KR101023508B1 - Apparatus and method for color correction of image including multiple light sources - Google Patents

Abstract

PURPOSE: A color correcting apparatus and a method thereof in order to have constancy of an object color are provided to supply an image having a color which is similar to a real color regardless of the kinds of light sources. CONSTITUTION: Each tone reproduction curve and a color compensation matrix are applied to an ICC profile tag. The color of a second image corresponding to each first image is compensated. A compensated-image generator(140) generates a plurality of compensation images. A synthetic image generator(150) generates one composite image by synthesizing each pixel according to a synthesis ratio.

Description

Apparatus and method for color correction of image including multiple light sources}

The present invention relates to an apparatus and method for correcting color of an image including a plurality of light sources, and more particularly, to an apparatus and method for accurately expressing colors of an image including a plurality of light sources of different types.

In digital technology, color is an important attribute for defining, understanding, and identifying objects. The color appears with the combined presence of the light source, the object and the observer. If any one of these three changes, a different color is obtained. The development of digital devices has increased the need to standardize the colors obtained from these devices for ease of handling, sharing and broadcasting. This is the main purpose of the Color Management System (CMS).

1 is a diagram illustrating a configuration of a color management system. Referring to FIG. 1, a color management system allows various digital devices to recognize an image in common. If there is no appropriate color management module (CMM), each of the devices shown in FIG. 1 will represent a given image differently. As a result, the color management system (1) assigns a specific meaning of color so that the image is not ambiguous, and (2) allows each imaging device to produce the same color.

The International Color Consortium (ICC) profile has been proposed to realize the functions of the color management system and to unify the color characteristics of various digital devices. The ICC defines standards for the configuration and device profiles of color operating systems that ensure that color images of the same quality are obtained on all hardware, regardless of the manufacturer of the digital device.

In ICC-based color management systems, profile connection space (PCS) assigns device-independent and color-specific values. Profiles also associate the color values of the PCS with the color format of the device. This is an important part of the overall color management process, as this connection provided by the ICC profile allows color conversion between devices. Finally, in the ICC-based color management system, the color management module performs the necessary calculation process using the profile connection relationship between the color values of the PCS and the color values dependent on the output device.

On the other hand, the problem of color constancy has been steadily studied in both human and computer vision, and most of the color uniformity conservation algorithms are based on limited assumptions such as constant illumination. However, this condition is not always met when both incandescent bulbs and natural light illuminate the object. To overcome these problems, various constraints have been introduced, such as the direction of the light source, the value of the color temperature, and the surface constraints.

The Retinex algorithm partially approaches the problem of changing lighting, eliminates the change in lighting, and calculates the surface brightness independently for each of the three color channels. The Retinex algorithm also assumes that a small change in one color channel represents a change in illumination, while a large change means a change in color. Thus, small changes in the color channel can be reduced, and large changes can be restored by summation. However, this method generates an error in classifying illuminance changes and can lead to serious inconsistencies in the restored results.

Another method of color identity uses the algebraic information of an image, which is statistical knowledge of light, color, and geometric model of color space. The solution of chromatic identity can be obtained by representing the surface and the light source in the form of a linear combination of finite dimensional based functions. In addition, the Bayesian approach has been used to solve the problem of color identity by probabilistic theory.

As described above, various studies have been conducted to maintain a constant color in an image regardless of lighting conditions. However, there is a need for a method that can implement color identity using a simpler method.

SUMMARY OF THE INVENTION The present invention provides a device and a method for correcting color of an image including a plurality of light sources capable of making the color of an object included in an image photographed under a plurality of different light sources have an identity. There is.

Another object of the present invention is to provide a method for correcting a color of an image including a plurality of light sources capable of causing the color of an object included in an image photographed under a plurality of different light sources to have identity. A computer readable recording medium having recorded thereon a program for execution is provided.

