JP2002016807A - Gradation conversion table creation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a processed image with improved image quality by eliminating influence by gradation characteristics for each digital camera model to image data acquired by a digital camera. SOLUTION: A model gradation characteristic curve C1 for absorbing the model gradation characteristics of the digital camera is independently created for each digital camera model from gradation correction curves C3 and C4, and the model gradation characteristic curve C1 is used to convert the image data acquired by the digital camera to what does not depend on the digital camera model. After that, another gradation modification processing is made, and a gradation conversion table T0 is created to obtain processed image data S2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method of creating a gradation conversion table used when performing a gradation changing process on image data obtained by a digital camera.

[0002]

2. Description of the Related Art In a digital camera, an image obtained by imaging is recorded as digital image data on a recording medium such as an internal memory or an IC card provided inside the digital camera, and based on the recorded digital image data, An image obtained by imaging can be displayed on a printer or a monitor. As described above, when an image obtained by a digital camera is printed, it is expected to have the same high-quality image quality as a photograph printed from a negative film.

A digital camera has an optical system (aperture, shutter, strobe), an imaging system (CCD, signal processing system),
Control system (AE, AWB, AF), recording / playback system (compression / expansion,
Memory control, display). Among these factors, the factors that affect the image quality of the reproduced image include strobe light color temperature, AE (auto exposure control) processing, AWB (auto white balance adjustment) processing, CCD color separation color filters, Examples include the number of pixels, gradation conversion, and matrix calculation processing for obtaining a luminance / color difference signal. In a digital camera, these factors are controlled to obtain digital image data that provides a high-quality reproduced image. I have.

For this reason, in digital cameras, AE
Function, AWB function, and image processing function such as gradation conversion.
Since the image processing has already been performed as described above, the image can be directly input to the copying apparatus to reproduce the image. However, in a normal digital camera, the digital image data is subjected to AE processing and AWB processing on the assumption that the image data is reproduced on a monitor.
E, AWB function may not be added. Furthermore, in many digital cameras, image data is formed due to improper exposure such as overexposure or underexposure or strobe light adjustment error.Therefore, when outputting to a printer, it is necessary to correct gradation and print. It is necessary to perform appropriate AE processing and AWB processing again. Here, the AE and AWB processes at the time of printer output are described as “Printer AE / A
WB processing ”.

Therefore, for example, as described in Japanese Patent Application Laid-Open No. H11-220619, an average value is obtained for each of RGB color signals constituting image data from a digital camera, and this average value is determined as a target value suitable for printing. A method has been proposed in which a correction value is determined so that the exposure amount and the white balance are corrected based on the correction value.

On the other hand, as described above, factors that affect the image quality of a reproduced image in a digital camera include color temperature of strobe light, AE (auto exposure control) processing, AWB
(Auto white balance adjustment) processing, CCD color separation color filter, number of pixels, gradation conversion, matrix calculation processing to obtain luminance / color difference signals, etc. These factors depend on the manufacturer and model of the digital camera. No,
Above all, gradation conversion is closely related to printer AE / AWB processing. In the above method, the printer AE / AW
The image data used for the B processing is, for example, an image obtained by converting original image data from a digital camera into a true number, or image data obtained by further converting the density, all depending on the type of digital camera. Image data having a gradation characteristic is analyzed by the printer AE / AWB processing.

For this reason, when performing printer AE / AWB processing, since image data to be analyzed depends on the type of digital camera, it is difficult to obtain a correction value suitable for printing, and It is very inefficient because it is necessary to print once and correct it by trial and error.

Accordingly, as a method of performing image processing according to the type of digital camera, for example, Japanese Patent Laid-Open No. 10-208034
As described in the publication, a method has been proposed in which an image is analyzed, a large number of gradations are assigned to a range where the luminance distribution density is large, and a gradation conversion table is automatically generated for each image.

