JP2003030646A - Light source discrimination method and device, object identification method and device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To discriminate the kinds of photographing light sources for which chromaticity points match on a chromaticity diagram. SOLUTION: By a single plate CCD 1 composed of a photoelectric conversion element having spectral sensitivity to four colors B, G, Ye and R, light emitted from the light source is detected and the image data S0 of four bands composed of chrominance signals B, G, Ye and R are obtained. In an extended chromaticity point calculation means 21, the average values t1 , t2 , t3 and t4 of the respective chrominance signals B, G, Ye and R are calculated and extended chromaticity points are calculated by the arithmetic operation of t1 /(t1 +t2 +t3 +t4 ), t2 /(t1 + t2 +t3 +t4 ) and t3 /(t1 +t2 +t3 +t4 ). The extended chromaticity point for a known light source is stored in a memory 22, distances between the calculated extended chromaticity points and the extended chromaticity point of the known light source are compared in a comparison means 23 and the known light source for which the extended chromaticity point in the shortest distance is calculated is discriminated as the photographing light source.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for discriminating a type of light source and a subject, and a program for causing a computer to execute these methods.

[0002]

2. Description of the Related Art In a digital camera, an image obtained by photographing can be stored as digital image data in a recording medium such as an internal memory or an IC card provided inside the digital camera. It can be reproduced as a hard copy such as a print or as a soft copy on a display. As described above, when the image data obtained by the digital camera is reproduced, it is expected that the image data will have the same high quality as a photograph printed from a negative film.

On the other hand, it is called chromatic adaptation to the human visual sense, and it is perceived as bluish or reddish even when seeing white paper under daylight or under fluorescent light or tungsten light whose spectral energy distribution is different from that of daylight. Instead, the mechanism for recognizing as white basically works. In consideration of such chromatic adaptation, an image representing the photographed image so that an image that does not depend on the light source color can be obtained even if the photographing light source is different, such as tungsten light, fluorescent light, or outdoor daylight. White balance adjustment processing (AWB processing) according to the photographing light source is performed on the data. Therefore, when performing the AWB processing, it is necessary to determine the type of the light source used for shooting. Therefore, by obtaining the chromaticity point of the photographing light source and comparing this chromaticity point with the known chromaticity point of the photographing light source,
The type of photographing light source is determined. here,
The chromaticity point is calculated as a two-dimensional value.

[0004]

By the way, in a digital camera, a CCD in which photoelectric conversion elements having spectral sensitivities for three colors of R, G, and B are arranged to detect a subject image, and RGB images are detected.
Image data composed of color signals of three colors is obtained. Therefore, the calculation of the chromaticity point for determining the photographing light source is performed based on the color signals of the RGB three colors. However, RGB3
It may not be possible to determine the type of light source from the color signal of the color. For example, it is not possible to discriminate the types of light sources whose chromaticity points match on the chromaticity diagram, such as F8 and F10 fluorescent lamps.

On the other hand, there are cases where it is desired to distinguish between two subjects that appear to be the same color but have different constituent substances when detected by a RGB three-band sensor. For example, if the flesh-colored area in the image is a human flesh-colored image, the soft-focus finish will improve the appearance of the image, but in the case of artificially created things such as engraving, try not to reduce the sharpness as much as possible. Treatment is preferable. However, 2
If the two subjects have brick color and skin color respectively,
Since the brick color and the flesh color have the same chromaticity point on the chromaticity diagram, their types cannot be identified.

The present invention has been made in view of the above circumstances, and a first object thereof is to accurately determine the type of photographing light source.

A second object of the present invention is to accurately identify a subject that looks like the same color to humans.

[0008]

A light source discriminating method according to the present invention detects light emitted from a predetermined light source or reflected light thereof by an image pickup device having spectral sensitivity to four colors, and outputs color signals of four colors. Acquiring, calculating an extended chromaticity point of the predetermined light source based on the color signals of the four colors, and comparing the distance between the extended chromaticity point and the extended chromaticity point of the known light source,
It is characterized in that the type of the predetermined light source is discriminated.

The color signals of the four colors are R, Ye,
G and B color signals are preferable.

The "extended chromaticity point" is t ₁ / t ₂ when the average value of each color signal in all pixels of the image represented by the color signals of 4 colors is t ₁ , t ₂ , t ₃ and t ₄ , respectively. (T ₁ + t ₂ +
t ₃ + t ₄ ), t ₂ / (t ₁ + t ₂ + t ₃ + t ₄ ), t ₃ / (t
It is calculated as a three-dimensional value by the calculation of ₁ + t ₂ + t ₃ + t ₄ ).

