CN109115701A - A kind of knitting woollen yarn Intelligent Selection color spelling cant method - Google Patents
A kind of knitting woollen yarn Intelligent Selection color spelling cant method Download PDFInfo
- Publication number
- CN109115701A CN109115701A CN201810852848.5A CN201810852848A CN109115701A CN 109115701 A CN109115701 A CN 109115701A CN 201810852848 A CN201810852848 A CN 201810852848A CN 109115701 A CN109115701 A CN 109115701A
- Authority
- CN
- China
- Prior art keywords
- color
- reflectivity
- formula
- knitting woollen
- woollen yarn
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—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
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)
- Spectrometry And Color Measurement (AREA)
Abstract
The present invention relates to a kind of knitting woollen yarns intelligently to spell cant method, includes the following steps；Step 1: measurement knitting woollen yarn sample reflectivity；Knitting woollen yarn sample reflectivity is measured using spectral luminosity instrument, test condition requires to include mirror-reflection, the reflectivity partial data for being 400~700nm comprising wavelength, 0~10nm of reflectivity wavelength measuring interval.
Description
Technical field
The present invention relates to a kind of knitting woollen yarn Intelligent Selection colors to spell cant method, belongs to weaving color technology field.
Background technique
In knitting woollen product, the kind of colour mixture is more and more after staple in bulk or wool top first dye, this colour mixture kind
Yarn is made of two or more kinds of different colours, identical or different fiber.In the production and processing of colour mixture knitting woollen yarn
Cheng Zhong spells the production link that hair is its most critical." spelling hair " is exactly by dyed different colours wool top or scattered wool fibre phase
Mutually collocation, color required for reaching and style.The spelling hair personnel that the prior art is generally in factory are by experience and try repeatedly
It spins to match colors, or by the permutation and combination method of computer, all possibility of exhaustion often provide many formulas, basic nothing
Method is quickly accurately identified its true color, so the production cycle is long, production efficiency is low, it is difficult to adapt to " small mount, multi items, fast
The market demand of delivery ".
Support vector machines (SVM) model has extensive application in multiple fields as a kind of mature identification model,
The principle is as follows: support vector machines (SVM) mainly carries out categorised decision to different classes of sample by supporting vector.Its decision according to
According to mainly an optimal classification surface (Optimal Hyper-plane) is found by Non-linear Kernel function algorithm, make its two sides sample
This class interval (Margin) is maximum.Therefore SVM algorithm itself is a kind of two-value classification method, that is, is not belonging to this kind of just belong to
In addition a kind of relationship.Its algorithm kernel is a symmetric function φ mapping k:X × X → F, therefore for all xiAnd x, all
There is k (xi, x) and={ φ (xi), φ (x) }, input space X is transformed into feature space F.
But its utilization in colour-spun yarns field is in the starting stage, such as China is disclosed one kind and supported based on least square
The invention of the colour-spun yarns color matching of vector machine, application No. is CN201710188008.9, are mentioned in that patent by SVM model use
It matches colors to colour-spun yarns, major technique is to be used to train using the reflectivity of standard sample and corresponding proportionate relationship as training sample
SVM model, can directly obtain the component color and ratio of the formula of prediction by trained SVM model, it is existing it is obvious not
Foot has:
(1) this method directly calculates ratio with least square method supporting vector machine, foundation be the reflectivity of standard sample and it is right
The relationship of ratio is answered, and SVM model is a sorting technique, for classifying in proportion, being equivalent to a ratio is exactly one point
Class will have application value, it is necessary to which as training sample, training sample amount is too big, feasible for exhaustion all composition proposal and ratio
Property is not high；
(2) it is exactly the sample being made of the reference colours that the invention, which requires target sample simultaneously, such as in the training in patent
Standard sample can forecast the ratio being closer to.And aim colour is not usually to be made of benchmark tinctorial pattern when practical application, to unknown benchmark
The sample of tinctorial pattern lacks Generalization Capability.
Summary of the invention
The purpose of the invention is to overcome problem above, a kind of knitting woollen yarn Intelligent Selection that quickly can accurately match colors is provided
Color spells cant method, simulates experience recognition capability of the human eye to color, the composition proposal and ratio all without exhaustion in Man-made Color Matching
Example is used as training sample, can reduce training sample amount, realizes intelligence color matching, and the present invention carries out group using support vector machines (SVM)
Monochromatizing identification, then matches colors on this basis, has Generalization Capability to the sample of unknown benchmark tinctorial pattern, improves color matching effect
The coincidence rate of fruit.