In order to achieve the above technical problem, a color correcting apparatus for an image including a plurality of light sources according to the present invention includes a case in which one light source of each of a plurality of different light sources is turned on and each of the plurality of light sources When it is turned on, a plurality of first images photographed by arranging color charts of red, green, blue, and gray scale colors in a preset photographing region and a plurality of second images photographing only the photographing region Image input unit for receiving a; Computing a color tone reproduction curve corresponding to each of the first images by comparing each color constituting a reference region, which is an area corresponding to the color chart extracted from the plurality of first images, with an actual color of the color chart. Curve calculator; A matrix calculator for calculating a color correction matrix for linearly converting each color constituting the reference area into each actual color constituting the color chart corresponding to each of the first images; A correction for generating a plurality of correction images by correcting a color of a second image corresponding to each of the first images based on the corrected ICC profile generated by applying each of the tone reproduction curves and the color correction matrix to an ICC profile tag. An image generator; And a synthesized image generator configured to synthesize each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio to generate one synthesized image.

In order to achieve the above technical problem, a color correction method of an image including a plurality of light sources according to the present invention includes a case in which one light source of each of a plurality of different light sources is turned on and the plurality of light sources are When all the lights are turned on, a plurality of first images photographed by arranging color charts of colors of red, green, blue, and gray scales within a preset photographing target area and a plurality of second photographing only the photographing subject areas An image input step of receiving an image; Computing a color tone reproduction curve corresponding to each of the first images by comparing each color constituting a reference region, which is an area corresponding to the color chart extracted from the plurality of first images, with an actual color of the color chart. Curve calculation step; A matrix calculation step of calculating a color correction matrix corresponding to each of the first images, the color correction matrix for linearly converting each color constituting the reference region to each actual color constituting the color chart; A correction for generating a plurality of correction images by correcting a color of a second image corresponding to each of the first images based on the corrected ICC profile generated by applying each of the tone reproduction curves and the color correction matrix to an ICC profile tag. Image generation step; And generating a composite image by synthesizing each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio.

According to the apparatus and method for correcting a color of an image including a plurality of light sources according to the present invention, a color tone reproduction curve and a color correction matrix are calculated from an image photographed when a plurality of light sources are turned on one by one and an image photographed when all are turned on. By correcting the color of the image by modifying the ICC profile based on this, an image showing a color close to the actual color can be obtained regardless of the type of light source. In addition, by using the method of modifying the ICC profile, it can be easily applied to the image, and a complicated calculation process is unnecessary. Furthermore, the texture of the color may be further improved by synthesizing the corrected image by the synthesis ratio based on the color difference value.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the apparatus and method for color correction of an image including a plurality of light sources according to the present invention.

2 is a block diagram showing the configuration of a preferred embodiment of the apparatus and method for color correction of an image including a plurality of light sources according to the present invention.

Referring to FIG. 2, the color correction apparatus according to the present invention includes an image input unit 110, a curve calculation unit 120, a matrix calculation unit 130, a correction image generation unit 140, and a composite image generation unit 150. Equipped.

The image input unit 110 includes a sample color including red, green, blue, and gray scale colors for each of a plurality of different types of light sources, each of which one light source is turned on and each of the plurality of light sources that are turned on. The plurality of first images photographed by arranging the chart in a preset photographing target region and a plurality of second images photographing only the photographing subject region are received.

In order to make the color of the object included in the image photographed under the plurality of light sources have an identity, information of the photographed image is required when only one light source is lit and all the light sources are lit. An image close to the color of the image photographed under the standard light source may be generated based on the collected information of each image.

The number of cases of the light source set for capturing the first image and the second image should include the case where a plurality of light sources are sequentially turned on one by one and the case where all of the plurality of light sources are turned on. The number of light sources is +1 '. In each case, the same image capturing region is photographed under a light source in which the lighting state is set to generate a first image and a second image. Here, the first image is an image photographed by positioning a color chart in a photographing target area, and is an image used to calculate parameters for color correction. Also, the second image is an image photographing only a photographing target region excluding a color chart, and corresponds to an image whose color is corrected by applying parameters calculated from the first image.

The color chart used for capturing the first image is composed of red, green, blue, and gray scale colors, and is used as a reference when correcting the color of the image. As the lighting state of the light source is set differently, the colors of the color charts included in each first image appear differently under the influence of the light source. Accordingly, when parameters of matching color charts that appear differently with colors of actual color charts are calculated and applied to the second image, the color of the object included in the second image is corrected to be close to the color of the actual object.

As a color chart, you can use the commonly used Qpcard to correct the color captured by the camera to be closer to the actual color. 3 is a diagram showing the color configuration of the Qpcard 201 used as a color chart. Among the plurality of colors constituting the Qpcard of FIG. 3, three colors 310 of red, green, and blue, and six colors 320 corresponding to gray scales are used to calculate a parameter for color correction.