The "extended chromaticity point of the known light source" is obtained by detecting the light emitted from the known light source or the reflected light of this light on the completely reflecting white plate by an image pickup device having spectral sensitivity to four colors. It is calculated in advance based on the obtained four color signals.

The "distance" is the distance in the three-dimensional space between the plotted chromaticity points when the calculated extended chromaticity point and the known expanded chromaticity point of the light source are plotted in the three-dimensional space. is there.

If the expanded chromaticity point matches the expanded chromaticity point of the known light source, it is possible to determine that the predetermined light source is the light source that calculated the matched expanded chromaticity point. Therefore, the calculated extended chromaticity point may not necessarily match the extended chromaticity point of the known light source. In this case,
The known light source that has calculated the expanded chromaticity point having the smallest distance from the expanded chromaticity point calculated for the predetermined light source is determined as the predetermined light source.

According to the subject identifying method of the present invention, reflected light of a plurality of subjects of light emitted from a predetermined light source is detected by an image pickup device having spectral sensitivity to four colors, and four colors of each subject are detected. Obtain the signal and
The extended chromaticity point of each subject is calculated based on the color signal of the color, and the distance between the extended chromaticity point of each subject and the extended chromaticity point of the known color is compared to determine each subject. It is characterized by identifying.

The color signals of the four colors are R, Ye,
G and B color signals are preferable.

The "extended chromaticity point of the subject" may be an extended chromaticity point for the color of the entire subject when the subject is a single color, but the color of a partial area of the subject is determined. When the user wants to do so, it means an extended chromaticity point for a partial area having a color to be discriminated in the subject.

The "extended chromaticity point of a known color" is obtained by detecting the reflected light of the light emitted from a predetermined light source on a subject whose color is known by an image pickup device having spectral sensitivity to four colors. It is calculated based on the color signals of the four colors.

The "distance" is the distance in the three-dimensional space between the plotted chromaticity points when the calculated extended chromaticity points and the known expanded chromaticity points are plotted in the three-dimensional space. .

A light source discriminating apparatus according to the present invention includes an image pickup device for detecting light emitted from a predetermined light source or reflected light thereof to obtain color signals of four colors, and the predetermined device based on the color signals of the four colors. An expanded chromaticity point calculating means for calculating an expanded chromaticity point of the light source, a storage means for storing the expanded chromaticity points of various known light sources, the expanded chromaticity point and the expanded chromaticity point of the known light source. And comparing means for determining the kind of the predetermined light source by comparing the distances.

The color signals of the four colors are R, Ye,
G and B color signals are preferable.

The subject identifying apparatus according to the present invention includes an image pickup device for detecting the reflected light of a plurality of subjects of light emitted from a predetermined light source, and acquiring color signals of four colors for each subject. Extended chromaticity point calculation means for calculating an extended chromaticity point of each subject based on color signals of colors, storage means for storing extended chromaticity points of various known colors, and the extension for each subject. It is characterized by further comprising a comparison unit for identifying each of the subjects by comparing the distance between the chromaticity point and the extended chromaticity point of the known color.

The color signals of the four colors are R, Ye,
G and B color signals are preferable.

Further, the light source discriminating method and the subject discriminating method according to the present invention may be provided as a program for causing a computer to execute.

[0024]

According to the light source discriminating method of the present invention, 4
The extended chromaticity point of a predetermined light source is calculated based on the color signal of the color, and the distance between the extended chromaticity point and the extended chromaticity point of the known light source is compared to determine the predetermined light source. Where RGB3
The chromaticity point calculated based on the color signal of the color is a two-dimensional value, whereas the expanded chromaticity point calculated by the present invention is 3
The value of the dimension. As described above, by calculating the chromaticity point as a three-dimensional value, the parameter for determining the chromaticity point can be expanded, and the type of light source that cannot be identified by the two-dimensional chromaticity point can be determined. it can.

According to the subject identifying method of the present invention, 4
The extended chromaticity points of a plurality of subjects are calculated based on the color signals of the colors, and the subjects are identified by comparing the distances between the extended chromaticity points of each subject and the extended chromaticity points of known colors. Here, the chromaticity point calculated based on the three color signals of RGB is a two-dimensional value, whereas the expanded chromaticity point calculated by the present invention is a three-dimensional value. In this way, by calculating the chromaticity point as a three-dimensional value, the parameters for identifying the chromaticity point can be expanded, and an object that looks like the same color that cannot be identified by the two-dimensional chromaticity point can be obtained. Can be identified.

[0026]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of a light source discriminating apparatus according to an embodiment of the present invention. Figure 1
As shown in FIG. 3, the light source discriminating apparatus according to the present embodiment includes the single-plate CCD 1 and the discriminating means 2.