Knitting woollen yarn Intelligent Selection color of the invention spells cant method, includes the following steps；
Step 1: measurement knitting woollen yarn sample reflectivity: it is measured using spectral luminosity instrument, knitting woollen yarn sample reflectivity,
The partial data that it, comprising mirror-reflection, is 400~700nm comprising wavelength that test condition, which is required, reflectivity wavelength measuring interval 0~
10nm；
Step 2: SVM identifies component color；First SVM model is trained, by the knitting woollen yarn of known color compositing formula
Sample identity is standard sample and carries out measuring reflectance, and the standard sample reflectivity data collected progress data prediction is obtained
To the first data set of feature vector, the first data set of feature vector is divided into the training set of supporting vector machine model training and is tested
Card collection；The sample reflectivity data of acquisition is subjected to data prediction, pretreatment forms the second data set of feature vector；To feature
The second data set of vector is identified that the color for obtaining the component color of sample knitting woollen yarn is constituted with trained SVM model；
Step 3: calculating the best of each component color of knitting woollen yarn with colourimetric matching algorithm or full spectral match algorithm
Portfolio ratio obtains the monochromatic reflectance of each component color according to the knitting woollen yarn sample component color that step 2 obtains, with list
Color reflectivity calculates the sample reflectivity that step 1 measures, and obtains spelling hair formula.
It is above-mentioned that color is selected to spell cant method, wherein it is C-SVC that the SVM types of models, which is selected, using RBF kernel function, punishment system
Number C is 3, and kernel functional parameter γ is that 0.0323, RBF kernel function formula (1) is as follows,
K(xi, x) and=exp (- γ | | xi- x | | 2), and γ > 0 ... formula (1).
It is above-mentioned that color is selected to spell cant method, wherein the SVM model, which is used, converts two Classification and Identifications for more Classification and Identifications,
Steps are as follows,
1) y is identified1When, contain y1For one kind, remaining y2…ymFor one kind, y is determined whether there is1；
2) y is identified2When, contain y2For one kind, y1, y3…ymIt is considered as one kind, determines whether there is y2；
……
M) y is identifiedmWhen, contain ymFor one kind, y1…ym-1It is considered as one kind, determines whether there is ym；
The single of two Classification and Identification formula (3) can be converted by multiple output equations of more Classification and Identification formula (2) in this way
Output,
f(xi)={ y1,y2,…ym, xi={ x1,x2,…,xl, y=± 1, i=1 ..., n ... ... formula (2)；
f(xi)=yj, xi={ x1,x2,…,xl, yj=± 1, j=1 ..., m ... ... ... formula (3)
N is sample total number in formula, and m is to constitute color sum, and l is the dimension of input data, and formula (2) shows input
xiFor l dimension data, export as { y1,y2,…ym,+1 indicates to contain, and -1 indicates to be free of, and formula (3) expression exports result each time
The judgement of corresponding a certain single color presence or absence after calculating m times, obtains the input color if having m kind color in monochromatic library
Spin xiColor composition.
It is above-mentioned that color is selected to spell cant method, wherein it is characterized in that, the data prediction, which uses, first expands 100 times for reflectivity,
Then joint L*a*b* value uses on this basis, forms x=[100*R (λ1),100*R(λ2) ..., 100*R (λn),L*,a*,
b*]T, wherein R (λn) represent λnWhen reflectance value, n value determines by reflectivity wavelength measuring interval, and when being divided into 10, n value is
31；L*, a*, b* are CIE1976LAB value, are calculated by R (λ) according to CIELAB formula.
It is above-mentioned that color is selected to spell cant method, wherein the step 3 further includes colour mixture Reflectivity Model, and colour mixture Reflectivity Model is full
It is enough lower two relational expressions:
Wherein: Rs(λ) indicates the reflectivity of knitting woollen yarn sample when wavelength is λ, Ri(λ) indicates i component list when wavelength is λ
The reflectivity of color, αiIndicate mass ratio shared by i component monochrome, M is model parameter, according to the type of wool yarn, yarn count tune
It is whole.
