CN107179292A - Different near infrared spectrum variable preferred result fusion methods and application - Google Patents
Different near infrared spectrum variable preferred result fusion methods and application Download PDFInfo
- Publication number
- CN107179292A CN107179292A CN201610136215.5A CN201610136215A CN107179292A CN 107179292 A CN107179292 A CN 107179292A CN 201610136215 A CN201610136215 A CN 201610136215A CN 107179292 A CN107179292 A CN 107179292A
- Authority
- CN
- China
- Prior art keywords
- variable
- preferred
- spectrum
- sensitive
- frequency
- 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
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 36
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 11
- 238000001228 spectrum Methods 0.000 claims abstract description 80
- 238000010219 correlation analysis Methods 0.000 claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 25
- 230000003595 spectral effect Effects 0.000 claims abstract description 22
- 238000012360 testing method Methods 0.000 claims abstract description 21
- 238000005457 optimization Methods 0.000 claims abstract description 9
- 230000001360 synchronised effect Effects 0.000 claims description 19
- 239000000284 extract Substances 0.000 claims description 5
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000000205 computational method Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 15
- 230000004927 fusion Effects 0.000 abstract description 5
- 238000012216 screening Methods 0.000 abstract 1
- 241000209140 Triticum Species 0.000 description 8
- 235000021307 Triticum Nutrition 0.000 description 8
- 230000002068 genetic effect Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 238000004497 NIR spectroscopy Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005283 ground state Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 239000000126 substance Substances 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
- 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
-
- 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
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
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 invention discloses a kind of different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis, this method includes:Step 1:Determine sample near infrared spectrum data and target testing concentration reference value;Step 2:Concentration perturbation dynamic spectrum is built, Two-dimensional Correlation Analysis is carried out;Step 3:Based on Two-dimensional Correlation Analysis, consider neighbouring spectral variables synteny, recognize the preferred sensitive variable of high frequency;Step 4:Based on Two-dimensional Correlation Analysis, the independent spectrum preferred sensitive variable of area's low frequency of identification;Step 5:Draw full spectrum regional complex variable preferred result.The method of the invention establishes the mutual ratio method and postsearch screening mechanism of algorithms of different variable preferred result, avoid unitary variant optimization algorithm limitation, strengthen crucial sensitive variable information on the basis of spectrum redundancy is reduced, solve the problems, such as the fusion of Different Results under a variety of preferred variable algorithms.
Description
Technical field
The present invention relates near infrared spectroscopy, Chemical Measurement field, more particularly to a kind of spectral variables selection fusion side
Method.
Background technology
Near infrared spectrum refers to 780~2526nm electromagnetic wave, its information come from molecule anharmonicity vibration make molecule from
Ground state absorbs to the frequency multiplication and sum of fundamental frequencies produced during high energy order transition.Molecular vibration frequency multiplication and the energy spacing of sum of fundamental frequencies transition are discrete,
Cause near infrared spectrum spectral peak wide.Identical molecular vibration frequency multiplication is various informative with sum of fundamental frequencies, causes near infrared spectrum spectral peak many.Work as inspection
When surveying object for multi-component complex sample, near-infrared spectra band is largely overlapping, and variable synteny is serious, causes near infrared spectrum to parse
It is difficult.Specifically, complex samples near infrared spectrum generally includes hundreds of or even thousands of wavelength or wave number variable, not all
Variable is beneficial to target to be measured, and wherein most variable height correlation, is redundant variables.Carried out based on chemometrics method
Variable preferably, while amount of compressed data and model calculation amount, can reduce noise signal influence, be improved beneficial to model performance.
However, near infrared spectrum variable optimization method is more in the prior art, and different variable optimization algorithm results are different, or even lance
Shield, puzzlement is brought for application.
Two-dimensional Correlation Analysis is a kind of by applying external disturbance to system, obtains a series of dynamic spectrum of changes, enters
And the chemometrics method that one-dimensional linear spectroscopic data is extended to two dimensional surface.The spectrogram extended through two-dimensional correlation has two
It is individual:Synchronous spectrum and asynchronous spectrum.Synchronous stave is levied at particular variables to be located at pair in dynamic spectral intensity synchronization and collaborative variation, synchronous spectrum
The peak of diagonal position is referred to as automatic peak, the overall degree that spectral intensity changes during representing to disturb at particular variables, reflects at this
The sensitivity that spectral intensity changes with external disturbance.Peak at off-diagonal is referred to as to intersect peak, and the synchronous peak that intersects represents two frequencies
The similitude that spectral intensity changes between rate.Asynchronous stave levies dynamic spectrum Strength Changes otherness at particular variables.Using two-dimensional phase
Pass method improves spectral resolution, and then separates the high superposed near-infrared spectral peak of complicated multicomponent sample.By correlation analysis,
Intermolecular and intramolecule interaction information is obtained, and then parses near-infrared variable synteny mechanism.Comprehensive Resolving Overlapped Peaks
With synteny variable resolution, possibility is provided for the fusion of different near infrared spectrum variable preferred results.
The content of the invention
Melt it is an object of the invention to provide a kind of different near infrared spectrum variable preferred results based on Two-dimensional Correlation Analysis
Conjunction method, to solve in the prior art that result is inconsistent when near infrared spectrum variable preferred with many algorithms/contradiction brought
The problem of.
To achieve the above object, the present invention proposes that a kind of different near infrared spectrum variables based on Two-dimensional Correlation Analysis are preferred
As a result fusion method, including step:
Step 1:Determine sample near infrared spectrum data and target testing concentration reference value;
Step 2:Concentration perturbation dynamic spectrum is built, Two-dimensional Correlation Analysis is carried out;
Step 3:Based on Two-dimensional Correlation Analysis, neighbouring spectral variables synteny, the preferred sensitive change of identification high frequency are considered
Amount;
Step 4:Based on Two-dimensional Correlation Analysis, the independent spectrum preferred sensitive variable of area's low frequency of identification;
Step 5:Draw full spectrum regional complex variable preferred result.
Wherein, in the step 1, sample near infrared spectrum X and target testing concentration reference value C is obtained, based near red
External spectrum X and target testing concentration reference value C, using the preferred spectrum sensitive variable of m kind variable optimization methods, every kind of variable is excellent
The sensitivity spectrum variable subset V that choosing method preferably goes outmInclude k sensitive variable γm1, γm2..., γmk, it is superimposed different variables
Preferred result, builds total preferred variable collection V.
Wherein, in the step 1, the target testing concentration reference value is the measurement of the material to be detected target to be measured
Value.
Wherein, in the step 2, equidistant target concentration y to be measured is setes, build target testing concentration distributed area
Interior n target testing concentration spacing is yesSample set, its corresponding light music score integrates as xj(γ)=x (γ, cj), j=1,
2 ..., n, γ represent spectral variables, cjThe target testing concentration of spectrum belonging to representing, j represents sample/spectra number.Wherein
The averaging spectrum of selected dynamic spectrum subset is reference spectrumComputational methods areUtilize son
Collect spectrum subtraction reference spectrum and constitute dynamic spectrumCircular is
Then carry out the synchronous correlation computations of two dimensionWhereinRepresent
γ1, γ2Synchronous coefficient correlation at variable,The γ of j samples/spectrum is represented respectively1、γ2Light at variable
Spectrum.Extract the synchronous automatic peak tangent line spectrum of spectrumWhereinRepresent γiAt variable
Synchronous coefficient correlation, i numbers for variable, and q is spectral variables number.
Wherein, in the step 3, define synteny and close on apart from d and repeat selection frequency f, it is assumed that spectral variables interval
The variable height correlation closed in synteny within distance, in total preferred variable collection V, variable is in the range of [i-d, i+d]
The preferred frequency of variable is defined as variable γiPreferred frequency fi, preferably frequency fi>=f variable and its variable within d are
Synteny closes on set of variables Vg, preferred unique variable is the preferred sensitive variable of high frequency in every group.
