JP4461892B2 - Electronic camera having color cast adjustment function by special light source, and program - Google Patents

Description

The present invention relates to an electronic camera having a function of reducing the color cast of an image.
The present invention also relates to a program for reducing the color cast of an image on a computer.

Conventionally, when photographing is performed under fluorescent lamp illumination, it is known that a green fog is generated in a captured image due to the wavelength characteristics of the fluorescent lamp.
Patent Document 1 describes a technique for correcting such a green fog by white balance adjustment. Furthermore, Patent Document 1 also describes a technique for weakening the correction of the green fog when the color of the image is analyzed and determined to be green of the plant.
JP 2003-264850 A (summary)

  By the way, in night scene shooting and indoor shooting, a complicated lighting environment in which fluorescent lights, light bulbs, and flashlights are mixed is often obtained. If there is unevenness in how each light hits in such an environment, only a part of the subject will have a green fog on the fluorescent lamp, and no green fog will occur on the light bulb or flashing part. .

For example, consider a case where a slow-synchronized photo is taken with a person illuminated by a light bulb in front of a skyscraper group (a large number of fluorescent lamps can be seen from the window) and the electronic flash device emitting light. Try. In such a case, a green fog caused by a fluorescent lamp occurs in the light of a high-rise building group that occupies most of the background. On the other hand, the person in the foreground is strongly illuminated by light bulbs and flashes, so there is almost no green fog.
If the green fog correction is performed on the captured image in this state as usual, the red-blue component of the person who does not cause the green fog rises too much, resulting in an unnatural magenta fog. .

  Conventionally, it is very difficult to solve such a problem automatically by image processing in an electronic camera because it is not possible to properly determine whether the occurrence of green fog is the entire screen or a part of the screen. . This is because the conventional technology only detects color fog in a limited area such as a low saturation area in the screen, and cannot determine whether or not the detected green fog occurs over the entire screen. It becomes.

Therefore, an object of the present invention is to provide a technique for appropriately determining whether or not a color cast such as a green cast is widely generated on the entire screen.
Another object of the present invention is to provide a technique for appropriately adjusting the color balance of a captured image by utilizing the determination of the extent of color cast.
Another object of the present invention is to provide a technique for appropriately suppressing an unnatural magenta fog generated in a screen when correcting a green fog of a fluorescent lamp.

<Claim 1>
An electronic camera according to a first aspect includes an imaging unit, an imaging control unit, a special light source determination unit, a uniform illumination determination unit, and a color balance adjustment unit.
The imaging unit captures a subject image.
The imaging control unit controls the imaging unit to generate a test image that is captured without flashing and a main image that is captured by performing flash emission.
The special light source determination unit determines a deviation from the black body locus for the low-saturation region of the main image, and detects a green fog caused by fluorescent lamp illumination.
The uniform illumination determination unit determines the uniformity of the illumination state by obtaining a histogram distribution for each of the test image and the main image and detecting the correlation between the two histogram distributions.
When the degree of histogram correlation is high and green fog is detected in the low saturation area, the color balance adjustment unit determines that the fluorescent lamp illumination is uniform, and the red-blue component of the main image is replaced with the green component. Raise it relatively to correct the green fog.
On the other hand, if the degree of histogram correlation is low and a green fog is detected in the low saturation area, the color balance adjustment unit determines that the fluorescent lamp illumination is non-uniform and performs the above correction for the green fog. Weaken and carry out.

<Claim 2>
When the electronic camera according to claim 2 satisfies the conditions that “the degree of histogram correlation is low” and “green fog is detected” in the electronic camera according to claim 1, the color balance adjustment unit sets the white balance adjustment value and the color The balance adjustment value is weighted and combined, and the color balance of the main image is adjusted using the weighted combination value.
Note that the white balance adjustment value here is a white balance adjustment value determined on the basis of the correlated color temperature of the main image (color temperature obtained by mapping on the black body locus). On the other hand, the color balance adjustment value is a color balance adjustment value for correcting the green fog.

<Claim 3>
According to a third aspect of the present invention, in calculating the weighted composite value in the electronic camera of the second aspect, the weighted ratio of the color balance adjustment value is lowered as the degree of histogram correlation is lower.

<Claim 4>
According to a fourth aspect of the present invention, in the electronic camera according to any one of the first to third aspects, the color balance adjustment unit is configured to determine a correspondence between a predetermined color of the low saturation region and the color balance adjustment value. Prepare multiple “data” for each color rendering difference of fluorescent lamps. And a color balance adjustment part calculates | requires a color balance adjustment value using corresponding data with high color rendering property, so that the degree of histogram correlation is low. The color balance adjustment unit adjusts the color balance of the main image based on the color balance adjustment value.

<Claim 5>
The electronic camera according to a fifth aspect includes an imaging unit, a special light source determination unit, a uniform illumination determination unit, and a color balance adjustment unit.
The imaging unit captures a subject image and generates a captured image.
The special light source determination unit detects a color cast by the special light source by determining a deviation from the black body locus in the low saturation region of the captured image.
The uniform illumination determination unit determines the uniformity of the illumination state of the captured image.
When it is determined that the illumination state is uniform illumination and a color cast is detected in the low saturation region, the color balance adjustment unit corrects the color cast in a direction that reduces the deviation from the black body locus of the captured image. On the other hand, when it is determined that the illumination state is non-uniform illumination and the color cast is detected in the low saturation region, the color balance adjustment unit performs the above correction for the color cast with a weakening.

<Claim 6>
A program according to a sixth aspect is a program for causing a computer to function as the special light source determination unit, the uniform illumination determination unit, and the color balance adjustment unit according to any one of the first to fifth aspects.

(Claim 1)
The electronic camera according to claim 1 generates a test image without flash and a main image with flash. Next, the electronic camera obtains the histogram distributions of these two images, and correlates to determine how similar the histogram distributions of both images are.

At this time, it can be determined that the more the histogram distributions of the two images are different, the more uneven the flashing method and the reflectance of the subject, and the non-uniform illumination state of the main image.
On the other hand, when the shapes of the histogram distributions of both images are similar, it can be determined that there is no unevenness in the way the flash reaches and the reflectance of the subject, and the illumination state of the main image is uniform.

  Furthermore, in this electronic camera, a low saturation area is extracted from the main image, and a green fog of fluorescent lamp illumination is detected by determining a deviation of the low saturation area from the black body locus.

  Here, when a green fog is detected from a part (low saturation region) of the main image determined to be uniform illumination, it can be determined that the green fog of the fluorescent lamp has occurred over a wide range of the main image. . In such a case, the electronic camera corrects the green fog by raising the red-blue component of the main image relative to the green component. As a result, it is possible to reliably reduce the green fog generated over a wide range of the screen and to obtain a natural main image with less green fog as a whole.

  On the other hand, if a green fog is detected from a part (low saturation area) of the main image determined to be non-uniform illumination, it can be determined that the green fog is generated only in a part of the main image. . In such a case, the electronic camera performs the above-described correction for the green fog with weakening. As a result, it is possible to solve the problem that magenta fog occurs in an area other than the green fog.

(Claim 2)
The electronic camera according to claim 2 is a weighted combination of a white balance adjustment value determined on the basis of a correlated color temperature (color temperature obtained by mapping on a black body locus) and a color balance adjustment value for correcting green fog. By doing so, the correction of the green fog is weakened.

  Green fog caused by fluorescent lamps appears in a form that deviates from the black body locus in the green direction. Accordingly, mapping on the black body locus cancels the deviation in the green direction, and a correlated color temperature is obtained in which the influence of the green fog is eliminated. The white balance adjustment value obtained on the basis of the correlated color temperature is an adjustment value that gives an appropriate color reproduction to a portion where green fog is not generated (in the above example, the main subject in front).

  The white balance adjustment value is weighted and synthesized with the color balance adjustment value for correcting the green fog, thereby shifting the original value of the color balance adjustment value to weaken the correction of the green fog. As a result, it is possible to reduce magenta fog that has been generated as a result of correcting the green fog. Furthermore, by including the white balance adjustment value in the weighted composite value, it is possible to improve the color reproducibility of a portion where no green fog is generated.

(Claim 3)
In calculating the weighted composite value, the electronic camera of claim 3 decreases the weight ratio of the color balance adjustment value as the degree of histogram correlation is lower.

  In general, the lower the degree of histogram correlation, the narrower and more limited the occurrence of green fog. In this situation, by lowering the weight ratio of the color balance adjustment value, it is possible to further weaken the green fog correction and to actively prevent magenta fog generated over a wide range in the screen.

(Claim 4)
The electronic camera of claim 4 prepares a plurality of correspondence data of color balance adjustment values for each color rendering difference of the fluorescent lamp. Since a fluorescent lamp with low color rendering properties has a noticeable green fog, the correction of the green fog is enhanced by using corresponding data that matches the fluorescent lamp. On the other hand, since a fluorescent lamp with high color rendering has little green fog, correction of the green fog is weakened by using corresponding data matched to the fluorescent lamp.

  Therefore, the electronic camera obtains the color balance adjustment value using the correspondence data of the fluorescent light group having higher color rendering properties as the degree of histogram correlation is lower. This is equivalent to further weakening the correction of the green fog as the histogram correlation is lower, that is, the portion where the green fog is generated is narrower. As a result, it is possible to positively prevent magenta fog that occurs over a wide range in the screen.

(Claim 5)
The electronic camera according to claim 5 detects a color cast by a special light source by extracting a low saturation area from the captured image and determining a deviation of the low saturation area from the black body locus.
Furthermore, this electronic camera determines the uniformity of the illumination state of the captured image. Such a determination is preferably performed by, for example, a process of comparing the luminance distribution of multi-pattern photometry or the average luminance of a plurality of locations of a captured image.

  Here, when a color cast is detected from a part of the captured image (low saturation region) determined to be uniform illumination, it can be determined that a color cast is generated by a special light source over a wide range of the captured image. . In this case, the electronic camera corrects the color cast by adjusting the color balance in a direction that reduces the deviation from the black body locus of the captured image. As a result, it is possible to reliably reduce the color cast over a wide range of the screen.

  On the other hand, when a color cast is detected from a part of the captured image determined to be non-uniform illumination (low saturation area), a color cast by a special light source occurs only in a part of the captured image. It can be judged. In this case, the electronic camera performs the above-described correction for the color cast with weakening. As a result, the problem that a new color cast occurs due to the correction of the color cast can be solved.

(Claim 6)
By executing the program according to claim 6 on a computer, the same color balance adjustment as that of the electronic camera according to any one of claims 1 to 5 can be performed on the computer.

<< Description of Configuration of this Embodiment >>
FIG. 1 is a diagram for explaining a component arrangement of the electronic camera 11 in the present embodiment.
In FIG. 1, the electronic camera 11 is provided with an electronic flash device 12 and a photographing lens 13. A quick return type mirror 15 is provided on the image space side of the photographic lens 13. A diffusion plate 16 is disposed at the imaging position of the subject light beam reflected by the mirror 15. The user observes the subject image formed on the diffusion plate 16 via the finder optical system 17. The light beam from the diffusion plate 16 is guided to the imaging surface of the photometric image sensor 18 provided at the corner of the finder optical system 17 and re-images the subject image on the imaging surface. Further, behind the mirror 15, a shutter 19a, a recording image sensor 19 and the like are arranged.

FIG. 2 is a block diagram relating to image processing of the electronic camera 11.
In FIG. 2, the output of the photometric image sensor 18 is input to the image memory 22 via the A / D converter 20. The output of the recording image sensor 19 is input to the image memory 22 via the A / D converter 21. The image memory 22 is connected to the bus 23. Connected to the bus 23 are an image processing unit 24, a compression recording unit 25, an imaging control unit 27, a flash emission control unit 28, a microprocessor 29, and the like.
Among these, the compression recording unit 25 compresses and records image data on a removable memory card 26. The microprocessor 29 is given an operation input from a release button 29a or the like.

<< Correspondence with Invention >>
Hereinafter, the correspondence between the description items of the claims and the present embodiment will be described. Note that the correspondence relationship here illustrates one interpretation for reference, and does not limit the present invention.
The imaging unit described in claims corresponds to the photometric imaging element 18 and the recording imaging element 19.
The imaging control unit described in the claims corresponds to the imaging control unit 27.
The special light source discriminating unit described in the claims is configured to extract a low saturation area from the main image generated by the recording image pickup device 19 by the microprocessor 29 and determine the deviation of the low saturation area from the black body locus. This corresponds to a function for detecting a color cast of a special light source (for example, a green cast of a fluorescent lamp).
The uniform illumination determination unit described in the claims corresponds to “a function for determining the uniformity of illumination by obtaining the histograms from the image data before and after the flash and detecting the correlation between both histograms” by the microprocessor 29.
The color balance adjusting unit described in the claims is “a function for adjusting the color balance of the main image according to the result of the detection of the color cast in the uniformity of the illumination state and the low saturation region” by the microprocessor 29 and the image processing unit 24. Corresponding to

<< Explanation of operation of this embodiment >>
FIG. 3 is a flowchart for explaining the operation of the electronic camera 11.
Hereinafter, the operation of the electronic camera 11 will be described along the step numbers shown in FIG.

[Step S1] Before starting the release operation, the mirror 15 is in the lowered state. Therefore, the subject luminous flux that has passed through the photographing lens 13 forms a subject image on the imaging surface of the photometric imaging element 18 via the diffusion plate 16 and the finder optical system 17.
In this state, the microprocessor 29 uses the imaging control unit 27 to give a drive signal to the photometric image sensor 18. A test image captured without flashing is output from the photometric image sensor 18. The test image is digitized in units of pixels via the A / D converter 20 and then temporarily stored in the image memory 22.

[Step S2] The microprocessor 29 accesses the image memory 22 and obtains a luminance component for each small region (for example, each pixel) of the test image.
It should be noted that the luminance components are generated by weighted addition of each color component constituting the test image according to the component ratio of the luminance component (for example, R: G: B = 0.29: 0.587: 0.114). Is preferred. Further, a color component (G color component or the like) containing a large amount of luminance component in the test image may be regarded as a luminance component as it is.
The microprocessor 29 classifies the luminance component of the test image obtained in this way into a rough luminance range, and obtains the frequency for each luminance range. In this way, the histogram distribution A of the test image illustrated in FIG. 4 is obtained.
It should be noted that data processing such as noise removal and large area extraction may be performed by removing a predetermined frequency value or less from the histogram distribution A.

[Step S3] The microprocessor 29 determines whether or not the release button 29a is fully pressed.
If the release button 29a is fully pressed, the microprocessor 29 proceeds to step S4.
On the other hand, if the release button 29a is not fully pressed (half pressed or not pressed), the microprocessor 29 returns the operation to step S1.

  [Step S4] The microprocessor 29 performs known preliminary light emission as follows. First, the microprocessor 29 raises the mirror 15 and projects a subject image on the shutter curtain of the shutter 19a. In this state, the microprocessor 29 causes the electronic flash device 12 to perform minute preliminary light emission using the flash light emission control unit 28. The microprocessor 29 measures the brightness of the shutter curtain during preliminary light emission using a photometric element (not shown) provided at a position where the shutter 19a is viewed. The microprocessor 29 determines the target light amount for the main light emission in accordance with the brightness of the shutter curtain during the preliminary light emission. Information on the target light amount of the main light emission is transmitted to the electronic flash device 12 side.

[Step S5] Subsequently, the microprocessor 29 opens the front curtain of the shutter 19a, and starts projecting a subject image on the imaging surface of the recording image sensor 19.
When single-shot main light emission is performed, the microprocessor 29 causes the electronic flash device 12 to start main light emission when the shutter 19a is fully opened. The electronic flash device 12 monitors the light emission amount of the main light emission, and stops the main light emission when reaching the target light amount for which information has been transmitted in advance.
On the other hand, when the set shutter time has elapsed, the microprocessor 29 closes the rear curtain of the shutter 19a. When the shutter 19a is completely closed in this way, the microprocessor 29 lowers the mirror 15 downward.

  [Step S <b> 6] The microprocessor 29 supplies a drive signal to the recording image sensor 19 using the imaging controller 27. The recording image sensor 19 outputs a main image captured with a flash. The main image is digitized in units of pixels via the A / D converter 21 and then temporarily stored in the image memory 22.

  [Step S7] The microprocessor 29 accesses the image memory 22 and reads out a low saturation area (area where the saturation is equal to or less than a predetermined value) from the main image. The microprocessor 29 obtains the chromaticity (R / G, B / G) for each small region (for example, each pixel) for the RGB color components in the low saturation region, and obtains the distribution center of these chromaticities. Subsequently, the microprocessor 29 determines the difference between the distribution center and the black body locus on the chromaticity coordinates.

[Step S8] The microprocessor 29 determines whether or not the deviation obtained in step S7 exceeds a threshold value. This threshold value is a value for distinguishing whether or not the degree of divergence can be perceived as a color cast, and can be determined from an image subjective evaluation experiment or the like.
Here, when the deviation is equal to or less than the threshold value, it can be determined that the color cast cannot be detected. In this case, the microprocessor 29 shifts the operation to normal color balance adjustment.
On the other hand, when the deviation exceeds the threshold, it is determined that the light source is a special light source (for example, a fluorescent lamp or a sodium lamp) that shows a wavelength distribution away from the black body locus, and the process shifts to the color balance adjustment of step S9.

[Step S <b> 9] The microprocessor 29 sets a color component adjustment gain in a direction that shortens the deviation from the black body locus, thereby obtaining a color balance adjustment value that brings the low saturation region closer to an achromatic color.
For example, in the case of a green fog of a fluorescent lamp, an adjustment gain that relatively increases the red-blue component is set to obtain a color balance adjustment value in order to shorten the deviation from the black body locus toward the green.

[Step S10] The microprocessor 29 accesses the image memory 22, reads the main image, and creates a processed image suitable for comparison with the test image.
In creating the processed image, it is preferable to perform image processing that suppresses the difference between the imaging elements as described below.

(1) Processing to cut out a processed image from the main image in accordance with the imaging field angle of the test image.
(2) The main image is divided into pixel blocks in accordance with the positions of the small areas (see step S4) of the photometric image sensor 18, and luminance components (including pseudo luminance components such as G color) are divided in pixel blocks. The requested process.
(3) The maximum gradation value, average gradation value, and most frequent gradation value of the main image according to the maximum gradation value, average gradation value, most frequent gradation value, and minimum gradation value of the test image. , And the process of normalizing and matching the minimum gradation value.
(4) A process of converting the number of gradation steps and gradation curve of the main image in accordance with the gradation characteristics of the test image.

[Step S11] The microprocessor 29 performs histogram analysis on the luminance component of the processed image created from the main image in the same manner as the test image, and creates a histogram distribution B as illustrated in FIG.
It should be noted that data processing such as noise removal and large area extraction may be performed by removing a predetermined frequency value or less from the histogram distribution B.

ただし、ＮｉとＭｉは、同一の輝度範囲ｉにおけるヒストグラム分布Ａ，Ｂの度数値である。
また例えば、ヒストグラム分布Ａ，Ｂの分布形状をパターンマッチングして、両者間の類似度からヒストグラム相関を検出してもよい。
また例えば、ヒストグラム分布Ａ、Ｂのピーク数、ピークの間隔、複数のピークの大小関係、またはピーク形状などについて一致度を判定することにより、ヒストグラム相関を簡易に求めてもよい。
このヒストグラム相関の高低から、次のような判断が可能になる。 [Step S12] The microprocessor 29 compares the distribution shapes of the two histogram distributions A and B, and obtains a histogram correlation between them.
For example, the Euclidean distance D between the histogram distributions A and B may be calculated according to the following equation, and the histogram correlation between them may be obtained. In this case, it can be determined that the smaller the Euclidean distance D is, the higher the degree of histogram correlation is.

However, Ni and Mi are frequency values of the histogram distributions A and B in the same luminance range i.
Further, for example, the distribution shapes of the histogram distributions A and B may be pattern-matched, and the histogram correlation may be detected from the similarity between the two.
Further, for example, the histogram correlation may be easily obtained by determining the degree of coincidence with respect to the number of peaks of the histogram distributions A and B, the interval between peaks, the magnitude relationship of a plurality of peaks, or the peak shape.
From the level of the histogram correlation, the following determination is possible.

(When histogram correlation is high)
First, when the histogram correlation is high, either the subject arrangement where the flash emission reaches uniformly or the subject arrangement where the flash emission does not reach uniformly is far away. With these subject arrangements, it can be determined that there is little unevenness in the illumination light and the illumination uniformity is high.
On the other hand, in step S8, the color cast has already been detected in the low saturation area in the screen. From these facts, it can be determined that the color cast by the special light source is also highly uniform and occurs over a wide range of the screen.

(When histogram correlation is low)
On the contrary, when the histogram correlation is low, it is determined that the subject arrangement is such that unevenness is caused in the way the flash emission reaches, and it can be determined that the luminance distribution has greatly changed before and after the flash emission. For example, there is a situation in which a near subject that is strongly illuminated by flash emission and a distant subject that is difficult to reach flash emission are mixed in one screen. In such a subject arrangement, it can be determined that the way the specific illumination light (sun, street lamp, flash emission, fluorescent lamp, etc.) reaches does not become uniform over the entire screen, and the illumination uniformity is low.
On the other hand, in step S8, the color cast has already been detected in the low saturation area in the screen. From these facts, it can be determined that the color cast by the special light source is uneven and occurs only in a part of the screen.

[Step S13] The microprocessor 29 changes the color balance adjustment value in such a direction that the correction of the color cast is weakened as the degree of the obtained histogram correlation is lower. For example, in the case of green fog correction, if the degree of histogram correlation is low, the green fog correction is weakened by slightly reducing the adjustment gain of the red-blue component.
Hereinafter, a measure for weakening the color cast correction will be described in detail.

(Color fog correction adjustment method 1)
The microprocessor 29 obtains the correlated color temperature by mapping the distribution center of the chromaticity in the low saturation area obtained in step S7 on the black body locus. This correlated color temperature is a color temperature on a black body locus, and is a color temperature ignoring a color cast in the direction of deviation. The microprocessor 29 obtains a white balance adjustment value for converting the correlated color temperature into the standard color temperature. The microprocessor 29 weights and synthesizes the white balance adjustment value and the previously obtained color balance adjustment value (see step S9). By this weighted synthesis, the color cast correction of the color balance adjustment value can be weakened. As a result, it is possible to suppress a new color cast (such as magenta fog) due to the color cast correction. Further, a white balance adjustment value is added to the weighted composite value. Therefore, the color temperature of an area where no color cast is generated (such as the subject in front of the flash illumination) can be brought close to the standard color temperature, and the color reproducibility of the main image can be improved.

(Color fog correction adjustment method 2)
The microprocessor 29 weights and synthesizes the “white balance adjustment value based on the correlated color temperature” and the “color balance adjustment value for correcting the color cast” in the same manner as the “adjustment method 1” described above. The microprocessor 29 changes the ratio of the weighted synthesis according to the histogram correlation obtained in step S12. That is, the weight ratio of the color balance adjustment value is lowered as the degree of histogram correlation is lower. As a result, when the degree of histogram correlation is low and the color cast in the screen is non-uniform, that is, in a local situation, the color cast correction is weakened, so that new color cast generation due to the color cast correction can be strongly suppressed. Conversely, the higher the degree of histogram correlation, the higher the weight ratio of the color balance adjustment value. As a result, when the degree of histogram correlation is high and the color cast in the screen is uniform, that is, in a wide range, it is possible to appropriately remove the color cast that occurs in the wide range of the screen by strengthening the color cast correction. It becomes possible.

(Color fog correction adjustment method 3)
For example, fluorescent lamps are divided into a plurality of groups from those having high color rendering properties to those having low color rendering properties.
The microprocessor 29 stores “correspondence data between the distribution center of the chromaticity obtained in step S7 and the color balance adjustment value for correcting the green fog” for each color rendering group. For example, the correspondence data can be determined by repeating a subjective evaluation experiment of color balance adjustment of the main image for each color rendering group. In this case, the green fog is removed more strongly as the correspondence data of the group having lower color rendering properties is used. On the other hand, the more the corresponding data of the group having higher color rendering properties is used, the more the green fog is removed. Therefore, as the degree of histogram correlation is lower, the microprocessor 29 uses the corresponding data having higher color rendering properties to determine the color balance adjustment value. As a result, when the degree of histogram correlation is low and the color cast in the screen is non-uniform, that is, in a local situation, the color cast correction can be weakened and the occurrence of a new color cast due to the color cast correction can be strongly suppressed.

  [Step S14] The microprocessor 29 adjusts the color balance of the main image in accordance with the adjustment value of the color fog correction adjusted in step S13.

  [Step S15] The main image for which the color balance adjustment in step S14 has been completed is compressed by the compression recording unit 25 and recorded and stored in the memory card 26.

<< Effects of this embodiment >>
As described above, in the present embodiment, by combining the uniformity determination of the illumination state and the color cast detection in the low saturation region, the color cast is generated over a wide range of the screen, or the color cast is generated. It can be discriminated whether it occurs only in a narrow area of the screen.

  Furthermore, in this embodiment, since the intensity of color fog correction is changed according to the extent of color fog determined in this way, it is possible to appropriately prevent problems such as the occurrence of a new color fog after the correction process.

  In the present embodiment, histogram distributions are taken for the test image without flash and the main image with flash. Even if the pattern changes to some extent due to the time difference between the test image and the main image, the histogram distribution can accurately determine the uniformity of the illumination state with almost no influence.

  Further, in the present embodiment, a test image is captured using the photometric imaging element 18 for use in split photometry. Since the photometric image sensor 18 has a smaller number of pixels than the recording image sensor 19, it takes less time to capture a test image and to perform image processing (histogram creation or the like). As a result, even if the above-described operation for determining the color balance adjustment value is taken into account, the increase in processing load is small, and the lightness of the imaging operation of the electronic camera 11 is not lost.

  In particular, in this embodiment, in step S10, the test image and the main image are processed so as to be suitable for comparison. As a result, despite the use of different image sensors 18 and 19, good comparison is possible, and the uniformity of the illumination state can be determined more accurately.

  Further, in the present embodiment, the histogram distributions A and B are compared with each other after the predetermined frequency values are rounded down from the histogram distributions A and B. Such truncation processing makes it possible to reduce the influence of noise, and to determine the uniformity of the illumination state more accurately. It is also possible to cut off a small degree value or less from the histogram distribution so that only a particularly high frequency value in the screen remains. In this case, it is possible to determine the uniformity of the illumination state only in the high-frequency part that stands out in the screen. As a result, it is possible to achieve color balance adjustment focused on a visually noticeable high-frequency portion.

<< Additional items of embodiment >>
In the above-described embodiment, the single-lens electronic camera 11 has been described. However, the present invention is not limited to this. For example, the present invention can be applied to a compact type electronic camera. In this case, it is preferable to capture the test image with a recording image sensor. In such a configuration, the photometric image sensor 18 is not required, and the configuration of the imaging unit can be further simplified.

  In the above-described embodiment, one color balance adjustment value is determined for the entire image. However, the present invention is not limited to this. For example, the screen may be divided into a plurality of areas, and the color balance adjustment value of the present invention may be determined for each area. In this case, it is possible to determine the uniformity of the illumination state for each region, and determine an appropriate color balance adjustment value for each region according to the uniformity determination.

  In the above-described embodiment, the case where the built-in type electronic flash device 12 is used has been described. However, the present invention is not limited to this. The present invention may be applied to a case in which an external type electronic flash device is used.

  Further, in the above-described embodiment, the case where the color balance adjustment is performed in the electronic camera has been described. However, the present invention is not limited to this. For example, a program for executing part or all of the processing shown in FIG. 3 may be created. By executing this program on an external computer, the above-described image processing can be performed on the computer.

  Note that non-uniform illumination during flash photography is often the case where the main subject in front is bright and the distant background is dark. Therefore, it may be determined that the illumination state is more uneven as the light / dark difference (for example, the standard deviation of the luminance distribution) of the flash image is larger.

In particular, the lower the ratio of ambient ambient light to flashlight, the more pronounced non-uniform illumination. Therefore, a value obtained by dividing the standard deviation of the luminance distribution by the luminance is preferably used as an indicator of the nonuniformity of illumination. With this index, even if the image brightness difference (standard deviation of the luminance distribution) is the same, the non-uniformity becomes lower as the ambient environment light becomes brighter. That is, when the ambient light hits the subject well and the luminance is high on average, such as during daytime synchronization, the non-uniformity is a low value. On the other hand, when the luminance is low on average as in flash photography in the dark, the non-uniformity is a high value.
It is possible to determine the extent of color fogging in the screen from these illumination uniformity determinations and color fog detection results.

  As described above, the present invention is a technique that can be used for an electronic camera, an image processing program, and the like.

It is a figure explaining component arrangement | positioning of the electronic camera 11 in this embodiment. 2 is a block diagram relating to image processing of the electronic camera 11. FIG. 3 is a flowchart for explaining the operation of the electronic camera 11. It is a figure which shows an example of two histogram distribution A and B. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Electronic camera 12 Electronic flash device 13 Shooting lens 16 Diffuser 17 Finder optical system 18 Photometric image sensor 19 Recording image sensor 19a Shutter 22 Image memory 23 Bus 24 Image processing unit 25 Compression recording unit 26 Memory card 27 Imaging control unit 28 Flash light emission control unit 29 Microprocessor 29a Release button

Claims (6)

An imaging unit that captures a subject image;
An imaging control unit that controls the imaging unit to generate a test image that is captured without flashing and a main image that is captured by flashing;
A special light source discriminating unit for determining a deviation from a black body locus for the low saturation region of the main image and detecting a green fog caused by fluorescent lamp illumination; and
A uniform illumination determination unit that determines the uniformity of the illumination state by obtaining a histogram distribution for each of the test image and the main image, and detecting a correlation between both histogram distributions;
When the histogram correlation is high and the green fog is detected in the low saturation area, it is determined that the fluorescent lamp illumination is uniform, and the red-blue component of the main image is relatively higher than the green component. To correct the green fog,
On the other hand, when the histogram correlation is low and the green fog is detected in the low saturation area, it is determined that the fluorescent lamp illumination is non-uniform, and the color balance adjustment unit weakens the correction of the green fog. An electronic camera characterized by comprising. The electronic camera according to claim 1,
The color balance adjustment unit
When the histogram correlation is low and the green fog is detected,
A weighted composite value of a white balance adjustment value determined based on a correlated color temperature (color temperature obtained by mapping on a black body locus) of the main image and a color balance adjustment value for correcting the green fog is obtained. An electronic camera that adjusts the color balance of the main image using the weighted composite value. The electronic camera according to claim 2,
In calculating the weighted composite value, the weight ratio of the color balance adjustment value is lowered as the histogram correlation is lower. The electronic camera according to any one of claims 1 to 3,
The color balance adjustment unit
Prepare a plurality of predetermined "correspondence data between the color of the low saturation region and the color balance adjustment value" for each color rendering difference of the fluorescent lamp, and the lower the degree of histogram correlation, the higher the color rendering property An electronic camera characterized by obtaining a color balance adjustment value using correspondence data and adjusting the color balance of the main image based on the color balance adjustment value. An imaging unit that captures a subject image and generates a captured image;
A special light source discriminating unit for detecting a color cast by a special light source by determining a deviation from a black body locus for a low saturation region of the captured image;
A uniform illumination determination unit for determining the uniformity of the illumination state of the captured image;
When the illumination state is determined to be uniform illumination and a color cast is detected in the low saturation region, the color cast is corrected in a direction to reduce the deviation from the black body locus of the captured image,
An electronic camera comprising: a color balance adjustment unit that weakens correction of the color cast when the illumination state is determined to be non-uniform illumination and color cast is detected in the low saturation region.   A program for causing a computer to function as the special light source determination unit, the uniform illumination determination unit, and the color balance adjustment unit according to any one of claims 1 to 5.

Images