Compared with prior art, the positive effect of the present invention are as follows:
It is of the invention to have the prominent advantages that the component color identification model constructed based on support vector machines, it is preparatory to component color
Identified, diminution select color range, reduce calculation amount, improve the coincidence rate of yarn effect, with traditional artificial color matching method with
And previous computer for colouring method is compared, and has speed fast, and color matching accuracy rate is high, and it is simple to operate, it is high-efficient.
Detailed description of the invention
Fig. 1 is the flow chart that Intelligent Selection color of the present invention spells hair.
Fig. 2 is the flow chart that SVM of the present invention identifies component color.
Fig. 3 is the system assumption diagram of SVM identification model of the present invention.
Fig. 4 is that HY-70291 of the present invention is formulated forecast result figure.
Fig. 5 is that 5M0040 of the present invention is formulated forecast result figure.
Fig. 6 is that HY-65511 of the present invention is formulated forecast result figure.
Fig. 7 is that the HY-65511 identified without SVM is formulated forecast result figure.
Specific embodiment
The present invention is described in detail combined with specific embodiments below.
Implement Intelligent Selection color spelling cant method of the invention and needs hardware components and software section.Wherein, hardware components are main
Including spectral luminosity instrument and computer；Software section is mainly that sample reflectivity obtains, SVM is identified and color matching operation program.Its
Working principle: as shown in Figure 1, measuring sample reflectivity as data input pin, with SVM pairs trained by spectral luminosity instrument
Pretreated sample reflectivity data carries out component color identification, anti-to sample color in conjunction with colour mixture model with the component color of identification
It penetrates rate and carries out full spectrum and colourimetric matching operation, output formula.
The present embodiment detailed step is as follows,
Step 1: the measurement of reflectivity；
Measuring reflectance: by Datacolor850 spectral luminosity instrument acquire, test condition be D65 light source, 10 ° of visual fields,
The aperture 20mm, chosen wavelength range 400-700nm are spaced 10nm.
Step 2: support vector machines (SVM) identifies component color；
1. the training of SVM；
Data: the knitting woollen yarn sample for spelling hair compositing formula known to 100 is had chosen from scale as standard sample, is shared
10 component colors (are labeled as y02,y06,y08,y10,y27,y31,y42,y47,y57,y58)。
The building and training of SVM model: SVM algorithm is by an integrated software support vector classification LIBSVM in Matlab
It is realized under R2011b version, the SVM type selected here is C-SVC, because it not only has outstanding performance, but also only two
A parameter (C, γ), C are penalty factor for the width for separating different classes of edge, the decision surface that γ determines for nuclear parameter to be arranged
Structure, the two values all determine by the grid data service that tool carries.The implementation case obtains optimal parameter are as follows: uses RBF
Function, penalty coefficient C are 3, and kernel functional parameter γ is 0.0323, and kernel function formula (1) is as follows,
K(xi, x) and=exp (- γ | | xi- x | | 2), and γ > 0 ... ... formula (1).
Training process as shown in Fig. 2, first by the knitting woollen yarn sample identity of known color compositing formula be standard sample, adopt
Collect obtained standard sample reflectivity data and carry out data prediction, data prediction is using first expanding 100 times for reflectivity, then
Combine L*, a* on this basis, b* value uses, and forms xi=[100*R (λ1),100*R(λ2) ..., 100*R (λn),L*,a*,
b*]T.The first data set of feature vector x is formed after pretreatmenti, the first data set is then divided into training set and verifying collection two
Part determines the weight and supporting vector of support vector cassification identification model using training set；Using in verification process
Determining Classification and Identification supporting vector machine model identifies that the input of verification process is the reflectivity number of verifying collection to verifying collection
According to the color exported as verifying collection is constituted, and compares the matching degree of output valve and actual value；If verification result accuracy is higher
Illustrate that the training is effective, otherwise reselects new parameter and be trained, until reaching satisfied result.Present invention training
The average trained forecast accuracy for obtaining 10 kinds of colors of knitting woollen yarn standard sample is 96%.