Wherein, in the step 3, preferred steps are:
1. as synteny closes on set of variables VgIn preferably frequency highest variable it is unique, then the variable-definition is that high frequency is preferred
Sensitive variable;
2. as synteny closes on set of variables VgIn preferably frequency highest variable it is not unique, calculate preferred frequency highest variable
Average, such as unique away from the nearest variable of the average, then the variable-definition is the preferred sensitive variable of high frequency;And
3. it is such as not unique away from the nearest variable of the average, then at the preferred frequency highest variable two-dimensional correlation synchronization coefficient correlation
Big variable is the preferred sensitive variable of high frequency.
Wherein, in the step 4, independent peak is composed based on the synchronous automatic peak tangent line of two-dimensional correlation, spectrum subinterval is divided, picks
Except the interval comprising the preferred sensitive variable of high frequency and with total interval of the preferred variable collection without common factor, it is remaining each interval in preferably only
One variable is the preferred sensitive variable of low frequency.
Wherein, in the step 4, preferred steps are:Preferred variable average in computation interval, such as away from the nearest variable of the average
Uniquely, then the variable is the preferred sensitive variable of low frequency;Such as not unique away from the nearest variable of the average, two dimension is same at selection preferred variable
It is the preferred sensitive variable of low frequency to walk the big variable of coefficient correlation.
Wherein, it is full spectrum regional complex with reference to the preferred sensitive variable of high frequency and the preferred sensitive variable of low frequency in the step 5
Variable preferred result.
Moreover, to achieve the above object, the invention also provides the change of the different near infrared spectrums based on Two-dimensional Correlation Analysis
Measure application of the preferred result fusion method in quality of agricultural product quick detection.
Different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis proposed by the invention, it is comprehensive
Conjunction considers synteny between near infrared spectrum peak overlap characteristic and variable, it is to avoid the limitation of unitary variant optimization algorithm,
Strengthen crucial sensitive variable information on the basis of spectrum redundancy is reduced, solve Different Results under a variety of preferred variable algorithms
Fusion problem.
Brief description of the drawings
Fig. 1 is the original near infrared spectrum of wheat seed.
Fig. 2 is the near infrared spectrum that wheat seed is pre-processed.
Fig. 3 is two methods preferably wheat seed near infrared spectrum sensitive variable result.
Fig. 4 is the automatic peak tangent line spectrum of wheat seed near-infrared two-dimensional correlation.
Fig. 5 is that synteny closes on variable.
Fig. 6 is the preferred sensitive variable of high frequency.
Fig. 7 divides for subinterval.
Fig. 8 is the variogram to be selected in interval without the preferred sensitive variable of high frequency.
Fig. 9 is the preferred sensitive variable of low frequency.
Figure 10 is full spectrum regional complex variable preferred result.
Embodiment
The present invention proposes a kind of different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis, its
Scheme is:
Step 1:Determine sample near infrared spectrum data and target testing concentration reference value
Sample near infrared spectrum X and target testing concentration reference value C is obtained, it is to be measured based near infrared spectrum X and target
Thing concentration reference value C, using the preferred spectrum sensitive variable of m kind variable optimization methods, it is quick that every kind of variable optimization method preferably goes out
Photosensitive spectrum variable subset VmInclude k sensitive variable γm1, γm2..., γmk, different variable preferred results are superimposed, are built total excellent
Select variables set V.
Wherein, target testing concentration reference value refers to the material to be detected (wheat seed in such as embodiment) target to be measured
The measured value (protein content of 42 ripe wheat seed samples in such as embodiment) of (protein content in such as embodiment).
Step 2:Concentration perturbation dynamic spectrum is built, Two-dimensional Correlation Analysis is carried out.
Set equidistant target concentration y to be measuredes, build n target determinand in target testing concentration distributed area dense
Degree spacing is yesSample set, its corresponding light music score integrates as xj(γ)=x (γ, cj), j=1,2 ..., n, γ represent spectrum
Variable, cjThe target testing concentration of spectrum belonging to representing, j represents sample/spectra number.Wherein select dynamic spectrum subset
Averaging spectrum is reference spectrumComputational methods areUtilize subset spectrum subtraction reference spectrum structure
Into dynamic spectrumCircular isThen carry out the synchronous phase of two dimension
Close and calculateWhereinRepresent γ1, γ2Synchronous phase at variable
Relation number,The γ of j samples/spectrum is represented respectively1、γ2Spectral value at variable.Extract synchronous spectrum automatic
Peak tangent line spectrumWhereinRepresent γiSynchronous coefficient correlation at variable, i is variable
Numbering, q is spectral variables number.