As described in Japanese Patent Application Laid-Open No. 11-220687,
When performing image processing on image data obtained by a digital camera, a system that determines different image processing conditions for the aforementioned image data according to the model of the digital camera and performs image processing based on the image processing conditions Has been proposed.

[0010]

SUMMARY OF THE INVENTION
In the system of No. 10-208034, when creating the gradation conversion table, only the analysis information of the image is used, so that an appropriate image may not be obtained as a result of the image processing.

In the system disclosed in Japanese Patent Application Laid-Open No. H11-220687, when performing gradation change on image data, gradation change processing conditions are determined according to the type of digital camera. Different tone characteristic correction and tone correction suitable for printing are simultaneously performed under the above-described tone change processing condition. However, the tone conversion processing condition (tone conversion table) used when performing tone conversion is as follows. Because it includes both elements of digital camera model gradation correction and print output gradation correction, printer AE
When performing / AWB processing, the original image data acquired by the digital camera must be analyzed.
The analysis performance of the printer AE / AWB becomes unstable, and trial and error by printout is inevitable.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has been made in consideration of the above-described circumstances, and is directed to a process of changing gradation so that a high-quality reproduced image can be efficiently obtained from image data obtained regardless of the type of digital camera. It is an object of the present invention to provide a method of creating a gradation conversion table for performing the following.

[0013]

A method for creating a gradation conversion table according to the present invention is used when image data acquired by a digital camera is subjected to a process of changing the gradation to obtain processed image data. And using a model gradation characteristic profile representing the gradation characteristic of the digital camera created for each model of the digital camera that has acquired the image data. It is.

Here, the "model gradation characteristic profile" is data which is a source of conversion data for correcting the gradation characteristic of each digital camera model. The model gradation characteristic profile is obtained by taking an actual signal of the subject (that is, a signal of γ = 1 for the subject) as an input signal and photographing the subject with a digital camera and acquiring the model gradation characteristic. May be used as an output signal, and if the data is used for back calculation, the actual signal of the subject is obtained from the signal obtained by photographing with the digital camera, that is, the model gradation characteristic of the digital camera. , And the signal can be converted into a signal having a gradation characteristic independent of the digital camera. From the viewpoint of convenience of use and simplicity of calculation, the input signal and the output signal are reversed from the above, that is, the signal obtained by the digital camera is used as the input signal, and the actual signal of the subject is used as the output signal. It is desirable to create a model gradation characteristic profile that can directly correct the gradation characteristic of a digital camera model. Further, the output signal and the input signal may be a gradation characteristic correction curve represented by a vertical axis and a horizontal axis, respectively, or a value in the gradation characteristic correction curve may be represented by an input signal and an output signal. Of course, a lookup table (LUT) represented by a corresponding table may be used.

The method of creating the model gradation characteristic profile is as follows. For each different photographing condition, a log chart obtained from a digital camera RGB value corresponding to each gray patch obtained by photographing a gray chart using each digital camera. And the value of
A logarithmic luminance measurement value obtained by logarithmizing a luminance measurement value obtained by measuring the luminance of each gray patch of the gray chart under each shooting condition is based on data obtained in correspondence with each gray patch. Preferably, it is created.

Under each of the different photographing conditions, when photographing a gray chart using each digital camera,
From the image data acquired under each exposure condition, there is a difference between the brightest gray patch and the second brightest gray patch, and the brightest gray patch is changed. The image data that is not saturated and is the most overexposed is selected, and the digital camera RGB value corresponding to each of the gray patches is obtained from the image data.
It is preferable to use the LogY value obtained from the GB value to create the approximate curve.

The model gradation characteristic profile is obtained by obtaining an approximate curve based on data obtained by associating the logarithmic luminance measurement value with the LogY value obtained under the same photographing conditions for each of the gray patches. Is more preferable.

Further, if the photographing conditions are strobe ON / O
It is preferable to include at least one of the FF condition, the type of shooting light source such as auto, clear sky, cloudy sky, fluorescent light, and tungsten, and sensitivity.

