CN107197225A - Color digital camera white balance correcting based on chromatic adaptation model - Google Patents

Abstract

The invention discloses a kind of color digital camera white balance correcting based on chromatic adaptation model, the present invention calculates color correction matrix using radical sign polynomial regression (Root Polynomial Regression) method, so that the equipment relevant response value RGB under several common light sources (hereinafter referred to as Calibrating source) is changed to same light source into device-independent CIE1931 tristimulus values XYZ.The RGB responses of unknown light source in actual scene (hereinafter referred to as testing light source) are changed into XYZ color space using the color correction matrix demarcated in advance, and its corresponding color under reference light source is calculated using CAT02 chromatic adaptation transformation models, the correspondence color is that observer produces the color that chromatic adaptation after-vision system is perceived to testing light source.

Description

White Balance Calibration Method of Color Digital Camera Based on Chromatic Adaptation Model

technical field

The invention relates to a method for adjusting the white balance correction result of a color digital camera by using a color adaptation model. The method can enable the color digital camera to realize color restoration more in line with human perception for shooting scenes.

Background technique

The same object often has different colorimetric parameters under different light sources. Due to the color constancy of the human visual system, these differences in chromaticity can be automatically compensated by the human eye and brain to a certain extent, so as to restore the "true" color of the object under different light sources. The white balance module in the image signal processing flow (ISP Pipeline) of a color digital camera corrects the color cast of objects under non-standard light sources by calculating the difference between the chromaticity of the actual light source and the chromaticity of the standard light source, thereby simulating the human Color constancy of the visual system.

The current color digital camera obtains the chromaticity of the light source mainly from two ways: 1) several types of light source modes are pre-set in the storage space of the camera, and the user specifies the light source type to which the scene belongs during actual shooting. This type of light source chromaticity acquisition method is called "manual white balance mode"; 2) Analyze the captured images, and predict the color of the light source through some light source estimation algorithms or with the help of external sensors. This type of light source chromaticity acquisition method is called "automatic white balance mode". Regardless of the working mode, the white balance correction module usually uses two or three gain coefficients to linearly adjust the red (Red), blue (Blue) or red, green (Green) and blue channels of the color cast image , so that the imaginary perfect reflective surface in the scene (perfect reflective surface with a constant spectral reflectance of 1 at any wavelength) has the same (or consistent with the reference white point) three-channel response value after white balance correction.

If any lighting source, especially some light sources whose chromaticity deviates significantly from the reference white point, uses a unified reference light source as the target of white balance correction, the following disadvantages will occur to the image after white balance correction: 1) The output image is too " 2) The rendering effect of the light source on the atmosphere of the scene is completely suppressed; 3) Forcibly correcting the light source with serious color cast to the reference light source will increase the subsequent processing of the image signal. The processing difficulty of modules (such as lens shading correction, color correction, etc.) may even cause image quality to deteriorate.

Contents of the invention

In order to make the white balance correction module of the digital camera realize more realistic scene color reproduction, the present invention utilizes the gain coefficient obtained in the original white balance correction module Calculates the color of the light source in the actual shooting scene. For the sake of uniform expression, in the present invention, the object color of the perfect reflective surface is used to represent the color of the light source, because the perfect reflective surface can reflect all the energy of the light source without wavelength selectivity. Use the CAT02 color adaptation transformation in the CIECAM02 color appearance model to calculate the corresponding color of the object color under the reference light source, so as to obtain the white balance correction gain coefficient after color adaptation In order to realize the white balance correction of the image that is more in line with the visual perception of the human eye.

The present invention uses the root-polynomial regression (Root-Polynomial Regression) method to calculate the color correction matrix, thereby converting the device-dependent response value RGB under several common light sources (hereinafter referred to as calibration light sources) to the device-independent CIE1931 tri-stimulus value under the same light source Value XYZ. Use the pre-calibrated color correction matrix to convert the RGB response value of the unknown light source to be corrected in the actual scene (hereinafter referred to as the test light source) to the XYZ color space, and use the CAT02 color adaptation transformation model to calculate its corresponding color under the reference light source , the corresponding color is the color perceived by the visual system of the observer after chromatic adaptation to the test light source.

The concrete technical scheme that the present invention adopts is as follows:

A color digital camera white balance correction method based on a chromatic adaptation model, the steps are as follows:

S1: Convert the device-related response value RGB under different calibration light sources to the device-independent tristimulus value CIE1931 XYZ under the same light source by using the root polynomial regression color correction method;

S2: Obtain the white balance correction gain coefficient of the image taken under the light source to be corrected, and calculate the Coordinates on the plane, at the camera Search for the calibration light source closest to the coordinate on the plane, call the color correction matrix corresponding to the calibration light source, convert the device-related camera response value of the light source into the CIE1931 XYZ space, and regard the light source color as the object color;

S3: Use the color adaptation transformation CAT02 in the CIECAM02 color appearance model to calculate the corresponding color of the object color under the standard light source after color adaptation:

S4: Remap the corresponding color back to the RGB space of the camera using the inverse matrix of the color correction matrix, and recalculate the white balance correction gain coefficient after color adaptation.

Based on the above technical solution, each step can adopt the following specific implementation methods:

As a preference, said S1 is specifically:

S101: For the calibration light source L whose spectral power distribution is P(λ), use the camera response value composition model to calculate the camera RGB values r _i , g _i and b _i of the i-th color block of the standard color card under the illumination of the light source:

In the formula, ρi (λ) represents the spectral reflectance of the _i -th color block, S ^k (λ) represents the spectral sensitivity function of the k-th channel of the camera, k=R, G, B, and Ω′ is the wavelength of the spectral response of the camera Range; for a standard color card with N color blocks, an N×3 camera response value matrix C(L) is calculated, where each row corresponds to the camera RGB value of a color block;

S102: Calculate the camera RGB values r ^ill , g ^ill and b ^ill of the perfect reflective surface under the calibration light source:

and record it in coordinates on the plane

S103: For the calibration light source with spectral power distribution P(λ), use CIE1931 2° standard observer color matching function Calculate the CIE1931 XYZ tristimulus value of the i-th color block of the standard color card under the illumination of this light source:

In the formula, Ω is the wavelength range of visible light; for a standard color card with N color blocks, an N×3 tristimulus value matrix T(L) is calculated, in which each row corresponds to the XYZ tristimulus value of a color block;

S104: Expand the dimension of the camera response value matrix C(L) from N×3 to N×q, q>3, wherein the 4th to q columns correspond to the root polynomials of the response values of each color block;

S105: Calculate the 6×3 color correction matrix M'(L) converted from C'(L) to T(L) by using the least square method or other correction matrix optimization methods with color difference as the objective function:

When the root mean square error between C′(L)·M′(L) and T(L) is taken as the optimization target, the pseudo-inverse method is used to calculate M′(L):

M'(L)＝[C'T(L)C'(L)] ^{-1 C'T} (L)· ^T (L),

When the color difference between C'(L)·M'(L) and T(L) is used as the optimization target, the nonlinear optimization method is used to calculate M'(L):

M'(L)=arg min △E(C'(L)·M'(L),T(L)),

Where △E(A,B) is a function used to calculate the color difference between A and B;

S106: Use the method in S105 to calculate the 3×3 color correction matrix M(L) under each calibration light source; S107: For all calibration light sources, use the method of S101 to S106 to calculate the respective color correction matrix M'(L) and M(L) and stored in the camera's built-in memory.

As preferably, said S2 is specifically:

S201: Manually set or use the existing automatic white balance algorithm to obtain the white balance correction gain coefficient of the image shot under the light source to be corrected

S202: Calculate the RGB value of the light source to be corrected in the raw domain of the camera:

in camera Plane search with The closest calibration light source L, and call its corresponding color correction matrix M'(L) from the camera's built-in memory;

S203: Transform the camera response value of the perfect reflective surface under the scene light source into the CIE1931 XYZ space by using the color correction matrix M′(L):

In the formula, X ^ill , Y ^ill , Z ^ill are the tristimulus values in XYZ space, respectively.

