CN104410850B - Colorful digital image chrominance correction method and system - Google Patents
Colorful digital image chrominance correction method and system Download PDFInfo
- Publication number
- CN104410850B CN104410850B CN201410819987.XA CN201410819987A CN104410850B CN 104410850 B CN104410850 B CN 104410850B CN 201410819987 A CN201410819987 A CN 201410819987A CN 104410850 B CN104410850 B CN 104410850B
- Authority
- CN
- China
- Prior art keywords
- subset
- packet
- chrominance information
- color sample
- light source
- 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.)
- Expired - Fee Related
Links
Landscapes
- Color Image Communication Systems (AREA)
- Image Processing (AREA)
- Spectrometry And Color Measurement (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention discloses a colorful digital image chrominance correction method and system. The colorful digital image chrominance correction method comprises the following steps: building a typical color sample spectral reflectivity data set, calculating chrominance information of each color sample in the data set under the condition with an original light source and grouping samples by taking main wavelength and color purity as basis, and solving the chrominance information of each color sample in each group of subset under the condition with a target light source; taking the chrominance information of each group of sample subset under the conditions with the original and target light sources as input-output end, and fitting and building a neural network; determining the corresponding neural network by a grouping and judging method on account of any chrominance information of the original light source, and forecasting the chrominance information under the corresponding target light source according to the neural network. By virtue of the colorful digital image chrominance correction method, the mapping accuracy of the colorful digital image chrominance information under different illumination conditions can be ensured; meanwhile, the method is convenient to implement.
Description
Technical field
The invention belongs to color digital image record and reproducing technology field are and in particular to a kind of be based on typical color sample
The color digital image chromaticity correction method and system of spectral reflectance data collection.
Background technology
Color digital image system is one of important carrier of objective things information record and reproduction.In actual applications,
Affected by different objective environment conditions and different itself otherness of chromatic image equipment, color digital image chrominance information record
There is diversity with the light conditions reproducing.For ensureing the standard of color digital image information record and colouring information in reproducing processes
Really property, needs by specific color correcting method to realize the accurate mapping of image chrominance information under different light conditions.
For this problem, the most commonly used solution of current industry is using chromatic adaptation transform method, same to realize
Accurate mapping under the conditions of different illumination scenes for the color information.The method is passed through to simulate human eye chromatic adaptation characteristic, by knot
Close different Illuminant chromaticity information, realize by object chrominance information under original light source to the simulation of object chrominance information under target light source
Prediction, and then ensure the accuracy of image object color information transmission.At present, in color digital image record and field of reproduction,
Industry proposes many classics chromatic adaptation transform methods, such as Von Kries method, Wrong Von Kries method, Bradford
Method, Helson method, Bartleson method and Hunt method etc..
Bibliography 1.H.R.Kang.Computational color technology [M] .Society of Photo
Optical, 2006.
Bibliography 2.M.R.Luo.A review of chromatic adaptation transforms [J]
.Review of Progress in Coloration and Related Topics, 2000.
Such method passes through the simulation of human eye chromatic adaptation mechanism, to some extent solves color digital image system
The problem that under different light conditions, chrominance information accurately maps.However, because the structure of above-mentioned chromatic adaptation transform method is all based on
Human eye vision psychophysics experiments, that is, said method is mainly basic for building with human eye vision subjective matching, therefore in colourity school
There is more obvious defect in positive objective and accurate property aspect.For this reason, in current research application, existing researcher is devoted to
Build chromaticity correction method from objective angle, to realize the color digital image chromaticity correction of higher precision, as bibliography 3 institute
State.
Bibliography 3.Rok Kreslin et al.Linear Chromatic Adaptation Transform
Based on Delaunay Triangulation [J] .Mathematical Problems in Engineering, 2014.
However, being restricted by subjective and objective factors such as theoretical method levels, above-mentioned objective method is in chromaticity correction accuracy side
Face equally exists the more obvious defect of the excessive grade of saturation color domain error.For problem above, academic circles at present with
Industrial quarters not yet proposes corresponding solution, to realize the accurate mapping of colors of image chrominance information under the conditions of different light scenes
With transmission.
