CN101159878B - Color matching method and image seizure device and electronic equipment using the same - Google Patents

Abstract

A color matching method includes: capturing of a plurality of major color signals in digital image; a conversion step, obtaining 3 is multiplied by 3 color matrix from a plurality of major color signals and converting a plurality of major color signals into a coordinate matrix with 3 is multiplied by 1 luminance, hue and saturation (LAB); and a color difference calculation step, according to a color correction type consisting of the color matrix, presetting weight matrix and correction matrix, performing iterative computation on a plurality of colors input and color difference recursively, until the color different is less than threshold value.

Description

Color matching method and use its image capturing device and electronic equipment

[technical field]

The present invention system is about color matching method, reach image capturing device that uses it or the electronic equipment that is provided with the image capturing device, especially in regard to the color matching method of handling the electronics digital image, and use its image capturing device or be provided with the electronic equipment of image capturing device.

[background technology]

In recent years, for example digital still camera, electronic type video camera, telephone bandset etc. to have an electronic installation of chromatic image trap setting widely popular.Generally speaking, be provided with the semiconductor image sensing device in these electronic equipments and become electric signal with video conversion with picked-up.Because for the inherent limitations of color sensing performance, making, image sensor proofreaies and correct according to the common needs of color of the image of its output.In addition, the color of image also can be looked illumination condition and difference, and for example, under moonlight and outdoor daylight, visually the perception to color is significantly different.So obtained electron image is on color is handled, the perception that meet the human vision sense organ is very important.

In these electronic installations, can handle through a series of signals by the image signal of image sensor output, wherein, match colors processing meeting is proofreaied and correct so that image output has distortionless color representation the color of the image obtained.

Handle about colour correction, known have a lot of prior arts to use the color of correction matrix with correcting image.For example, May in 2006 bulletin on the 30th U.S. Patent number 7053935, disclosed the matching result between human visual system and the unique spectral response curve group and only used default correction matrix to handle in colour correction, but, because what use is the correction matrix of being scheduled to, so the result of its treatment for correcting also can't dynamically adjust, thereby be difficult to obtain colour correction result fast and accurately, and high-quality image can't be provided.Because prior art system uses fixing correction matrix to proofread and correct color, so its aberration is big and can't converge to predetermined value apace.

Therefore, need dynamically to adjust aberration and also converge to the color matching method of predetermined value apace to obtain the color demonstration of pin-point accuracy.

[summary of the invention]

Take into account above-mentioned, the correction matrix of the present invention dynamically to adjust, high accuracy ground and reach the aberration convergence apace to show high color image true to nature.

One of the present invention purpose system provides color matching method, can carry out colour correction to digital image, and it can be handled the image of obtaining is carried out colour correction according to a plurality of main colors in the image of obtaining.

The aspect one of according to the present invention provides color matching method, comprising: a plurality of main color signals in the acquisition digital image; Switch process is obtained 3 * 3 color matrices from these a plurality of main color signals, and converts these a plurality of main color signals to 3 * 1 illuminance, tone, degree of saturation (LAB) coordinate matrix respectively; And, the aberration calculation procedure, according to the colour correction formula that this color matrices, presetting weight and correction matrix are formed, the output and the aberration of ground these a plurality of colors of iterative computation of pulling over are till this aberration is less than critical value.

In addition, in the correction matrix value of each element be can according to present aberration and before aberration and dynamically change to obtain aberration apace less than critical value.And, can not limit the pulling over computing of number of times and obtain the chromatic image demonstration true to nature of the very little height of aberration the multidimensional color.

Another aspect according to the present invention provides image capturing device and electronic equipment, and it can carry out fast the image of obtaining, the colour correction of high precision, shows so that the little height of aberration chromatic image true to nature to be provided.

[brief description of drawingsfig]

Fig. 1 is a flow chart, and the digital image handling process is described;

Fig. 2 is a flow chart, and the color matching method according to embodiments of the invention is described;

Fig. 3 is a chromaticity diagram, is used for the color matching method of explanation according to embodiments of the invention;

Fig. 4 repeatedly calculates number of times and maximum excess when showing different PID coefficients; And

Fig. 5 is before demonstration is handled according to match colors of the present invention and chromaticity diagram and color coordinate figure afterwards.

[execution mode]

Will be with reference to the accompanying drawings, illustrate according to embodiments of the invention.

Fig. 1 is a flow chart, and general digital image handling process is described.

As shown in Figure 1, the processing of numerical digit clamp is accepted in the raw video signal output meeting that is obtained by the image sensor in the image capturing device in step 101.In step 101, as the black reference data in the color photo sensors, the value of for example red, green, blue colors such as (hereinafter to be referred as the RGB) channel of obtaining and optics black position standard are subtracted each other and the base value that makes each color channels does not alter a great deal to the physics performance of different color because of sensor or causes appreciable error with optics black.

Now proceeds to step 102 to carry out the channel migration processing.In step 102, the color channels of handling through clamp such as for example RGB are carried out the channel migration processing, so that having, these color channels values equate or extremely close position standard.The color channel (RGB) of these processing of process can have identical color base value and its numeric value analysis degree can become maximum.

Then, proceed to step 103.In step 103, carry out the color white balance and handle so that colouristic properties is identical with GTG.

Afterwards, proceed to step 104 carrying out image interpolarting, the three primary colors image of the image obtained is combined and form the image frame that will show.

Behind image interpolarting, the color of the image of obtaining may be only accurate, thereby need match colors to handle so that image need be adjusted to color accurately, thereby can carry out step 105 and handle to carry out match colors.

At last, the image signal of the above-mentioned processing of process can be accepted the Gamma correction of step 106, and the picture that benefit is grasped can show on harmony ground, then with image signal output (step 107).

General digital image is handled all and need be finished through above-mentioned treatment step, wherein, relevant match colors is handled, and prior art is all carried out match colors with predetermined fixed correction matrix and handled, thereby is difficult to obtain rapidly and accurately the high-quality accurately image of color.Will be described in detail in the following according to the match colors in the image processing of the present invention and handle.

Fig. 2 is a flow chart, and the color matching method according to embodiments of the invention is described.

As shown in Figure 2, at first obtain three bluish-green red signals of main spectral color in step 201, that is the green of the blueness of the 3rd row first row as shown in Figure 3, same column second row, and the redness of same column the third line.Fig. 3 system by the image capturing dress take the photograph chromaticity diagram to be used for image processing.

Now, proceed to step 202, (CommisionInternationale De L ' Eclairage) changes, and obtains 3 * 3 matrixes (1) that are used for colour correction from RGB (RGB) three color signals, simultaneously three color RGB is converted to corresponding 3x1 matrix (2) according to famous CIE.Matrix (1) and (2) are as described below:

$\left[\begin{array}{ccc}{a}_{11}& a_{12}& a_{13}\\ a_{21}& a_{22}& a_{23}\\ a_{31}& a_{32}& a_{33}\end{array}\right]$ Matrix (1)

$\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]$ Matrix (2)

Wherein, matrix (1) is according to CIE and the color image signal obtained, calculate and get, and in matrix (2), L _InIllumination, a of expression color signal _InExpression tone, b _InThe expression saturation is commonly referred to as the Lab coordinate.

In the present embodiment, as follows respectively according to the red bluish-green matrix of obtaining (1) and (2):

$\left[\begin{array}{ccc}{a}_{11}& a_{12}& a_{13}\\ a_{21}& a_{22}& a_{23}\\ a_{31}& a_{32}& a_{33}\end{array}\right]=\left[\begin{array}{ccc}0.94& -0.08& 0.14\\ 0.13& 0.81& 0.08\\ -0.27& 0.32& 0.60\end{array}\right]$

For blue $\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]=\left[\begin{array}{c}25.408\\ -17.46\\ -52.8\end{array}\right]$

For green $\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]=\left[\begin{array}{c}57.993\\ -43.63\\ 42.28\end{array}\right]$

For red $\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]=\left[\begin{array}{c}53.75\\ 95.18\\ 9.35\end{array}\right]$

As mentioned above,, RGB is changed the back substantially linear color matrix that is produced to be used for color correction described later according to the CIE standard.

Now proceeds to step 203, with the above-mentioned matrix application that obtains to following equation (1) to carry out the interative computation of colour correction.

$\left[\begin{array}{c}{L}_{\mathrm{out}}\\ a_{\mathrm{out}}\\ b_{\mathrm{out}}\end{array}\right]=(\left[\begin{array}{ccc}{a}_{11}& a_{12}& a_{13}\\ a_{21}& a_{22}& a_{23}\\ a_{31}& a_{32}& a_{33}\end{array}\right]+\left[\begin{array}{ccc}{w}_{11}& w_{12}& w_{13}\\ w_{21}& w_{22}& w_{23}\\ w_{31}& w_{32}& w_{33}\end{array}\right]\left[\begin{array}{ccc}{k}_{11}& k_{12}& k_{13}\\ k_{21}& k_{22}& k_{23}\\ k_{31}& k_{32}& k_{33}\end{array}\right])x\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]$

Equation (1)

Wherein, $\left[\begin{array}{ccc}{w}_{11}& w_{12}& w_{13}\\ w_{21}& w_{22}& w_{23}\\ w_{31}& w_{32}& w_{33}\end{array}\right]$ Be called weight matrix

$\left[\begin{array}{ccc}{k}_{11}& k_{12}& k_{13}\\ k_{21}& k_{22}& k_{23}\\ k_{31}& k_{32}& k_{33}\end{array}\right]$ Be called correction matrix

The value of each element is a preset value in the weight matrix, is in the present embodiment

$\left[\begin{array}{ccc}{w}_{11}& w_{12}& w_{13}\\ w_{21}& w_{22}& w_{23}\\ w_{31}& w_{32}& w_{33}\end{array}\right]=\left[\begin{array}{ccc}0.98& 0.01& 0.01\\ 0.01& 0.98& 0.01\\ 0.01& 0.01& 0.98\end{array}\right]$

According to the present invention, that each element definition in the correction matrix is as follows:

k _Ij=P _Ij+ D _Ij+ I _Ij, about P _Ij, D _Ij, and I _IjTo in aftermentioned, describe in detail

Above-mentioned Δ E is the aberration function, and according to CIEI 976, Δ E system is defined as

$\mathrm{\ΔE}=\sqrt{{(L_{\mathrm{out}}-L_{\mathrm{Ref}})}^2+(a_{\mathrm{out}}-a_{\mathrm{Ref}})+{(b_{\mathrm{out}}-b_{\mathrm{Ref}})}^2}$

Wherein, L _Out, a _Out, b _OutSystem calculate and the Lab coordinate of output color signal, and L _Ref, a _Ref, b _RefThe Lab coordinate fiducial value of the color signal that system obtains.In other words, aberration mean the color of output and take the photograph true color between difference, color Δ E can represent with the distance between the Lab coordinate.

In step 203,, each color is carried out interative computation and the operation result of each color is compared with fiducial value, and obtain value of chromatism according to equation 1.According to present embodiment, be that as mentioned above redgreenblue is carried out computing, still,, also can carry out computing to more different colours according to the present invention, that is, can carry out the color computing of multidimensional.

Now proceeds to step 204, obtains value of chromatism and default critical value in the comparison step 203, if value of chromatism is less than critical value, represent that then colour correction has reached required, then finish match colors and handle, if greater than critical value, then proceed to step 205,, calculate the P that makes new advances with according to following equation _Ij, D _Ij, and I _Ij,

Pij＝Ep＝K _pxΔE(t)

D _ij＝Ed＝K _dx(ΔE(t)-ΔE(t-1))

I _ij＝Ei＝K _i∫ _{_∞}ΔE(t)dt

Wherein, K _p, K _d, and K _iBe respectively proportional control factor, derivative control coefficient, reach integral control coefficient, Δ E (t) represents present value of chromatism, the previous stored value of chromatism of Δ E (t-1) representative.

Now is back to step 203, the P that will obtain _Ij, D _Ij, and I _IjIn the correction matrix of substitution equation (1), carry out interative computation.

So, carry out interative computation till aberration is less than critical value.

According to this example, K _p, K _d, and K _iValue, in i=j and i ≠ J situation, be different, that is, as follows according to the correction matrix of present embodiment:

$\left[\begin{array}{ccc}1.21\mathrm{Ep}+2.13\mathrm{Ed}+0.53\mathrm{Ei}& 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}& 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}\\ 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}& 1.21\mathrm{Ep}+2.13\mathrm{Ed}+0.53\mathrm{Ei}& 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}\\ 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}& 1.1\mathrm{Ep}+2\mathrm{Ed}+0.5\mathrm{Ei}& 1.21\mathrm{Ep}+2.13\mathrm{Ed}+0.53\mathrm{Ei}\end{array}\right]$

According to above-mentioned, with different K _p, K _d, and K _iCoefficient is carried out match colors respectively and is handled, and that obtains the results are shown among Fig. 4.Fig. 4 shows the match colors processing according to present embodiment, at different K _p, K _d, and K _iDuring value, reach the out of range value of repeatedly calculating number of times and maximum of required aberration.According to the result of present embodiment, error heals when big as can be known, the big K of healing _pValue means response faster, and it is bigger that this also means the feedback that will compensate.Bigger K _dCan reduce the excess of the compensation of being adjusted and can distinguish the aberration direction.K _iMean more greatly and in stable state, eliminate error, but cost is to cause bigger excess and slow down the convergence target value when compensation.

Via aforesaid calculation step,, can obtain the required correction matrix of match colors that is used for image processing with three kinds of main color RGB in the reference colour coloured picture as a reference.Handle take the photograph arbitrarily the match colors of image the time, the aberration that each element value in this correction matrix can be looked each interative computation changes and becomes, thereby can dynamically change, so that aberration can dwindle apace, and obtains the chromatic image of high fidelity.Herein, though with three colors of RGB present embodiment is described,, the present invention can use the color more than three to carry out match colors and handle, and the number of times of the interative computation of execution can be unrestricted.

Under tabulate 1, show among the embodiment of color matching method the effect that is produced during each control coefrficient modulation according to the present invention.

Table 1

After the aforesaid processing of process, the last colour correction that obtains the results are shown among Fig. 5.Image and CIE LAB coordinate figure thereof before Fig. 5 a system proofreaies and correct, image and CIE LAB coordinate figure thereof after Fig. 5 b system proofreaies and correct, from Fig. 5 b, obviously the aberration after overcorrect is restrained significantly as can be known.

According to the present invention, owing to can carry out the aberration matching treatment to the multidimensional color, and the iterations that reaches Optimal Error value (that is aberration minimum) is not limited, that is, according to the present invention, owing to can adopt multidimensional color iteration, the contents value in 3 * 3 matrixes can be carried out always and upgrade to iteration, till under the critical value that error (aberration) is being preset.

According to the present invention, according to present aberration information that measures and previous aberration information, and each element value of modulation correction matrix dynamically to be used for next iterative computation and further to calculate, can make aberration converge to optimum value apace.

Color matching method according to the present invention can be implemented with software or firmware mode.

Can be used for the electronic equipment that for example digital still camera, numerical digit video recorder, mobile phone etc. are provided with chromatic image trap setting or display according to color matching method of the present invention, with processing take the photograph chromatic image, or be used for computer system, handle chromatic image is carried out match colors.

Though with reference to being considered as preferred embodiment at present, the present invention is described,, should be appreciated that to the invention is not restricted to the embodiment that disclosed.On the contrary, invention system revises and isotropism in order to contain the spirit and the difference in the scope that are contained in the attached claim in back.The scope of following claim system comprises all this kind modifications and equivalent structure and function with the explanation of broad sense.

Claims (7)

1. a color matching method that is used for digital image comprises the following steps:

A plurality of main color signals are as input signal in the acquisition digital image;

Switch process, obtain 3 * 3 color matrices from these a plurality of main color signals, and convert these a plurality of input color signals to 3 * 1 illuminance, tone, degree of saturation (LAB) coordinate matrix respectively, the LAB coordinate matrix of this color matrices and input is expressed as follows respectively:

Color matrices: $\left[\begin{array}{ccc}{a}_{11}& a_{12}& a_{13}\\ a_{21}& a_{22}& a_{23}\\ a_{31}& a_{32}& a_{33}\end{array}\right]---\left(1\right)$

Input matrix: $\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]---\left(2\right)$

Each element a of color matrices ₁₁To a ₃₃Be real number, L _In, a _In, b _InRepresent the LAB coordinate after each color conversion respectively; And

The aberration calculation procedure, according to the colour correction formula that this color matrices, weight matrix and correction matrix are formed, the output and the aberration of ground these a plurality of colors of iterative computation of pulling over, less than critical value, this aberration system is defined as until this aberration

Wherein, L _Out, a _Out, and b _OutBe the LAB coordinate of this color output, L _Ref, a _Ref, and b _RefBe the LAB benchmark coordinate of this color, and this colour correction formula (1) is

$\left[\begin{array}{c}{L}_{\mathrm{out}}\\ a_{\mathrm{out}}\\ b_{\mathrm{out}}\end{array}\right]=(\left[\begin{array}{ccc}{a}_{11}& a_{12}& a_{13}\\ a_{21}& a_{22}& a_{23}\\ a_{31}& a_{32}& a_{33}\end{array}\right]+\left[\begin{array}{ccc}{w}_{11}& w_{12}& w_{13}\\ w_{21}& w_{22}& w_{23}\\ w_{31}& w_{32}& w_{33}\end{array}\right]\left[\begin{array}{ccc}{k}_{11}& k_{12}& k_{13}\\ k_{21}& k_{22}& k_{23}\\ k_{31}& k_{32}& k_{33}\end{array}\right])x\left[\begin{array}{c}{L}_{\mathrm{in}}\\ a_{\mathrm{in}}\\ b_{\mathrm{in}}\end{array}\right]---\left(1\right)$

Wherein, $\left[\begin{array}{c}{L}_{\mathrm{out}}\\ a_{\mathrm{out}}\\ b_{\mathrm{out}}\end{array}\right]$ Be the LAB coordinate matrix of this color output, $\left[\begin{array}{ccc}{w}_{11}& w_{12}& w_{13}\\ w_{21}& w_{22}& w_{23}\\ w_{31}& w_{32}& w_{33}\end{array}\right]$ Be this weight matrix,

$\left[\begin{array}{ccc}{k}_{11}& k_{12}& k_{13}\\ k_{21}& k_{22}& k_{23}\\ k_{31}& k_{32}& k_{33}\end{array}\right]$ Be this correction matrix, each element definition is as follows in this correction matrix:

k _ij＝P _ij+D _ij+I _ij

Wherein, P _Ij=Ep=K _pX Δ E (t)

D _ij＝Ed＝K _dx(ΔE(t)-ΔE(t-1))

$I_{\mathrm{ij}}=\mathrm{Ei}=K_i{\∫}_{-\∞}\mathrm{\ΔE}\left(t\right)\mathrm{dt}$

Wherein, K _p, K _d, and K _iBe respectively proportional control factor, derivative control coefficient, reach integral control coefficient.

2. as the method for the 1st of claim, wherein, this switch process system carries out according to international lighting association (CIE) standard.

3. as the method for the 1st of claim, wherein, the pass of each element is that each element of each element of i=j and i ≠ j is unequal in this correction matrix, and still, each element of i=j is equal to each other, and each element of i ≠ j also is equal to each other.

4. chromatic image trap setting, use as the 1st to 3 of claim in each method so that the chromatic image that captures is carried out the match colors processing.

5. a digital still camera has chromatic image catcher and display, wherein, this chromatic image catcher use as the 1st to 3 of claim in each method, to take the photograph chromatic image carry out the match colors processing.

6. a mobile phone has chromatic image catcher and display, wherein, this chromatic image catcher use as the 1st to 3 of claim in each method, to take the photograph chromatic image carry out the match colors processing.

7. computer system, use as the 1st to 3 of claim in each method, chromatic image is carried out the match colors processing.

Images