Preferably, said S3 is specifically:

Use the color adaptation transformation CAT02 in the CIECAM02 color appearance model to calculate the corresponding color of the object color [X ^ill , Y ^ill , Z ^ill ] under the standard light source after color adaptation:

In the formula, the four inputs of the color adaptation transformation model f _CAT02 are the tristimulus value of the object color to be calculated, the tristimulus value of the light source to be adapted, the tristimulus value of the reference light source and the environmental brightness factor L _A .

Further, the ambient brightness factor is calculated using two sigmoid functions:

In the formula, the light source chromaticity distance d is obtained by calculating the Euclidean distance between the actual light source and the reference light source chromaticity in the CIELUV uniform color space, and a ₁ , b ₁ , K ₁ , a ₂ , b ₂ , and K ₂ are used to adjust the shape of the sigmoid function The undetermined parameters of .

As preferably, said S4 is specifically:

Invert the 3×3 color correction matrix M(L) corresponding to the calibration light source L, and use the perfect reflector tristimulus value after color adaptation Remap back into camera RGB space:

From this, calculate the white balance correction gain coefficient after color adaptation

The invention converts the chromatic-adapted CIE1931 XYZ tristimulus value back into the camera RGB color space by inverting the calibrated color correction matrix, thereby determining the chromatic-adapted white balance correction gain coefficient.

In order to have a more intuitive understanding of the implementation process of the present invention, an embodiment is given below and described in detail with the accompanying drawings.

Description of drawings

Fig. 1 is the calibration light source used in the embodiment of the present invention by and As a plane composed of horizontal and vertical coordinates (hereinafter referred to as Coordinate distribution on the plane).

Fig. 2 is a flow chart of calibrating the color correction matrices of several calibration light sources in the present invention.

3 is a schematic diagram of the corresponding relationship between the _CAT02 color adaptation model input parameter LA (environmental brightness factor), d (Euclidean distance in CIELUV uniform color space), and E (shooting scene illuminance) used in the embodiment of the present invention.

Fig. 4 is a flow chart of the calculation of the white balance correction gain coefficient after chromatic adaptation is performed on a test light source in the present invention.

detailed description

The present invention will be further elaborated and illustrated below in conjunction with the accompanying drawings and specific embodiments.

At present, the white balance correction module of most color digital cameras corrects the neutral point under any light source to the driving value under the reference light source. Makes the corrected color appear noticeably distorted. The present invention proposes a method for adjusting the color adaptation of the original gain coefficient of the digital camera white balance correction module by using the CAT02 color adaptation transformation in the CIECAM02 color appearance model, so that the image after white balance correction is more in line with the color perception of human eyes .

The present invention uses a root-polynomial regression color correction (Root-Polynomial Regression ColorCorrection) method to convert device-related response values RGB under several common light sources to device-independent tristimulus values CIE1931 XYZ under the same light source. Since the color correction matrix used in the square root polynomial regression color correction depends on the spectral power distribution function of the light source, the present invention needs to pre-calibrate the color correction matrix for several typical light sources.

Figure 1 shows a feasible calibration light source selection method, and draws 39 calibration light sources in the camera Coordinate distribution on the plane.

Fig. 2 is a flow chart of calibrating the color correction matrices of several calibration light sources in the present invention. Among them, the number and types of calibration light sources can be flexibly selected, and in some application scenarios with a large storage cost limitation, only the D65 light source can be selected as the only calibration light source to calculate the color correction matrix.

1. The calibration process of the present invention comprises the following steps:

In order to convert the camera response value RGB to the device-independent tristimulus value XYZ, the present invention adopts a root polynomial regression color correction method.

For the calibration light source L whose spectral power distribution is P(λ), use the camera response value construction model to calculate the camera RGB value of the i-th color block of the standard color card under the illumination of the light source:

In the formula, ρ _i (λ) represents the spectral reflectance of the i-th color block, S ^k (λ) represents the spectral sensitivity function of the k-th channel of the camera (k=R, G, B), which can be obtained from the standard of the camera when it leaves the factory It is said to be obtained from the data or calculated by using the relevant spectral sensitivity estimation algorithm, and Ω' is the wavelength range of the camera's spectral response. For a standard color card with N color patches, an N×3 camera response value matrix C(L) can be calculated, where each row corresponds to the camera RGB value of a color patch.