Hereinafter, for convenience of description, a case in which the first image and the second image are taken when two light sources such as an incandescent lamp and a fluorescent lamp are used will be described as a representative embodiment. However, the type of light source is not limited thereto, and the present invention may be applied to any kind of light source.

4 is a diagram illustrating an example of first images photographed by different types of light sources for the same photographing target area. (A) of FIG. 4 is a first image photographed when only an incandescent lamp is lit, (b) is a first image photographed when only a fluorescent lamp is lit, and (c) is a first image photographed when both an incandescent lamp and a fluorescent lamp are lit. Display the image. In addition, in the first image shown in (a) to (c) of FIG. 4, the image in which the Qpcard is excluded becomes the second image.

As can be seen with reference to FIG. 4, when the types of light sources are different, the lighting states of the light sources are set to be different from each other, and the color chart is positioned on a path of light emitted from the light sources. You can do that. In fact, when shooting is performed under a plurality of light sources of different types, a plurality of light sources emit light in different directions, so that each part of the object is affected by different light sources.

That is, referring to FIG. 4A, since the incandescent lamp is located to the left of the photographing target area, the Qpcard is also disposed perpendicular to the path of the light irradiated by the incandescent lamp toward the point where the incandescent lamp is located. In addition, referring to FIG. 4B, since the fluorescent lamp is located on the right side of the photographing target area, the Qpcard is disposed to be perpendicular to the path of light irradiated by the fluorescent lamp toward the point where the fluorescent lamp is located. However, referring to FIG. 4C, since the incandescent lamp and the fluorescent lamp irradiate light from both the left and the right side of the shooting target area, the Qpcard is disposed in front of the camera shooting the shooting target area.

As described above, when photographing a plurality of first and second images with different lighting states, other matters except for the arrangement of the light source and the Qpcard, for example, the arrangement of the object in the shooting target area and the shooting conditions of the camera The settings and the like are all the same. However, if the exposure exposure value of the object varies according to the type of light source, and the range of colors represented in each first image is too different, colors outside the range of color gamut of the ICC profile are generated. Inconvenience to replace with other colors occurs. Therefore, even if the type of light source is changed, the exposure value of the camera is set so that the exposure exposure value of the object is always maintained, thereby capturing the first image and the second image. For example, an exposure value may be set to −0.3 when capturing an image as shown in FIG. 4A, and an exposure value may be set as −0.5 when capturing an image as shown in FIG. 4B.

The curve calculating unit 120 compares each color constituting the reference area, which is an area corresponding to the color chart extracted from the plurality of first images, with the color of the color chart, and generates a tone reproduction curve corresponding to each first image. Calculate. The reference area is an area in which only an area corresponding to the color chart is selected and extracted from the first image as shown in FIG. 4. The process of extracting the reference region from the first image may be performed by a user input.

The Tone Reproduction Curve (TRC) provides the information needed to convert between the CIE XYZ or Lab encoding of the image output device and the Profile Connection Space (PCS). The curve calculator 120 calculates a color tone reproduction curve corresponding to the red color channel by comparing a point corresponding to red among the colors constituting the reference region extracted from the first image and the actual red color of the color chart. Tonal reproduction curves corresponding to the green and blue color channels are calculated. In this case, the actual red, green, and blue colors of the color chart will use the red, green, and blue portions 310 shown on the color chart of FIG. 3. In addition, the six gray scale colors 320 displayed on the color chart of FIG. 3 are used to calculate a tone reproduction curve in contrast with the corresponding gray scale color in the reference region.

The relationship between the tone reproduction curves corresponding to the red, green, and blue color channels is a function of matching the red, green, and blue colors constituting the reference region extracted from each first image with the actual red, green, and blue colors of the color chart. It is expressed and estimated using Equation 1 below using curve regression.

Where g (x) is a function of the hue reproduction curve, and a _n (n = 1,2,3,4) is the least-squares dirrerence between each color in the reference area and the color value of the actual color of the color chart. ) Is a normalized color value, and f ₁ (x) to f ₄ (x) are defined as in Equation 2 below.

The color tone reproduction curves calculated for the red, green, and blue color channels by the processes described above are applied to the ICC profile to be used for color correction, which will be described in detail later.

