature data point that main instrument measures is more, high resolution.When with PDS timing, due to target instrument
Spectral signature data point is few, and information content is insufficient, so that the spectrum of main instrument is suitble to some abnormal phenomenon occur after correction.And this law
The principal component for directly being shifted collection spectrum with main instrument using target instrument spectrum is associated with (see SMt=XTtFTt), wherein SMtIt combines
The bulk information of main instrument transfer collection sample, when with SMt=XTtFTtIt is associated, as much as possible by the spectral information of target instrument
It is transmitted, calculated result is under least square principle to SMtApproximate evaluation, therefore preferably remain the transfer of main instrument
Collect the information of spectrum.Meanwhile utilizing formula Xt Mv=XTvFTtVMt' when obtaining being suitble to main instrument forecast set spectrum, VMt' it is by main instrument
Device transfer collection spectrum obtains, which also carries the spectral information of main instrument transfer collection, therefore Xt MvIt can preferably remain
The spectral information of main instrument.As seen from Figure 7, the forecast set spectrum that this law Model transfer obtained be suitble to main instrument is smooth, does not go out
Existing abnormal phenomenon.Astragalus Root P.E shown in table 2 is through the effect assessment (reversed Model Transfer) before and after this law Model transfer.
2 Astragalus Root P.E of table is through the effect assessment (reversed Model Transfer) before and after this law Model transfer
As can be seen from Table 2, the Model transfer effect of this law is preferable, predictive ability compared with being improved before Model transfer, tie by prediction
Fruit and main instrument result are close, improve the success rate of Model transfer；And traditional PDS method is existing since exception occurs in map
As, therefore Model transfer is unsuccessful.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (10)
1. a kind of Model transfer method based on spectroscopic data, characterized in that include:
Step (1): the index determining and spectra collection of sample: for several calibration samples, using professional standard detection method
Measure n index；It is utilized respectively the acquisition that main instrument and target instrument carry out spectrum to several described calibration samples, is led
The spectrum X of instrument acquisitionMWith the spectrum X of target instrument acquisitionT；
Step (2): the pretreatment of spectrum: to the spectrum X of main instrument acquisitionMWith the spectrum X of target instrument acquisitionTIt is respectively adopted same
The mode of sample is pre-processed；
Step (3): several described calibration samples are divided into calibration samples and forecast sample, according to calibration samples and pre- test sample
This spectrum X for acquiring main instrumentMIt is divided into calibration set sample and forecast set sample；According to calibration samples and forecast sample by mesh
The spectrum X of nonius instrument acquisitionTIt is divided into calibration set sample and forecast set sample；
Step (4): Model transfer, i.e. forward transfer are carried out from main instrument to target instrument:
Step (401): the sample of setting quantity is selected to collect sample as transfer in main instrumental correction collection sample, then in target
Instrumental correction concentration selects the sample of same sequence number to shift collection sample, the transfer collection sample and target of main instrument as target instrument
Instrument transfer collection sample serial number corresponds；Main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage adjustment
To shift spectrum X in collection sample with target instrumentTSpectral variables it is consistent with wave-length coverage；
Step (402): main instrument is shifted to the spectrum intensity data of collection sample and the spectral intensity of target instrument transfer collection sample
Data are associated to obtain relational model to get transition matrix is arrived；
Step (403): substituting into transition matrix for the spectroscopic data of main instrumental correction collection sample, obtains being suitble to target instrument calibration set
The spectroscopic data of sample；
Step (404): the spectroscopic data based on suitable target instrument calibration set sample establishes target instrument calibration model: will be suitble to
The spectroscopic data of target instrument calibration set sample is as independent variable, using n testing index as dependent variable, using Multivariate Correction side
The spectroscopic data of suitable target instrument calibration set sample is associated with by method with n achievement data, establishes n calibration model；
Step (405): carrying out the acquisition of spectrum using target instrument to sample to be tested, then pre- using target instrument calibration model
The index value of test sample sheet.
2. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that step (404) it
Afterwards, before step (405), further includes: the spectrum of target instrument forecast set is substituted into calibration model, is calculated by calibration model
The predicted value of n index of each forecast set sample out, predicted value and actual measured value are compared, and carry out model evaluation.
3. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that the step (4) is replaced
It is changed to: carrying out Model transfer, i.e. reverse transition from target instrument to main instrument:
Step (411): the sample of setting quantity is selected to collect sample as transfer in main instrumental correction collection sample, then in target
Instrumental correction concentration selects the sample of same sequence number to shift collection sample, the transfer collection sample and target of main instrument as target instrument
Instrument transfer collection sample corresponds；Main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage be adjusted to
Spectrum X in target instrument transfer collection sampleTSpectral variables it is consistent with wave-length coverage；
Step (412): target instrument is shifted into collection sample spectrum intensity data and main instrument transfer collection sample spectrum intensity data
It is associated to obtain relational model to get transition matrix is arrived；
Step (413): carrying out the acquisition of spectrum using target instrument to sample to be tested, based on relational model that target instrument is to be measured
The spectrum correction of sample is the sample to be tested spectrum for being suitble to main instrument；
Step (414): calibration model is established using main instrumental correction collection sample spectrum: by the spectrum number of main instrumental correction collection sample
According to as independent variable, using n testing index as dependent variable, n calibration model is established using multivariate calibration methods；By suitable master
The calibration model that the spectrum of instrument sample to be tested substitutes into main instrument predicts the index value of sample to be predicted.
4. a kind of Model transfer method based on spectroscopic data as claimed in claim 3, characterized in that step (414) it
Afterwards, it may also include that the forecast sample spectrum that the spectrum of target instrument forecast set is corrected to using relational model and is suitble to main instrument
Calibration model is substituted into, the predicted value of n index of each forecast set sample is calculated by calibration model, by predicted value and practical survey
Definite value compares, and carries out model evaluation.
5. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that in forward transfer process
In, target instrument is shifted into collection spectrum XTtPrincipal component decomposition is carried out, is obtained:
XTt=STtVTt'； (1)
Wherein, STtIndicate principal component scores matrix, VTtExpression principal component load matrix, symbol " ' " indicate principal component load matrix
Transposition, subscript " T " indicate target instrument, subscript " t " indicate transfer collection；
Set number of principal components lT；Main instrument is shifted into collection spectrum XMtWith STtAssociation:
STt=XMtFMt； (2)
Wherein, FMtIndicate that transition matrix, symbol " M " indicate that main instrument, " t " indicate transfer collection；
According to FMt=X+ Mt STtIt calculates and solves, symbol "+" indicates broad sense inverse operation；X+ MtIndicate XMtGeneralized inverse matrix；
Formula (2) are substituted into formula (1), the spectrum after obtaining model conversion:
Xt T=XMFMtVTt'； (3)
Wherein, Xt TIndicate that the spectrum matrix for being suitable for target instrument sample after converting, symbol " t " indicate conversion, symbol " T " indicates
Target instrument, XMIndicate the spectroscopic data based on main Instrument measuring；
According to formula (3), the spectrum of main instrumental correction collection is corrected as:
Xt Tc=XMcFMtVTt'； (4)
By formula (4) by the spectrum X of main instrumental correction collectionMcIt is corrected to the spectrum X of suitable target instrumentt Tc, wherein symbol " c " table
Show calibration set.
6. a kind of Model transfer method based on spectroscopic data as claimed in claim 3, characterized in that in reverse transition process
In, main instrument is shifted into collection spectrum XMtPrincipal component decomposition is carried out, is obtained:
XMt=SMtVMt'； (5)
Wherein, SMtIndicate that the score matrix of principal component, " M " indicate that main instrument, " t " indicate transfer collection；VMtIndicate principal component load
Matrix, symbol " ' " transposition of representing matrix；
Set number of principal components lM, target instrument is shifted into collection spectrum XTtWith SMtAssociation:
SMt=XTtFTt； (6)
Wherein, FTtIndicate transition matrix, symbol " T " indicates target instrument, and " t " indicates transfer collection；
Wherein according to FTt=X+ Tt SMtIt calculates and solves；X+ TtIndicate XTtGeneralized inverse matrix；
Formula (6) are substituted into formula (5), the spectrum after model conversion:
Xt M=XTFTtVMt'； (7)
Wherein, Xt MIndicate that the spectrum matrix for being suitable for main instrument sample after converting, symbol " t " indicate conversion, symbol " M " indicates master
Instrument；XTIndicate the spectroscopic data measured based on target instrument, symbol " T " indicates target instrument；
According to formula (7), the spectrum correction of the forecast set sample of target instrument are as follows:
Xt Mv=XTvFTtVMt'； (8)
Pass through formula (8), the forecast set spectrum X of target instrumentTvIt is corrected to the forecast set spectrum X for being suitble to main instrument requirementst Mv, symbol
Number " v " indicates forecast set；
For the sample to be tested of target instrument measurement, converted according to formula (9):
Xt Mu=XTuFTtVMt'； (9)
Symbol " u " indicates sample to be tested, XTuIndicate the spectrum of the sample to be tested of target instrument measurement, Xt MuFor XTuIt is suitble to after conversion
The spectrum of the sample to be tested of main instrument.
7. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that n in the step (1)
More than or equal to 1；Main instrument and target instrument belong to the spectrometer of spectrum of the same race；The spectrum include: infrared spectroscopy, it is ultraviolet-
Visible spectrum, Raman spectrum or NMR spectrum；The infrared spectroscopy, including middle infrared spectrum and near infrared spectrum；
The pretreatment mode of the step (2), comprising: smoothing processing, first derivative calculate, second dervative calculates, at standardization
Reason, baseline drift processing, standard normal variable processing, multiplicative scatter correction processing and go trend handle in any one or it is more
The combination of kind.
8. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that in the step (3)
The quantity of the calibration set is greater than or equal to the quantity of forecast set；Several described calibration samples are divided in the step (3)
For calibration samples and forecast sample, division mode include: KS method, Rank-KS method, SPXY method, Rank-SPXY method and
Any one in concentration gradients method.
9. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that
Transfer collection quantity is more than or equal to 10 in the step (4), is less than or equal to calibration set quantity；
The selection criteria of transfer collection is according to the minimum principle of predicted root mean square error RMSEP, from main instrument in the step (4)
The calibration set of some index selects several samples that the RMSEP of the identical index of target instrument is made to reach minimum, corresponding at this time
Sample is to shift collection sample；For main instrument and target instrument, the serial number of transfer collection sample is consistent.
10. a kind of Model transfer method based on spectroscopic data as described in claim 1, characterized in that
It is described that main instrument is shifted into spectrum X in collection sampleMSpectral variables and wave-length coverage be adjusted to and target instrument transfer collection sample
Spectrum X in thisTSpectral variables and the consistent specific steps of wave-length coverage are as follows:
The spectrum X that main instrument is acquiredMSpectral variables be adjusted to target instrument spectrum XTSpectral variables；
Calculate the spectrum X of main instrument acquisitionMWave-length coverage and target instrument acquisition spectrum XTWave-length coverage intersection；It is right
The spectrum X of main instrument acquisition in wave-length coverage intersectionMWith the spectrum X of target instrument acquisitionTRetained, in wave
The spectrum X of main instrument acquisition outside long range intersectionMWith the spectrum X of target instrument acquisitionTIt is deleted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811271230.6A CN109444066B (en) | 2018-10-29 | 2018-10-29 | Model transfer method based on spectral data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811271230.6A CN109444066B (en) | 2018-10-29 | 2018-10-29 | Model transfer method based on spectral data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109444066A true CN109444066A (en) | 2019-03-08 |
| CN109444066B CN109444066B (en) | 2020-04-14 |
Family
ID=65549356
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811271230.6A Active CN109444066B (en) | 2018-10-29 | 2018-10-29 | Model transfer method based on spectral data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109444066B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111198165A (en) * | 2020-01-14 | 2020-05-26 | 重庆理工大学 | Method for measuring water quality parameters based on spectral data standardization |
| CN111337452A (en) * | 2020-04-08 | 2020-06-26 | 四川虹微技术有限公司 | Method for verifying feasibility of spectral data model transfer algorithm |
| CN111563436A (en) * | 2020-04-28 | 2020-08-21 | 东北大学秦皇岛分校 | Infrared spectrum measuring instrument calibration migration method based on CT-CDD |
