JPH07210682A - Color recognition device - Google Patents

Abstract

PURPOSE:To provide the color recognition device capable of stably recognizing colors without storing the reference color even when the environment such as illumination is changed at the time of taking in the color information. CONSTITUTION:The device is provided with a means 103 calculating the chromaticity correcting amount by inputting the color information of the achromatic color sample and obtaining the change amount of the chromaticity affected by the illumination, means 104 correcting the chromaticity inputting the color information of the object to be measured and the chromaticity correction amount and outputting the corrected chromaticity of the object to be measured. On the lightness, a means 105 calculating the brightness correction amount obtaining the change amount of the brightness affected by the illumination and calculating the brightness correction amount at the output and a means 104 correcting the brightness which outputs the corrected brightness of the object to be measured are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color recognition apparatus for recognizing color information of a color image, and more specifically, it is capable of stably displaying a color even when an environment such as lighting when capturing color information of a color image is different. The present invention relates to a color recognition device for recognition.

[0002]

2. Description of the Related Art In recent years, pattern extraction using color information of color images and recognition of color information have been frequently performed. In general, a specific color to be recognized is stored in advance as a reference color, and the color information of the object and the stored color information are compared to determine whether or not they are the same color.

As a color information acquisition means for obtaining color information from a color image, a desired area is designated for color image data obtained from a color camera and the color information of that portion is used. A color camera has three light-receiving elements
An object that obtains color image data of an object formed by forming three color filters corresponding to the primary colors and an object that forms an image of the object passing through the three filters on three different light receiving surface elements respectively. is there.

Further, as a color information acquiring means for directly obtaining color information by designating a position from plane information of an image as well as specifying a position, a filter corresponding to three primary colors of light is used. A method of directly guiding the reflected light at a desired position of an object to the color sensor, which is a set of three light receiving elements for measuring the amount of light transmitted through, is also widely used.

In order to compare the color information, the color information is divided for each element expressing a color, a coordinate space having each element as a coordinate axis is considered, and the color information is represented by a coordinate value indicating a position in the color space. The color information is compared by comparing the coordinate values.
Hereinafter, a coordinate space having an element representing a color as a coordinate axis is referred to as a color space, and color information indicated by a position in the color space is referred to as color space coordinates.

The color space is expressed in a rectangular coordinate space of RGB using the three primary colors of light, red (R) green (G) blue (B) as elements, and brightness (brightness) hue (hue). ) It is widely known that there are three elements of color intensity (saturation), which are expressed in a cylindrical coordinate system in which lightness is taken in the height direction, hue is taken in the angular direction, and saturation is taken in the radial direction. Further, the color information can be roughly divided into lightness and a dark element excluding lightness among elements expressing color, for example, a color element expressed by combining a hue and a saturation, which is called chromaticity. .

[0007]

However, in the recognition of color information of a color image, if the environment for capturing the color image such as lighting conditions changes, the obtained color information changes even for the same color. , It makes an erroneous judgment as to whether or not it matches the previously stored reference color. In general, when an environment such as a lighting condition for capturing a color image changes, all the reference colors are stored again under the environment to avoid erroneous recognition. This is the same when the color information is obtained by the color sensor.

An object of the present invention is to provide a color recognizing device which can stably recognize a color without re-storing the reference color even if the environment such as lighting when capturing color information changes. Is.

[0009]

In order to achieve the above object, a color recognition apparatus according to the present invention measures an environment in which color information is captured when performing color recognition, and based on the measurement result, an ideal environment is obtained. The amount of correction to the color information acquired in step 1 is obtained, the color information of the reference color is corrected and stored with this amount of correction, and the color information of the recognized color is corrected and then compared with the stored reference color. . At this time, a color sample, which is a specific color, is used to measure the environment in which the color information is captured, and this color sample is called a color sample.

In the present invention, achromatic color samples having different lightness are used in order to obtain the correction amount of chromaticity. The chromaticity obtained from the color information of this color sample is considered to be the chromaticity of the environment itself that has changed, that is, the color of the illumination itself or the colored reflected light that reaches the color sample from the surroundings. In the present invention, the chromaticity obtained from this achromatic color sample is considered as the correction amount, and by subtracting this, the color information under the ideal achromatic illumination environment can be obtained.

Due to the non-linearity of the color information acquisition means itself or the non-linearity of the color space itself, the correction amount of the chromaticity differs depending on the lightness, and simple proportional distribution cannot cope with the actual change. In the present invention, the correction amount of chromaticity obtained from a certain achromatic color sample is considered as the correction amount of chromaticity at the lightness of the color sample, and is obtained by preparing a plurality of achromatic color samples having different lightness. Interpolation of the correction amount of each chromaticity is used as the correction amount for arbitrary brightness.

In the present invention, chromatic color samples having different hues are used to correct the lightness. A mapping that converts the lightness of the chromatic color sample into the lightness under an ideal environment, for example, the lightness when the reference color is stored, is obtained as the lightness correction amount. By using this brightness correction amount, the brightness of the recognized color can be corrected to the brightness when the reference color is stored for comparison.

[0013]

[Function] When the environment for capturing color information changes, if the color space coordinates of the achromatic and chromatic color samples described above are obtained, the correction amount of the color information in that environment can be obtained. There is no need to reset the color information of the standard color of.

Only the achromatic color sample is required for the chromaticity correction, and the number may be determined according to the lightness width of the reference color and appropriately interpolated, and the chromatic color sample used for the lightness correction is also used. Since a small number is sufficient, a relatively small number of color samples are required for correction.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a color recognition method and apparatus according to an embodiment of the present invention. The configuration of the embodiment of the present invention will be described below with reference to FIG.

The color information acquisition means 101 takes in a color sample, a reference color or a recognized color and outputs color information 1011.

The color space conversion means 102 is the color information acquisition means 1
The color information output by 01 is input, and the color information 1012 converted into an expression that is easy to use for correction of color space coordinates and color recognition is output.

The chromaticity correction amount calculation means 103 calculates the chromaticity correction amount 10 from the color information obtained by the color information acquisition means by the color information of the achromatic color sample and converted by the color space conversion means 102.
13 is calculated and output.

The chromaticity correction means 104 inputs the color information 1012 obtained from the reference color or the color to be recognized and converted by the color space conversion means 102 and the chromaticity correction amount 1013, and outputs the corrected chromaticity 1014.

The lightness correction amount calculation means 105 uses the color information acquisition means 1 as m color samples having mutually different hues.
The color information 1011 acquired by 01 and the color information 1012 obtained by converting the color space coordinates of this color information by the color space conversion means 102 are input and a lightness correction amount 1015 is output.

The lightness correction means 106 uses the color information 1011 obtained from the reference color or the recognized color and the lightness correction amount 1015.
Is input and the corrected brightness 1016 is output.

When the color information of the reference color is taken in by the color information acquisition means 101, the color information storage means 107 receives the corrected chromaticity 1014 and the corrected lightness 1016,
Further, by inputting the information 1019 for specifying the color from the color specifying information input means 109, the information 101 for specifying the color is obtained.
9 and the corrected chromaticity 1014 and the corrected lightness 1016, which are information representing the reference color, are stored in association with each other. Also,
The color information storage unit 107 stores the recognition range together with the color information of the reference color. That is, the recognition range information 1020 input from the recognition range setting unit 110 when the information 1019 for specifying the color is input is the information 1019 for specifying the color.
It is stored as recognition range information of the reference color corresponding to.

Further, when the color call information 1018 is input, the color information storage means 107 outputs the corresponding color information and the recognition range as the reference color information 1017 and the recognition range information 1021.

When the corrected hue 1014 of the reference color and the corrected lightness 1016 are input to the color information storage means 107, the specific information input means 109 is a color for specifying which color it is. The information 1019 for specifying is given to the color information storage means.