At the same time, calculate the camera RGB value of the perfect reflective surface under the calibration light source:

and record it in coordinates on the plane

For the calibration light source with spectral power distribution P(λ), use CIE1931 2° standard observer color matching function Calculate the CIE1931 XYZ tristimulus value of the i-th color block of the standard color card under the illumination of this light source:

where Ω is the wavelength range of visible light. For a standard color card with N color blocks, an N×3 tristimulus value matrix T(L) can be calculated, where each row corresponds to the XYZ tristimulus value of a color block.

Expand the dimension of the camera response value matrix C(L) from N×3 to N×q(q>3), where the 4th to q columns correspond to the root sign polynomials of the response values of each color block. Take the quadratic root sign polynomial as an example, at this time, q=6, the i-th behavior of the extended camera response value matrix C′(L)

Using the least squares method or other correction matrix optimization methods with color difference as the objective function, calculate the 6×3 color correction matrix M'(L) converted from C'(L) to T(L):

When the root mean square error between C′(L)·M′(L) and T(L) is used as the optimization target, the pseudo-inverse method can be used to calculate M′(L):

M'(L)＝[C'T(L)C'(L)] ^{-1 C'T} (L)· ^T (L).

When the color difference between C′(L)·M′(L) and T(L) is used as the optimization target, M′(L) can be calculated by using nonlinear optimization methods such as the Gauss-Newton method:

M'(L)=arg min △E(C'(L)·M'(L),T(L)).

Where △E(A,B) is a function used to calculate the color difference between A and B;

At the same time, using a similar method, the 3×3 color correction matrix M(L) under each calibration light source is calculated. The difference between M(L) and M'(L) is that M'(L) is applicable to the response value matrix C'(L) after the root sign polynomial expansion, and M(L) is applicable to the original response value matrix C( L).

For all calibration light sources, the respective color correction matrices M'(L) and M(L) are calculated by the above method and stored in the built-in memory of the camera.

2. The present invention carries out the process of correcting the white balance based on the color adaptation model to the image taken under any unknown light source as follows:

Manually set or use the existing automatic white balance algorithm to obtain the white balance correction gain coefficient of the image

Calculate the RGB value of the light source in the scene on the camera's raw domain:

in camera Plane search with The nearest calibration light source L, and its corresponding color correction matrix M'(L) is called from the built-in memory of the camera.

Use the color correction matrix M′(L) to transform the camera response value of the perfect reflective surface under the scene light source into the CIE1931 XYZ space:

Use the color adaptation transformation CAT02 in the CIECAM02 color appearance model to calculate the corresponding color of the object color [X ^ill , Y ^ill , Z ^ill ] under the standard light source after color adaptation:

In the formula, the four inputs of the chromatic adaptation transformation model f _CAT02 are the tristimulus value of the object color to be calculated, the tristimulus value of the light source to be adapted, the tristimulus value of the reference light source, and the L _A environmental brightness factor. Since the present invention needs to calculate the perceived color of the test light source after chromatic adaptation, which is equivalent to calculating the corresponding color of the perfect reflective surface under the test light source, the first two inputs of the model are the CIE1931 XYZ tristimulus values of the test light source. In this embodiment, the CIE D65 illuminant is selected as the standard illuminant, so

The environmental brightness factor L _A can comprehensively consider the two factors of the chromaticity distance d between the actual light source and the reference light source and the scene illuminance E. In the present invention, two sigmoid functions are used to calculate LA _:

In the formula, the light source chromaticity distance d can be obtained by calculating the Euclidean distance between the actual light source and the reference light source chromaticity in the CIELUV uniform color space, and a ₁ , b ₁ , K ₁ , a ₂ , b ₂ , and K ₂ are used as the adjusted sigmoid function The undetermined parameters of the shape can be calibrated according to actual needs. A possible corresponding relationship between the ambient brightness factor and d, E is shown in Figure 3.