| CN111579526A (en) * | 2020-06-08 | 2020-08-25 | 贵州中烟工业有限责任公司 | Method for representing difference and correction of near-infrared instrument |
| US20210025815A1 (en) * | 2018-04-03 | 2021-01-28 | Shenzhen Institute For Drug Control (Shenzhen Testing Center Of Medical Devices | Near-infrared spectroscopy-based method for chemical pattern recognition of authenticity of traditional chinese medicine gleditsiae spina |
| CN112461770A (en) * | 2020-11-17 | 2021-03-09 | 山东省科学院海洋仪器仪表研究所 | Method for acquiring performance of spectrometer |
| CN112617824A (en) * | 2020-12-09 | 2021-04-09 | 中南民族大学 | Ice drug addict detection method and system based on multichannel fNIRS signals |
| CN112683816A (en) * | 2020-12-25 | 2021-04-20 | 中船重工安谱（湖北）仪器有限公司 | Spectrum identification method for spectrum model transmission |
| WO2023046216A1 (en) * | 2021-09-22 | 2023-03-30 | 山东大学 | Graph conversion method for near infrared spectrum and feature graph, and use thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101832922A (en) * | 2010-05-19 | 2010-09-15 | 中国农业大学 | Method for transferring near infrared model of organic fertilizer product |
| CN103018178A (en) * | 2012-12-06 | 2013-04-03 | 江苏省质量安全工程研究院 | LPP (Local Preserving Projection)-based Infrared spectrometer calibration method |
| CN104089911A (en) * | 2014-06-27 | 2014-10-08 | 桂林电子科技大学 | Spectral model transmission method based on unary linear regression |
| CN105738317A (en) * | 2016-02-16 | 2016-07-06 | 广州纤维产品检测研究院 | Textile near-infrared model transfer method |
| CN105842190A (en) * | 2016-03-17 | 2016-08-10 | 浙江中烟工业有限责任公司 | Near-infrared model transfer method based on spectral regression |
| CN106845548A (en) * | 2017-01-24 | 2017-06-13 | 清华大学 | Model transfer method based on high-spectral data |
-
2018
- 2018-10-29 CN CN201811271230.6A patent/CN109444066B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101832922A (en) * | 2010-05-19 | 2010-09-15 | 中国农业大学 | Method for transferring near infrared model of organic fertilizer product |
| CN103018178A (en) * | 2012-12-06 | 2013-04-03 | 江苏省质量安全工程研究院 | LPP (Local Preserving Projection)-based Infrared spectrometer calibration method |
| CN104089911A (en) * | 2014-06-27 | 2014-10-08 | 桂林电子科技大学 | Spectral model transmission method based on unary linear regression |
| CN105738317A (en) * | 2016-02-16 | 2016-07-06 | 广州纤维产品检测研究院 | Textile near-infrared model transfer method |
| CN105842190A (en) * | 2016-03-17 | 2016-08-10 | 浙江中烟工业有限责任公司 | Near-infrared model transfer method based on spectral regression |
| CN106845548A (en) * | 2017-01-24 | 2017-06-13 | 清华大学 | Model transfer method based on high-spectral data |
Non-Patent Citations (1)
| Title |
|---|
| 王家俊 等: "基于拓展光谱的近红外光谱模型转移方法", 《中国烟草学报》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210025815A1 (en) * | 2018-04-03 | 2021-01-28 | Shenzhen Institute For Drug Control (Shenzhen Testing Center Of Medical Devices | Near-infrared spectroscopy-based method for chemical pattern recognition of authenticity of traditional chinese medicine gleditsiae spina |
| US11656176B2 (en) * | 2018-04-03 | 2023-05-23 | Shenzhen Institute For Drug Control (Shenzhen Testing Center Of Medical Devices) | Near-infrared spectroscopy-based method for chemical pattern recognition of authenticity of traditional Chinese medicine Gleditsiae spina |
| CN111198165A (en) * | 2020-01-14 | 2020-05-26 | 重庆理工大学 | Method for measuring water quality parameters based on spectral data standardization |
| CN111337452A (en) * | 2020-04-08 | 2020-06-26 | 四川虹微技术有限公司 | Method for verifying feasibility of spectral data model transfer algorithm |
| CN111563436B (en) * | 2020-04-28 | 2022-04-08 | 东北大学秦皇岛分校 | Infrared spectrum measuring instrument calibration migration method based on CT-CDD |
| CN111563436A (en) * | 2020-04-28 | 2020-08-21 | 东北大学秦皇岛分校 | Infrared spectrum measuring instrument calibration migration method based on CT-CDD |
| CN111579526A (en) * | 2020-06-08 | 2020-08-25 | 贵州中烟工业有限责任公司 | Method for representing difference and correction of near-infrared instrument |