2. sample reflectivity pre-processes: i.e. data value amplification, 1 order derivative, 2 ranks can be used in the extraction process of spectral signature
The methods of derivative, principal component analysis, increase characteristic value.Data prediction of the present invention is using first expanding 100 times for reflectivity, then
Combine L*, a* on this basis, the method that b* value uses forms xi=[100*R (λ1),100*R(λ2) ..., 100*R (λn),
L*,a*,b*]T, wherein R (λn) represent λnWhen reflectance value, n value determines by reflectivity wavelength measuring interval, when being divided into 10,
N value is 31；L*, a*, b* are CIE1976LAB value, are calculated by R (λ) according to CIELAB formula, which is existing maturation
Technology is not done excessively repeat herein.
By the knitting woollen yarn sample reflectivity data R of acquisitionSReflectivity is first equally expanded 100 times, it is then basic herein
Upper joint L*, a*, b* value use, and wherein L*, a*, b* value can be obtained from Datacolor850 spectral luminosity instrument, by pretreatment
The second data set of feature vector is formed afterwards, and the amplification data obtained after Rs processing is as shown in table 1.
1 sample reflectivity of table amplifies 100 times of data and L*a*b* value
3. SVM selects color
When knitting woollen yarn only there are two types of composition color when, SVM color two Classification and Identification model structures such as Fig. 3 institute of building
Show, when the color of knitting woollen yarn forms more than two kinds of colors, which is exactly Classification and Identification model more than one, that is, is defined
The a certain corresponding multiple composition color outputs of sample input, as shown in formula (2),
f(xi)={ y1,y2,…ym, xi={ x1,x2,…,xl, y=± 1, i=1 ..., n ... ... formula (2)；
N is sample total number in formula, and m is to constitute color sum, and l is the dimension of input data, and formula (2) shows input
xiFor l dimension data, export as { y1,y2,…ym, symbol expression is made of corresponding m color ,+1 indicates to contain, and -1 indicates not
Contain.
Due to no upper real parallel algorithm of strictly looking like, two classification knowledge is converted into using by more Classification and Identifications here
Not, its step are as follows,
1) y is identified1When, contain y1For one kind, remaining y2…ymFor one kind, y is determined whether there is1；
2) y is identified2When, contain y2For one kind, y1, y3…ymIt is considered as one kind, determines whether there is y2；
……
M) y is identifiedmWhen, contain ymFor one kind, y1…ym-1It is considered as one kind, determines whether there is ym；
So multiple output equations of formula (2) can be converted to formula (3) individually to export:
f(xi)=yj, xi={ x1,x2,…,xl, yj=± 1, j=1 ..., m ... ... ... formula (3)
Formula (3) indicates to export the judgement that result corresponds to a certain single color presence or absence each time, if in monochromatic library
There is m kind color, then after calculating m times, obtains input colour-spun yarns xiColor composition.
By taking HY-70291 sample as an example, judgement includes y02,y06,y08,y10,y27,y31,y42,y47,y57,y58This 10 kinds of colors
In which kind, SVM formula (3) after using training calculate 10 times after output result for,
Y={ -1, -1, -1 ,+1, -1 ,+1 ,+1, -1, -1 ,+1 }, indicates the colour-spun yarns by color y10, y31, y42, and y58。
Sample reflectivity data is identified with trained SVM model, acquires the component color of the knitting woollen yarn
It is as shown in table 2 to constitute range.
The component color that 2 knitting woollen yarn sample of table is identified through SVM model
| Sample | Component color |
| HY-70291 | B-058,B-010,B-042,B-031 |
| 5M0040 | B-42,B-047,B-027 |
| HY-65511 | B-010,B-057 |
Step 3: calculating the best of each component color of knitting woollen yarn with colourimetric matching algorithm or full spectral match algorithm
Portfolio ratio obtains the monochromatic reflectance of each component color according to the knitting woollen yarn sample component color that step 2 obtains, with list
Color reflectivity calculates the sample reflectivity that step 1 measures, obtain spell hair formula, wherein monochromatic reflectance be it is existing
Primary data, this will not be repeated here.Wherein matching algorithm using optimization colour mixture Reflectivity Model, colour mixture Reflectivity Model meet with
Lower two relational expressions,
Wherein: Rs(λ) indicates the reflectivity of knitting woollen yarn sample when wavelength is λ, Ri(λ) indicates i component list when wavelength is λ
The reflectivity of color, xiIndicate mass ratio shared by i component monochrome, M is model parameter, according to the type of wool yarn, yarn count tune
It is whole.