Next, the matrix calculator 130 calculates a color correction matrix for linearly converting each color constituting the reference area into each actual color constituting the color chart, corresponding to each first image.

The linear transformation from each color constituting the reference area to the actual color of the color chart is defined as in Equation 3 below.

Here, M _XYZ is a matrix of actual color values of the color chart, M _Crr is a color correction matrix, and M ^' _XYZ is a matrix of color values constituting the reference area, respectively.

,

,

,

,

That is, X _i , Y _i and Z _i represent actual color values of the color chart, and X ^' _i , Y ^' _i and Z ^' _i represent color values constituting the reference area.

By simple relocation using Singular Value Decomposition (SVD), Equation 3 may be represented as Equation 4 to Equation 6 below.

The values of each element constituting the color correction matrix can be calculated from equations (4) to (6), and the completed color correction matrix is applied to the ICC profile and used for color correction like the tone reproduction curve.

The corrected image generating unit 140 corrects the colors of the second image corresponding to each of the first images based on the corrected ICC profile generated by applying each tone reproduction curve and the color correction matrix to the ICC profile tag. Create a correction image. Since the color tone reproduction curve and the color correction matrix are calculated for each first image, the corrected ICC profile is also generated in the same number as the first image.

Based on the reference region extracted from the first image, a color tone reproduction curve and a color correction matrix for obtaining the color of the reference region closer to the actual color of the color chart were obtained. In the second image, the color chart included in the first image is excluded, but the same photographed subject region as the first image is photographed under the same photographing conditions. Therefore, the color tone correction curve and the color correction matrix obtained from the reference region are applied to the second image. If you do, you can get colors close to the actual colors that make up the color chart.

To this end, the corrected image generating unit 140 applies a plurality of tonal reproduction curves and color correction matrices to the ICC profile representing the second image, exactly the ICC profile of the camera photographing the second image, and corrects the plurality of modified ICCs. Create a profile. The corrected ICC profile generates a corrected image in which color is corrected from the second image, and the color of each object included in the second image is corrected to be close to the actual color and included in the corrected image. Table 1 below shows some of the ICC profile tags included in the ICC profile.

Tag Name sign Data type mediaWhitePointTag wtpt XYZType mediaBlackPointTag bkpt XYZType CopyrightTag cprt multiLocalizedUnicodeTag profileDescriptionTag desc multiLocalizedUnicodeTag redMatrixColumnTag rXYZ XYZType greenMatrixColumnTag gXYZ XYZType blueMatrixColumnTag bXYZ XYZType redTRCTag rTRC curveType greenTRCTag gTRC curveType blueTRCTag bTRC curveType

ICC profile tags shown in Table 1 are ICC. It is described in detail in the specification at 1: 2004-10. Among them, the mediaWhitePointTag is set to white under a D65 light source, that is, a standard light source. mediaBlackPointTag is black.

The color correction matrix calculated by the matrix calculator 130 is applied to redMatrixColumnTag, greenMatrixColumnTag, and blueMatrixColumnTag among the ICC profile tags of Table 1 to generate a modified ICC profile tag. edMatrixColumnTag, greenMatrixColumnTag and blueMatrixColumnTag represent each column of the matrix used for TRC / matrix conversion. In the present invention, each element of the color correction matrix is applied to each color tag as shown in Equation 7 below.

The corrected image generator may generate a corrected ICC profile by applying a chromatic adaptation transformation defined by the ICC standard to the color correction matrix. Since the light source of the color device does not match the profile connection space (PCS), color adaptive transformation should be considered when applying the color correction matrix. ICC.1: 2004-10 proposes a linear Bradford model for color adaptive transformation. This is expressed as in Equation 8 below.

Therefore, a new color correction matrix after color adaptive conversion is obtained as shown in Equation 9 below.

The corrected image generator 140 may apply the color correction matrix of Equation 9 in which the color adaptive transformation is considered when generating the corrected ICC profile.

Meanwhile, as described above, the tone reproduction curves calculated for the red, green, and blue color channels are respectively applied to redTRCTag, greenTRCTag, and blueTRCTag among the ICC profile tags of Table 1. Three color tone reproduction curves are used to map input color pixels to output color pixels by means of a one-dimensional look-up table (LUT). The lookup table is expressed as in Equation 10 below.