The recognition range setting means 110 gives the magnitude of the difference between each element of the reference color information and each element of the color information of the recognized color, which is regarded as "the element matches". , As recognition range information 1020.

When the color information of the color to be recognized is taken in by the color information acquisition means 101, the color information comparison means 108 continuously outputs the color recall information 1018 and the reference color information 1017 obtained from the color information storage means 107. And recognition range information 10
Enter 21. The corrected chromaticity 1014 and the corrected luminosity 1016 input at this time are compared with the reference color information 1017, and from the color recall information when the recognition range information 1021 determines that the two match. The recognition result 1022 is output.

The operation of each component according to this embodiment will be described below. In the color information acquisition unit 101 according to the present embodiment, the output color information 1011 is color information represented by (R, G, B) which is a color space coordinate system widely used in devices that handle color information.

In the color space conversion means 102 according to the present embodiment, the color information 1011 of the (R, G, B) expression is input, and the color of the (Y, H1, H2) expression which is the psychological characteristic value expressing the color. Information 1012. At this time, Y is brightness and H1
And H2 represent Cartesian coordinates in a plane spanned by hue and saturation. Hereinafter, in this embodiment, the color information represented by the remaining elements excluding the brightness, such as the color information represented only by H1 and H2, will be referred to as “chromaticity”.

Chromaticity correction amount calculation means 103 according to this embodiment
Then, the chromaticity correction amount is obtained from the n pieces of color information 1012-31 to 1012-3n obtained by the color information acquisition means 101 capturing the color information of n achromatic color samples having different lightness from each other. 1013 is calculated and output.

FIG. 2 shows an example of the chromaticity correction amount calculated by the chromaticity correction amount calculation means 103 according to this embodiment. Since it is considered that the change in the chromaticity correction amount due to the difference in lightness is continuous, the color information obtained from the n achromatic color samples is third-ordered in the color space of (Y, H1, H2) expression. A chromaticity correction amount of 10 is obtained by obtaining a curve interpolated by the spline interpolation method.
13 The interpolation method is not limited to the cubic spline interpolation method shown here, and any mathematically known interpolation method may be used. In addition, if the brightness range is small for all the standard colors, an achromatic color sample of average brightness is set to 1
You may use only one and consider it as the correction amount of the whole brightness.

FIG. 3 shows an example of chromaticity correction in this embodiment. Color space coordinates (Yc,
H1c, H2c), and the achromatic color space coordinates (Yc, H1) at the lightness Yc are obtained from the curve having the chromaticity correction amount 1013.
g, H2g) is regarded as the position of the lightness axis at that lightness, and the corrected color space coordinates of the recognized color are (Yc, H1c-H1).
g, H2c-H2g). (H1c-H1g, H2
c-H2g) is the corrected chromaticity 1014.

Brightness correction amount calculation means 105 according to the present embodiment
Is color information 1011 obtained from color samples of three chromatic colors having different hues (R1, G1, B
1), (R2, G2, B2), (R3, G3, B3) are converted into certain constant values Y1, Y2, Y3 by a mapping F in the form of a matrix with 1 row and 3 columns by the equation (1) of the equation 1. This is defined and this is set as the brightness correction amount 1015. At this time, Y1, Y2, and Y3 are color information (Y1, H11, H2) obtained by color space conversion of color information obtained from the three chromatic color samples.
1), (Y2, H12, H22), (Y3, H13, H
23) is the lightness Y1, Y2, Y3.

[0033]

[Equation 1]

The lightness correction means 106 according to the present embodiment uses the color information 1011 obtained from the reference color or the recognized color as (R, G, B) and a mapping F which is a matrix showing the lightness correction amount 1015. It corrects by the formula (2) of 2.

[0035]

[Equation 2]

The color information storage means 107 stores the color information of the reference color and the recognition range of the reference color in association with the information for specifying the color. Color information storage means 10 according to the present embodiment
In FIG. 7, a numerical value called a color number is used as the information 1019 for specifying the color. The corrected chromaticity 1014 and the corrected lightness 1016, which are color information representing the reference color, are stored in the elements of the table whose position is determined by the color number. At this time, the recognition range information 1020 input from the recognition range setting means 110
Is added to the element of the table whose position is determined by the color number which is the information 1019 for specifying the color. The recognition range in the present embodiment is a numerical value representing the width of each element of the digitized color information.

Further, when the color calling information 1018 is input, the color information storage means 107 outputs the corresponding color information and the recognition range as the reference color information 1017 and the recognition range information 1021. In this embodiment, the color calling is performed. The information 1018 uses the same numerical value as the color number as the information 1019 for specifying the color,
The reference color information and the recognition range information are obtained and output from the elements of the table whose position is determined by the color number.

FIG. 4 shows the structure of the table of the color information storage means 107 according to an embodiment of the present invention. There are k table elements, 401-1, 401-2, and 401-k, respectively. The element read / written when the color number i is designated is 401-i.

Each element is a set of four pieces of information. For example, in the i-th element, the chromaticity of the reference color 4011-
i, reference color lightness 4012-i, chromaticity recognition range 401
3-i, and the brightness recognition range 4014-i. A combination of the chromaticity 4011-i of the reference color and the lightness 4012-i of the reference color is the reference color information and the recognition range 401 of the chromaticity.
The recognition range information is a combination of 3-i and the brightness recognition range 4014-i.

When the corrected hue 1014 of the reference color and the corrected lightness 1016 are input to the color information storage means 107, the specific information input means 109 outputs information for specifying which color it is. Although it is given to the color information storage means,
In this embodiment, a numerical value called a color number is used as information for specifying a color, and a method of inputting a numerical value from a key switch provided on the operation panel is general. Of course, the information for specifying the color is not limited to the numerical value, and a character string representing the color name can be used.

The recognition range setting means 110 gives the magnitude of the difference between each element of the reference color information and each element of the color information of the color to be recognized, which is regarded as "the element matches". , Is output as recognition range information 1020,
In the present embodiment, the chromaticity represented by (H1, H2) is represented as the chromaticity represented by (S, H) in terms of saturation and hue for the convenience of recognizing color information and setting the recognition range. The relationship between the saturation S and H1 and H2 is expressed by the formula (3) of Equation 3, and the hue H
And the relationship between H1 and H2 is obtained by the equation (4) of the equation 3.

[Equation 3]

Further, in this embodiment, the recognition range information 1
020 is represented by a numerical value. Numerical values are generally set using key switches provided on the operation panel. Alternatively, since the sensory amount of color is set, the resistance value of the variable resistor independently provided for each element of color information is electrically read, and the numerical value obtained by the analog / digital conversion means is set to each value. A method of sensuously setting the recognition range can be considered by setting the recognition range information 1020 by regarding the size of the difference given to each element. At this time, if a rotary or slide type variable resistor knob is provided on the operation panel corresponding to each of the three elements (brightness, saturation, and hue) that represent color, the position of this knob can be recognized. You can set the range intuitively.

In this embodiment, the color call information 1018 uses a numerical value called a color number, and the recognition result 1022 has the same numerical value as the color number. The corrected chromaticity 1014, which is the color information of the color to be recognized, and the corrected brightness 1016, and the reference color information 101, which is the stored color information of the reference color.
The color information is compared with that of No. 7, the absolute value of the difference is obtained for each element, and if the absolute value of the difference for each element is smaller than the value of the corresponding element of the recognition range information for all the elements, the recognized It is determined that the color matches the reference color information.

Although the embodiment of the present invention has been described above, the present invention is not limited to the configuration of this embodiment. For example, the color space coordinates expressing color information are color space coordinates based on psychological characteristics. Anything will do. Further, the interpolation method used by the hue correction amount calculation means 103 is not particularly limited. The lightness correction mapping calculation means 105 has no particular fixed value for converting the chromaticity coordinates of the chromatic color sample, and it is decided according to the required number of colors. The left side of the equation (2) of the equation 2 is also for adjusting the scale, and it may be omitted or may be a trigonometric function or an exponential function.

[0045]

As described above, according to the present invention, when the environment for capturing color information such as the illumination condition changes and the color information captured for the same color also changes, the same environment is maintained. By correcting the color space coordinates of the recognized color based on the color space coordinates of the color sample and comparing it with the color space coordinates of the specific reference color stored in advance, the environment for capturing color information changes. Can also stably recognize colors. Also,
Since the number of color samples used for correction is small, the correction amount can be easily obtained, and the correction amount of chromaticity with respect to the change of lightness also takes into consideration the actual change.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a diagram illustrating calculation of a hue correction amount according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating hue correction according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating a configuration of a color information storage unit according to an exemplary embodiment of the present invention.

[Explanation of symbols]

 101 Color Information Acquisition Means 102 Color Space Conversion Means 103 Chromaticity Correction Amount Calculation Means 104 Chromaticity Correction Means 105 Lightness Correction Mapping Calculation Means 106 Lightness Correction Means 107 Color Information Storage Means 108 Color Information Comparison Means 109 Specific Information Input Means 110 Recognition Range Setting means

Claims (7)

[Claims] 1. A color information acquisition means for obtaining color information of an object from a color image, and a color space conversion for converting an expression of color information obtained by the color information acquisition means into color information of an expression suitable for color recognition processing. Means, color information storage means for storing color information of a specific color to be recognized in advance before the recognition processing, color information obtained from the object to be recognized, and colors stored in the color information storage means. A color recognition device having color information comparison means for comparing information with each other to determine whether or not the colors are the same, which is represented by the remaining elements excluding brightness among the elements expressing color. For chromaticity, which is the color information that depends on The chromaticity correction amount calculation means for outputting the chromaticity correction amount and the color information of the measured object. The chromaticity correction means for inputting the chromaticity correction amount and the corrected chromaticity of the object to be measured, and the brightness representing the brightness among the elements expressing the color, the change in the brightness due to the influence of lighting. In order to correct the amount, input the color information of the chromatic color sample, obtain the amount of change in lightness due to the influence of lighting, calculate the lightness correction amount, and output the lightness correction amount calculation means; And a lightness correction amount that is input, and outputs a corrected lightness of the object to be measured. 2. The achromatic color sample for correcting chromaticity is at least two achromatic color samples having different lightness, and the chromaticity obtained from color information of the color sample is the color sample. Considering this is the amount of chromaticity displacement in the lightness of the sample, this is the chromaticity correction amount,
The color recognition device according to claim 1, further comprising chromaticity correction amount calculation means for interpolating a displacement amount of each lightness to obtain a chromaticity correction amount for the entire lightness range. 3. An achromatic color sample for correcting the chromaticity when the difference in lightness between the specific colors to be recognized is small, one achromatic color sample representing the specific color to be recognized is used. The chromaticity correction amount calculation means for calculating a chromaticity obtained from the color information of the color sample as a chromaticity correction amount over the entire lightness range, the chromaticity correction amount calculating means being provided. Color recognition device. 4. The chromaticity correction amount and the color information according to claim 2 or 3 are input, and the chromaticity displacement amount in the lightness of the color information is calculated from the chromaticity correction amount and the lightness of the color information. The color recognition device according to claim 1, further comprising: a chromaticity correction unit that calculates the chromaticity of the color information and outputs the corrected chromaticity by subtracting the chromaticity displacement amount from the chromaticity. 5. The color sample for correcting the lightness is at least one color sample, and a map for converting the lightness of the color sample into the lightness of the color sample itself is used as the lightness correction amount. The color recognition device according to claim 1, further comprising: a lightness correction amount calculation unit. 6. The lightness correction amount according to claim 4 and the color information are input, and the corrected lightness is calculated and output from the color information using a map which is the lightness correction amount. The color recognition device according to claim 1, further comprising a brightness correction unit. 7. The color recognition device according to claim 1, further comprising color information acquisition means for directly reading color information of a desired point on the recognition target object.