Step 3:Based on Two-dimensional Correlation Analysis, neighbouring spectral variables synteny, the preferred sensitive change of identification high frequency are considered
Amount.
Define synteny to close on apart from d and repeat selection frequency f, it is assumed that spectral variables are spaced in synteny and closed on apart from it
Interior variable height correlation, in total preferred variable collection V, the preferred frequency of variable of the variable in the range of [i-d, i+d] is defined as
Variable γiPreferred frequency fi, preferably frequency fi>=f variable and its variable within d are that synteny closes on set of variables
Vg, preferred unique variable is the preferred sensitive variable of high frequency in every group.Preferred criteria is:1. as synteny closes on set of variables VgIn
It is preferred that frequency highest variable is unique, then the variable-definition is the preferred sensitive variable of high frequency;2. as synteny closes on set of variables Vg
In preferably frequency highest variable it is not unique, calculate the average of preferred frequency highest variable, such as away from the nearest variable of the average only
One, then the variable-definition is the preferred sensitive variable of high frequency;3. it is such as not unique away from the nearest variable of the average, then preferred frequency highest
The big variable of two-dimensional correlation synchronization coefficient correlation is the preferred sensitive variable of high frequency at variable.
Step 4:Based on Two-dimensional Correlation Analysis, the independent spectrum preferred sensitive variable of area's low frequency of identification
Independent peak is composed based on the synchronous automatic peak tangent line of two-dimensional correlation, spectrum subinterval is divided, rejects preferably quick comprising high frequency
Feel variable interval and with total interval of the preferred variable collection without common factor, it is remaining each interval in preferably unique variable be that low frequency is preferred
Sensitive variable.Preferred steps are:Preferred variable average in computation interval, such as unique away from the nearest variable of the average, then the variable is
The preferred sensitive variable of low frequency;Such as not unique away from the nearest variable of the average, two dimension synchronous coefficient correlation is big at selection preferred variable
Variable is the preferred sensitive variable of low frequency.
Step 5:Draw full spectrum regional complex variable preferred result
It is full spectrum regional complex variable preferred result with reference to the preferred sensitive variable of high frequency and the preferred sensitive variable of low frequency.
The different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis of the present invention can be applied to
The fast spectrum of the qualities such as protein content, the moisture of the agricultural product such as wheat seed evaluates field, for different samples, treats
Surveying parameter in target, this method can be adjusted correspondingly.
Embodiment:
First, sample near infrared spectrum data and target testing concentration reference value are determined;
Choose 42 ripe wheat seed samples, the near infrared light that collection 1000-1700nm scopes intrinsic resolution is 1.6nm
Spectrum determines its protein content 10.51~16.94%, average is 14.09% (see Fig. 1) according to GB/T 24899-2010,
Standard deviation 1.45%.Spectrum is through 9 points smooth, variable standardization, normalization pretreatment is (see Fig. 2).It is preferred from two kinds of variables
Algorithm:Elastic network(s) (Elastic Net) and genetic algorithm (Genetic Algorithm), respectively preferably 22 and 23 sensitive changes
Amount is (see Fig. 3).
2nd, concentration perturbation dynamic spectrum is built, Two-dimensional Correlation Analysis is carried out
Select protein content be 10.51,11.07,11.66,12.26,12.85,13.43,14.12,14.64,
15.21st, 15.74,16.37,16.94% 12 sample spectrums, carry out two-dimensional correlation calculating and extract its automatic peak tangent line spectrum
(see Fig. 4).