Where i is a number assigned to the plurality of color pixels, g _R (x _i ) is the color tone reproduction curve calculated for the red color channel, g _G (x _i ) is the color tone reproduction curve calculated for the green color channel, and g _B (x _i ) is the hue reproduction curve calculated for the blue color channel.

Through the above process, the corrected image generation unit 140 may obtain a corrected ICC profile, and the corrected ICC profile is applied to the plurality of second images through the profile connection space PCS, and as a result, the corrected color is corrected. A plurality of corrected images are obtained.

The composite image generator 150 generates one composite image by synthesizing each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio.

The plurality of corrected images generated by the corrected image generating unit 140 are obtained by applying a corrected ICC profile to the plurality of second images, and the plurality of second images are light sources of one of a plurality of different light sources. In the case where the respective light sources and the plurality of light sources are all lit, the respective photographs are photographed. Therefore, the synthesized image obtained by synthesizing all the plurality of correction images is an image photographed under a plurality of different light sources, and becomes an image representing a color close to the actual color of the object.

The composite image generator 150 uses a color mixture model to generate one composite image from the plurality of correction images. The color mixture model is generated based on the color difference value between the color corresponding to white among the colors constituting each corrected image and the color corresponding to white among the actual colors constituting the color chart, and generates a composite image. The ratio is also set on this basis. To this end, the synthesized image generator 150 converts the plurality of corrected images into a lab color space. Therefore, the color difference value is obtained by calculating the distance between the point corresponding to white included in the corrected image and the point corresponding to white included in the color chart in the Lab color space.

FIG. 5 is a graph illustrating an example in which a plurality of colors constituting a corrected image generated from a second image corresponding to the first image illustrated in FIG. 4A and actual colors of a color chart correspond to each other in a lab color space. . Among the colors shown in FIG. 5, white is used by the synthesized image generator 150 to generate a color mixing model.

For convenience of description, the corrected images generated by the corrected image generating unit 140 are respectively converted into 'T images' and' from the second images corresponding to the three first images shown in FIGS. 4A to 4C. In the case of the F image 'and the' P image ', a color difference value between white included in each corrected image and actual white included in a color chart is calculated by the following equation (11).

Where dE _T , dE _F and dE _P are the color difference values between white included in the T image, F image, and P image, and white included in the color chart, respectively, and Wp _R is defined as white color included in the color chart in the Lab color space. The positions of the corresponding points, and Wp _T , Wp _F and Wp _P are the positions of the points corresponding to the white color included in the T image, the F image and the P image in the Lab color space, respectively.

The composition ratio is set based on the color difference value calculated by Equation 11, and among the pixels corresponding to the same position in each corrected image, a pixel having the smallest color difference value is included in the synthesized image at a larger ratio. This is to make the color of the object shown in the composite image appear closer to the actual color of the color chart. The composition ratio is set for each pixel constituting the corrected image, and the generation of the composite image is also performed in units of pixels. For example, when generating a synthesized image based on the synthesis ratio set from Equation 11, the pixel value of each pixel constituting the synthesized image is determined by the following equation (12).

은 합성영상을 구성하는 화소의 화소값이고, TCorr, FCorr 및 PCorr은 각각 T 영상, F 영상 및 P 영상에서 해당 화소와 동일한 위치에 해당하는 화소의 화소값이다.here,

Are pixel values of pixels constituting the composite image, and TCorr, FCorr, and PCorr are pixel values of pixels corresponding to the same positions in the T image, the F image, and the P image, respectively.

과 같은 관계가 성립하게 된다.In the ideal case, each pixel of the correction image is mixed in the same ratio.

This relationship is established.

FIG. 6 is a diagram illustrating an example in which pixel values at the same position of a synthesized image are determined by color difference values calculated for pixels corresponding to the same position in a T image, an F image, and a P image, and a synthesis ratio set based on the same. . Referring to FIG. 6, since the color difference value calculated for the pixel of the T image is the largest, the synthesis ratio is set to the smallest, and the color ratio value calculated for the pixel of the P image is the smallest, so that the composite ratio is set to the largest. You can check it.

