CN105259135B - Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method - Google Patents
Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method Download PDFInfo
- Publication number
- CN105259135B CN105259135B CN201510826933.0A CN201510826933A CN105259135B CN 105259135 B CN105259135 B CN 105259135B CN 201510826933 A CN201510826933 A CN 201510826933A CN 105259135 B CN105259135 B CN 105259135B
- Authority
- CN
- China
- Prior art keywords
- temperature
- online
- real
- infrared
- compensating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
The present invention relates to a kind of real-time online without measuring point temperature-compensating near-infrared spectrometers measuring method, and the physical parameter normal data of testing sample is measured using laboratory method；Same sample gathers near infrared spectrum under different temperatures level；The spectroscopic data of collection is pre-processed and pivot analysis, the implicit variable modeling by the use of offset minimum binary using temperature as non-separation, to obtain the physical parameter measured value at current time；New spectroscopic data collection is obtained online, constructs online recursive algorithm, physical parameter measured value is updated, and is completed with the near infrared online measurement without measuring point temperature compensation function.The present invention is participated in temperature as the implicit variable factors of non-separation in near-infrared modeling process, and recursive algorithm is formed based on physical parameter to be measured, so as to rely on model in itself to the physical measurement under the adaptability completion different temperatures of temperature, direct temperature metrical information and correlation computations are not needed so that the model established has more preferable versatility.Change of the online recursive algorithm invented to sample temperature is insensitive, while measurement result has smaller global error.
Description
Technical field
The present invention relates to real-time online without measuring point temperature-compensating near-infrared spectrometers measuring method, suitable for by ring
The physical parameter that border temperature influences, as fluid viscosity, material density, constituent concentration, food quality, agricultural product composition, medicine have
The on-line real-time measuremen of effective component content, oil product of gasoline quality etc..
Background technology
Near-infrared spectrum technique because its analyze speed it is fast, it is small to sample broke, without chemical contamination, almost be adapted to all kinds of samples
The features such as product analysis, multicomponent multichannel determine simultaneously, turns into a bright spot in on-line analysis instrument.And existing major part is near
Infrared detection method is all off-line measurement, on the one hand can not provide more comprehensive, real-time sample message for production and quality testing department,
Another aspect off-line measurement can not possibly realize computer on-line monitoring and the purpose controlled in real time.Therefore how near infrared light is utilized
Analyzer e measurement technology is composed, realization automatically analyzes detection in real time online, has weight to the economic benefit and social benefit for improving enterprise
The meaning wanted.
When near-infrared spectrometers real-time online is applied, measurement result can be by such environmental effects.Research shows, warm
The change of degree can produce the skew of vibrational spectrum so that the measurement result of near infrared spectrum under specified temp, be only applicable to the temperature
Sample quality analysis under degree, and the on-line analysis effect for sample quality is undesirable, this shortcoming greatly limit near-infrared
The application of spectroanalysis instrument real-time online measuring technology.Therefore, it is real-time to study that thermal adaptability is strong, precision is high, robustness is good
Line Measurement Technique, turn near infrared technology can effective application on site key.
The content of the invention
Method proposed by the present invention, during for on-line measurement, temperature change has large effect to near-infrared measuring, establishes
Online recursive algorithm with temperature-compensating mechanism.Purpose is that the change to sample temperature is not when being measured using near infrared online
Sensitivity, and measurement error is small.
The present invention to achieve the above object, adopts the following technical scheme that：
Step of the present invention is divided into three parts.Part I, the experimental design of modeling data, spectral collection and initial near red
Outer parameter models of physical is established；Part II, the pretreatment of near infrared spectrum and the foundation of calibration model；Part III, construction
Online recursive algorithm, complete with the near infrared online measurement without measuring point temperature compensation function.
The experimental facilities of modeling data includes, and the sample cell (2) that (1) sample temperature can be adjusted can displays temperature change
Temperature meter (3) the near infrared spectrum collection instrument (4) of change does not produce the optic probe significantly affected to sample temperature.(5)
The computer tape deck connected near infrared spectrum collection instrument.
Present invention experiment and data collection step are as follows：
Experimental procedure one：Minimum and maximum temperature value under the conditions of confirmatory sample is online.Temperature range is divided into multiple levels
Value.Each temperature levels are typically greater than 5 times of thermometric instruments resolution ratio, to reach effective discrimination.
Experimental procedure two：Under a normal temperature of defined, primary standard number is obtained to all samples physical parameter
According to.
Spectroscopic data is collected respectively under different temperatures level to same sample.Record corresponding sample temperature simultaneously
Value.This temperature value is used for the change of confirmatory sample temperature.In the present invention, temperature is as an implicit variable, and temperature value is in itself
Accurate record is not essential.
Temperature is established as the implicit variable factors compensation model of non-separation and correction algorithm is as follows：
Step 1：Target optical spectral data set is formed with spectrum of the different temperatures under horizontal, target optical spectrum set is carried out
Pretreatment using physical property parameter mode to be measured as target.These pretreatments include the superposition of one or more of following algorithms：
First derivative, second dervative, maximum-minimum sandards, basic bottom line correction, scatter correction, constant bias correction, etc..This
The determination for locating Preprocessing Algorithm is different with the state of physical parameter to be measured and sample.Fig. 2 examples are that a kind of high molecular polymer exists
Original spectrum under different temperatures.Fig. 3 is that the first derivative under different temperatures pre-processes local spectrum.
Step 2：Pivot analysis (PCA) is done to spectrum after caused pretreatment above, rejects statistics exceptional value so that whole
The pivot pattern of individual pre-processed spectrum data is all within a statistical certainty.Fig. 4 examples, it is a kind of high molecular polymer
PCA ideographs.
Step 3：Based on more than pre-process after spectrum, with physical parameter to be measured a set point of temperature original analysis value
As predictive variable, spectrum wave number is as independent variable after pretreatment.Physical parameter is established with partial least squares algorithm (PLS) to correct
Model：
P=D1y1+D2y2+…Dnyn
Herein, P is measured value of the physical property variable at a temperature of required standard, Di, i=1,2 ... n are regression coefficients, yiIt is
Spectrum is in wave number i=1,2 after pretreatment ... the numerical value at n.
Step 4：New spectroscopic data collection, and laboratory initial data corresponding to acquisition simultaneously are obtained online.Using following
Method forms recurrence correction algorithm：
(1) calculate and measure in next step：Pr(k+1)=P (k)+K [L (k-1)-P (k-1)]
(2) by current revised predicted value, Pr(k) the measured value P (k-1) of last moment is assigned to, repeats to walk above
Suddenly, recurrence assignment operation is done.
P hereinr(k) it is the current near-infrared physical measurement correction value with temperature-compensating, P (k-1) is that previous step does not have
There is the near-infrared physical measurement value of amendment, L (k-1) is the actual physical parameter value used in last computation, and K is modifying factor or number
Word wave filter.
In above-mentioned steps four, modifying factor or digital filter, can be more compared with general Statistic analysis and logic judgment,
Or combinations thereof calculates.
In above-mentioned steps four, when each step calculates, physical parameter near infrared correction used can be by updating
Spectroscopic data regenerates.Whole computational algorithm forms recursive form.
The present invention is participated in near-infrared modeling process using temperature as the implicit variable factors of non-separation, thus is being used
During near-infrared measuring, model can be relied in itself to the physical measurement under the adaptability completion different temperatures of temperature, it is not necessary to straight
Jointing temp metrical information and correlation computations so that the model established has more preferable versatility.The recursive algorithm tool invented
There is the preferable adaptability to sample temperature and the change of other measuring conditions.
Brief description of the drawings
Fig. 1 is without measuring point temperature-compensating experimental provision
A kind of original spectrum of the high molecular polymers of Fig. 2 in different temperatures
First derivative under Fig. 3 different temperatures pre-processes local spectrum
A kind of high molecular polymer pivot analysis of Fig. 4 and pattern abnormity point
A kind of near-infrared measuring models of high molecular polymer viscosity of Fig. 5
Spectrum wave-number range used in a kind of high molecular polymer modelings of Fig. 6
There is Fig. 7 the online recurrence result of implementation without measuring point temperature-compensating to compare
Fig. 8 on-line measurement implementation steps block diagrams
Embodiment
Below by taking a kind of high molecular polymer viscosity measurement as an example, illustrate specific implementation method.This example is not formed pair
The scope limitation of the inventive method.
Whole implementation steps block diagram is as shown in Figure 8.
Step 1：Gather representative sample, it is ensured that the physical parameter to be measured of sample can cover measurement request
Scope.Total number of samples is at 40-60.
Step 2：Using the laboratory equipment shown in Fig. 1, respectively in 24 DEG C, 35 DEG C, 50 DEG C, 60 DEG C, 70 DEG C of five differences
The near infrared spectrum of each sample is gathered under temperature levels, while records experiment condition such as temperature etc..The original spectrum of collection is shown in
Fig. 2.
Step 3：Different pretreatments is carried out to spectrum and is compared, to determine finally applicable preprocess method.Example
In, first derivative processing has been carried out to macromolecule high viscosity sample.Treatment effect is as shown in Figure 3.The effect of first derivative processing
The drift up and down of spectrum is mainly eliminated, while remains the information that temperature change influences on spectral shape.