Details are provided below with full spectral match algorithm for colourimetric matching algorithm；
(1) tristimulus values matching algorithm
According to the group monochromatizing (R selected in step 2α、Rβ…Rγ), operation calculates knitting woollen yarn using tristimulus algorithm
The composition ratio of each component color, because tristimulus values color matching need to only meet X simultaneously(s)=X(m),Y(s)=Y(m),Z(s)=Z(m)；Compared to full spectrum algorithm, this algorithm only there are three equation,
Flexibility ratio is higher, can obtain more accurate calculated result than full spectral match；
Wherein k is normaliztion constant, and Δ λ is wavelength interval.X, Y, Z are tristimulus values, RλIt is the spectral reflectivity of sample,
SλIt is the relative spectral power of standard illuminants；For standard colorimetric observer's tristimulus values；
In order to facilitate the realization of computerized algorithm, lower column matrix and vector are first defined:
Tristimulus values equation is expressed in matrix as kTSR(s)=kTSR(m)；Because monochromatic spectrum matching algorithm is chosen, therefore not
There can be the spectrum isomerism of especially severe, i.e., the respective value of reflectivity and standard color sample under any one wavelength with tinctorial pattern
It differs less big, therefore can relatively accurately be expressed as,
Definition
Wherein
R can be obtained(s)-R(m)=D [F(s)-F(m)], therefore TSDF(s)=TSDF(m)If F(m)=F × x,
WhereinF (R) represents formula (3),
Obtain TSDF(s)=TSDFx solves x=(TSDF)-1TSDF(s),
Screening obtains most always being formulated
(2) spectral match algorithm
Using least square method, make by several monochrome (R1,R2…RnAny combination) be mixed to get the curve of spectrum matching visitor
Family sample (Rs) the curve of spectrum when difference reach minimum；
I.e. setting makes to match color and sample chroman footIn formulaIndicate sample in the reflectivity of af at wavelength lambda；Indicate matching sample in the reflectivity of af at wavelength lambda.Wave-length coverage selects 400~700nm of visible light, is spaced 10nm；
Thus
I.e.Wherein f (R)λRepresent formula (3).
If sampleMatch sampleThen F(S)=F(t),
That is F(S)=F × X,
Wherein
Since above-mentioned equation is 31 equation solutions, 3 unknown quantitys, least square method solve system of equation is used
Up to X=(FT×F)-1×FT×F(S)It is corresponding monochromatic for (Rx…Rn), select the wherein lesser 2-3 group of difference
Close (Rα、Rβ…Rγ)。
In formula: subscript " T " representing matrix transposition, subscript " -1 " representing matrix are inverted；RxIndicate R1-RnIn any one
Monochrome, Rα、Rβ, RγRepresent R1-RnIn specific some is monochromatic.
Both tristimulus algorithm and full spectral match algorithm select one to carry out operation, the forecast obtained through step 3
It is formulated as shown in table 3, from table 3 it can be seen that forecast that formula and reality formula are close consistent, sample color matching forecast result such as Fig. 4-
Shown in 6, sample HY-70291,5M0040, HY-65511 are all only a kind of with number formulary, precise formulations can be locked, so this method
Forecast that accuracy is high.
The forecast formula of 3 knitting woollen yarn sample of table and practical formula
By taking HY-65511 as an example, formula and technology is carried out to it by active computer color technology.Active computer color matching
Technology is not in addition to color identification process uses SVM identification composition color model, remaining implementation condition and above-mentioned steps one, step 3 are all
It is identical.Formula and technology result without the formula and technology of SVM model identification component color as shown in fig. 7, as can be seen from Figure 7 give
Formula out is very more, has 18 kinds with number formulary, can not quick lock in precise formulations.
Specific embodiments of the present invention are described above.It is to be appreciated that the invention is not limited to above-mentioned
Particular implementation, those skilled in the art can make various deformations or amendments within the scope of the claims, this not shadow
Ring substantive content of the invention.
