KR20150087688A - System and method for adjusting auto white balance(AWB) - Google Patents

Abstract

According to an embodiment of the present invention, a white balance adjusting system includes: a first achromatic color determining unit which receives RGB information to determine whether a color is achromatic or not; a first correction unit which corrects data determined as an achromatic color in the first achromatic color determining unit; a color correction unit which color-corrects the data; a second achromatic color determining unit which determines whether the colors of the data determined as an achromatic color by the first achromatic color determining unit and the color-corrected data are achromatic or not; and a second correction unit which corrects the data determined to have the achromatic color by the second achromatic color determining unit.

Description

[0001] The present invention relates to an automatic white balance adjustment system and an automatic white balance adjustment method,

The present invention relates to an automatic white balance adjustment system and an automatic white balance adjustment method, and more particularly, to a system and method for more accurate automatic white balance adjustment.

In general, the white balance adjustment is performed in the photographing apparatus to prevent an achromatic subject from being photographed due to the influence of the light source. The white balance adjustment detects the white color having the largest change in the object color by the light source from the input image data, determines the color temperature by the R, G, and B ratio of the detected white color, Refers to a color correction function that shifts the overall color by correcting R (Red) and B (Blue) according to the temperature to adjust the color balance.

When the same object is photographed using a video device such as a digital still camera or a digital video camera, the apparent color of the photographed color varies depending on conditions of various light sources such as a white lamp, a fluorescent lamp, and a sunlight in the room. The human eye has a color constancy that can adapt to each light source and make white feel white. However, when the color temperature of the light source is high, the white light including the blue color is reproduced by reflecting the R, G, and B components included in the light source having different color temperature, and the red It reproduces white containing color.

Therefore, it is necessary to adjust the white balance so that white appears as white when the color temperature is changed by the light source. For example, in the case of white with blue color, the gain of R (red) is increased and the gain of B (blue) is adjusted to be decreased. In the case of white containing red color, It is necessary to adjust to reduce the gain of the antenna.

A typical Auto White Balance algorithm is a method of accumulating R, G, and B values or U and V values of an achromatic sample in an image and accumulating the ratio of the R integrated value and the G integrated value to the B integrated value Adjust the AWB gain so that the ratio of the integrated value of G and G becomes 1, or adjust the AWB gain so that the U and V integrated value becomes 0. In order to select an achromatic sample in the image, AWB sample window is set in R / G, B / G or U, V and only sample values included in the set AWB sample window are accumulated.

Typically, the AWB sample window is set to include all of the color temperature ranges for which AWB adjustment should be made. However, when an object having a different color of light is included in the image in this setting, AWB adjustment is not properly performed by being included in the AWB sample window even if it is not an achromatic object.

The embodiment of the present invention is characterized in that white balance adjustment can be accurately performed even if an object having a different color of light is included in an AWB sample window in order to solve the above-described conventional problems.

According to an aspect of the present invention, there is provided an automatic white balance adjustment system including a first achromatic color determination unit for receiving R, G, and B data and determining whether the color is achromatic; A color correction unit for performing color correction on the corrected data; and a color correction unit for correcting the data determined to be achromatic in the first achromatic color determination unit on the basis of a predetermined threshold value, A second achromatic color determining section for determining whether or not the data is achromatic; and a second correcting section for correcting data determined to be achromatic in the second achromatic color determining section.

The color determination threshold value adjuster may further include a color determination threshold value adjuster configured to increase the color determination threshold value if the number of samples that pass the second achromatic color determination is less than a preset threshold value When the number of samples passing through the second achromatic color determination is greater than a preset threshold value, the color determination threshold value is adjusted to be small.

Further, the first correcting unit and the second correcting unit may include an integrating unit for integrating the R, G, and B data and the achromatic color determination result generated by the first achromatic color determining unit, An AWB gain calculation unit for calculating an AWB gain by the B integrated value, and an AWB correction unit for performing AWB correction using the AWB gain calculated value.

The first achromatic color determining unit applies an achromatic color sample window. If the R, G, B data values are located in the sample window, the first achromatic color determining unit determines achromatic color. If the R, G, B data values are located outside the sample window, .

According to another aspect of the present invention, there is provided an automatic white balance adjustment method comprising the steps of: receiving R, G and B data and determining whether the color is achromatic; performing a primary AWB correction on data determined to be achromatic; Determining whether the achromatic color of the data and the color corrected data are achromatic or not based on the predetermined color determining threshold value; And performing a secondary AWB correction.

In addition, the step of performing the secondary AWB correction may include accumulating data determined to be a second achromatic color, calculating a secondary AWB gain through the accumulated data, Further comprising a step of performing AWB correction.