When it is necessary to adjust the gray balance, the model gradation characteristics obtained by the above method are assumed to correspond to G, and R and G are determined based on the difference between R and G and the difference between B and G.
It is preferable to create a model gradation characteristic profile corresponding to B and B.

[0020]

According to the method for creating a gradation conversion table of the present invention, image data is acquired because there is an independent model gradation characteristic profile for acquiring the image data and absorbing the model gradation characteristic of the digital camera. According to the model of the digital camera, first, preprocessing for absorbing only the gradation characteristics of the model can be performed. Therefore, the printer AE / AWB processing can analyze image data having gradation characteristics independent of the model of the digital camera, thereby improving the performance of the printer AE / AWB. In addition, since data to be processed such as standard gradation processing for printout and / or nonlinear correction of a highlight portion and a shadow portion can be converted into image data independent of a model, each process can be performed quickly and quickly. Can be done accurately. Therefore, it is possible to efficiently obtain a high-quality reproduced image regardless of the type of the digital camera.

Further, if the model gradation conversion table according to the method of the present invention is used, the gradation characteristic absorption processing of the digital camera can be separated from the processing of changing the gradation so as to be suitable for the printer. However, it is not necessary to change the gradation characteristic absorption processing portion of the digital camera, so that the amount of calculation for correction and the amount of data associated therewith can be reduced, and the cost can be reduced.

Further, in the tone conversion table creating method of the present invention, when creating an independent model tone characteristic profile, a gray chart is set using each digital camera, and an exposure condition from underexposure to overexposure is set. In the image data obtained under the respective exposure conditions, there is a difference between the brightest gray patch and the second brightest gray patch, and the brightest gray patch is not saturated, and Over image data is selected, and the digital camera RGB density values corresponding to the respective gray patches are determined from the image data, so that the tone characteristics of the model of the digital camera can be absorbed and the CCD and lens of each digital camera can be used. Can be corrected. Furthermore, since the image data selected by the above method does not have the highlight portion crushed and the shadow portion is not crushed,
Since reliable data can be obtained, the performance of gradation conversion processing can be improved.

In the present invention, when a model gradation characteristic profile is created, the strobe condition (ON / OF)
F), and more reliable data can be obtained by creating for each shooting light source (auto, sunny, cloudy, fluorescent, tungsten, etc.). In addition, since tone correction is performed inside the digital camera to make noise less noticeable, if a model tone characteristic profile is created for each sensitivity (to the extent that noise is less noticeable), the digital camera model The correction of the tonal characteristic becomes more accurate.

Further, as the model tone characteristic profile created by the tone conversion table creating method according to the present invention, the same profile is used for each of R, G, and B in order to maintain the gray balance. However, if the gray balance is changed, the digital camera usually only controls the RGB gain in the antilog space, so the difference between the R and G densities and the B and G densities Since a model gradation characteristic profile corresponding to R and B can be created based on the difference between, the gray balance can be easily adjusted as needed.

[0025]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of an embodiment of a gradation conversion table creation method according to the present invention. As can be seen from FIG. 1, in the present embodiment, the gradation conversion table T0 for performing the process of changing the gradation on the image data obtained by the digital camera is the color data constituting the image data S1. R1, G1, B1 data with gradation conversion processing
This is for converting color data R2, G2, and B2 constituting S2, and is created from the first quadrant to the fourth quadrant. In the processing from the first quadrant to the fourth quadrant, all the processing is performed in the RGB density space.

In the calculation in each quadrant, a negative density value is also stored.