Claims (5)
1. a kind of knitting woollen yarn Intelligent Selection color spells cant method, which is characterized in that include the following steps；
Step 1: measurement knitting woollen yarn sample reflectivity；Knitting woollen yarn sample reflectivity, test-strips are measured using spectral luminosity instrument
The reflectivity partial data that it, comprising mirror-reflection, is 400~700nm comprising wavelength that part, which is required, reflectivity wavelength measuring interval 0~
10nm；
Step 2: support vector machines identifies component color；First supporting vector machine model is trained, by known color compositing formula
Knitting woollen yarn sample identity be standard sample and to carry out measuring reflectance, the standard sample reflectivity data collected is counted
Data preprocess obtains the first data set of feature vector, and the first data set of feature vector is divided into supporting vector machine model training
Training set and verifying collection；The knitting woollen yarn sample reflectivity data of acquisition is subjected to data prediction, pretreatment formed feature to
Two data set of flow control；The second data set of feature vector is identified with trained supporting vector machine model, obtains knitting woollen
The color of the component color of yarn sample is constituted；
Step 3: calculating the optimal combination of each component color of knitting woollen yarn with colourimetric matching algorithm or full spectral match algorithm
Ratio obtains the monochromatic reflectance of each component color according to the knitting woollen yarn sample component color that step 2 obtains, with monochromatic anti-
It penetrates rate to calculate the knitting woollen yarn sample reflectivity that step 1 measures, obtains spelling hair formula.
2. a kind of knitting woollen yarn Intelligent Selection color as described in claim 1 spells cant method, which is characterized in that the support vector machines
Types of models is C-SVC, and using RBF kernel function, penalty coefficient C is 3, and kernel functional parameter γ is 0.0323, RBF kernel function formula
(1) as follows；
K(xi, x) and=exp (- γ | | xi-x||2), γ > 0 ... ... formula (1).
3. a kind of knitting woollen yarn Intelligent Selection color as described in claim 1 spells cant method, which is characterized in that the support vector machines
Model uses the method for converting two Classification and Identifications for more Classification and Identifications, and its step are as follows；
1) y is identified1When, contain y1For one kind, remaining y2…ymFor one kind, y is determined whether there is1；
2) y is identified2When, contain y2For one kind, y1, y3…ymIt is considered as one kind, determines whether there is y2；
……
M) y is identifiedmWhen, contain ymFor one kind, y1…ym-1It is considered as one kind, determines whether there is ym；
So multiple output equations of more Classification and Identification formula (2) can be converted to the single output of two Classification and Identification formula (3)；
f(xi)={ y1,y2,…ym, xi={ x1,x2,…,xl, y=± 1, i=1 ..., n ... formula (2)；
f(xi)=yj, xi={ x1,x2,…,xl, yj=± 1, j=1 ..., m ... ... ... formula (3)；
N is sample total number in formula, and m is to constitute color sum, and l is the dimension of input data, and formula (2) shows to input xiFor l
Dimension data exports as { y1,y2,…ym,+1 indicates to contain, and -1 indicates to be free of, and formula (3) indicates that output result is corresponding each time
The judgement of a certain single color presence or absence calculates m times if there is m kind color in supporting vector machine model monochromatic data library
Afterwards, input colour-spun yarns x is obtainediColor composition.
4. a kind of knitting woollen yarn Intelligent Selection color as described in claim 1 spells cant method, which is characterized in that the data prediction
Using reflectivity is first expanded 100 times, then joint L*a*b* value is used on this basis, forms x=[100*R (λ1),100*
R(λ2) ..., 100*R (λn),L*,a*,b*]T, wherein R (λn) represent λnWhen reflectance value, n value by reflectivity wavelength measure between
Every decision, when being divided into 10, n value is 31；L*, a*, b* are CIE1976LAB value, are calculated by R (λ) according to CIELAB formula
It arrives.