In the present invention, after performing the primary AWB correction, the chromatic color sample is further emphasized through the color correction, and then the secondary AWB correction is performed, so that more accurate samples can be selected and AWB correction can be performed. Thus, it is possible to obtain an effect of performing accurate and precise white balance adjustment.

It will be apparent to those skilled in the art that various modifications, additions, and substitutions are possible, and that various modifications, additions and substitutions are possible, within the spirit and scope of the appended claims. As shown in Fig.

1 is a system diagram showing a white balance adjustment system according to an embodiment of the present invention.
2 is a flowchart illustrating a white balance adjustment method according to an embodiment of the present invention.
3A and 3B are graphs showing an achromatic sample distribution according to an embodiment of the present invention.
4A and 4B are graphs showing an achromatic sample window according to an embodiment of the present invention.
5 is a graph showing an achromatic color determination criterion according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a system diagram showing a white balance adjustment system according to an embodiment of the present invention.

1, an automatic white balance (AWB) adjustment system includes a first achromatic color determining unit 100, a first correcting unit 103, a color correcting unit 120, A determination unit 125, a color determination threshold value adjuster 130, and a second correction unit 133. [ The first correction unit 103 includes a first integration unit 105, a first AWB gain calculation unit 110, and a first AWB correction unit 115. The second correction unit 133 includes a second integration unit 135, a second AWB gain calculation unit 140, and a second AWB correction unit 145.

First, the first achromatic color determining unit 100 receives red (R), green (G), and blue (B) data to determine whether or not achromatic color is present, and outputs the achromatic color determination result to the first integrating unit 105 and the second achromatic color And transmits it to the determination section 125. The first achromatic color determining unit 100 determines a achromatic color when the sample window is set in the R / G, B / G domain, or the U and V domains, and when the value is set in the set sample window, do.

The first integrating unit 105 receives the R, G, and B data before AWB correction and the achromatic color determination result generated by the first achromatic color determining unit 100. If it is determined that the achromatic color determination result is achromatic, the first integrating unit 105 integrates the input R, G, and B data. At this time, the integration cycle is usually set to be one frame interval, but the user may select the integration cycle according to the register control variable. For example, in the case of a low pixel imaging system, accumulating achromatic samples over several frames secures the number of samples suitable for AWB adjustment. On the other hand, in the case of the high-pixel imaging system, the number of samples suitable for AWB adjustment can be secured even if only one frame is accumulated. Therefore, the integration period is controlled by the user and selected.

The first AWB gain calculator 110 calculates the AWB gain using the R integrated value, the G integrated value, and the B integrated value integrated in the first integrating unit 105. The AWB gain is calculated by R gain (G total / R total) and B gain (G total / B total) and means the value to be corrected.

 The first AWB correction unit 115 receives the R, G, and B data and the AWB gain values calculated by the first AWB gain calculation unit 110, and performs the first AWB correction. The R gain is multiplied by the R gain calculated by the primary AWB gain calculation unit 115, and the AWG correction is performed by multiplying the B gain by the B gain.

The color correction unit 120 receives the primary AWB-corrected data and performs color correction. In general, an image captured by a photographing device such as an image sensor and not subjected to color correction has a value different from that of a human eye. When the color correction is performed, the color of the image can be realized with a value similar to that of the human eye. In other words, a more accurate achromatic color determination can be performed when achromatic color determination is performed on the data subjected to the color correction.

The second achromatic color determining unit 125 determines whether the primary AWB corrected data and the color-corrected data are secondarily achromatic. The second achromatic color determination is performed only in the sample that has passed the first achromatic color determination, and the color determination threshold value adjuster 130 receives the color determination threshold value and proceeds.

The color determination threshold value adjuster 130 is for performing AWB correction by selecting only the most achromatic sample in the image. When the number of samples that pass the second achromatic color determination during one frame is counted, 2 color determination threshold value gth used in the achromatic color determination section 125 is made smaller and the threshold value is made larger when the color determination threshold value gth is smaller than the appropriate sample number range. At this time, by setting the maximum value at which the threshold value can be increased to the register control variable, the threshold value becomes too large to prevent the chromatic color sample from being used for the AWB correction. That is, the color determination threshold value adjuster 130 may perform AWB correction only on samples close to achromatic color so that accurate AWB correction can be performed. The appropriate sample number range used for the color judgment threshold value is set as a register control variable so as to be able to correspond to various pixel counts.

The second integrating unit 135 receives the R, G, and B data before AWB correction, and integrates the R, G, and B data of the achromatic color in the second achromatic color determining unit 125, respectively. At this time, the accumulation period may be set as a register control variable such that the user can set an appropriate value as in the first accumulation unit 105.

The second AWB gain calculator 140 calculates the second AWB gain based on the R, G, and B accumulated values accumulated in the second accumulator 105. The first AWB gain calculator 110 outputs the R gain and the B gain as in the first AWB gain calculator 110, and the calculation method of the R gain and the B gain is also the same as the first AWB gain calculator 140. However, it is preferable that the R and B gain values output from the second AWB gain calculation unit 140 are not outputted at one time but are divided into several frames to reach the final AWB gain value. This is to prevent AWB gain value from being changed suddenly during AWB correction to prevent the image from jarring and perform AWB correction more smoothly. At this time, the speed at which the AWB gain value reaches the final gain value is set as a register control variable, so that the user can set it.

The second AWB correction unit 145 receives the R, G, and B data before the AWB correction and the AWB gain calculated by the second AWB gain calculation unit 140, and performs the secondary AWB correction. The secondary AWB correction is performed by multiplying the R value calculated by the second AWB calculation section 140 by the R value and multiplying the B gain by the B gain to perform the AWB correction. The data with the secondary AWB correction is the final output of this AWB system.

2 is a flowchart illustrating a white balance adjustment method according to an embodiment of the present invention.

Referring to FIG. 2, R, G, and B data are received before auto white balance (AWB) correction is performed to determine whether the image is achromatic (step S100). The step of determining whether or not the first achromatic color is to be performed may proceed through the achromatic color sample window of the first achromatic color determining unit 100. [ G / B, G / R, U, and V values are calculated using the input R, G, and B data values that have not undergone AWB correction, are determined to be achromatic when they are located in the sample window, and chromatic colors are determined when they are located outside the sample window .

3A and 3B, an achromatic sample distribution according to a color temperature of a light source in color coordinates will be described below. Here, the x axis represents the relative color coordinate (G / B) of G (Green) with respect to B (Blue) and the y axis represents the relative color coordinate (G / R) of G (Green) with respect to R (Red).

Generally, achromatic samples with low color temperature in G / R and G / B coordinates are distributed in a position with a low G / R value and a high G / B value (upper left). As the color temperature increases, the sample is distributed in the direction of the arrow in FIG. 3A (lower right direction). Also, in U and V coordinates, achromatic samples of low color temperature are generally distributed at a position with a low U value and a high V value (upper left). As the color temperature increases, the sample is distributed in the direction of the arrow in FIG. 3B (lower right).

4A and 4B, a sample window of the first achromatic color determining unit 100 is a G / R, a G / B domain or U, V domain. For example, the sample window of the first achromatic color determining section 100 is set to include all the light sources for which the photographing apparatus should perform AWB correction. As shown in FIGS. 4A and 4B, the sample window is set so that the achromatic points of the A, B, and C light sources are all included when the light source to which the photographing apparatus should perform correction is the A, B, and C light sources. The sample window of the first achromatic color determining unit 100 may be set as a register variable so that the user can set it according to the situation.

In addition, the sample window of the first achromatic color determining unit 100 is not limited to the rectangular shape as shown in Figs. 4A and 4B, but may take other forms.

Next, the data determined to be the primary achromatic color are accumulated, and the primary AWB gain is calculated through the accumulated data (step S110). The data determined as the first achromatic color are accumulated through the first accumulator 105, and the first AWB gain is calculated in the first AWB gain calculator 110.

Next, the first AWB correction is performed with the calculated first AWB gain value (step S120). Thereafter, the data subjected to the primary AWB correction is subjected to color correction (step S130).

The color correction method is shown in Equation (1) below.

Here, cm00 to cm22 are set as register control variables so that the user can set an appropriate value. When color correction is performed, the color of the image can be realized with a value similar to that of the human eye. Therefore, when the achromatic color determination is performed with the data subjected to the color correction, more accurate achromatic color determination can be performed.

Then, the secondary achromatic color determination is performed on the data for which the color correction has been performed and the data that has passed the primary achromatic color determination (step S140). The second achromatic color determination is performed only in the sample that has passed the first achromatic color determination, and the color determination threshold value is input from the color determination threshold value adjuster 130 to determine the achromatic color.

The second achromatic color determination method will be described with reference to FIG. The X-axis of the graph shown in FIG. 5 represents the color, and the Y-axis represents the threshold value. For example, if there are samples A, B, and C, the threshold is set to a small value for min (R, G, B) ) For larger samples, the threshold value should be set larger. Here, a sample having a small value of min (R, G, B) has a relatively dark color, and a sample having a large value of min (R, G, B) has a relatively bright color. That is, the color determination threshold value is set to be smaller as the color becomes darker.

If the difference between the maximum value and the minimum value of the R, G, B values of the input sample is below the threshold value, it is determined as achromatic. Here, since the difference between the maximum value and the minimum value of the sample A and the sample B is equal to or less than the threshold value, it is determined as achromatic. On the other hand, since the difference between the maximum value and the minimum value is equal to or larger than the threshold value, the sample C is determined to be a chromatic color.

Next, the number of samples that passed the second achromatic color determination is counted to adjust the color determination threshold value (step S145). If the number of samples that pass the second achromatic color determination exceeds the optimum sample number range, the color determination threshold value gth used by the second achromatic color determining unit 125 is adjusted to be small. If the number of samples passes the second achromatic color determination unit 125, do. At this time, the maximum value that can increase the threshold value is set as a register control variable so that the user can set it, thereby preventing the chromatic color sample from being used for AWB correction because the threshold value becomes too large. In addition, the user can set the appropriate number of samples in accordance with the number of pixels by setting the number of registers to be adjusted.

Next, the data determined to be achromatic in the second order are accumulated, and the secondary AWB gain is calculated through the R, G, and B integrated values (step S150).

The R gain and the B gain are outputted as in the first AWB gain calculator 110 and the calculation method of the R gain and the B gain is also the same as the first AWB gain calculator 110. However, it is preferable that the R and B gain values output from the second AWB gain calculation unit 140 are not outputted at one time but are divided into several frames to reach the final AWB gain value. This is to prevent AWB gain value from being changed suddenly during AWB correction to prevent the image from jarring and perform AWB correction more smoothly.

Next, the secondary AWB correction is performed with the secondary AWB gain value (step S160).

As described above, the present invention obtains the primary AWB gain with an achromatic sample by the primary AWB sample window, and performs AWB correction. Thereafter, the chromatic color sample is further emphasized through the color correction, and the final AWB gain value is obtained only through the appropriate sample close to the achromatic color, and the AWB correction is performed, so that more accurate samples can be selected and AWB correction can be performed. Therefore, there is an effect of solving the problem that AWB adjustment is not properly performed because samples of other light source colors are included in the image.

100: first achromatic color determining unit 103: first correcting unit
105: first integrating unit 110: first AWB gain calculating unit
115: first AWB correction unit 120: color correction unit
125: second achromatic color determination unit 130: color determination threshold value adjustment unit
133: second correction unit 135: second integration unit
140: second AWB gain calculation unit 145: second AWB correction unit
150: AWB corrected data

Claims (7)

A first achromatic color determining unit that receives R, G, and B data and determines whether the color is achromatic;
A first correcting unit for correcting data determined as achromatic in the first achromatic color determining unit;
A color correction unit for performing color correction on the corrected data;
A second achromatic color determining unit for determining whether achromatic color of the data determined as achromatic color and the color corrected data of the first achromatic color determining unit is based on a predetermined threshold value; And
A second correction unit for correcting data determined to be achromatic in the second achromatic color determining unit,
And an automatic white balance adjustment system. The method according to claim 1,
And a color judgment threshold value adjuster for generating the color judgment threshold value. The method of claim 2,
Wherein the color determination threshold value adjuster adjusts the color determination threshold value to a larger value if the number of samples passed through the second achromatic color determination is smaller than a predetermined threshold value and if the number of samples that pass the second achromatic color determination exceeds a predetermined threshold value, And the color judgment threshold value is adjusted to be small. The method according to claim 1,
The first correcting unit and the second correcting unit are
An integrating unit for integrating the R, G, and B data and an achromatic color determination result generated by the first achromatic color determining unit;
An AWB gain calculation unit for calculating an AWB gain by integrating the R, G, and B values integrated in the integration unit; And
An AWB correction unit for performing AWB correction using the AWB gain calculation value,
And an automatic white balance system. The method according to claim 1,
The primary achromatic color determining unit applies an achromatic color sample window. If the R, G, B data values are located in the sample window, the primary achromatic color determining unit determines achromatic color. If the R, G, B data values are located outside the sample window, Wherein the automatic white balance system comprises: Receiving R, G, and B data and determining whether the color is achromatic;
Performing primary AWB correction on the data determined to be achromatic;
Performing color correction on the corrected data;
Determining whether the data determined as achromatic color and the data subjected to color correction are achromatic based on a predetermined color determination threshold value; And
Performing a secondary AWB correction on the data determined to be achromatic
Wherein the automatic white balance method comprises: The method of claim 6,
The step of performing the secondary AWB correction
Accumulating data determined to be a second achromatic color;
Calculating a second AWB gain through the accumulated data; And
The AWB correction is performed with the calculated second-order AWB gain value
Further comprising the steps of:

Images