Step 4：Pivot analysis (PCA) is done to spectrum after caused pretreatment above, rejects statistics exceptional value so that whole
The pivot pattern of pre-processed spectrum data is all within a statistical certainty.As shown in figure 4, have in pivot ideograph three it is different
Chang Dian, modeling process should be no longer used in rejecting.
Step 5：Based on more than pre-process after spectrum, using physical parameter to be measured defined temperature original analysis value as
Predictive variable, spectrum wave number is as independent variable after pretreatment.Physical parameter straightening die is established with partial least squares algorithm (PLS)
Type：
P=D1y1+D2y2+…Dnyn
Herein, P is measured value of the physical property variable at a temperature of required standard, Di, i=1,2 ... n are regression coefficients, yiIt is
Spectrum is in wave number i=1,2 after pretreatment ... the numerical value at n.Fig. 5 is a kind of viscosity near-infrared measuring mould of high molecular polymer
Type.The correlation of model predication value and measured value is 0.997, model accuracy R2For 0.995.Spectrum wave number model used in Fig. 6 modelings
Enclose.Wavelength band 7243-4497cm used in modeling-1。
Step 6：10 new spectroscopic data collection, and laboratory initial data corresponding to acquisition simultaneously are obtained online.
Step 7：Error E (k)=L (k)-P (k) of 10 samples in the past is calculated, and forms an error time sequence
E(k-1),E(k-2),…E(k-10)。
Step 8：Low pass dynamic filter computing is done to above-mentioned error time sequence, one-step prediction value is obtained, is designated as B.
Step 9：Calculate viscosity correction measured value：Pr=P+B
P is the current near-infrared physical measurement value with temperature-compensating herein.
Step 10：By current correction value Pr(k) the measured value P (k-1) of last moment is assigned to, does recursive operation, weight
Multiple above step 6-9.
Fig. 7 is the results contrast example of algorithms of different.Sample temperature changes between 24-70 degrees Celsius.Fixed temperature mould
Type is the physical measurement model using 50 degrees Celsius of establishment of spectrum.As marked in figure, No. 64 sample sequence moment, start
Run the recursive algorithm of the present invention.It can be seen that change of the fixed temperature model to temperature is sensitive.The inventive method is to sample
The change of product temperature degree has preferable insensitivity, while recursive calculation causes measurement result to have smaller global error.
Claims (7)
1. a kind of real-time online without measuring point temperature-compensating near-infrared spectrometers measuring method, it is characterised in that including with
Lower step：
Step 1：The physical parameter normal data of multiple testing samples is obtained, and near infrared light is carried out under different temperatures level
The collection of spectrum；
Step 2：The near infrared spectrum gathered in step 1 is pre-processed and counted outlier processing；
Step 3：Using physical property Experiment Parameter room normal data to be measured as predictive variable, with the spectrum ripple treated in step 2
Number is independent variable, establishes geophysical parameter prediction model, obtains the measured value under current time set point of temperature；
Step 4：New spectroscopic data collection is obtained online, and renewal is modified to measured value using online recursive algorithm；
Online recurrence correction algorithm is described in step 4：
Pr(k+1)=Pr(k)+K[L(k-1)-P(k-1)]
Wherein Pr(k) it is measurement correction value current time with temperature-compensating, P (k-1) is the measured value of last moment, L (k-
1) it is actual physical property reference value used in last computation, K is modifying factor or lower order filter, and Pr (k+1) is subsequent time tool
There is the measurement correction value of temperature-compensating.
2. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：The horizontal interval of distinguishing of different temperatures described in step 1 is greater than 5 times of temperature measuring set resolution ratio.
3. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：Pretreated spectra described in step 2 includes the superposition of one or more of following algorithms：First derivative, two
Order derivative, maximum-minimum sandards, basic bottom line correction, scatter correction.
4. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：Statistics outlier processing method described in step 2 is principle component analysis so that derivative spectrum data and pre-
The pivot pattern of spectroscopic data is handled all in statistical certainty.
5. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：Geophysical parameter prediction model described in step 3, which is established, uses partial least squares algorithm to enter trip temperature to imply variable
Linear regression.
6. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：K is Statistic analysis and logic judgment, or the combination of Statistic analysis and logic judgment calculates.