3rd, based on Two-dimensional Correlation Analysis, consider neighbouring spectral variables synteny, recognize the preferred sensitive variable of high frequency
Synteny is defined to close on apart from d=8nm and repeat selection frequency f=2.Merge elastic network(s) and genetic algorithm preferably becomes
Result is measured, total preferred variable collection V is built.Wherein variable and its synteny close on distance range γi-8~γi+8It is interior, preferably select frequency
Secondary >=f=2 variable has 28 (see Fig. 5), is closed on according to synteny apart from group forming criterion, is divided into 10 groups.Wherein 1273.4nm
Chosen with 1298.4nm variables by elastic network(s) and genetic algorithm, be used as the preferred sensitive variable of high frequency;1026.6nm、
1257.8nm, 1364.1nm, 1403.1nm and 1476.6nm are near away from average in group, are the preferred sensitive variable of high frequency;1010.9nm、
The 1242.2nm and 1326.6nm synchronous coefficient correlation in the automatic peak of two-dimensional correlation is high, preferably the preferred sensitive variable of high frequency.High frequency
It is preferred that sensitive variable result is as shown in Figure 6.
4th, based on Two-dimensional Correlation Analysis, the independent spectrum preferred sensitive variable of area's low frequency of identification
Subinterval (see Fig. 7) is divided based on the automatic peak tangent line spectrum of two-dimensional correlation, extracts interval without the preferred sensitive variable of high frequency.
There are 14 variables without the preferred sensitive variable interval of high frequency (see Fig. 8).Wherein 1023.4nm, 1126.6nm, 1146.9nm,
1187.5nm, 1259.4nm, 1354.7nm, 1584.4nm and 1693.8nm are near away from average in group, and preferably low frequency is preferably sensitive
Variable;The synchronous coefficient correlation of 1167.2nm two-dimensional correlation is high, preferably the preferred sensitive variable of low frequency.The preferred sensitive variable of low frequency
As a result it is as shown in Figure 9.
5th, full spectrum regional complex variable preferred result is drawn
With reference to the preferred sensitive variable of high frequency and the preferred sensitive variable result of low frequency, the preferred sensitive change of comprehensive algorithms of different is drawn
Measure result (see Figure 10).
The effect of the present invention:
Different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis proposed by the invention, it is comprehensive
Conjunction considers synteny between near infrared spectrum peak overlap characteristic and variable, it is to avoid the limitation of unitary variant optimization algorithm,
Increase crucial sensitive variable information on the basis of redundancy is reduced, the fusion for solving Different Results under a variety of preferred variable algorithms is asked
Topic.
Certainly, the present invention can also have other various embodiments, ripe in the case of without departing substantially from spirit of the invention and its essence
Various corresponding changes and deformation, but these corresponding changes and deformation can be made according to the present invention by knowing those skilled in the art
The protection domain of the claims in the present invention should all be belonged to.
Claims (10)
1. a kind of different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis, it is characterised in that bag
Include step:
Step 1:Determine sample near infrared spectrum data and target testing concentration reference value;
Step 2:Concentration perturbation dynamic spectrum is built, Two-dimensional Correlation Analysis is carried out;
Step 3:Based on Two-dimensional Correlation Analysis, consider neighbouring spectral variables synteny, recognize the preferred sensitive variable of high frequency;
Step 4:Based on Two-dimensional Correlation Analysis, the independent spectrum preferred sensitive variable of area's low frequency of identification;
Step 5:Draw full spectrum regional complex variable preferred result.
2. according to the method described in claim 1, it is characterised in that in the step 1, obtain sample near infrared spectrum X and mesh
Testing concentration reference value C is marked, it is preferably square using m kinds variable based near infrared spectrum X and target testing concentration reference value C
The preferred spectrum sensitive variable of method, the sensitivity spectrum variable subset V that every kind of variable optimization method preferably goes outmInclude k sensitive variable
γm1, γm2..., γmk, different variable preferred results are superimposed, total preferred variable collection V is built.
3. according to the method described in claim 1, it is characterised in that in the step 1, the target testing concentration reference value
For the measured value of the material to be detected target to be measured.