A method of creating a model gradation characteristic profile for absorbing the model camera gradation characteristics will be described later. As shown in FIG. 1, however, as shown in FIG. In order to obtain, in the first quadrant, the model gradation characteristic profile independently created (here, the model gradation characteristic curve C1)
Is used to set processing conditions for absorbing the gradation characteristics of the model of the digital camera. Here, the horizontal axis is the RGB density value of the image acquired by the digital camera (hereinafter, referred to as the RGB density value).
Digital camera RGB density). The image data acquired by the digital camera is ITU-R BT. 709 (REC 70
9), color data R, G, and B constituting image data are represented by R ′, based on the following equations (1) to (3).
G 'and B' can be converted. LogR ', LogB', and LogC 'obtained by logarithmizing these R', G ', and B' are digital cameras R on the horizontal axis.
Use the value of GB concentration.

[0029] Pr = R / 255 Pg = G / 255 (1) Pb = B / 255 R '= ((Pr + 0.099) /1.099) 2.222 G' = ((Pg + 0.099) /1.099) 2. ²²² (Pr, Pg, Pb ≧ 0.081) (2) B ′ = ((Pb + 0.099) /1.099) ^2.222 R ′ = Pr / 4.5 G ′ = Pg / 4.5 (Pr, Pg, Pb <0.081) (3) B ′ = Pb / 4.5 With this gradation curve C1, a logarithmic exposure can be obtained from the digital camera RGB density of the image data obtained from the color data R1, G1, and B1 constituting the image data S1. It will be. Here, the logarithmic exposure amount is a logarithmic value of the exposure amount captured by the digital camera.

Here, using the normal gradation curve C1, R,
Performs the process of absorbing the model gradation characteristics for each of G and B, but if it is necessary to change the gray balance, the model gradation characteristics absorption curve corresponding to each of R, G, and B described later C
The model gradation characteristic absorption processing may be performed using 1 _R , C 1, and C 1 _B.

In the second quadrant, a straight line C2 for correcting the exposure amount is set. Straight line C2 that corrects this exposure
Is basically a straight line passing through the origin, but the exposure amount is corrected by translating this curve C2 in the direction of arrow A. The exposure correction here includes correction of white balance (printer AE / AWB). As the correction amount required for performing the printer AE / AWB process, for example,
As described in JP-A-220619, an average value may be obtained for each of the RGB color signals constituting the image data, and a correction value obtained such that the average value becomes a target value suitable for printing may be used.

In the third quadrant, a reference gradation curve is set. Here, it is assumed that the standard gradation curve C3 has been set. This standard gradation curve C3 is an S-shaped curve, and the middle portion corresponds to γ = 1.6. Here, in the present embodiment, the conversion by the gradation curve C3 is γ
This is called conversion. The γ-conversion is intended to enhance the gradation of the output of a printer by making the gradation hard when printing out.

In the fourth quadrant, a curve C4 for nonlinearly correcting a highlight portion and a shadow portion of an image is set. Normally, the print has a narrow reproduction range of density and jumps to a highlight portion of an image, but is liable to be crushed in a shadow portion. Therefore, here, for example, a method described in Japanese Patent Application Laid-Open No. If the print density increases due to the AWB process, the gradation on the highlight side is hardened and the gradation on the shadow side is softened. Conversely, if the print density decreases, the highlight A correction amount is determined so as to soften the density on the image side and to harden the gradation on the shadow side, and reflect this in the correction curve C4.

Since the data obtained after the processing in the fourth quadrant is the gradation-converted RGB density values (LogR2 ', LogG2', LogB2 '), this data is subjected to antilogarithmic conversion processing.
The calculation procedure of the above equations (1) to (3) is reversed, and the color data R2, G2,
Find B2.

From the first quadrant to the fourth quadrant, the floor for performing the gradation changing process on the image data S1 composed of R1, G1, and B1 based on the set curves. Image data S1 to be used for pre-processing by setting tone conversion conditions
Calculation of digital camera density value, color data R2, G to be post-processed
2. A gradation conversion table T0 for converting the image data S1 into gradation-processed image data S2 is created together with the calculation of B2 and the like.