5. a kind of knitting woollen yarn Intelligent Selection color as described in claim 1 spells cant method, which is characterized in that the step 3 is also wrapped
Colour mixture Reflectivity Model is included, colour mixture Reflectivity Model meets following two relational expressions；
Wherein；Rs(λ) indicates the reflectivity of knitting woollen yarn sample when wavelength is λ, Ri(λ) indicates that i component is monochromatic when wavelength is λ
Reflectivity, αiIndicate mass ratio shared by i component monochrome, M is model parameter, is adjusted according to the type of wool yarn, yarn count.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810852848.5A CN109115701B (en) | 2018-07-30 | 2018-07-30 | Intelligent color-selecting wool-splicing method for wool knitting yarns |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810852848.5A CN109115701B (en) | 2018-07-30 | 2018-07-30 | Intelligent color-selecting wool-splicing method for wool knitting yarns |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109115701A true CN109115701A (en) | 2019-01-01 |
| CN109115701B CN109115701B (en) | 2021-01-15 |
Family
ID=64863656
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810852848.5A Active CN109115701B (en) | 2018-07-30 | 2018-07-30 | Intelligent color-selecting wool-splicing method for wool knitting yarns |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109115701B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750403B (en) * | 2012-05-28 | 2016-06-15 | 嘉兴学院 | The recipe determination of a kind of colour-spun yarns color matching and modification method |
| CN107103181A (en) * | 2017-03-27 | 2017-08-29 | 东华大学 | A kind of colour-spun yarns color matching method based on least square method supporting vector machine |
| CN107766603A (en) * | 2017-04-24 | 2018-03-06 | 东华大学 | A kind of colour-spun yarns computer is measured color method |
-
2018
- 2018-07-30 CN CN201810852848.5A patent/CN109115701B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750403B (en) * | 2012-05-28 | 2016-06-15 | 嘉兴学院 | The recipe determination of a kind of colour-spun yarns color matching and modification method |
| CN107103181A (en) * | 2017-03-27 | 2017-08-29 | 东华大学 | A kind of colour-spun yarns color matching method based on least square method supporting vector machine |
| CN107766603A (en) * | 2017-04-24 | 2018-03-06 | 东华大学 | A kind of colour-spun yarns computer is measured color method |
Non-Patent Citations (4)
| Title |
|---|
| BANDPAY, M. GORJI,ET AL: "Mathematical and empirical evaluation of accuracy of the Kubelka–Munk model for color match prediction of opaque and translucent surface coatings", 《J. COAT. TECHNOL. RES.》 * |
| 张永昌: "基于最小二乘拟合的数值分析方法在织物染色配色中的应用研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
| 李含春: "数值分析和支持向量机在织物染色配色中的对比研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 沈加加 等: "基于约束最小二乘法的色纺纱配色方法", 《针织工业》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109115701B (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107766603B (en) | A kind of colour-spun yarns computer is measured color method | |
| CN107103181A (en) | A kind of colour-spun yarns color matching method based on least square method supporting vector machine | |
| CN101668109B (en) | Soft-proofing method for color-spinning yarn by tri-stimulus value matching | |
| CN100510703C (en) | Textile fibre identification method | |
| CN101833593A (en) | Computer color-matching method of color-blended wool fabrics based on full-spectrum color matching | |
| CN105787559A (en) | Colored spun yarn color matching method based on neural network | |
| CN110082310A (en) | A kind of near infrared band EO-1 hyperion diagnostic method of rubber tree LTN content | |
| CN109582994A (en) | A kind of colour-spun yarns Intelligent Selection color color matching method | |
| CN106934070A (en) | A kind of color-woven fabric design producing method of fast reaction | |
| CN105628708A (en) | Quick nondestructive testing method for multi-parameter quality of south Xinjiang red dates | |
| Fortier et al. | Near-infrared classification of cotton lint, botanical and field trash | |
| CN109671113A (en) | A kind of color textile fabric based on image technique is measured color method | |
| CN114235153A (en) | Digital camera-based fabric color digital measurement method | |
| CN108960193A (en) | A kind of across component infrared spectroscopy model transplantations method based on transfer learning | |
| CN109115701A (en) | A kind of knitting woollen yarn Intelligent Selection color spelling cant method | |
| CN109146270B (en) | Establishment method of vegetable seedling strengthening index | |
| CN104865218B (en) | The quick method for determining cotton ramie blended spinning fiber content | |
| CN105787275B (en) | A kind of computer for colouring method of leather coloring | |
| CN108828147A (en) | A kind of rigid bamboo poison moth hazard detection method coupling remote sensing response characteristic | |
| Xu et al. | Cotton color grading with a neural network | |
| Afjeh et al. | Colorimetric and spectral properties of natural colorants used in handmade traditional Persian carpets | |
| Di | Research on Supplier Evaluation Based on Entropy Method and TOPSIS | |
| CN101894174A (en) | Computer-assisted twisting and color-matching method for color-mixed wool fabric | |
| CN109813431A (en) | A kind of one-step method color matching method of color textile fabric | |
| CN109636861A (en) | A kind of prediction technique of colour-spun yarns fabric appearance benefit |
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 |