4. according to the method described in claim 1, it is characterised in that in the step 2, set equidistant target concentration y to be measuredes, build
N target testing concentration spacing is y in target testing concentration distributed areaesSample set, its corresponding light music score integrates as xj(γ)
=x (γ, cj), j=1,2 ..., n, γ represent spectral variables, cjThe target testing concentration of spectrum belonging to representing, j represents sample/spectrum
Numbering.The averaging spectrum for wherein selecting dynamic spectrum subset is reference spectrumComputational methods are
Dynamic spectrum is constituted using subset spectrum subtraction reference spectrumCircular
Then carry out the synchronous correlation computations of two dimensionWhereinRepresent
γ1, γ2Synchronous coefficient correlation at variable,The γ of j samples/spectrum is represented respectively1、γ2Light at variable
Spectrum.Extract the synchronous automatic peak tangent line spectrum of spectrumWhereinRepresent γiAt variable
Synchronous coefficient correlation, i numbers for variable, and q is spectral variables number.
5. synteny according to the method described in claim 1, it is characterised in that in the step 3, is defined to close on apart from d and again
Frequency f is selected in final election, it is assumed that spectral variables are spaced in the variable height correlation that synteny is closed within distance, in total preferred variable collection
In V, the preferred frequency of variable of the variable in the range of [i-d, i+d] is defined as variable γiPreferred frequency fi, preferably frequency fi≥f
Variable and its variable within d close on set of variables V for syntenyg, preferred unique variable is preferred for high frequency in every group
Sensitive variable.
6. method according to claim 5, it is characterised in that
In the step 3, preferred steps are:
As synteny closes on set of variables VgIn preferably frequency highest variable it is unique, then to be that high frequency is preferred sensitive become the variable-definition
Amount;
As synteny closes on set of variables VgIn preferably frequency highest variable it is not unique, calculate the average of preferred frequency highest variable,
Such as unique away from the nearest variable of the average, then the variable-definition is the preferred sensitive variable of high frequency;And
It is such as not unique away from the nearest variable of the average, then the change of two-dimensional correlation synchronization coefficient correlation greatly at preferred frequency highest variable
Measure as the preferred sensitive variable of high frequency.
7. according to the method described in claim 1, it is characterised in that in the step 4, cut based on the synchronous automatic peak of two-dimensional correlation
Line spectrum independence peak, divides spectrum subinterval, rejects the interval comprising the preferred sensitive variable of high frequency and with total preferred variable collection without friendship
Preferably unique variable is the preferred sensitive variable of low frequency in the interval of collection, remaining each interval.
8. method according to claim 7, it is characterised in that in the step 4, preferred steps are:It is excellent in computation interval
Mean variable value is selected, such as unique away from the nearest variable of the average, then the variable is the preferred sensitive variable of low frequency;Such as become recently away from the average
Amount is not unique, and the big variable of the synchronous coefficient correlation of two dimension is the preferred sensitive variable of low frequency at selection preferred variable.
9. according to the method described in claim 1, it is characterised in that
It is preferred for full spectrum regional complex variable with the preferred sensitive variable of low frequency with reference to the preferred sensitive variable of high frequency in the step 5
As a result.