As can be understood from the above description, when the gradation conversion table T0 is created in the present embodiment, the model gradation characteristic profile (C1 in this embodiment) is the reference gradation curve.
C3 is independent of the curve C4 for performing non-linear correction on the highlight portion and the shadow portion. Therefore, first, the process of absorbing the model gradation characteristics of the digital camera can be performed using the model gradation characteristic profile C1. Therefore, after the printer
The image data used when determining the processing conditions for the AE / AWB processing, γ conversion, highlight part and shadow part gradation non-linear correction processing does not depend on the model of the digital camera, and is used for each processing. Processing conditions can be accurately determined, and as a result, more appropriate image processing can be performed.

Next, a method of creating a model gradation characteristic profile used in the present invention will be described.

In order to create a model gradation characteristic profile,
First, the luminance measurement and photographing of the gray chart are performed. For each model, auto, fluorescent lamp, tungsten, cloudy sky, outdoor clear sky, strobe ON / OFF condition, and each desired combination of digital camera sensitivity and shooting conditions,
The luminance of the gray chart is measured, and the measured luminance value of each gray patch of the gray chart is logarithmized to obtain a logarithmic luminance measurement value. The photographing for the gray chart is performed under the exposure condition (slope photographing) that is changed stepwise from underexposure to overexposure for each photographing condition in order to obtain accurate data.

For each photographing condition, the brightest gray patch is not saturated and the brightest gray patch is selected from the image data obtained by the above-described slope photographing.
Select the image data that is the most overexposed so that there is a difference in the second brightest gray patch and that the shadow part is not crushed. As described above, the image data acquired by the digital camera is based on ITU-R BT. 709 (REC.709), the color data R, G, and B constituting this image data are converted to CIE1931 based on the following equations (1) to (4).
Convert to tristimulus values X, Y, Z.

[0040] Pr = R / 255 Pg = G / 255 (1) Pb = B / 255 R '= ((Pr + 0.099) /1.099) 2.222 G' = ((Pg + 0.099) /1.099) 2. ²²² (Pr, Pg, Pb ≧ 0.081) (2) B ′ = ((Pb + 0.099) /1.099) ^2.222 R ′ = Pr / 4.5 G ′ = Pg / 4.5 (Pr, Pg, Pb <0.081) (3) B ′ = Pb / 4.5 X R ′ Y = | A | · G ′ (4) Z B ′ where the matrix | A | is the color data R ′, G ′,
This is a matrix for converting B 'into tristimulus values X, Y, and Z. For example, the following values can be used.

0.4124 0.3576 0.1805 | A | = 0.2126 0.7152 0.0722 (5) 0.0193 0.1192 1.0571 Instead of the matrix | A |, the tristimulus values X, Y, and Z may be obtained by a look-up table.

From the values of X, Y, and Z obtained by the above calculation, a LogY value corresponding to each gray patch is obtained. Or R ',
The LogG 'value obtained by logarithmizing G' and B 'may be used as the LogY value.

A model gradation characteristic profile of a digital camera is created based on the logarithmic luminance measurement value and LogY value of each gray patch obtained here. For example, the following method can be used.

The logarithmic luminance measurement value corresponding to the brightest gray patch is normalized to a log exposure value of 0, the LogY value corresponding to the brightest gray patch is normalized to a digital camera RGB density of 0, and the logarithmic luminance measurement value for each patch is normalized. And the LogY value are made to correspond to the logarithmic exposure amount and the RGB density of the digital camera, and an approximate curve C1 ′ shown in FIG. 2 is obtained by a cubic polynomial using the least squares method.
It should be noted that the approximate curve C1 ′ may be obtained using another method.

Using this approximate curve C1 'as the basis of the model gradation characteristic profile, C1' is translated in parallel as shown by arrow B in FIG. 2 to create a model gradation characteristic profile of the digital camera. For example, in a digital camera, the AE of the digital camera is usually controlled so as to reproduce the skin of a person at an appropriate density because it is desired to most appropriately represent the skin of the person. Therefore, the human skin density, in this case,
Considering 0.75 as an appropriately corrected density value, the approximate curve C
Digital camera RGB density as indicated by arrow B based on 1 '
A curve C1 obtained by translating C1 ′ so that 0.75 becomes the log exposure amount 0.75 is defined as a model gradation characteristic curve. By using the curve C1 created in this way, the AE characteristics of the digital camera can be absorbed at the stage of processing the gradation characteristics of the model while maintaining the gray balance of the image.
/ AWB processing is simplified and efficiency is improved.

As described above, normally, in order not to lose the gray balance, R, R,
The process of absorbing the model gradation characteristics is performed for each of G and B. However, when it is necessary to change the gray balance, usually, inside the digital camera, the RGB gain is controlled in an exact number space. If you create a model gradation characteristic profile for the R and B colors based on the difference between the density of R and G, and the difference between the density of B and G, you can easily adjust the gray balance as needed. it can.

The model gradation characteristic profile created in this manner can accurately correct the model gradation characteristic of each digital camera under each photographing condition without depending on image data obtained by imaging with a digital camera. Therefore, it is possible to prevent the gradation correction performance from being unstable due to the analysis for each image.

In the present embodiment, the curve corresponding to the digital camera RGB density value and the logarithmic exposure amount is used as the model gradation characteristic profile. Any form can be used as long as it can be made independent and can perform processing for absorbing the gradation characteristics of the model of the digital camera.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a method of creating a gradation conversion table according to an embodiment of the present invention.

FIG. 2 is a view for explaining a method of creating a model gradation characteristic profile in the embodiment.

[Explanation of symbols]

 C1 Model gradation characteristic profile C2 Exposure amount correction straight line C3 γ conversion curve C4 Non-linear correction curve

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference) TT02 TT09

Claims (6)

[Claims] 1. A method of creating a gradation conversion table used when performing gradation change processing on image data acquired by a digital camera to obtain processed image data, comprising: A method for creating a gradation conversion table, wherein a model gradation characteristic profile representing a model gradation characteristic of a camera is used. 2. A LogY value obtained from a digital camera RGB value corresponding to each gray patch obtained by photographing a gray chart using each digital camera, for each of the model gradation characteristic profiles for different photographing conditions. And a logarithmic luminance measurement value obtained by logarithmizing a luminance measurement value obtained by measuring the luminance of each gray patch of the gray chart under the respective photographing conditions to data obtained in correspondence with each of the gray patches. 2. The method according to claim 1, wherein the table is created based on the image data. 3. In each of the photographing conditions, when photographing a gray chart using each digital camera, photographing is performed while changing exposure conditions from underexposure to overexposure, and image data acquired under the respective exposure conditions is obtained. Among them, there is a difference between the brightest gray patch and the second brightest gray patch, and the brightest gray patch is not saturated, and selects the most overexposed image data. 3. The method according to claim 2, wherein RGB values of the digital camera corresponding to the patch are obtained. 4. The model gradation characteristic profile is obtained under the same photographing condition as a luminance measured value obtained by measuring the luminance of each gray patch of the gray chart for each photographing condition and each digital camera. LogY obtained
4. The gradation conversion table creating method according to claim 2, wherein the approximate curve is an approximate curve based on data obtained by associating a value with each of the gray patches. 5. The photographing condition is as follows: strobe ON / OFF
5. The method according to claim 2, wherein at least one of the condition, the type of the imaging light source, and the sensitivity is included.
Method for creating a gradation conversion table as described in item 1. 6. A model gradation characteristic profile corresponding to G color, and a model gradation characteristic profile corresponding to each of R and B is created based on a difference between R and B and a difference between B and G. 6. One of claims 2 to 5, characterized in that:
Method of creating gradation conversion table described in section.