Finally, the 3×3 color correction matrix M(L) corresponding to the test light source is inverted, and the perfect reflector tristimulus value after color adaptation Remap back into camera RGB space:

Thus, the white balance correction gain coefficient after color adaptation is calculated:

By using the gain coefficient pair, the white balance correction of the image that is more in line with the visual perception of the human eye can be realized.

The flow chart of calculating the white balance correction gain coefficient after chromatic adaptation to a test light source is shown in FIG. 4 .

The above-mentioned embodiment is only a preferred solution of the present invention, but it is not intended to limit the present invention. Various changes and modifications can be made by those skilled in the relevant technical fields without departing from the spirit and scope of the present invention. Therefore, all technical solutions obtained by means of equivalent replacement or equivalent transformation fall within the protection scope of the present invention.

Claims (6)

1. A color digital camera white balance correction method based on color adaptation model, it is characterized in that, the steps are as follows: S1: Convert the device-related response value RGB under different calibration light sources to the device-independent tristimulus value CIE1931XYZ under the same light source by using the root polynomial regression color correction method; S2: Obtain the white balance correction gain coefficient of the image taken under the light source to be corrected, and calculate the Coordinates on the plane, at the camera Search the calibration light source closest to the coordinate on the plane, call the color correction matrix corresponding to the calibration light source, convert the device-related camera response value of the light source into the CIE1931XYZ space, and regard the light source color as the object color; S3: Use the color adaptation transformation CAT02 in the CIECAM02 color appearance model to calculate the corresponding color of the object color under the standard light source after color adaptation: S4: Remap the corresponding color back to the RGB space of the camera using the inverse matrix of the color correction matrix, and recalculate the white balance correction gain coefficient after color adaptation. 2. the color digital camera white balance correction method based on color adaptation model as claimed in claim 1, is characterized in that, described S1 is specifically: S101: For the calibration light source L whose spectral power distribution is P(λ), use the camera response value composition model to calculate the camera RGB values r _i , g _i and b _i of the i-th color block of the standard color card under the illumination of the light source: <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>r</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>R</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>g</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>G</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>b</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>B</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> </mtable> </mfenced> In the formula, ρi (λ) represents the spectral reflectance of the _i -th color block, S ^k (λ) represents the spectral sensitivity function of the k-th channel of the camera, k=R, G, B, and Ω′ is the wavelength of the spectral response of the camera Range; for a standard color card with N color blocks, an N×3 camera response value matrix C(L) is calculated, where each row corresponds to the camera RGB value of a color block; S102: Calculate the camera RGB values r ^ill , g ^ill and b ^ill of the perfect reflective surface under the calibration light source: <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msup> <mi>r</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>R</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msup> <mi>g</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>G</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msup> <mi>b</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <msup> <mi>&amp;Omega;</mi> <mo>&amp;prime;</mo> </msup> </munder> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <msup> <mi>S</mi> <mi>B</mi> </msup> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> </mtable> </mfenced> and record it in coordinates on the plane S103: For the calibration light source with spectral power distribution P(λ), use CIE1931 2° standard observer color matching function Calculate the CIE1931XYZ tristimulus value of the i-th color block of the standard color card under the illumination of this light source: <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>X</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <mi>&amp;Omega;</mi> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mover> <mi>x</mi> <mo>&amp;OverBar;</mo> </mover> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>Y</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <mi>&amp;Omega;</mi> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mover> <mi>y</mi> <mo>&amp;OverBar;</mo> </mover> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>Z</mi> <mi>i</mi> </msub> <mo>=</mo> <munder> <mo>&amp;Integral;</mo> <mi>&amp;Omega;</mi> </munder> <msub> <mi>&amp;rho;</mi> <mi>i</mi> </msub> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>P</mi> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mover> <mi>z</mi> <mo>&amp;OverBar;</mo> </mover> <mrow> <mo>(</mo> <mi>&amp;lambda;</mi> <mo>)</mo> </mrow> <mi>d</mi> <mi>&amp;lambda;</mi> <mo>,</mo> </mrow> </mtd> </mtr> </mtable> </mfenced> In the formula, Ω is the wavelength range of visible light; for a standard color card with N color blocks, an N×3 tristimulus value matrix T(L) is calculated, in which each row corresponds to the XYZ tristimulus value of a color block; S104: Expand the dimension of the camera response value matrix C(L) from N×3 to N×q, q>3, wherein the 4th to q columns correspond to the root polynomials of the response values of each color block; S105: Calculate the 6×3 color correction matrix M'(L) converted from C'(L) to T(L) by using the least square method or other correction matrix optimization methods with color difference as the objective function: <mrow> <msup> <mi>C</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>L</mi> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <msup> <mi>M</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>L</mi> <mo>)</mo> </mrow> <mo>&amp;cong;</mo> <mi>T</mi> <mrow> <mo>(</mo> <mi>L</mi> <mo>)</mo> </mrow> <mo>,</mo> </mrow> When the root mean square error between C′(L)·M′(L) and T(L) is taken as the optimization target, the pseudo-inverse method is used to calculate M′(L): M'(L)＝[C'T(L)C'(L)] ^{-1 C'T} (L)· ^T (L), When the color difference between C'(L)·M'(L) and T(L) is used as the optimization target, the nonlinear optimization method is used to calculate M'(L): M'(L)=argmin△E(C'(L)·M'(L),T(L)), Where △E(A,B) is a function used to calculate the color difference between A and B; S106: Use the method in S105 to calculate the 3×3 color correction matrix M(L) under each calibration light source; S107: For all the calibration light sources, calculate the respective color correction matrices M'(L) and M(L) by using the method of S101-S106, and store them in the built-in memory of the camera. 3. the color digital camera white balance correction method based on color adaptation model as claimed in claim 1, is characterized in that, described S2 is specifically: S201: Manually set or use the existing automatic white balance algorithm to obtain the white balance correction gain coefficient of the image shot under the light source to be corrected S202: Calculate the RGB value of the light source to be corrected in the raw domain of the camera: <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <msup> <mi>R</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <mfrac> <mn>1</mn> <msubsup> <mi>R</mi> <mrow> <mi>g</mi> <mi>a</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>A</mi> <mi>W</mi> <mi>B</mi> </mrow> </msubsup> </mfrac> <mo>,</mo> </mtd> </mtr> <mtr> <mtd> <msup> <mi>G</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <mn>1</mn> <mo>,</mo> </mtd> </mtr> <mtr> <mtd> <msup> <mi>B</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>=</mo> <mfrac> <mn>1</mn> <msubsup> <mi>B</mi> <mrow> <mi>g</mi> <mi>a</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>A</mi> <mi>W</mi> <mi>B</mi> </mrow> </msubsup> </mfrac> <mo>&amp;CenterDot;</mo> </mtd> </mtr> </mtable> </mfenced> in camera Plane search with The closest calibration light source L, and call its corresponding color correction matrix M'(L) from the camera's built-in memory; S203: Transform the camera response value of the perfect reflective surface under the scene light source into CIE1931XYZ space by using the color correction matrix M′(L): <mrow> <msup> <mrow> <mo>&amp;lsqb;</mo> <msup> <mi>X</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Y</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Z</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>=</mo> <msup> <mi>M</mi> <mo>&amp;prime;</mo> </msup> <mrow> <mo>(</mo> <mi>L</mi> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <msup> <mi>R</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>G</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>B</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msqrt> <mrow> <msup> <mi>R</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <msup> <mi>G</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> </mrow> </msqrt> <mo>,</mo> <msqrt> <mrow> <msup> <mi>G</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <msup> <mi>B</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> </mrow> </msqrt> <mo>,</mo> <msqrt> <mrow> <msup> <mi>R</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <msup> <mi>B</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> </mrow> </msqrt> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>.</mo> </mrow> In the formula, X ^ill , Y ^ill , Z ^ill are the tristimulus values in XYZ space, respectively. 