Meanwhile, the generated synthetic image may generate an error such as block phenomenon or color mismatch. Block phenomena are common when using compression standards such as JPEG, and color mismatches are common at edge boundaries. Therefore, in order to reduce the error, the synthesized image generator 150 replaces the L channel of the synthesized image with the L channel of the second image photographed when all the light sources are turned on. This is because the L channel of the second image maintains the brightness ratio regardless of the change in color tone.

7 is a diagram illustrating a composite image in which an error and an L channel are replaced in the composite image. FIG. 7A illustrates a composite image generated from a T image, an F image, and a P image, and (b) is a second image corresponding to the P image. Referring to FIG. 7C, it may be confirmed that color mismatch occurs because the color of the object included in the composite image is not uniform. 7 (d) shows the composite image and the second image with respect to the L channel, the a channel, and the b channel. When the a channel, the b channel of the synthesized image, and the L channel of the second image are combined, FIG. ), You can obtain a composite image with the L channel replaced. The final resultant image shown in (e) of FIG. 7 shows a constant color close to the actual color, even though it is photographed under a plurality of light sources, as captured by a single light source.

8 is a flowchart illustrating a process of performing a preferred embodiment of the apparatus and method for color correction of an image including a plurality of light sources according to the present invention.

Referring to FIG. 8, the image input unit 110 may display red, green, blue, and gray scales in a case where one light source among each of a plurality of light sources of different types is turned on and each light source is turned on. A plurality of first images photographed by arranging a color chart composed of colors in a preset photographing subject area and a plurality of second images photographing only photographing subject regions are received (S810). In this case, the matters regarding the adjustment of the exposure value according to the arrangement direction of the color chart or the type of the light source in the photographing area are the same as those described above.

Next, the curve calculating unit 120 compares each color constituting the reference area, which is an area corresponding to the color chart extracted from the plurality of first images, to a color corresponding to each first image in comparison with the actual color of the color chart. A reproduction curve is calculated (S820). Tonal reproduction curves are calculated for the red, green and blue color channels, respectively. In addition, the matrix calculator 130 calculates a color correction matrix for linearly converting each color constituting the reference area into each actual color constituting the color chart in response to each first image (S830).

The corrected image generating unit 140 corrects the colors of the second image corresponding to each of the first images based on the corrected ICC profile generated by applying each tone reproduction curve and the color correction matrix to the ICC profile tag. A corrected image is generated (S840). Since there are a plurality of tonal reproduction curves and color correction matrices, a plurality of modified ICC profiles are also generated, which are applied to each second image. Finally, the synthesized image generating unit 150 generates one composite image by synthesizing each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio (S850). The composition ratio is set based on the color difference value, and the color difference value is calculated by the distance between the point corresponding to the white of the corrected image and the point corresponding to the white of the color chart in the lab color space.

Experiments were conducted to confirm the performance of the present invention. Experiments were performed on images taken under incandescent lamps emitting light in one direction and fluorescent lamps emitting light in the other direction, including natural light. A color scatter plot represented on the ab color space was used for the qualitative evaluation of the experimental results.

9 is a view showing images obtained as a result of the experiment using incandescent and fluorescent lamps. (A) of FIG. 9 is a second image photographed when only an incandescent lamp is lit, (b) is a second image photographed when only a fluorescent lamp is lit, and (c) is a second image photographed when both an incandescent lamp and a fluorescent lamp are lit. It is shown. (D) to (f) of FIG. 9 show color mixing models obtained from the second images of FIGS. 9 (a) to (c), respectively, by applying a composition ratio to each pixel constituting the second image. will be. The image of (d) shows the largest error on the left side, and the image of (e) shows the largest error on the right side, which corresponds to the direction of the light source irradiating light when each image is taken. Therefore, it can be seen that the error caused by the light source disappears due to the color correction and the synthesis of the corrected image. 9 (g) shows a synthesized image, and (h) shows a synthesized image in which the L channel is replaced. As the L channel is replaced, errors such as color mismatch are eliminated.

FIG. 10 is a diagram illustrating color scatter plots of a second image and a corrected image obtained when all of a plurality of light sources are turned on. FIG. 10A illustrates a P image, which is a second image photographed when both an incandescent lamp and a fluorescent lamp are turned on, and shows a low color texture of the image because the scatter plot is not uniformly distributed in the ab color space. On the other hand, (b) of FIG. 10 shows that the color texture of the corrected image after color correction is performed on the P image, and the color texture of the image is increased due to the uniform color distribution.

FIG. 11 is a diagram for visually comparing before and after color correction with respect to a portion where color gradients or changes appear in an image. In FIGS. 11A to 11C, an upper image is a P image photographed when both an incandescent lamp and a fluorescent lamp are turned on among second images that are not color corrected, and a lower image is a modified ICC profile in the second image. This is a correction image generated by applying. It can be seen that in FIG. 11 (a) to (c), the color scatter diagram shows more even distribution when the color correction is performed.

12 is a diagram illustrating the results of testing the performance of the present invention on various test images. 12 (a) is an image taken indoors, (b) is an image taken outdoors, (c) is a synthetic image (synthetic image). All three cases are based on images taken under incandescent and fluorescent lamps by Canon's D-10 model camera. 12 (a) and 12 (c), it can be seen that the present invention exhibits excellent performance in color correction even with respect to an image photographed when natural light is present. 12 (c) also shows an improved color texture as in the previous experimental results.

The performance of the present invention was tested on images taken under other light sources besides incandescent and fluorescent lamps used in the previous experiment. The performance of the present invention is also compared with the existing color correction techniques PROFILE MAKER ^™ and I-1 in the form of root-mean-square error. The results are shown in Table 2 below.

Light source PROFILE MAKER ^TM I-1 Invention A_CWF 119.4761 67.7193 53.0244 CWF-HORIZON 127.2063 81.3686 53.2583 CWF-U30 121.8296 69.035 59.539 D65-A 106.7238 39.8982 35.9207 D65-CWF 103.0539 35.5892 31.9188 D65-U30 108.719 45.3516 42.1452 HORIZON-A 122.0171 70.1249 41.2256 HORIZON-U30 154.8729 123.3306 52.9711 U30-A 130.1626 83.9268 53.8381

As shown in Table 2, the experimental environments were HORIZON (1750 lux, 6319 K), CWF (1606 lux, 3970 K), A (1969 lux, 2863 K), U30 (1675 lux, 2843 K) and D-65 (1660 lux, 6319 K). It was set to a combination of various light sources, including. Referring to Table 2, it can be seen that the present invention shows excellent performance compared to the existing techniques for all kinds of light sources.

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and may be implemented in the form of a carrier wave (for example, transmission via the Internet) . The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

Although the preferred embodiments of the present invention have been shown and described above, the present invention is not limited to the specific preferred embodiments described above, and the present invention belongs to the present invention without departing from the gist of the present invention as claimed in the claims. Various modifications can be made by those skilled in the art, and such changes are within the scope of the claims.

1 is a diagram illustrating a configuration of a color management system;

2 is a block diagram showing the configuration of a preferred embodiment of the apparatus and method for color correction of an image including a plurality of light sources according to the present invention;

3 is a view showing a color configuration of the Qpcard 201 used as a color chart,

4 is a diagram illustrating an example of first images photographed by different types of light sources for the same photographing target area;

FIG. 5 is a graph illustrating an example in which a plurality of colors constituting a corrected image generated from a second image corresponding to the first image illustrated in FIG. 4A and actual colors of a color chart are mapped in a Lab color space;

FIG. 6 is a diagram illustrating an example in which pixel values at the same position of a synthesized image are determined by color difference values calculated for pixels corresponding to the same position in a T image, an F image, and a P image, and a synthesis ratio set based on the same;

7 is a view illustrating a composite image in which an error and an L channel are replaced in the composite image;

8 is a flowchart illustrating a process of performing a preferred embodiment of the apparatus and method for color correction of an image including a plurality of light sources according to the present invention;

9 is a view showing images obtained as a result of experiments using incandescent and fluorescent lamps,

10 is a view showing a color scatter diagram of a second image and a corrected image obtained when all of a plurality of light sources are turned on;

FIG. 11 is a diagram for visually comparing before and after color correction with respect to a portion where color gradients or changes appear in an image; and

12 is a diagram illustrating the results of testing the performance of the present invention on various test images.

Claims (15)

When a light source of each of a plurality of light sources of different types is turned on and each of the plurality of light sources is turned on, a color chart composed of colors of red, green, blue, and gray scale is set for each target. An image input unit configured to receive a plurality of first images photographed by being placed in an area and a plurality of second images captured only by the photographing target area; Computing a color tone reproduction curve corresponding to each of the first images by comparing each color constituting a reference region, which is an area corresponding to the color chart extracted from the plurality of first images, with an actual color of the color chart. Curve calculator; A matrix calculator for calculating a color correction matrix for linearly converting each color constituting the reference area into each actual color constituting the color chart corresponding to each of the first images; A correction for generating a plurality of correction images by correcting a color of a second image corresponding to each of the first images based on the corrected ICC profile generated by applying each of the tone reproduction curves and the color correction matrix to an ICC profile tag. An image generator; And And a synthesized image generation unit configured to synthesize each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio to generate one synthesized image. The method of claim 1, Directions of light irradiated from the plurality of light sources are different from each other, and when the plurality of light sources are turned on one by one, the color chart is positioned on a path of light irradiated from the light source within the photographing target area. Color correction device. 3. The method of claim 2, The plurality of first and second images may be color corrected according to an exposure value set to maintain a constant reflection exposure value of an object included in the photographing target area when the plurality of light sources are turned on one by one. Device. The method according to any one of claims 1 to 3, The curve calculating unit reproduces a color tone corresponding to a red, green, and blue color channel by comparing red, green, and blue, respectively, among the plurality of colors constituting the reference area with red, green, and blue of actual colors of the color chart. Color correction apparatus, characterized in that for calculating. The method according to any one of claims 1 to 3, And the corrected image generating unit generates the corrected ICC profile by applying a color adaptive transformation model defined by an ICC standard to the color correction matrix. The method according to any one of claims 1 to 3, The composition ratio may be set based on a distance between a point corresponding to white among the colors constituting the respective corrected images in a lab color space and a point corresponding to white among the actual colors constituting the color chart. Color correction device. The method according to any one of claims 1 to 3, The synthesis image generating unit is a color correction device, characterized in that for replacing the L channel of the synthesized image generated in the lab color space with the L channel of the second image photographed when all of the plurality of light sources are turned on. When a light source of each of a plurality of light sources of different types is turned on and each of the plurality of light sources is turned on, a color chart composed of colors of red, green, blue, and gray scale is set for each target. An image input step of receiving a plurality of first images photographed by being placed in an area and a plurality of second images photographing only the photographing target area; Computing a color tone reproduction curve corresponding to each of the first images by comparing each color constituting a reference region, which is an area corresponding to the color chart extracted from the plurality of first images, with an actual color of the color chart. Curve calculation step; A matrix calculation step of calculating a color correction matrix corresponding to each of the first images, the color correction matrix for linearly converting each color constituting the reference region to each actual color constituting the color chart; A correction for generating a plurality of correction images by correcting a color of a second image corresponding to each of the first images based on the corrected ICC profile generated by applying each of the tone reproduction curves and the color correction matrix to an ICC profile tag. Image generation step; And And generating a composite image by synthesizing each pixel corresponding to the same position in the plurality of corrected images according to a preset synthesis ratio. The method of claim 8, Directions of light irradiated from the plurality of light sources are different from each other, and when the plurality of light sources are turned on one by one, the color chart is positioned on a path of light irradiated from the light source within the photographing target area. How to color correct. The method of claim 9, The plurality of first and second images may be color corrected according to an exposure value set to maintain a constant reflection exposure value of an object included in the photographing target area when the plurality of light sources are turned on one by one. Way. The method according to any one of claims 8 to 10, In the step of calculating the curve, a color tone corresponding to the red, green, and blue color channels by comparing red, green, and blue, among red, green, and blue colors among the plurality of colors constituting the reference area, respectively, among the actual colors of the color chart. A color correction method comprising calculating a reproduction curve. The method according to any one of claims 8 to 10, And generating a corrected ICC profile by applying a color adaptive transformation model defined by an ICC standard to the color correction matrix. The method according to any one of claims 8 to 10, The composition ratio may be set based on a distance between a point corresponding to white among the colors constituting the respective corrected images in a lab color space and a point corresponding to white among the actual colors constituting the color chart. How to color correct. The method according to any one of claims 8 to 10, In the synthesis image generation step, the color channel correction method, characterized in that for replacing the L channel of the synthesized image generated in the Lab color space with the L channel of the second image photographed when the plurality of light sources are all turned on. A computer-readable recording medium having recorded thereon a program for executing the color correction method according to any one of claims 8 to 10.

Images