7. real-time online according to claim 1 without measuring point temperature-compensating near-infrared spectrometers measuring method, its
It is characterised by：For the online recurrence correction algorithm when each step calculates, physical parameter calibration model used is the light by updating
Modal data regenerates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510826933.0A CN105259135B (en) | 2015-11-24 | 2015-11-24 | Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510826933.0A CN105259135B (en) | 2015-11-24 | 2015-11-24 | Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105259135A CN105259135A (en) | 2016-01-20 |
| CN105259135B true CN105259135B (en) | 2018-02-23 |
Family
ID=55098905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510826933.0A Active CN105259135B (en) | 2015-11-24 | 2015-11-24 | Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105259135B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105300924A (en) * | 2015-11-24 | 2016-02-03 | 江南大学 | Measuring point-free compensation modeling method of near infrared calibration model under effect of temperature |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426167A (en) * | 2013-07-09 | 2013-12-04 | 哈尔滨工程大学 | Hyperspectral real-time detection method based on recursive analysis |
| CN104732215A (en) * | 2015-03-25 | 2015-06-24 | 广西大学 | Remote-sensing image coastline extracting method based on information vector machine |
-
2015
- 2015-11-24 CN CN201510826933.0A patent/CN105259135B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426167A (en) * | 2013-07-09 | 2013-12-04 | 哈尔滨工程大学 | Hyperspectral real-time detection method based on recursive analysis |
| CN104732215A (en) * | 2015-03-25 | 2015-06-24 | 广西大学 | Remote-sensing image coastline extracting method based on information vector machine |
Non-Patent Citations (4)
| Title |
|---|
| Assessing the temperature influence on the soluble solids content of watermelon juice as measured by visible and near-infrared spectroscopy and chemometrics;20130531;《Journal of Food Engineering》;20130531;第119卷(第1期);第22-27页 * |
| Influence and correction of temperature on optical measurement for fat and protein contents in a complex food model system;Xuedian Zhang 等;《Infrared Physics & Technology》;20091130;第53卷(第3期);第177-181页 * |
| Prediction of titratable acidity, malic acid, and citric acid in bayberry fruit by near-infrared spectroscopy;Lijuan Xie 等;《Food Research International》;20111231;第44卷(第7期);第2198-2204页 * |
| 苹果糖度近红外光谱分析模型的温度补偿;王加华 等;《光谱学与光谱分析》;20090630;第29卷(第6期);第1517-1520页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105259135A (en) | 2016-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105300923B (en) | Without measuring point model of temperature compensation modification method during a kind of near-infrared spectrometers application on site | |
| CN105486658B (en) | A kind of near-infrared physical parameter measurement method with no measuring point temperature compensation function | |
| CN110161013B (en) | Laser-induced breakdown spectroscopy data processing method and system based on machine learning | |
| CN101929951A (en) | Method for distinguishing milk doped with ewe's milk by near infrared spectrum | |
| CN103983595A (en) | Water quality turbidity calculating method based on ultraviolet-visible spectroscopy treatment | |
| CN107703097A (en) | Utilize the method and its application of decay total reflection probe and the model of near infrared spectrometer structure fast prediction oil property | |
| CN105466885B (en) | Based on the near infrared online measuring method without measuring point temperature-compensating mechanism | |
| CN109324013A (en) | A method of it is quickly analyzed using Gaussian process regression model building oil property near-infrared | |
| CN105334185A (en) | Spectrum projection discrimination-based near infrared model maintenance method | |
| CN108760647A (en) | A kind of wheat content of molds line detecting method based on Vis/NIR technology | |
| CN105259136B (en) | Near infrared correction without measuring point temperature correction | |
| CN107860743A (en) | Utilize the method and its application of the model of reflective near infrared fibre-optical probe structure fast prediction oil property | |
| CN104516991A (en) | Gamma sensor full-temperature range compensation method | |
| CN116985183A (en) | Quality monitoring and management method and system for near infrared spectrum analyzer | |
| CN105628646B (en) | A kind of online tar prediction of cigarette and method for early warning | |
| CN103760130B (en) | The method of Tween-80 content near infrared ray compound Moschus injection | |
| CN114112978A (en) | Wavelet function calibration modeling method for detecting concentration of solution in crystallization process by using in-situ infrared spectroscopy | |
| CN105259135B (en) | Suitable for real-time online without measuring point temperature-compensating near-infrared measuring method | |
| CN112651173B (en) | Agricultural product quality nondestructive testing method based on cross-domain spectral information and generalizable system | |
| EP3702758A1 (en) | Analysis method, analysis apparatus, printer, and print system | |
| CN111141809B (en) | Soil nutrient ion content detection method based on non-contact type conductivity signal | |
| CN107356556A (en) | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis | |
| CN105300924A (en) | Measuring point-free compensation modeling method of near infrared calibration model under effect of temperature | |
| CN114441469A (en) | Calibration method and device of moisture meter and computer equipment | |
| CN108226082B (en) | Calibration method for damping response time of moisture meter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |