JP2874657B2 - Image color cast correction method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a technique for correcting an input image recorded in a color cast state in a digital image processing apparatus.

[0002]

2. Description of the Related Art A conventional image processing apparatus of this type uses RGB.
Is changed interactively, and the operator corrects while checking the processing result.

In addition, Japanese Patent Application Laid-Open No. 2-94893 discloses a device for performing correction in accordance with the spectral characteristics of different types of light sources.

A conventional image processing apparatus will be described with reference to FIG. FIG. 7 is a block diagram created based on the description of Japanese Patent Application Laid-Open No. 2-94893.

[0005] The spectral characteristic storage means 102 stores correction amounts of red, green, and blue color components corresponding to spectral characteristics of different types of light sources. The correction amount search unit 103 searches the correction amount of the specified light source from the spectral characteristic storage unit 102 and inputs the correction amount to the color correction unit 101. The data in the input image buffer 100 is color-corrected by the color correction unit 101 and output to the output image buffer 104. Assuming that u input pixel values are R, G, and B and output pixel values are R ', G', and B ', the color correction method is represented by Expression (1).

R ′ = R / r0, G ′ = G / g0, B ′ = B / b0 (1) That is, as a correction method, an RGB value is multiplied by a constant term.

[0007]

Many conventional image processing apparatuses interactively correct the RGB tone curve, and have a large number of adjustment parameters, which cannot be intuitively corrected. was there. In addition, since the operation requires skill, there is a problem that the correction cannot be performed as expected by an amateur.

Further, there is a problem that it is impossible to completely correct various light source colors only by correcting the RGB value by a constant term, and it is not possible to correct the color to a good color. Further, there is a problem that color fogging caused not only by the difference in the color of the light source but also by the characteristics of the film, the characteristics of the lens, and the exposure characteristics cannot be corrected only by the characteristics of various light sources.

[0009]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the method for correcting color cast of an image according to the present invention employs a method of highlighting an RGB tone curve in response to a defect in an input image.
Selecting a correction pattern consisting of control information of color , halftone, and shadow; specifying an amount to be corrected; obtaining a control parameter from the correction pattern; and RGB from the correction amount and the control parameter.
And the step of performing table conversion for all pixels using the tone curve data.

Further, the image color cast correction apparatus of the present invention comprises an input image buffer for holding input image data ;
Highlights, midtones and shadows on the GB tone curve
Correction of correction patterns consisting of control information
A correction pattern storage unit that stores a plurality of correction patterns in accordance with the correction amount; a correction amount storage unit that stores the correction amount of the color cast; a tone curve generation unit that generates RGB tone curve data from the correction amount in the correction pattern storage unit; It has an LUT storage buffer for storing curve data, an LUT conversion unit for performing table conversion on image data, and an output image buffer for holding output data.

The principle of the image color cast correction apparatus of the present invention will be described.

The processing contents of the present invention will be described with reference to the flowchart of FIG.

First, an operator checks an input image on a monitor or the like, and selects one pattern to be processed from correction patterns prepared in advance (step A1).

[0014] Regarding the color defect of the image, there are several patterns such as green cast of fluorescent lamp, yellow cast of electric lamp, red cast of snow scene, and blue cast of north window light. This is not only due to the type of light source, but also photographic film and CC
It also relates to the characteristics of D.

The present invention holds the control parameters of the tone curve for each of such defective patterns. The correction pattern means a set of the control parameters.

Next, the operator specifies a correction amount as a degree of correction (step A2).

Then, a control parameter for the correction pattern is obtained (step A3), and an actual RGB tone curve is calculated from the control parameter and the correction amount (step A4).
Output to the UT storage buffer (step A5). next,
A color conversion process is performed on the input image data using the tone curve stored in the LUT, and the processing is terminated (step A).
6).

FIG. 5 shows an example of the correction pattern.
The data example 40 of the correction pattern storage unit includes an item of a defect pattern such as a fluorescent lamp green cast, and a tone curve control parameter 41 for correcting the defect. As examples of the tone curve control parameters 41, there are highlight (hl), halftone (m), and shadow (sd) data as shown in FIG. Each of these has three RGB control parameters 42. For example, hl is (R
h, Gh, Bh).

Using highlights, halftones and shadows,
An example of tone curve control will be described with reference to FIG. The correction amount grad previously specified and hl, m, s
The tone curve is controlled by the parameter d.

In the example shown in the figure, the M control point 33 is controlled by halftone (m) and grad data, and is used to generate a halftone correction curve 30. HL control point 35
Is controlled by highlight (hl) and grad data, and the SD control point 34 determines the shadow (sd) and grad
It is controlled by d data. HL control point 35 and SD
The control point 34 is used to generate a highlight / shadow correction curve 31.

The halftone correction curve 30 moves the M control point 33 on the tone curve upward when the halftone correction amount is +, and moves downward when the halftone correction amount is-. Then, the starting point (0, 0) on the low luminance side of the tone curve, the M control point 33, and the starting point (255, 255) on the high luminance side of the tone curve are smoothly interpolated by curve interpolation to obtain a tone curve.

The highlight / shadow correction curve 31 is
When the highlight control amount and the shadow control amount are +-,
As shown, the HL control point 35 and the SD control point 34 move. A straight line is interpolated between the control points HL and SD.

Finally, a composite curve 32 is obtained by combining the halftone correction curve 30 with the highlight and shadow correction curves 31. The above processing is performed using the RGB
This is performed for each of the RGB tone curves using the control parameter 42.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The following describes an apparatus for correcting a color cast of an image, which is an output of the present invention, with reference to the drawings. 1 and 2 are block diagrams showing an embodiment of an image color cast correction apparatus according to the present invention.

In the embodiment of the present invention, an input image buffer 1 for holding input image data and a correction pattern storage unit 2 for storing a correction pattern for correcting a color cast of the input image.
A correction pattern specifying means 3 for the operator to specify a correction pattern, a correction amount storage unit 4 for storing a correction amount of color cast, a correction amount specifying means 5 for the operator to specify a correction amount, Tone curve generating means 6 for generating RGB tone curve data from the data in the storage unit and the correction amount, and L for storing the tone curve data.
It comprises a UT storage buffer 7, LUT conversion means 8 for performing table conversion on image data, and an output image buffer 9 for holding output data.

The tone curve generating means 6 further comprises a highlight control amount calculating section 20 for calculating a control amount of each highlight section of RGB from the data of the correction pattern storage section and the correction amount, and a halftone control section of each of RGB. A halftone control amount calculation unit 21 for calculating a control amount, a shadow control amount calculation unit 22 for calculating a control amount for each of the RGB shadow units, and a tone curve data calculation for calculating tone curve data from the three control amounts. A unit 25 includes a straight line generating module 23 and a spline generating module 24.

First, the operator selects which correction pattern is to be applied to the image in the input image buffer 1 by the correction pattern designating means 3. At the same time, the color cast correction amount is input to the correction amount storage unit 4 by the operator. Note that a fixed value can be stored in the correction amount storage unit 4 in advance.

As an example of the correction pattern designating means 3, there is an apparatus having a monitor for displaying an image and an I / F for selecting a correction pattern with a mouse. FIG. 6 shows an example of the correction pattern specifying means 4 and the correction amount specifying means 4. The correction pattern designating means 50 displays items such as a green color cast of a fluorescent light, a yellow color cast of an electric light, a red color cast of a snow scene, and a blue color cast of a north window light. Enter what you want to do. Further, a slider bar for designating the strength of the correction amount is displayed as the correction amount designating means 51, and the slider bar can also be used to designate the correction amount using the mouse 52.

Next, the tone curve generating means 6 obtains tone curve data of the specified correction pattern from the correction pattern storage unit 2 and calculates RGB tone curve data by combining the specified correction amounts.
The tone curve data is input to the LUT storage buffer 7.

In the tone curve generating means 6, first, control parameters are obtained from the correction pattern storage unit 2. FIG. 5 shows an example of the control parameters.

Hereinafter, the operation of the apparatus will be described assuming that the data shown in FIG. 5 is used. The highlight control amount calculation unit 20
First, a highlight control parameter hl is obtained from the correction pattern storage unit 2 as a control parameter. The control parameter hl is stored in the form of the RGB control parameter 42, which is (Rh, Gh, Bh). Next, the correction amount grad is obtained from the correction amount storage unit 4. And g
A highlight control amount is calculated from rad and hl. The highlight control amount HL = (HLr, HLg, HLb) can be obtained by HLr = Rh × grad (2) HLg = Gh × grad (3) HLb = Bh × grad (4)

Similarly, the halftone control amount calculator 21 calculates a halftone control amount. Halftone parameter m =
(R _m , G _m , B _m ), the halftone control amount M = (M
_r , M _g , M _b ) can be obtained by the _following equation: M _r = R _m × grad (5) M _g = G _m × grad (6) M _b = B _m × grad (7)

Similarly, the shadow control amount calculator 22 calculates the shadow control amount. Shadow parameter s
If d = (R _s , G _s , B _s ), the shadow control amount S
D = (SD _r , SD _g , SD _b ) can be obtained by SD _r = R _s × grad (8) SD _g = G _s × grad (9) SD _b = B _s × grad (10)

That is, the control amount per unit correction amount is described in the correction pattern storage unit 2, and each control amount calculation unit multiplies the unit control amount by the data in the correction amount storage unit 4 to freely perform correction. The degree is increasing.

Next, tone curve data is calculated by the tone curve data calculating section 25 based on the three control amounts. FIG. 4 shows an example of tone curve data calculation. Here, it is assumed that the RGB data is 8 bits and has a pixel value from 0 to 255. The starting point on the low luminance side of the tone curve is (0,
0), the starting point on the high luminance side of the tone curve is (255, 25)
5).

First, the halftone correction curve 30 is calculated from the halftone control amount M. The M control point 33 is a control point in the middle of the brightness of the tone curve. In FIG. 4, the M control point 33 is a point (128, 128). When the halftone control amount M is +, the M control point 33 moves upward, and when it is-, moves downward.

An example of the movement of the M control point 33 will be described below. When the R component of the halftone control quantity M and M _r, M control point 33 of the R tone curve from (128, 128), (128,
128 + M _r ). That is, when M _r = 100, the M control point 33 of the R tone curve is (128,
From (128) to (128,228). Similarly, the M control points 33 of the G tone curve and the B tone curve can be moved.

Therefore, when the value is +, a curve like the conversion characteristic 36 is obtained. For the actual calculation, a curve can be generated by performing curve interpolation on the point (0, 0), the M control point 33, and the point (255, 255) by spline interpolation or the like. Curve interpolation is based on the starting point (0, 0) and point (255, 2) of the tone curve.
55) and the M control point 33 need only be smoothly connected, and an arbitrary interpolation function can be used. At the time of curve interpolation, the curve generation module 24 is started.

Next, the highlight control amount HL and the shadow control amount SD are used to determine the highlight and shadow correction curve 31.
Is calculated. The HL control point 35 is a control point on the higher tone curve of the tone curve. Highlight control amount HL
Is +, the HL control point 35 is (255, 25
It is assumed that the movement is from (5) to (0, 255).
When the highlight control amount HL is-, the HL control point 35
Moves from (255, 255) to (255, 0). The SD control point 34 is a control point on the lower side of the luminance of the tone curve. Shadow control amount SD
Is +, the SD control point 34 moves from (0, 0) to (255, 0). When the shadow control amount SD is-, the SD control point 34
It is assumed that the object moves from (0,0) to (0,255).

An example of the movement of the SD control point 34 and the HL control point 35 will be described below. When shadow control amount SD of the R component and SD _r, when the SD _r> 0 is the R tone curve SD
The control point 34 moves from (0, 0) to (SD _r , 0). When SD _r <0, the R tone curve S
The D control point 34 moves from (0, 0) to (0, -SD _r ).

[0041] When the high light control amount HL of the R component and HL _r, when HL _r> 0 is the R tone curve HL
The control point 35 is (255, −255) from (255, −255).
HL _r , 255). When HL _r <0, the HL control point 35 of the R tone curve is (255, 2
55) to (255, 255 + HL _r ).

Similarly, the SD control point 34 and the HL control point 35 of the G tone curve and the B tone curve can be moved.

A highlight and a shadow correction curve 31 are obtained by interpolating the thus set control points with a straight line. At the time of linear interpolation, the linear generation module 23 is started.

Next, the tone curve data calculator 25 combines these two tables. The halftone correction curve 30 obtained above is highlighted in Table A [].
Assuming that the shadow correction curve 31 is stored in TableB [], the composite curve 32 is expressed as Table [i] = TableB [TableA [i]] (i = 0, 1, 2,... 255) (11) It can be obtained by the formula. FIG. 4 shows a conversion characteristic 38 as an example of the composite curve. Note that the composite curve 32 is represented by R
It is calculated independently for each of GB.

The calculated composite curve 32 is output to the LUT storage buffer 7 as RGB tone curve data.

After the table is set in the LUT storage buffer 7, the RGB image data is sequentially read by the LUT conversion unit 8, and according to the data in the LUT storage buffer 7, the RGB data is converted by referring to the same table. . The values after the table conversion are stored in the output image buffer 9. Thus, the process ends.

[0047]

As described above, the apparatus for correcting color cast of an image according to the present invention merely designates a correction pattern corresponding to a defect in an image and a correction amount thereof. And snow scene correction.

Since there are few adjustment parameters and an intuitive designation method, there is an effect that the time required for the correction is short, the man-hour for the operator is reduced, and the correction can be performed as expected even by an operator.

The present apparatus can correct not only the correction due to the difference in the light source color but also the color cast caused by the film sensitivity and the exposure processing by registering an appropriate correction pattern.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an image color cast correction apparatus of the present invention.

FIG. 2 is a block diagram illustrating an embodiment of an image color cast correction apparatus according to the present invention.

FIG. 3 is a flowchart of an image color cast correction method according to the present invention.

FIG. 4 is a conceptual diagram of generation of a color cast correction table.

FIG. 5 is an example of data in a correction pattern storage unit.

FIG. 6 is an example of a correction parameter specifying device.

FIG. 7 is a block diagram showing an embodiment of a conventional image color cast correction apparatus.

[Description of Signs] 1 Input image buffer 2 Correction pattern storage unit 3 Correction pattern designation unit 4 Correction amount storage unit 5 Correction amount designation unit 6 Tone curve generation unit 7 LUT storage buffer 8 LUT conversion unit 9 Output image buffer 20 Highlight control Amount calculation unit 21 Halftone control amount calculation unit 22 Shadow control amount calculation unit 23 Straight line generation module 24 Curve generation module 25 Tone curve data calculation unit 30 Halftone correction curve 31 Highlight, shadow correction curve 32 Synthetic curve 33 M control point 34 SD control point 35 HL control point 36 Conversion characteristic 37 Conversion characteristic 38 Conversion characteristic 40 Data example of correction pattern storage unit 41 Tone curve control parameter 42 RGB control parameter 50 Correction pattern specifying means 51 Correction amount specifying means 52 Mouse 53 CRT monitor 1 0 input image buffer 101 color corrector 102 spectral characteristics storage 103 correction amount retrieval unit 104 outputs the image buffer

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06T 5/00-5/50 G06T 1/00 H04N 1/56-1/62 H04N 9/69

Claims (6)

(57) [Claims] [Claim 1] to calculate the RGB tone curve from a correction amount indicating a degree of correction pattern and the correction is selected as failure of the input image, performs color conversion processing on all the pixels using the tone curve image Color cast correction method,
The correction pattern is a high-resolution image on an RGB tone curve.
A color cast correction method for an image, wherein the control information is color , halftone, and shadow control information . 2. A control point H in a high luminance range in the tone curve.
L and the control point SD in the low luminance region are interpolated by a straight line,
Control point M at the midpoint in the tone curve and the height of the tone curve.
Performs curve interpolation that smoothly connects the starting point on the luminance side and the starting point on the low luminance side.
2. The color cast correction of an image according to claim 1, wherein
Method. 3. A correction selected according to a defect of an input image.
From the pattern and the correction amount indicating the degree of correction, the RGB
Tone curve and calculate all pixels using the tone curve.
A color cast correction device for performing an image color conversion process,
Input image buffer to hold input image data , highlight, halftone, shadow on RGB tone curve
Of the correction pattern consisting of the control information of
A correction pattern storage unit that stores a plurality of colors according to
A correction amount storage unit for storing the correction amount of the RGB, and an RGB tone mask based on the correction pattern and the correction amount.
Curve generating means for generating a curve, an LUT storage buffer for storing the tone curve, and an LUT for performing table conversion on the input image data
A conversion unit, and an output image buffer for holding the converted image data.
A color cast correction device for images, characterized in that: 4. The method according to claim 1, wherein said tone curve generating means is adapted to output said correction curve.
From the turn and the correction amount, control of each highlight portion of RGB is performed.
A highlight control amount calculation unit for calculating the control amount;
A halftone control amount calculation unit for calculating a halftone control amount;
Shadow control amount calculation for calculating the control amount of each shadow portion of B
The tone curve is calculated from the output portion and the three control amounts.
And a tone curve calculation unit that performs
An apparatus for correcting color cast of an image according to claim 3. 5. The tone curve generating means according to claim 1 , wherein
Control point HL in the high luminance area and control point SD in the low luminance area
A straight line generation module that interpolates
Control point M in the middle of the curve,
Curve generation that performs curve interpolation that smoothly connects to the starting point on the low luminance side
A raw module is provided.
Image color cast correction device. 6. A correction pattern for correcting an image defect.
A correction pattern designating means designated by the operator, and an operator
And a correction amount specifying means for specifying a correction strength.
6. The color of an image according to claim 3, 4 or 5, wherein
Fogging correction device.