| CN111579526B (en) * | 2020-06-08 | 2023-05-30 | 贵州中烟工业有限责任公司 | Method for representing difference and correction of near infrared instrument |
| CN112461770A (en) * | 2020-11-17 | 2021-03-09 | 山东省科学院海洋仪器仪表研究所 | Method for acquiring performance of spectrometer |
| CN112461770B (en) * | 2020-11-17 | 2022-11-29 | 山东省科学院海洋仪器仪表研究所 | Method for acquiring performance of spectrometer |
| CN112617824A (en) * | 2020-12-09 | 2021-04-09 | 中南民族大学 | Ice drug addict detection method and system based on multichannel fNIRS signals |
| CN112617824B (en) * | 2020-12-09 | 2021-10-01 | 南京睿蜂健康医疗科技有限公司 | Ice drug addict detection system based on multichannel fNIRS signals |
| CN112683816B (en) * | 2020-12-25 | 2021-08-06 | 中船重工安谱（湖北）仪器有限公司 | Spectrum identification method for spectrum model transmission |
| CN112683816A (en) * | 2020-12-25 | 2021-04-20 | 中船重工安谱（湖北）仪器有限公司 | Spectrum identification method for spectrum model transmission |
| WO2023046216A1 (en) * | 2021-09-22 | 2023-03-30 | 山东大学 | Graph conversion method for near infrared spectrum and feature graph, and use thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109444066B (en) | 2020-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109444066A (en) | Model transfer method based on spectroscopic data | |
| Pozzetti et al. | The VIMOS VLT Deep Survey-The assembly history of the stellar mass in galaxies: from the young to the old universe | |
| Chen et al. | The Las Campanas Infrared Survey. IV. The Photometric Redshift Survey and the Rest-frame R-band Galaxy Luminosity Function at 0.5≤ z≤ 1.5 | |
| CN104062257B (en) | A kind of based on the method for general flavone content near infrared ray solution | |
| CN106815643B (en) | Infrared spectroscopy Model Transfer method based on random forest transfer learning | |
| Ni et al. | Screening wavelengths with consistent and stable signals to realize calibration model transfer of near infrared spectra | |
| CN106680238B (en) | Method based on infrared spectrum analysis material component content | |
| CN102590129B (en) | Method for detecting content of amino acid in peanuts by near infrared method | |
| CN110455722A (en) | Rubber tree blade phosphorus content EO-1 hyperion inversion method and system | |
| CN104897607A (en) | Food modeling and rapid detecting integration method and system adopting portable NIRS (near infrared spectroscopy) | |
| CN104089911A (en) | Spectral model transmission method based on unary linear regression | |
| CN102435556A (en) | Accurate spectrum quantitative analysis method used for complex heterogeneous mixture system | |
| CN101413885A (en) | Near-infrared spectrum method for rapidly quantifying honey quality | |
| CN1982872A (en) | Near-infrared diffuse reflection spectral method for fastly inspecting drop water content | |
| CN103854305A (en) | Module transfer method based on multiscale modeling | |
| Zhang et al. | The method of calibration model transfer by optimizing wavelength combinations based on consistent and stable spectral signals | |
| CN101216419A (en) | Method for quickly detecting yellow wine quality index | |
| CN105784672A (en) | Drug detector standardization method based on dual-tree complex wavelet algorithm | |
| CN106769981B (en) | A kind of pitch wax content infrared spectroscopy calibration model bi-directional method | |
| CN106018331B (en) | The method for estimating stability and pretreatment optimization method of multi-channel spectral system | |
| CN106596464A (en) | Near-infrared autocorrelation spectrum detection method for melamine doped in milk powder | |
| Shen et al. | Rapid and real-time detection of moisture in black tea during withering using micro-near-infrared spectroscopy | |
| CN109374556A (en) | Moisture content rapid detection method in compound fertilizer based on visible-near-infrared spectrum | |
| Boysworth et al. | Aspects of multivariate calibration applied to near-infrared spectroscopy | |
| CN106126879B (en) | A kind of soil near-infrared spectrum analysis prediction technique based on rarefaction representation technology |
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 |