10. the claim 1-9 different near infrared spectrum variable preferred result fusion methods based on Two-dimensional Correlation Analysis are in agriculture
Application in product quality quick detection.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610136215.5A CN107179292B (en) | 2016-03-10 | 2016-03-10 | Different near infrared spectrum variable optimization result fusion method and application |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610136215.5A CN107179292B (en) | 2016-03-10 | 2016-03-10 | Different near infrared spectrum variable optimization result fusion method and application |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107179292A true CN107179292A (en) | 2017-09-19 |
CN107179292B CN107179292B (en) | 2020-03-27 |
Family
ID=59829916
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610136215.5A Active CN107179292B (en) | 2016-03-10 | 2016-03-10 | Different near infrared spectrum variable optimization result fusion method and application |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107179292B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116242799A (en) * | 2023-03-14 | 2023-06-09 | 合肥工业大学 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100090845A1 (en) * | 2007-05-29 | 2010-04-15 | Mark Leon Polak | Infrared gas detection and spectral analysis method |
CN101915744A (en) * | 2010-07-05 | 2010-12-15 | 北京航空航天大学 | Near infrared spectrum nondestructive testing method and device for material component content |
CN103308463A (en) * | 2013-06-28 | 2013-09-18 | 中国农业大学 | Characteristic spectrum area selection method for near infrared spectrum |
CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
CN105138834A (en) * | 2015-08-18 | 2015-12-09 | 浙江中烟工业有限责任公司 | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering |
-
2016
- 2016-03-10 CN CN201610136215.5A patent/CN107179292B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100090845A1 (en) * | 2007-05-29 | 2010-04-15 | Mark Leon Polak | Infrared gas detection and spectral analysis method |
CN101915744A (en) * | 2010-07-05 | 2010-12-15 | 北京航空航天大学 | Near infrared spectrum nondestructive testing method and device for material component content |
CN103308463A (en) * | 2013-06-28 | 2013-09-18 | 中国农业大学 | Characteristic spectrum area selection method for near infrared spectrum |
CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
CN105138834A (en) * | 2015-08-18 | 2015-12-09 | 浙江中烟工业有限责任公司 | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering |
Non-Patent Citations (1)
Title |
---|
吕程序: ""基于二维相关NIRS/NIRM的蛋白饲料原料判别方法研究"", 《中国博士学位论文全文数据库》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116242799A (en) * | 2023-03-14 | 2023-06-09 | 合肥工业大学 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
CN116242799B (en) * | 2023-03-14 | 2023-08-18 | 合肥工业大学 | Base oil detection device and detection method based on deep learning infrared multidimensional fusion algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN107179292B (en) | 2020-03-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105630743B (en) | A kind of system of selection of spectrum wave number | |
CN104158611B (en) | Wireless signal Interference Detection system and method based on spectrum analysis | |
CN106560697A (en) | Method for identifying producing area of Wuyi rock tea through combination of near infrared spectroscopy and trace element detection | |
CN103235095B (en) | Water-injected meat detection method and device | |
CN102305772A (en) | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method | |
CN103134765A (en) | Chinese medicine sample authenticity preliminary screening method based on terahertz time-domain spectrum | |
CN103235179A (en) | Frequency mask trigger with non-uniform bandwidth segments | |
CN103278473B (en) | The mensuration of pipering and moisture and method for evaluating quality in white pepper | |
CN106560704A (en) | Wuyi rock tea production place identification method through combination of isotope detection and trace element detection | |
CN106560691A (en) | Identification method for producing area of Wuyi rock tea and with deep learning function | |
CN103293118A (en) | Hogwash oil identification method based on near infrared reflectance spectroscopy | |
CN105067650A (en) | Method for calculating characteristic peak of derivative detection spectrum through using wavelet | |
CN106560698A (en) | Identification method for producing area of plant based on multiple detection technologies | |
CN110108644A (en) | A kind of maize variety identification method based on depth cascade forest and high spectrum image | |
Liu et al. | Discrimination of the fruits of Amomum tsao-ko according to geographical origin by 2DCOS image with RGB and Resnet image analysis techniques | |
CN107860845A (en) | The method that automatic parsing GC MS overlap peaks accurately identify compound | |
CN105784628A (en) | Method for detecting chemical composition of soil organic matter with mid-infrared spectra | |
CN107219184A (en) | Meat identification method and device applied to origin tracing | |
CN105758819A (en) | Method for detecting organic components of soil by utilizing near infrared spectrum | |
CN106560699A (en) | Combined detection method used for identification of producing area of Wuyi rock tea | |
CN103940767A (en) | Gas concentration inversion method based on multi-manifold learning | |
CN105784672A (en) | Drug detector standardization method based on dual-tree complex wavelet algorithm | |
CN107247033B (en) | Identify the method for Huanghua Pear maturity based on rapid decay formula life cycle algorithm and PLSDA | |
CN105628708A (en) | Quick nondestructive testing method for multi-parameter quality of south Xinjiang red dates | |
CN107632010A (en) | A kind of quantitative approach of combination LIBS to steel samples |
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 |