4. the color digital camera white balance correction method based on color adaptation model as claimed in claim 1, is characterized in that, described S3 is specifically: Use the color adaptation transformation CAT02 in the CIECAM02 color appearance model to calculate the corresponding color of the object color [X ^ill , Y ^ill , Z ^ill ] under the standard light source after color adaptation: <mrow> <msup> <mrow> <mo>&amp;lsqb;</mo> <msubsup> <mi>X</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>Y</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>Z</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>=</mo> <msub> <mi>f</mi> <mrow> <mi>C</mi> <mi>A</mi> <mi>T</mi> <mn>02</mn> </mrow> </msub> <mrow> <mo>(</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <msup> <mi>X</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Y</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Z</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>,</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <msup> <mi>X</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Y</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Z</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>,</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <msup> <mi>X</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Y</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msup> <mo>,</mo> <msup> <mi>Z</mi> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> </mrow> </msup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>,</mo> <msub> <mi>L</mi> <mi>A</mi> </msub> <mo>)</mo> </mrow> <mo>,</mo> </mrow> In the formula, the four inputs of the color adaptation transformation model f _CAT02 are the tristimulus value of the object color to be calculated, the tristimulus value of the light source to be adapted, the tristimulus value of the reference light source and the environmental brightness factor L _A . 5. the color digital camera white balance correction method based on color adaptation model as claimed in claim 4, is characterized in that, described environment brightness factor uses two sigmoid functions to calculate: <mrow> <msub> <mi>L</mi> <mi>A</mi> </msub> <mo>=</mo> <mfrac> <msub> <mi>K</mi> <mn>1</mn> </msub> <mrow> <mn>1</mn> <mo>+</mo> <msup> <mi>e</mi> <mrow> <mo>(</mo> <msub> <mi>a</mi> <mn>1</mn> </msub> <mi>d</mi> <mo>-</mo> <msub> <mi>b</mi> <mn>1</mn> </msub> <mo>)</mo> </mrow> </msup> </mrow> </mfrac> <mo>&amp;CenterDot;</mo> <mfrac> <msub> <mi>K</mi> <mn>2</mn> </msub> <mrow> <mn>1</mn> <mo>+</mo> <msup> <mi>e</mi> <mrow> <mo>(</mo> <msub> <mi>b</mi> <mn>2</mn> </msub> <mo>-</mo> <msub> <mi>a</mi> <mn>2</mn> </msub> <mi>E</mi> <mo>)</mo> </mrow> </msup> </mrow> </mfrac> <mo>,</mo> </mrow> In the formula, the light source chromaticity distance d is obtained by calculating the Euclidean distance between the actual light source and the reference light source chromaticity in the CIELUV uniform color space, and a ₁ , b ₁ , K ₁ , a ₂ , b ₂ , and K ₂ are used to adjust the shape of the sigmoid function The undetermined parameters of . 6. the color digital camera white balance correction method based on color adaptation model as claimed in claim 3, is characterized in that, described S4 is specifically: Invert the 3×3 color correction matrix M(L) corresponding to the calibration light source L, and use the perfect reflector tristimulus value after color adaptation Remap back into camera RGB space: <mrow> <msup> <mrow> <mo>&amp;lsqb;</mo> <msubsup> <mi>R</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>G</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>B</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>=</mo> <msup> <mi>M</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mrow> <mo>(</mo> <mi>L</mi> <mo>)</mo> </mrow> <mo>&amp;CenterDot;</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <msubsup> <mi>X</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>Y</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>,</mo> <msubsup> <mi>Z</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <mo>&amp;rsqb;</mo> </mrow> <mi>T</mi> </msup> <mo>,</mo> </mrow> From this, calculate the white balance correction gain coefficient after color adaptation <mrow> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msubsup> <mi>R</mi> <mrow> <mi>g</mi> <mi>a</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>C</mi> <mi>A</mi> <mi>T</mi> </mrow> </msubsup> <mo>=</mo> <mfrac> <msubsup> <mi>G</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <msubsup> <mi>R</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> </mfrac> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msubsup> <mi>B</mi> <mrow> <mi>g</mi> <mi>a</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mi>C</mi> <mi>A</mi> <mi>T</mi> </mrow> </msubsup> <mo>=</mo> <mfrac> <msubsup> <mi>G</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> <msubsup> <mi>B</mi> <mi>a</mi> <mrow> <mi>i</mi> <mi>l</mi> <mi>l</mi> </mrow> </msubsup> </mfrac> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>.</mo> </mrow> 3