As shown in FIG. 2, the single-plate CCD 1 is a photoelectric conversion element (hereinafter referred to as a sensor) having a spectral sensitivity with respect to each color of B (blue), G (green), Ye (yellow), and R (red). 4) image data S0 composed of B, G, Ye, and R color signals can be obtained by photographing a subject. The B, G, and
By performing interpolation calculation on each of the Ye and R color signals,
Color signals of all four colors can be acquired at the position of each sensor. Further, instead of the single-plate CCD 1, four CCDs for receiving the respective lights of B, G, Ye, and R separated by the prism are provided, and B obtained from each CCD is
The image data S0 may be composed of G, Ye, and R color signals. Also, by switching the four spectral filters that transmit the respective colors of B, G, Ye, and R, one CC is selected.
Color signals of B, G, Ye, and R4 colors may be obtained from D.

The discriminating means 2 is provided for each color signal B, G, Ye, R.
Average value t _{1 of} all pixels represented by the image data S0 of
t ₂ , t ₃ , and t ₄ are calculated, and the average values t ₁ ,
Expanded chromaticity point calculation means for calculating expanded chromaticity points t ₁ / T, t ₂ / T, t ₃ / T based on t ₂ , t ₃ and t ₄ by the following equations (1) to (3) 21, a memory 22 that stores extended chromaticity points calculated in advance for various known light sources, extended chromaticity points t ₁ / T, t ₂ / T, t ₃ / T, and extended chromaticity points of known light sources. Comparing means 2 to determine the type of the photographing light source used for photographing and output discrimination information D representing the discrimination result.
3 and 3. t ₁ / T = t ₁ / (t ₁ + t ₂ + t ₃ + t ₄ ) (1) t ₂ / T = t ₂ / (t ₁ + t ₂ + t ₃ + t ₄ ) (2) t ₃ / T = t ₃ / (T ₁ + t ₂ + t ₃ + t ₄ ) (3)

In the comparison means 23, the calculated expanded chromaticity point and the expanded chromaticity point of the known light source are t ₁ / T,
The distances between the calculated expanded chromaticity points and the expanded chromaticity points of a plurality of known light sources are plotted on a three-dimensional space composed of the t ₂ / T and t ₃ / T axes and compared. Then, the light source that has calculated the expanded chromaticity point that is the same as or is closest to the calculated expanded chromaticity point is determined as the imaging light source.

The memory 22 detects the light emitted from various known light sources or the reflected light of the light by a perfect reflection white plate by the single-plate CCD 1 and further expands the colors based on the detected four color signals. The extended chromaticity point obtained by calculation in the degree point calculation means 21 is stored in association with the type of light source. In this embodiment, it is assumed that the extended chromaticity point of the known light source is obtained based on the light reflected by the perfect reflection white plate. The calculated extended chromaticity points are shown in Table 1 below. Here, in Table 1, F6, F8, and F10 fluorescent lamps are used as known light sources, and the extended chromaticity points for each light source are shown. In addition to the extended chromaticity point, Table 1 also shows the chromaticity point (x, y) calculated by the conventional method.

[0031]

[Table 1] Here, the calculation of chromaticity points (hereinafter referred to as standard chromaticity points) by the conventional method will be described. First, CIE1931 is used by using three sensors whose spectral sensitivity characteristics match the color matching function.
XYZ tristimulus values are obtained, and the tristimulus values X, Y, and Z are converted by the following equation (4) to obtain a standard chromaticity point (x, y). x = X / (X + Y + Z) y = Y / (X + Y + Z) (4)

Here, F6, F8, and F10 are used as photographing light sources, and samples 1 of the light emitted from the photographing light sources are
Expansion for each light source of Samples 1 and 2 obtained by detecting reflected light for No. 2 by the single-plate CCD 1 and further calculating by the expanded chromaticity point calculation means 21 based on the color signal output from the single-plate CCD 1. The chromaticity points are shown in Table 2 below along with the standard chromaticity points.

[Table 2] Regarding the extended chromaticity points and standard chromaticity points calculated for Samples 1 and 2, the calculation results of the distances to the extended chromaticity points and standard chromaticity points calculated for known F6, F8, and F10 light sources are shown below. The results are shown in Table 3 and Table 4. Table 3
And in Table 4, for the value with the smallest distance,
The mark is given.

[Table 3]

[Table 4] In Sample 1, when the photographic light sources are F6 and F10, the distances between the standard chromaticity point and the extended chromaticity points to the extended chromaticity points for F6 and F10 are the shortest, and therefore the type of the photographing light source can be determined. However, when the photographing light source is F8, the light source with the shortest distance is F10 when the standard chromaticity point is used, so the light source is erroneously determined. On the other hand, when the extended chromaticity point is used, the light source with the shortest distance is F8, so that the type of the photographing light source can be accurately determined.

Also in Sample 2, when the photographing light sources are F6 and F10, both the standard chromaticity point and the extended chromaticity point have the shortest distances from the extended chromaticity points for F6 and F10. The type of can be determined. However, when the photographing light source is F8, the light source with the shortest distance is F10 when the standard chromaticity point is used, so the light source is erroneously determined. On the other hand, when the extended chromaticity point is used, the light source with the shortest distance is F8, so that the type of the photographing light source can be accurately determined.

Next, the operation of this embodiment will be described. FIG. 3 is a flowchart showing the operation of this embodiment. First, the subject is photographed by the single-plate CCD 1, and B,
The 4-band image data S0 including the color signals of G, Ye, and R4 is acquired (step S1). The image data S0 is input to the expanded chromaticity point calculation means 21, where the expanded chromaticity point is calculated (step S2). The calculated expanded chromaticity points are input to the comparison means 23, where the distances between the expanded chromaticity points of the plurality of known light sources stored in the memory 22 are compared (step S3), and the determination is made based on the comparison result. Discrimination information D indicating the type of the light source is output (step S
4), the process ends.

As described above, the embodiment of the light source discriminating device is
The extended chromaticity point is calculated based on the four color signals B, G, Ye, R, and the light source used for photographing is determined based on the distance between the extended chromaticity point and the known extended chromaticity point of the light source. It is something that is done. Here, when the standard chromaticity point calculated based on the color signals of the RGB three colors is used, as described above, F
It is not possible to distinguish between 8 and F10. On the other hand, if the extended chromaticity point is used, F8 and F10 can be discriminated. Therefore, according to the present embodiment, it is possible to determine the type of light source that could not be identified by the conventional standard chromaticity point.

Next, an embodiment of the subject identifying apparatus according to the present invention will be described. FIG. 4 is a schematic block diagram showing the configuration of the subject identifying apparatus according to the embodiment of the present invention.
As shown in FIG. 3, the subject identifying apparatus according to the present embodiment includes a single-plate CCD 1 similar to the embodiment of the light source identifying apparatus,
And a discriminating means 4.

The discriminating means 4 is for each color signal B, G, Ye, R.
The average values ct ₁ , ct ₂ , ct ₃ , ct ₄ of a certain color region in the image represented by the image data S0 are calculated, and based on the average values ct ₁ , ct ₂ , ct ₃ , ct ₄ , Extended chromaticity points ct ₁ / T, ct ₂ / by the following equations (5) to (7)
Expanded chromaticity point calculation means 41 for calculating T, ct ₃ / T,
The memory 42 storing the expanded chromaticity points calculated in advance for various known colors, and the expanded chromaticity points ct ₁ / T, ct ₂ / T,
ct ₃ / T is compared with the extended chromaticity point of a known color to determine the subject and to output the discrimination information C indicating the discrimination result. ct ₁ / T = ct ₁ / (ct ₁ + ct ₂ + ct ₃ + ct ₄ ) (5) ct ₂ / T = ct ₂ / (ct ₁ + ct ₂ + ct ₃ + ct ₄ ) (6) ct ₃ / T = ct ₃ / (Ct ₁ + ct ₂ + ct ₃ + ct ₄ ) (7) In the comparison means 43, the calculated extended chromaticity point and the extended chromaticity point of the known color are ct ₁ / T, ct ₂ / T, c.
The distances between the calculated extended chromaticity points and the extended chromaticity points of a plurality of known colors are plotted on a three-dimensional space composed of the t ₃ / T axes and compared. Then, the color for which the expanded chromaticity point closest to the calculated expanded chromaticity point is calculated is determined to be the color of the photographed subject.

In the memory 42, the reflected light of the light emitted to the subject having various known colors is detected by the single-plate CCD 1, and further calculated by the expanded chromaticity point calculation means 41 based on the detected color signal. The extended chromaticity point obtained by doing so is stored in association with the color type. In the present embodiment, the same light source as in the case of calculating the extended chromaticity point of a known color is used to determine the color of the subject.

Next, the operation of this embodiment will be described. FIG. 5 is a flowchart showing the operation of this embodiment. First, the subject is photographed by the single-plate CCD 1, and B,
The 4-band image data S0 including the color signals of G, Ye, and R4 is acquired (step S11). Image data S0
Is input to the expanded chromaticity point calculation means 41, where the expanded chromaticity point is calculated (step S12). The calculated expanded chromaticity points are input to the comparison means 43, where the distances to the expanded chromaticity points of a plurality of known colors stored in the memory 42 are compared (step S13), and determination is made based on the comparison result. Discrimination information C representing the selected color is output (step S14),
The process ends.

As described above, the embodiment of the subject identifying apparatus calculates the extended chromaticity points for a plurality of subjects based on the four color signals B, G, Ye, and R, and the extended color points for each subject are calculated. The subject is identified based on the distance between the degree point and the extended chromaticity point of the known color. Here, the chromaticity point calculated based on the color signals of the three colors RGB has a two-dimensional value, whereas the extended chromaticity point calculated according to the present embodiment has a three-dimensional value. As described above, by calculating the chromaticity point as a three-dimensional value, the parameter for identifying the chromaticity point can be expanded, and a subject that cannot be identified by the two-dimensional chromaticity point can be identified.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing a configuration of a light source discrimination device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an image pickup surface of a single-chip CCD.

FIG. 3 is a flowchart showing the operation of the embodiment of the light source determination device.

FIG. 4 is a schematic block diagram showing the configuration of a subject identifying device according to an embodiment of the present invention.

FIG. 5 is a flowchart showing an operation of the embodiment of the subject identifying apparatus.

[Explanation of symbols]

1 Single CCD 2,4 Discrimination means 21,41 Extended chromaticity point calculation means 22,42 memory 23,43 Comparison means

Claims (10)

[Claims] 1. Color light signals of four colors are obtained by detecting light emitted from a predetermined light source or reflected light thereof by an imaging device having spectral sensitivity to four colors, and based on the color signals of the four colors. A light source characterized in that the type of the predetermined light source is determined by calculating the extended chromaticity point of the predetermined light source and comparing the distance between the extended chromaticity point and the extended chromaticity point of the known light source. How to determine. 2. The color signals of the four colors are R, Ye, G,
The light source discrimination method according to claim 1, wherein the light source is a B color signal. 3. The reflected light of a plurality of subjects of light emitted from a predetermined light source is detected by an imaging device having spectral sensitivity to four colors, and color signals of four colors are obtained for each subject, The extended chromaticity point of each subject is calculated based on the four color signals, and the distance between the extended chromaticity point of each subject and the extended chromaticity point of the known color is compared to obtain the subject. A method for identifying a subject, which comprises: 4. The color signals of the four colors are R, Ye, G,
4. The object identifying method according to claim 3, wherein the object is a B color signal. 5. An imaging device for detecting light emitted from a predetermined light source or reflected light thereof to obtain color signals of four colors, and extended chromaticity of the predetermined light source based on the color signals of the four colors. Expanded chromaticity point calculation means for calculating points, storage means for storing expanded chromaticity points for various known light sources, and comparison of the distance between the expanded chromaticity point and the expanded chromaticity point of the known light source. Accordingly, the light source discriminating device is provided with a comparing unit that discriminates the type of the predetermined light source. 6. The color signals of the four colors are R, Ye, G,
The light source discriminating device according to claim 5, wherein the light source discriminating device is a B color signal. 7. An imaging device for detecting reflected light of a plurality of subjects of light emitted from a predetermined light source to obtain color signals of four colors for each subject, and an imaging device based on the color signals of the four colors. Extended chromaticity point calculating means for calculating the extended chromaticity point of each subject, storage means for storing the extended chromaticity point of various known colors, and the extended chromaticity point of the respective subject and the known A subject identifying apparatus, comprising: a comparing unit that identifies each subject by comparing a distance with an extended chromaticity point of a color. 8. The color signals of the four colors are R, Ye, G,
The subject identifying apparatus according to claim 7, wherein the subject identifying apparatus is a B color signal. 9. A procedure of detecting light emitted from a predetermined light source or its reflected light by an imaging device having spectral sensitivity to four colors to obtain color signals of four colors, and A procedure for calculating the extended chromaticity point of the predetermined light source based on the above, and a procedure for determining the type of the predetermined light source by comparing the distance between the extended chromaticity point and the extended chromaticity point of the known light source. A program for causing a computer to execute a light source determination method having the above. 10. Reflected light from a plurality of subjects of light emitted from a predetermined light source is detected by an image pickup device having spectral sensitivity to four colors, and four lights for each subject are detected.
A procedure for obtaining a color signal of a color; a procedure for calculating an extended chromaticity point of each of the subjects based on the color signals of the four colors; and an extended chromaticity point of the subject and an extended chromaticity of a known color. A program for causing a computer to execute a subject identification method including a procedure for identifying each subject by comparing the distances with points.