Content of the invention
The invention aims to problem described in solution background technology, one kind is proposed based on typical color sample spectrum
The color digital image chromaticity correction method and system of reflectivity data collection.
A kind of color digital image chromaticity correction method of offer is provided, comprises the following steps：
Step 1, chooses M typical color sample, with the spectral reflectivity number in each typical case's color sample visible-range
According to the typical color sample spectrum reflectivity data collection G of compositions；
Step 2, based on the spectral reflectance data of each typical case's color sample in step 1, public using following colorimetry
Formula calculates each sample respectively in source light source LsUnder the conditions of chrominance information, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ,
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye
Vision matching function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra
Reflectivity S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is parameter；
Step 3, calculates typical color sample chroma-data set GcIn each sample dominant wavelength and excitation information, with dominant wavelength
For packet according to typical color sample chroma-data set GsIt is grouped first, subsequently with excitation for packet according to first
Packet gained set carries out secondary packet, and the final number of packet of note is T, obtains T subset；
Step 4, for gained typical case's color sample chroma-data set G after two packets of step 3sT subset, respectively with
Purpose light source LtFor lighting source E (λ), using the purpose light source L of colorimetry equations described in step 2tUnder the conditions of each in subset
The chrominance information of sample；
Step 5, for each subset, is grouped the source light source L of this subset of gained respectively with step 3sUnder the conditions of this colourity of various kinds
Information is input, and solve the purpose light source Lt of this subset of gained with step 4 under the conditions of, this chrominance information of various kinds, as output end, is instructed
Practice corresponding BP neural network；
Step 6, for a certain chrominance information Cs under the light conditions of source, calculates dominant wavelength and excitation information, according to step 3
Packet mode, determine corresponding subset and BP neural network, and using this BP neural network prediction chrominance information Cs in purpose
Corresponding chrominance information Ct under light conditions.
And, with dominant wavelength to data set G in step 3sWhen being grouped first by dominant wavelength for negative value all samples
As one group, afterwards other samples are according to carrying out average packet with dominant wavelength；Subsequently the above-mentioned gained of packet first is owned
Packet subset, with excitation for secondary packet foundation, carries out average packet.
The present invention provides a kind of color digital image chromaticity correction system, including with lower module：
Typical color sample spectrum data set builds module, for choosing M typical color sample, with each typical case's color sample
Spectral reflectance data in this visible-range constitutes typical color sample spectrum reflectivity data collection Gs；
Typical color sample chroma-data set computing module, for being built in module with typical color sample spectrum data set
Based on the spectral reflectance data of each typical case's color sample, calculate each sample respectively in source light source using following colorimetry formula
LsUnder the conditions of chrominance information, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ,
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye
Vision matching function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra
Reflectivity S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is parameter；
Data set grouping module, for calculating typical color sample chroma-data set GcIn each sample dominant wavelength and excitation
Information, with dominant wavelength for packet according to typical color sample chroma-data set GsIt is grouped first, with excitation be subsequently
According to carrying out secondary packet to being grouped gained set first, the final number of packet of note is T, obtains T subset for packet；
Packet subset chrominance information solves module, for for gained typical case's color after the secondary packet of data set grouping module
Sample chroma-data set GsT subset, respectively with purpose light source LtFor lighting source E (λ), asked using described colorimetry formula
Solution purpose light source LtUnder the conditions of in subset each sample chrominance information；
Neural metwork training module, for for each subset, being grouped this subset of gained with data set grouping module respectively
Source light source LsUnder the conditions of this chrominance information of various kinds be input, be grouped subset chrominance information solve module solve this subset of gained
Purpose light source LtUnder the conditions of this chrominance information of various kinds be output end, train corresponding BP neural network；
Chromaticity correction module, is believed with excitation for for a certain chrominance information Cs under the light conditions of source, calculating dominant wavelength
Breath, according to the packet mode of data set grouping module, determines corresponding subset and BP neural network, and utilizes this BP neural network
Corresponding chrominance information Ct under purpose light conditions of prediction chrominance information Cs.
And, with dominant wavelength to data set G in data set grouping modulesWhen being grouped first dominant wavelength is negative value
Other samples, as one group, are according to carrying out average packet with dominant wavelength by all samples afterwards；Subsequently it is grouped above-mentioned first
Gained all packets subset, with excitation for secondary packet foundation, carries out average packet.
A kind of color digital image colourity school based on typical color sample spectrum reflectivity data collection proposed by the present invention
Positive technical scheme, in the premise determining typical color sample spectrum reflectivity data collection and chromaticity correction source and purpose light source
Under, in conjunction with dominant wavelength and excitation group technology, same color card is built under different lighting conditions by BP neural network
The relevance model of colourity difference, so achieve under the conditions of different light scenes the accurate mapping of colors of image chrominance information with
Transmission.What the method was ideal solves problem described in background section, thereby may be ensured that color digital image information
The accuracy of transmittance process, and then meet high-quality chromatic image information record and the demand reproducing.Therefore, the present invention solves
The problem of the accurate transmission of colors of image chrominance information under different lighting conditions, and easy to implement, color digital image record with
Field of reproduction has stronger applicability.Because technical solution of the present invention has important application meaning, by multiple project supports：
1. fund 2014M5606253.2. National Nature fund project 61275172.3. State Cultural Relics Bureau historical relic on China's post-doctors face
Protection field Science and Technology research general problem 2013-YB-HT-034.4. country 973 basic research sub-projects
2012CB725302.Technical solution of the present invention is protected, will have weight to China's relevant industries competition first place in the world
Want meaning.
Brief description
Fig. 1 is the flow chart of the embodiment of the present invention.
Specific embodiment
In conjunction with accompanying drawing, the embodiment of the present invention is provided to be described in detail below.
A kind of colorful digital shadow based on typical color sample spectrum reflectivity data collection that embodiment provides as shown in Figure 1
As chromaticity correction method, the ideal chromaticity correction solving the problems, such as under the conditions of different images object and light scene, can
To ensure the accuracy of color digital image information transmittance process, and then meet high-quality chromatic image information record and reproduction
Demand.Embodiment adopts 9297 color samples to build typical color sample set, with 1250 Meng Saier tarnish color samples
As experimental check sample set, with D65 standard illuminants for source light source, with A standard illuminants for purpose light source, with institute of the present invention
The method of stating carries out chromaticity correction.And by Von Kries method, Wrong Von Kries method, Bradford method, Helson
Method, Bartleson method, in 6 kinds of chromatic adaptation conversion such as Hunt method and bibliography 3 Kreslin method totally 7 kinds existing
Method is as comparison.It should be noted that the invention is not limited in above-mentioned image object and light source type, for other images
Object and light source type, this method is equally applicable.
Computer software technology can be adopted to realize automatically by those skilled in the art when technical solution of the present invention is embodied as
Run.The method flow that embodiment provides comprises the following steps：
1) choose M typical color sample, typical color is constituted with the spectral reflectance data in each sample visible-range
Color sample spectrum reflectivity data collection Gs；
Those skilled in the art can voluntarily preset the value of M.When being embodied as, sample set color gamut should be ensured as far as possible
Maximization.Visible-range is generally 380nm 780nm.When being embodied as, can be in advance with each sample of spectrophotometer measurement
Corresponding spectrum reflectivity information, take 380nm 780nm wave band data.In embodiment, by uniform sampling in printing device colour gamut
6000 color samples of preparation, 1687 Japanese typical pigments color samples and 1600 Meng Saier gloss color samples are altogether
As typical sample collection (M=9297), this sample set has wide colour gamut to 9297 samples, and sample distribution is uniform.Specifically
During enforcement, typical color sample spectrum reflectivity data collection G can be previously generatedsAnd input.
2) by 1) in each sample spectral reflectance data based on, calculate each sample respectively using following colorimetry formula
In source light source LsUnder the conditions of chrominance information, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ, formula one
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye
Vision matching function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra
Reflectivity S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is the parameter being determined by y (λ), E (λ)；
In an embodiment, this step realizes M=9297 typical color sample in source light source LSUnder chromatic value solve, respectively
Chromatic value forms typical color sample chroma-data set Gc.Wherein, source light source LSIt is set to D65 standard illuminants, that is, E (λ) adopts
D65 standard illuminants corresponding relative spectral power distributions curve.For each sample, S (λ) is respectively adopted corresponding visible-range
Interior spectral reflectance data.
3) typical color sample chroma-data set G is calculated with the existing dominant wavelength of colourity theory and excitation computing formulac
In each sample dominant wavelength and excitation information, with dominant wavelength for packet according to data set GsIt is grouped first, subsequently with color
Purity be packet according to carrying out secondary packet to being grouped gained set first, the final number of packet of note is T；
In an embodiment, with dominant wavelength to data set GsWhen being grouped first, by dominant wavelength for negative value all samples
As one group, afterwards other samples are that (concrete packet count can be by people in the art according to carrying out average packet with dominant wavelength
Member sets), the dominant wavelength of embodiment is 5 on the occasion of average number of packets, is divided into 6 groups together with negative value sample group；Subsequently by above-mentioned head
, with excitation for secondary packet foundation, (concrete packet count can be by this area to carry out average packet for secondary 6 groups of packet subsets of packet gained
Technical staff sets), the average number of packets of the secondary packet of embodiment is 2, final that T=12 is grouped.Its dominant wavelength and colour purity
Degree scope is respectively
First group：380nm≤dominant wavelength≤480nm, excitation≤0.31；
Second group：380nm≤dominant wavelength≤480nm, excitation ＞ 0.31；
3rd group：480nm ＜ dominant wavelength≤503nm, excitation≤0.21；
4th group：480nm ＜ dominant wavelength≤503nm, excitation ＞ 0.21；
5th group：503nm ＜ dominant wavelength≤569nm, excitation≤0.23；
6th group：503nm ＜ dominant wavelength≤569nm, excitation ＞ 0.23；
7th group：569nm ＜ dominant wavelength≤588nm, excitation≤0.39；
8th group：569nm ＜ dominant wavelength≤588nm, excitation ＞ 0.39；
9th group：588nm ＜ dominant wavelength≤780nm, excitation≤0.41；
Tenth group：588nm ＜ dominant wavelength≤780nm, excitation ＞ 0.41；
11st group：Dominant wavelength ＜ 0, excitation≤0.18；
12nd group：Dominant wavelength ＜ 0, excitation ＞ 0.18；
Wherein, the computational methods of dominant wavelength and excitation can be found in J.Schanda.CIE colorimetry [M] .Wiley
Online Library, 2007, it will not go into details for the present invention.
4) be directed to 3) secondary packet after the data obtained collection GsEach subset, be utilized respectively 2) described in method solve purpose
Light source LtUnder the conditions of in each group subset each sample chrominance information, bag is with purpose light source LtCarry out by formula one for lighting source E (λ)
Solve；
In an embodiment, for 3) the data obtained collection Gs12 subsets, be utilized respectively 2) Chinese style one, with standard illuminants
A is purpose light source Lt, solve the chrominance information of each sample in each group subset under purpose light conditions.
5) be directed to each subset, respectively with 3) packet this group subset of gained source light source LsUnder the conditions of this chrominance information of various kinds be
Input, with 4) solve this group subset of gained purpose light source LtUnder the conditions of this chrominance information of various kinds be output end, training is corresponding
BP neural network；
In embodiment, with 3) in each subset of obtaining of packet in source light source LsUnder the conditions of chrominance information as input number
According to 4) in solve each group subset in target light source LtUnder the conditions of chrominance information as output data, build BP nerve net
Network.Wherein, for 3) in gained 12 packet subsets, 12 BP neural networks need to be built altogether.When being embodied as, can be found in BP god
Realize through network prior art.
So, based on each packet samples collection under packet samples collection chrominance information each under the light conditions of source and purpose light conditions
Chrominance information, can build neutral net for each corresponding grouped data matching, for chrominance information under subsequent source light conditions
Chrominance information under purpose light conditions is obtained based on group-discriminate.
6) it is directed to a certain chrominance information Cs under the light conditions of source, calculated using the existing dominant wavelength of colourity theory and excitation
Formula calculates its dominant wavelength and excitation information, and combines 3) described packet situation, determine corresponding subset and BP nerve
Network, and predict its corresponding chrominance information Ct under purpose light conditions using this BP neural network.
In an embodiment, taking certain color sample chrominance information Cs under the light conditions of source as a example, its CIEXYZ value for (84,
89,99) calculate its dominant wavelength and excitation information using the existing dominant wavelength of colourity theory and excitation computing formula, obtain its master
Wavelength is 477nm, and excitation is 0.01, then by embodiment 3) understand that it belongs to first packet, therefore utilize first group of data
Corresponding BP neural network predicts its chrominance information Ct under the conditions of target light source, solve its CIEXYZ value for (97,88,
33), very close with theoretical value (98,89,32).
When being embodied as, pre-build son to any purpose light conditions according to step 1～5 for any source light conditions
Collection divides and corresponding BP neural network, you can for the prediction of corresponding chrominance information.
For being further characterized by the advantage in chromaticity correction precision aspect for the inventive method, with 1250 Meng Saier tarnish colors
Color sample as experimental check sample set, with D65 standard illuminants for source light source, with A standard illuminants for purpose light source, with this
Invention methods described carries out chromaticity correction.And by Von Kries method, Wrong Von Kries method, Bradford method,
Helson method, Bartleson method, Kreslin method totally 7 in 6 kinds of chromatic adaptation conversion such as Hunt method and bibliography 3
Plant existing method as comparison.Experimental result shows, Von Kries method, Wrong Von Kries method, Bradford side
Method, Helson method, Bartleson method, Hunt method and Kreslin method are represented with colour difference formula CIEDE2000
Chromaticity correction precision is respectively 4.67,2.86,5.38,3.36,4.76,4.13,2.11, and chromaticity correction precision of the present invention is
1.53, accuracy benefits are obvious.Wherein, CIEDE2000 colour difference formula can be found in Ming R Luo.CIE 2000 color
difference formula:CIEDE2000[A].In 9th Congress of the International Color
Association [C], Year：It will not go into details for the 554-9. present invention.
The present invention correspondingly provides a kind of color digital image chromaticity correction system, including with lower module：
Typical color sample spectrum data set builds module, for choosing M typical color sample, with each typical case's color sample
Spectral reflectance data in this visible-range constitutes typical color sample spectrum reflectivity data collection Gs；
Typical color sample chroma-data set computing module, for being built in module with typical color sample spectrum data set
Based on the spectral reflectance data of each typical case's color sample, calculate each sample respectively in source light source using following colorimetry formula
LsUnder the conditions of chrominance information, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ,
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye
Vision matching function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra
Reflectivity S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is parameter；
Data set grouping module, for calculating typical color sample chroma-data set GcIn each sample dominant wavelength and excitation
Information, with dominant wavelength for packet according to typical color sample chroma-data set GsIt is grouped first, with excitation be subsequently
According to carrying out secondary packet to being grouped gained set first, the final number of packet of note is T, obtains T subset for packet；
Packet subset chrominance information solves module, for for gained typical case's color after the secondary packet of data set grouping module
Sample chroma-data set GsT subset, respectively with purpose light source LtFor lighting source E (λ), asked using described colorimetry formula
Solution purpose light source LtUnder the conditions of in subset each sample chrominance information；
Neural metwork training module, for for each subset, being grouped this subset of gained with data set grouping module respectively
Source light source LsUnder the conditions of this chrominance information of various kinds be input, be grouped subset chrominance information solve module solve this subset of gained
Purpose light source LtUnder the conditions of this chrominance information of various kinds be output end, train corresponding BP neural network；
Chromaticity correction module, is believed with excitation for for a certain chrominance information Cs under the light conditions of source, calculating dominant wavelength
Breath, according to the packet mode of data set grouping module, determines corresponding subset and BP neural network, and utilizes this BP neural network
Corresponding chrominance information Ct under purpose light conditions of prediction chrominance information Cs.
Wherein, in data set grouping module with dominant wavelength to data set GsWhen being grouped first dominant wavelength is negative value
Other samples, as one group, are according to carrying out average packet with dominant wavelength by all samples afterwards；Subsequently it is grouped above-mentioned first
Gained all packets subset, with excitation for secondary packet foundation, carries out average packet.
Each module implements with each step accordingly, and it will not go into details for the present invention.
Specific embodiment described herein is only explanation for example to present invention spirit.The affiliated technology of the present invention is led
The technical staff in domain can be made various modifications or supplement or replaced using similar mode to described specific embodiment
Generation, but the spirit without departing from the present invention or surmount scope defined in appended claims.
Claims (4)
1. a kind of color digital image chromaticity correction method is it is characterised in that be based on typical color sample spectrum reflectivity data
Chromaticity correction realized by collection, comprises the following steps：
Step 1, chooses M typical color sample, with the spectral reflectance data structure in each typical case's color sample visible-range
Become typical color sample spectrum reflectivity data collection Gs；
Step 2, based on the spectral reflectance data of each typical case's color sample in step 1, is divided using following colorimetry formula
Do not calculate each sample in source light source LsUnder the conditions of chrominance information, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ,
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye vision
Adaptation function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra reflects
Rate S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is parameter；
Step 3, calculates typical color sample chroma-data set GcIn each sample dominant wavelength and excitation information, with dominant wavelength for point
Group is according to typical color sample spectrum reflectivity data collection GsIt is grouped first, subsequently with excitation for packet according to head
Secondary packet gained set carries out secondary packet, and the final number of packet of note is T, obtains T subset；
Step 4, for gained typical case color sample spectrum reflectivity data collection G after two packets of step 3sT subset, respectively
With purpose light source LtFor lighting source E (λ), using the purpose light source L of colorimetry equations described in step 2tUnder the conditions of in subset
The chrominance information of each sample；
Step 5, for each subset, is grouped the source light source L of this subset of gained respectively with step 3sUnder the conditions of this chrominance information of various kinds be
Input, solves the purpose light source L of this subset of gained with step 4tUnder the conditions of this chrominance information of various kinds be output end, training is corresponding
BP neural network；
Step 6, for a certain chrominance information Cs under the light conditions of source, calculates dominant wavelength and excitation information, dividing according to step 3
Group mode, determines corresponding subset and BP neural network, and using this BP neural network prediction chrominance information Cs in purpose light source
Under the conditions of corresponding chrominance information Ct.
2. according to claim 1 color digital image chromaticity correction method it is characterised in that：With dominant wavelength pair in step 3
Typical color sample spectrum reflectivity data collection GsWhen being grouped first using dominant wavelength for negative value all samples as one group,
Afterwards other samples are according to carrying out average packet with dominant wavelength；Subsequently by all for the above-mentioned gained of packet first be grouped subset with
Excitation is secondary packet foundation, carries out average packet.
3. a kind of color digital image chromaticity correction system is it is characterised in that be used for based on typical color sample spectrum reflectivity
Data set realizes chromaticity correction, including with lower module：
Typical color sample spectrum data set builds module, for choosing M typical color sample, can with each typical case's color sample
See that the spectral reflectance data in optical range constitutes typical color sample spectrum reflectivity data collection Gs；
Typical color sample chroma-data set computing module, for building each allusion quotation in module with typical color sample spectrum data set
Based on the spectral reflectance data of type color sample, calculate each sample respectively in source light source L using following colorimetry formulasBar
Chrominance information under part, and form typical color sample chroma-data set Gc,
X=k ∫ x (λ) E (λ) S (λ) d λ,
Y=k ∫ y (λ) E (λ) S (λ) d λ,
Z=k ∫ z (λ) E (λ) S (λ) d λ,
K=100/ [∫ y (λ) E (λ) d λ],
Wherein, X, Y, Z represent colourity tristimulus values, and λ represents each band wavelength of visible ray；X (λ), y (λ), z (λ) are human eye vision
Adaptation function, lighting source E (λ) adopts source light source LsCorresponding relative spectral power distributions curve, color object spectra reflects
Rate S (λ) adopts the spectral reflectance data in the corresponding visible-range of sample, and k is parameter；
Data set grouping module, for calculating typical color sample chroma-data set GcIn each sample dominant wavelength and excitation information,
With dominant wavelength for packet according to typical color sample spectrum reflectivity data collection GsIt is grouped first, with excitation be subsequently
According to carrying out secondary packet to being grouped gained set first, the final number of packet of note is T, obtains T subset for packet；
Packet subset chrominance information solves module, for for gained typical case's color sample after the secondary packet of data set grouping module
Spectral reflectance data collection GsT subset, respectively with purpose light source LtFor lighting source E (λ), using described colorimetry formula
Solve purpose light source LtUnder the conditions of in subset each sample chrominance information；
Neural metwork training module, for for each subset, being grouped the source light of this subset of gained respectively with data set grouping module
Source LsUnder the conditions of this chrominance information of various kinds be input, be grouped subset chrominance information solve module solve this subset of gained mesh
Light source LtUnder the conditions of this chrominance information of various kinds be output end, train corresponding BP neural network；
Chromaticity correction module, for for a certain chrominance information Cs under the light conditions of source, calculating dominant wavelength and excitation information, root
According to the packet mode of data set grouping module, determine corresponding subset and BP neural network, and using the prediction of this BP neural network
Corresponding chrominance information Ct under purpose light conditions of chrominance information Cs.
4. according to claim 3 color digital image chromaticity correction system it is characterised in that：In data set grouping module with
Dominant wavelength is to typical color sample spectrum reflectivity data collection GsWhen being grouped first by dominant wavelength for negative value all samples
As one group, afterwards other samples are according to carrying out average packet with dominant wavelength；Subsequently the above-mentioned gained of packet first is owned
Packet subset, with excitation for secondary packet foundation, carries out average packet.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410819987.XA CN104410850B (en) | 2014-12-25 | 2014-12-25 | Colorful digital image chrominance correction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410819987.XA CN104410850B (en) | 2014-12-25 | 2014-12-25 | Colorful digital image chrominance correction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104410850A CN104410850A (en) | 2015-03-11 |
| CN104410850B true CN104410850B (en) | 2017-02-22 |
Family
ID=52648436
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410819987.XA Expired - Fee Related CN104410850B (en) | 2014-12-25 | 2014-12-25 | Colorful digital image chrominance correction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104410850B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106053024B (en) * | 2016-06-27 | 2018-08-10 | 武汉大学 | A kind of LED light source preference degree prediction technique towards monochromatic system object |
| CN107093031A (en) * | 2017-05-10 | 2017-08-25 | 广东溢达纺织有限公司 | Color data Internet management method and system |
| CN107507250B (en) * | 2017-06-02 | 2020-08-21 | 北京工业大学 | Surface color and tongue color image color correction method based on convolutional neural network |
| CN108519157B (en) * | 2018-03-16 | 2019-07-09 | 武汉大学 | A kind of generation of metamerism spectrum and evaluation method and system for light source detection |
| CN109274945B (en) * | 2018-09-29 | 2020-05-22 | 合刃科技（深圳）有限公司 | Method and system for self-adaptively performing true color restoration on image |
| CN110487403A (en) * | 2019-09-02 | 2019-11-22 | 常州市武进区半导体照明应用技术研究院 | A kind of prediction technique of LED light spectral power distributions |
| CN116678839B (en) * | 2023-07-13 | 2023-11-10 | 季华实验室 | Luminescent material detection method, device, terminal equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101447020A (en) * | 2008-12-12 | 2009-06-03 | 北京理工大学 | Pornographic image recognizing method based on intuitionistic fuzzy |
| CN102075666A (en) * | 2009-11-25 | 2011-05-25 | 惠普开发有限公司 | Method and device used for removing background colors from image |
| WO2011103377A1 (en) * | 2010-02-22 | 2011-08-25 | Dolby Laboratories Licensing Corporation | System and method for adjusting display based on detected environment |
| CN102572450A (en) * | 2012-01-10 | 2012-07-11 | 中国传媒大学 | Three-dimensional video color calibration method based on scale invariant feature transform (SIFT) characteristics and generalized regression neural networks (GRNN) |
| CN104092919A (en) * | 2014-07-14 | 2014-10-08 | 武汉大学 | Chromatic adaptation transformation optimizing method and system for color digital imaging system |
-
2014
- 2014-12-25 CN CN201410819987.XA patent/CN104410850B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101447020A (en) * | 2008-12-12 | 2009-06-03 | 北京理工大学 | Pornographic image recognizing method based on intuitionistic fuzzy |
| CN102075666A (en) * | 2009-11-25 | 2011-05-25 | 惠普开发有限公司 | Method and device used for removing background colors from image |
| WO2011103377A1 (en) * | 2010-02-22 | 2011-08-25 | Dolby Laboratories Licensing Corporation | System and method for adjusting display based on detected environment |
| CN102572450A (en) * | 2012-01-10 | 2012-07-11 | 中国传媒大学 | Three-dimensional video color calibration method based on scale invariant feature transform (SIFT) characteristics and generalized regression neural networks (GRNN) |
| CN104092919A (en) * | 2014-07-14 | 2014-10-08 | 武汉大学 | Chromatic adaptation transformation optimizing method and system for color digital imaging system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104410850A (en) | 2015-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104410850B (en) | Colorful digital image chrominance correction method and system | |
| CN105069234B (en) | The spectrum dimension reduction method and system of a kind of view-based access control model Perception Features | |
| CN104217409B (en) | A kind of image color correction method based on simulated annealing optimization algorithm | |
| CN103954362B (en) | A kind of digital method for measuring color based on imaging device | |
| CN104485068A (en) | Luminance-chrominance correction method and system of LED (Light Emitting Diode) display screen | |
| CN106531060A (en) | Luminance correcting method and device for LED display device | |
| CN103854261B (en) | The bearing calibration of colour cast image | |
| US10902801B2 (en) | Driving method and apparatus for display apparatus | |
| CN103612483A (en) | Printing ink color matching method based on spectral matching | |
| CN113673389B (en) | Painting illumination visual evaluation method related to spectrum power distribution of light source | |
| CN104574371A (en) | Characterization calibration method for high dynamic digital color camera | |
| CN104092919B (en) | Chromatic adaptation transformation optimizing method and system for color digital imaging system | |
| CN107680142A (en) | Improve the method for the overlapping mapping of overseas color | |
| CN103870689B (en) | A kind of printing print system Forecast of Spectra method | |
| CN107424197A (en) | It is a kind of that the method across media color reproduction is realized based on spectrum domain mapping | |
| CN109272463A (en) | A kind of mural painting color recovery method | |
| CN103209331A (en) | System and method for strengthening image color saturation | |
| CN101594545A (en) | A kind of color domain expanding system and method | |
| US7659982B2 (en) | Quantitative evaluation of a color filter | |
| CN110324476B (en) | Method for representing color generation performance of mobile phone screen | |
| CN102355544B (en) | Color material selection assisting device and color material selection assisting method | |
| CN104182627A (en) | Method for accurately predicting and representing colors for display device | |
| CN108844632B (en) | Method for evaluating metamerism difference of observers among different display devices | |
| CN107766896B (en) | Spectrum dimensionality reduction method based on hue clustering | |
| US10242461B1 (en) | Method to improve overlay mapping of out-of-gamut |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170222 Termination date: 20201225 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |