JP2016194449A - Coloring checkup device, and coloring checkup method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To clearly and simply quantify such visual impressions as metallic feeling and glitter of pearls, and thereby rationalize comparative checkup of the tested object with a reference object.SOLUTION: Illuminating units 3 are scattered in an arc shape on the inner peripheral part of the inner wall of a dome-shaped housing 4 and indirectly illuminate the inner wall face above. The illuminating units 3, using LED illumination of 92 or more in color rendering index, are high color rendering white light sources for indirect illumination and, in spite of their small size and light weight, satisfy the requirements of light sources for color evaluation and are free from the problem of color aberration, which usually occurs to LED. The dome-shaped housing 4 is equipped with a dome-shaped cover 41, a round illumination aperture 42, an L-sectioned light blocking wall 43 constituting a coaxial round shape with the periphery of the cover 41 to form the illumination aperture 42 and a fitting part 44 for fixing a camera in a position inclined from the upper center of the inside. The camera 2 is arranged on the fitting part 44, and the illuminating units 3 are arranged inside the light blocking wall 43.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a color inspection apparatus and method for inspecting the color distribution of an image in order to evaluate the coloration of a product.

  There are many products that require painting or coloring, such as electrical appliances, automobiles, housing / building products, electrical appliances, clothing, etc., and they are painted and colored during the manufacturing process, and the appearance color during the sales process The product is often selected by.

  For example, taking an automobile as an example, the inventions described in Non-Patent Documents 1 to 5 have been proposed, but countermeasures such as flip-flops are not yet sufficient. In addition, with the introduction of new models, the color of the outer panel continues to increase, and the demand for color by consumers is diversifying. New colors are often set even for minor changes and special specification cars. The new colors announced by Japanese automakers are dozens of colors per year, and color variations are infinitely wide, making inspection difficult.

  Incidentally, there is an RGB color system camera as a conventional means for acquiring coloring information. The RGB color system has been proposed by the International Commission on Illumination (CIE), and uses the three primary colors of specific wavelengths obtained from the actual spectral spectrum to add these colors to obtain the same color for the desired color. Is. However, in the RGB color system, there is a negative part in the RGB color matching function that expresses the spectral sensitivity corresponding to the human eye, so it is not possible to subtract light depending on additive color mixing. It is difficult to handle negative values of spectral sensitivity as they are. Therefore, the RGB color system camera approximates and expresses the negative part generated in the RGB color matching function by modifying and correcting it. However, this approximation process makes it impossible to accurately capture colors in the color gamut of human eyes, causing color shifts and color collapse of images and videos. On the other hand, there is a CIE XYZ color matching function (hereinafter referred to as XYZ color matching function) as a color space coordinate-converted so as not to cause a negative part in the RGB color matching function, and as means for acquiring color information using this, There are spectral colorimetry method and tristimulus value direct reading method.

  The spectrocolorimetric method directly measures the emission spectrum emitted from the light source by a large number of sensors, or measures the reflectance for each wavelength in the reflection spectrum of the sample, and calculates the sensitivity using the XYZ color matching function. By doing this, a tristimulus value XYZ with high measurement accuracy is obtained. On the other hand, the tristimulus value direct reading method is a method of directly reading tristimulus values XYZ which are colorimetric values by an optical sensor (color sensor or photoelectric colorimeter) equipped with three types of filters.

  Among such coloring information acquisition means, there is a demand for means for acquiring and analyzing color information as described above, and Patent Document 1 is cited as a related art related thereto. It is an object of the present invention to provide a color unevenness inspection method capable of easily performing color unevenness inspection, and an inspection image data generation apparatus used in the color unevenness inspection method. Further, the color unevenness inspection method includes an inspection image display step for displaying an image for color unevenness inspection on the projector 2, a color space conversion characteristic acquisition step for acquiring RGB / XYZ conversion characteristics of the projector 2, and a color unevenness. The second color space format of the captured image data is set to the first of the projector 2 based on the imaging process of acquiring the captured image data by capturing the inspection image by the imaging means and the RGB / XYZ conversion characteristics of the projector 2. Based on the color space conversion step for generating the converted image data converted into the color space format, the converted image display step for displaying the converted image data by the projector 2, and the converted image for color unevenness inspection, the color unevenness inspection is performed. A color unevenness inspection step. Japanese Patent Application Laid-Open No. 2004-228620 has an invention for determining whether or not the quality is poor from the overlap rate between the document image data and the masked inspection image data.

  Regarding the evaluation of coloring, an invention for measuring the appearance of automobile painting as shown in Non-Patent Document 1 has been proposed. The appearance of the paint has color, gloss, smoothness (red skin), metallic feeling, depth, and other defects, such as bumps, bumps, dents, scrapes, and repellency. Measurements are proposed, of which texture measurement is scratch, color, metallic feeling, fleshiness, deepness, but the background of the changes in the color trend of automobiles and increasing interest in sensitivity and so on.

  In recent years, dark-colored systems such as red and blue have been increasing instead of light-colored systems typified by white in automotive topcoats. It has been empirically known that the dark color system is a color in which dirt and scratches on the surface of the coating film are conspicuous, and it has become necessary to measure scratch resistance and develop a paint with improved scratch resistance.

  Regarding the color measurement, since the optical properties of the coating film are different between the solid and metallic systems, the color measurement method and measuring machine are also different between the two.

  In the solid coating film, the distribution of diffuse reflection light in the layer having color information is isotropic, and therefore, it can be measured with a normal colorimeter. The colorimeter includes a spectral colorimeter and a colorimeter. The spectrocolorimeter measures the spectral reflectance ρ (λ) of the object, and the equation (from the spectral distribution P (λ) of the illumination light and the spectral tristimulus values x (λ), y (λ), z (λ) ( The tristimulus value XYZ is obtained by calculation using 2).

  The colorimetric color difference meter directly measures tristimulus values XYZ, and its principle is similar to the process of human perception of color by looking at an object (sample). The integrating sphere corresponding to the eyeball collects diffusely reflected light from the sample and guides it to the light receiver, and has sensitivity characteristics of spectral tristimulus values x (λ), y (λ), and z (λ) in the retina. The equivalent of the three types of photoreceptor cells is replaced with a light receiver that combines a color filter and a photodiode, and the amount of stimulation transmitted to the cerebrum through the optic nerve is proportional to the output generated by the light receiver. , Displayed as a Z value.

  As described above, the colorimetric color difference meter is similar to the human eye in the colorimetric principle, and many small and handy products are suitable for use in the field. However, care must be taken because there are coatings that exhibit the same color (metamerism) as the light source. Even though the two coatings look the same in natural daylight, they may look different under incandescent lighting. In this way, even if the colors match under a specific light source, a phenomenon in which the same color shifts when the light source changes is called a condition equal color by the light source. Conditional color matching is caused by the difference in pigment type, that is, the difference in spectral reflectance between the two types of coating films. Therefore, when measuring the color of such a coating film, it is necessary to use a spectrocolorimeter.

  Since the metallic coating film includes a glittering material in the base layer, the diffuse reflection light distribution in the layer exhibits anisotropy. Therefore, it shows a phenomenon in which colors appear different when the incident angle and the light receiving angle are different, that is, geometric metamerism. A variable angle spectrophotometer that can change the incident angle and the light receiving angle is used to measure the color of such a metallic coating film. Currently, it has been proposed to obtain the tristimulus values XYZ by measuring the spectral reflectance ρ (λ) under angular conditions.

  Regarding the measurement of metallic feeling, metallic feeling is a sensation caused by the appearance of aluminum flakes in the coating film, and is also referred to as glitter, sparkle or glitter. As a measuring method, a method of scanning a coating film with a microgloss meter and analyzing a reflected light intensity curve has been proposed, but it is not so popular. Recently, a laser type metallic feeling measuring device (ALCOPE) has been developed and gradually becoming popular.

JP 2010-145097 A JP 2005-238817 A

"Measurement of car paint appearance" Toyota Central R & D Review Vol.29No.2 (1994.6) JAMAGAZINE April 2007 Changes in pigments used in automotive “paints” and automotive paints Masaaki Arimoto "Actual product colors in color coordination" Issue: Tokyo Chamber of Commerce and Industry "Techno Cosmos" No. 15 "Color Design and Paint Color Design" Publisher: Nippon Paint “Paint Research” No. 140 “Development of“ Cosmo Silver ”for Toyota and Soarer” Published by Kansai Paint

  However, according to the invention of Patent Document 1, the invention of Patent Document 2, the invention of Non-Patent Document 1, and the description of Non-Patent Documents 2 to 5, it is possible to objectively evaluate the color without depending on the visual inspection of the inspector. However, the accuracy and method of obtaining color information, that is, the camera as an imaging means is the same as the conventional one, and there are still problems in the coloring evaluation accuracy and method required in recent years.

  Conventionally, a hue such as a metallic feeling is discriminated by human eyes, but it is difficult to discriminate the color, and a solution is required. In particular, accurate evaluation associated with colored flip-flops of vehicles and the like is difficult, and in such a case, analysis and evaluation of accurate color distribution is required.

  In recent years, metallic coatings and the like have been newly developed, and there are coatings in which various materials such as mica are mixed in addition to aluminum. It is only necessary that the color form intended by the manufacturer be used. Depending on other manufacturing conditions such as the operation of the air gun, there are variations in coloring, and unexpected situations may occur. The problem with the phenomenon called flip-flops is the phenomenon in which shading and brightness change depending on the angle at which the appearance color of the vehicle is viewed when the underlying material is different. There may be a combination of different materials such as doors and fenders, bonnets and fenders in the same position.

  In addition, there are various methods of quality control relating to color, but even today, visual inspections relying on limit samples by experienced inspectors are performed at many sites. There are several problems in the industry where quality is strict. Problem 1 is lack of objectivity. When it is desired to determine how the color management purpose is reflected in the human eye, the visual inspection that is also determined by the human eye can be inspected with a certain degree of accuracy. In situations such as when you want to guarantee quality to customers, it is scarce to say objective data. Problem 2 cannot reproduce colors. While the trend for traceability is increasing, the only data that can be left by visual inspection is the insufficiency of the inspector, which is insufficient. It is important to be able to retrieve the test results accumulated every day in a form that can be understood at any time. Problem 3 is the hurdle to securing human resources. Visual inspection can be inspected with a certain level of accuracy and speed if only a sample sample is prepared, but it takes time and money to develop and hire personnel. Since the person in charge is human, there are also unstable factors. There is a possibility that the inspection accuracy may change depending on the physical condition. There are various reasons for leaving the job, individual differences among inspectors, and human error. In addition, in the visual inspection, since there is a problem of labor costs and stability of quality, there is a product defect loss and a transportation cost and time loss due to a round trip of the product between the factory and the user.

  Further, the weak points of color management in the conventional spectrophotometer are: The imaging range is narrow (for example, a circle of about 5 mm). For color measurement, the average X, Y, and Z values of this range are measured, and the pattern (texture and metallic feeling) is averaged, so it is unknown. The current situation is that the colors cannot be taken accurately, so the colors and textures look slightly different from the appearance. 2. It is difficult to actually see the color image with numerical data alone. Even if a standard color sample is taken with a conventional camera, the color appears slightly different from the appearance. 4. The color sample itself will fade out over time and will be different from the original color. There is a problem that the sense of color seen by people varies.

  Since the conventional camera can only detect color data in the color gamut surrounded by RGB triangles, the color accuracy is low and colors cannot be displayed accurately. The color data of the displayed image is also displayed within the sRGB color gamut, and accurate color information is lost. The reason why the color of an image captured by a conventional camera is displayed differently from the actual color is that the color is displayed in the sRGB region. Also, pigments containing lame and pearl components are difficult to measure color because total reflection is added. Conventional spectrophotometers have a very narrow colorimetric range, and the numerical values vary considerably depending on the measurement location.

  Furthermore, depending on the brightness of the lighting, the colors seem to differ, but this was not taken into account.

  Therefore, the present invention responds to the demands for the acquisition, reproduction, and analysis of faithful color information in the coloring field, and also obtains an accurate image color distribution that is faithful to the human eye. It is an object of the present invention to provide a color inspection apparatus and method for accurately and simply inspecting subtle shades such as textures such as pearly feeling and metallic feeling and unevenness.

  In view of the above-described problems, the present invention is directed to a camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function, and an object to be measured indirectly. A dome-shaped housing having an illumination section that illuminates automatically, an opening for illumination, a peripheral portion of the camera and an illumination section, and an illumination surface that irradiates illumination light toward the inner surface of the dome; and An arithmetic processing unit that converts the acquired three-band visual sensitivity image into tristimulus values XYZ in the CIE XYZ color system, wherein the arithmetic processing unit is obtained by imaging the measurement object. Among these, the specified inspection area is set, and as the measurement object, the inspection object and the reference object are respectively calculated for the XYZ value of each pixel of the inspection area or the xy value normalized from the XYZ value, xy coordinates of xy chromaticity diagram Creates an xy chromaticity histogram distribution or an XYZ chromaticity histogram distribution by dividing the inspection area of the XYZ coordinates of the XYZ chromaticity diagram with a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid. A color inspection apparatus for inspecting a color by calculating a color distribution coincidence index indicating an overlapping ratio of two xy chromaticity histogram distributions or XYZ chromaticity histogram distributions of the inspection object and the reference object is there.

  Since x + y + z = 1, the calculation of z can be omitted. In place of the two-dimensional grid, if all XYZ are calculated to form a three-dimensional grid, the calculation time increases. The overlapping ratio of the two XYZ chromaticity histogram distributions may be an index in a two-dimensional overlap or an index in a three-dimensional overlap.

  In another aspect of the invention, a camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to a CIE XYZ color matching function, and an object to be measured are provided. An illuminating unit that indirectly illuminates, a dome-shaped housing that includes an illumination opening, and the camera and illuminating unit are provided in the center, and an irradiation surface that irradiates illumination light toward the inner surface of the dome, and the camera A three-band visual sensitivity image obtained by the calculation processing unit for converting the tristimulus value XYZ in the CIE XYZ color system, wherein the arithmetic processing unit is obtained by imaging the measurement object. In the image, the specified inspection area is set, and the XYZ values of each pixel in the inspection area are converted into Lab for the inspection object and the reference object as measurement objects, respectively, and the Lab coordinates of the Lab chromaticity diagram Divide the inspection area with grids, and A Lab chromaticity histogram distribution is created by integrating the number of pixels of the inspection object and the reference object to which the inspection object belongs, and a color distribution coincidence index indicating the overlapping ratio of the two Lab chromaticity histogram distributions of the inspection object and the reference object is obtained. The color inspection apparatus is characterized by inspecting a color by calculation.

  In the chromaticity histogram distribution, the pixel integration number of the two pixels is compared in a grid unit, the smaller pixel integration number is specified, the pixel integration number is integrated, and the total number of pixels in the inspection region is It is preferable to calculate the color distribution coincidence index by calculating a ratio of the accumulated number.

  The lighting unit is divided into a plurality of areas, and each lighting unit is turned on independently to calculate a three-band visual sensitivity image using projection light from different directions, and whether the measurement object has a flip-flop. It is preferable to check whether.

  The lighting unit is divided into left and right regions, and the left and right lighting units are lit independently, so that a 3-band visual sensitivity image is calculated by projection light from the left and right directions, and the measurement object has a flip-flop. It is preferable to check whether.

  It is preferable that data indicating the color distribution coincidence index is transmitted and received between computers via a communication line.

  According to another aspect of the present invention, there is provided a color inspection method using a camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to a CIE XYZ color matching function. And illuminating the object to be imaged with indirect illumination light from the opening of the dome-shaped housing, and converting an image having three spectral sensitivities acquired by imaging with the camera into tristimulus values XYZ in the CIE XYZ color system. A step of setting the specified inspection region among the three-band visual sensitivity images obtained by imaging the measurement object; and the inspection object and the reference object as the measurement object, respectively, A step of calculating an XYZ value of each pixel or an xy value normalized from the XYZ value with respect to the inspection region, and an xy coordinate of the xy chromaticity diagram or an inspection region of the XYZ coordinate of the XYZ chromaticity diagram Are divided by grids, and the number of pixels of the test object and reference object belonging to each grid is integrated to create an xy chromaticity histogram distribution or an XYZ chromaticity histogram distribution; and 2 of the test object and the reference object A color inspection method comprising: a step of inspecting colors by calculating a color distribution coincidence index indicating an overlapping ratio of two xy chromaticity histogram distributions or XYZ chromaticity histogram distributions. The lattice may be a planar lattice or a three-dimensional lattice.

  Yet another embodiment of the present invention relates to a color inspection method using a camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to a CIE XYZ color matching function. Illuminating an object to be imaged with indirect illumination light from the opening of the dome-shaped housing, and converting an image having three spectral sensitivities acquired by imaging with the camera into tristimulus values XYZ in the CIE XYZ color system And a step of setting the specified inspection area among the three-band visual sensitivity images obtained by imaging the measurement object, and the inspection object and the reference object as the measurement object, respectively, A step of converting the XYZ values of the pixels into Lab, and an inspection region of Lab coordinates in the Lab chromaticity diagram is partitioned by a grid, and the number of pixels of the inspection object and the reference object belonging to each grid is integrated A step of creating a Lab chromaticity histogram distribution, a step of inspecting colors by calculating a color distribution coincidence index indicating an overlapping ratio of the two Lab chromaticity histogram distributions of the inspection object and the reference object, and A color inspection method characterized by comprising:

  One or more cameras may be used. The image may be taken with the camera fixed or moved. When there are a plurality of cameras, they can be installed at locations corresponding to imaging angles.

  The camera according to the present invention is preferably imaged with three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)), that is, the observation object is divided into three channels. Can be used regardless of whether it is an optical filter, a dichroic mirror or a dichroic prism set to obtain these spectral sensitivities.

  The spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)) of the camera is a mountain shape having a single peak that does not have a negative value from the CIE XYZ spectral characteristics, and the peak of each spectral sensitivity curve. It is equivalently converted from the condition that the values are equal and the overlapping of spectral sensitivity curves is minimized, and the spectral characteristic S1 curve has a peak wavelength of 582 nm and a half-value width of 523 to 629 nm. The 1/10 width is 491 to 663 nm. The curve of the spectral characteristic S2 has a peak wavelength of 543 nm, a full width at half maximum of 506 to 589 nm, and a 1/10 width of 464 to 632 nm. The curve of the spectral characteristic S3 has a peak wavelength of 446 nm, a full width at half maximum of 423 to 478 nm, and a 1/10 width of 409 to 508 nm.

  The present invention responds to demands for the acquisition, reproduction, and analysis of faithful color information in the coloring field, acquires accurate image color distribution that is faithful and accurate to the human eye, and produces lamé, pearl, and metallic sensations of products and human bodies. Subtle shades such as texture and unevenness can be inspected accurately and easily. The calculation time can be shortened, the application examples are wide, and the industrial utility value is great.

1 is a block diagram of a color inspection apparatus 1 according to Embodiment 1 of the present invention. (A) is the longitudinal cross-sectional view of the coloring test | inspection apparatus 1 of this invention Embodiment 1, (b) is the XX sectional view taken on the line. 1 is a perspective view of a color inspection apparatus 1 according to Embodiment 1 of the present invention. It is a perspective view which shows the use condition of the coloring test | inspection apparatus 1 of this invention Embodiment 1. FIG. It is a function which shows the spectral sensitivity of the camera 2 which is an XYZ color system camera in Embodiment 1 of this invention. This is a specific example of a method for acquiring image information according to three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) in Embodiment 1 of the present invention. (A) is explanatory drawing in the case of using a dichroic mirror. (B) is explanatory drawing in the case of using a filter turret. (C) is explanatory drawing at the time of attaching optical filter 22a, 22b, 22c to the image pick-up element 23 microscopically. It is a flowchart in the camera 2 of Embodiment 1 of the present invention. It is a flowchart in the arithmetic processing unit 7 of Embodiment 1 of this invention. It is a sub chart in the arithmetic processing unit 7 of Embodiment 1 of this invention. (A) is explanatory drawing which shows the test | inspection area | region T in the arithmetic processing unit 7 of Embodiment 1, (b) is xy chromaticity diagram which shows the chromaticity area | region K on the chromaticity diagram corresponding to the test | inspection area | region T, (c) ) Is an explanatory diagram of a chromaticity region K partitioned by a grid G, (d) is a schematic diagram showing a state of chromaticity overlapping on an xy two-dimensional chromaticity diagram, and (e) is an explanatory diagram showing a minimum distribution, (F) is explanatory drawing which shows an example of xy chromaticity histogram distribution. (A) is explanatory drawing which shows the metallic degree of a measuring object, (b) is xy chromaticity histogram distribution map, (c) is a three-dimensional image figure of xy chromaticity histogram distribution. It is a sub chart in the arithmetic processing unit 7 of Embodiment 2 of this invention. (A) is a conventional system and (b) is an application example of the color inspection apparatus 301 of the third embodiment.

  A coloring inspection apparatus 1 according to a preferred embodiment 1 of the present invention will be described with reference to FIGS.

  The coloring inspection apparatus 1 includes a camera 2 that captures an image of a vehicle J that is an object to be imaged, a plurality of illumination units 3 that have LEDs that illuminate the vehicle J, and a dome-shaped housing in which the camera 2 and the illumination unit 3 are mounted. 4, a gripping unit 5, a switch 6 for turning on / off an illumination power source, an arithmetic processing device 7 connectable to the camera 2, and a display device 8, and capable of functioning as a two-dimensional colorimeter It is. Details will be described below.

The spectral sensitivity of the camera 2 satisfies the router condition, and its normalized sensitivity (S1 (λ), S2 (λ), S3 (λ)) is negative from the XYZ color matching functions as shown in FIG. The peak value of each spectral sensitivity curve is the same, and the equivalent spectral conversion is performed under the condition that the overlapping of spectral sensitivity curves is as small as possible. Specifically, the normalized sensitivity (S1 (λ), S2 (λ), S3 (λ)) has the following characteristics.
Peak wavelength Half width 1/10 width S1 582nm 523-629nm 491-663nm
S2 543nm 506-589nm 464-632nm
S3 446 nm 423-478 nm 409-508 nm

  The peak wavelength of the spectral characteristic S1 can be handled as 580 ± 4 nm, the peak wavelength of the spectral characteristic S2 is 543 ± 3 nm, and the peak wavelength of the spectral characteristic S3 can be handled as 446 ± 7 nm.

Conversion between the three normalized sensitivities (S1 (λ), S2 (λ), S3 (λ)) and the tristimulus values XYZ is obtained using the following Equation 1. For details on the spectral characteristics themselves, refer to Japanese Patent Application Laid-Open No. 2005-257827.

  For an image, color information obtained primarily is a three-band visual sensitivity image S1i based on three normalized sensitivities (S1 (λ), S2 (λ), S3 (λ)) based on functions equivalent to XYZ color matching functions. , S2i, S3i (i = 1 to m, where m is the number of pixels), it is faithful to the sensitivity of the human eye and is highly accurate compared to the case of acquiring by RGB. Since the overlap of the standardized sensitivity (S1 (λ), S2 (λ), S3 (λ)) is small, the S / N ratio is sufficient, and the curve of the standardized sensitivity curve also changes naturally. Errors are kept to a minimum.

  The specifications of the camera are, for example, an effective frequency value of about 5 million pixels, an effective area of 9.93 mm x 8.7 mm, an image size of 3.45 μm x 3.45 μm, a video output of 12 bits, a camera interface GigE, the number of frames (when adjusting focus), 3 to 7 frames / sec, shutter speed 1 / 15,600 sec to 1/15 sec, integration time up to 3 seconds, S / N ratio 60 dB or more, lens mount F mount, operating temperature 0 ° C. to 40 ° C., operating humidity 20% to 80%.

  As shown in FIG. 1, the camera 2 is disposed behind the photographing lens 21, the three optical filters 22a, 22b, and 22c disposed behind the photographing lens 21, and the optical filters 22a, 22b, and 22c. An image sensor 23 (CCD, CMOS, etc.). The three normalized sensitivities (S1 (λ), S2 (λ), and S3 (λ)) of the camera 2 are given by the product of the spectral transmittance of the optical filters 22a, 22b, and 22c and the spectral sensitivity of the image sensor 23. Is. The arrangement relationship between the optical filters 22a, 22b, and 22c and the image sensor 23 in FIG. 1 is merely shown schematically. Specific examples of the method of acquiring the three-band visual sensitivity images S1i, S2i, and S3i according to the three spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) will be given below. Any of these methods can be adopted, and other methods can be adopted.

  FIG. 6A shows a system using a dichroic mirror. This is because light of a specific wavelength is reflected by the dichroic mirror 22c ′, and the remaining light that has been transmitted is further reflected by another dichroic mirror 22a ′ to be spectrally separated, and the image pickup devices 23a, 23b. , 23c are read in parallel. Here, the dichroic mirror 22a ′ corresponds to the optical filters 22a and 22b, and the dichroic mirror 22c ′ corresponds to the optical filter 22c. Light incident from the photographic lens 21 is reflected by the dichroic mirror 22c ′ according to the spectral sensitivity S3, and the remaining light is transmitted. The light reflected by the dichroic mirror 22c ′ is reflected by the reflecting mirror 26, and the spectral sensitivity S3 is obtained by the imaging device 23c. On the other hand, the light transmitted through the dichroic mirror 22c ′ is reflected by the dichroic mirror 22a ′, and the light according to the spectral sensitivity S1 is reflected, and the remaining light according to the spectral sensitivity S2 is transmitted. Therefore, the light is captured by the image sensor 23a and the image sensor 23b, respectively. To obtain the spectral sensitivities S1 and S2. Instead of the dichroic mirror, a dichroic prism having the same characteristics may be used to split the light into three, and the image sensors 23a, 23b, and 23c may be bonded to the positions where each light is transmitted.

  FIG. 6B shows a system using a filter turret 27. Optical filters 22a, 22b, and 22c are provided on a filter turret 27 having the same direction as the incident light from the photographing lens 21 as a rotation axis, and these are mechanically rotated. Sensitivity images S1i, S2i, S3i are obtained.

  FIG. 6C shows a system in which the optical filters 22 a, 22 b, and 22 c are microscopically attached to the image sensor 23. The optical filters 22a, 22b, and 22c on the image sensor 23 are provided in a Bayer array type. This arrangement is such that the optical filter 22b is provided in half of the area on the image sensor 23 divided into a grid, and the optical filter 22a and the optical filter 22c are equally arranged in the remaining half of the area. That is, the arrangement amount is optical filter 22a: optical filter 22b: optical filter 22c = 1: 2: 1. The arrangement of the optical filters 22a, 22b, and 22c other than the Bayer arrangement is not particularly hindered in the first embodiment. Each of the optical filters 22a, 22b, and 22c is very fine and is attached to the image sensor 23 by printing. However, in the present invention, this arrangement is not meaningful, but a filter having characteristics of spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)) is attached to the image sensor.

  Camera 2 is equipped with a measurement head that can accurately measure color and texture without being affected by outside light, and adopts a color visual reproduction camera system to provide tracking that does not fluctuate. ing.

  The camera 2 acquires 3-band visual sensitivity images S1i, S2i, S3i based on the spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)), records them in the arithmetic processing unit 24, and displays them on the display unit 25. The 3-band visual sensitivity images S1i, S2i, S3i are transmitted from the camera 2 to the arithmetic processing unit 7, converted into the tristimulus values XYZ in the XYZ color system by the arithmetic processing unit 7, and image data obtained by the tristimulus values XYZ is obtained. An arithmetic process for normalizing and converting to xy is performed, and an image visualized by the xyz-RGB conversion process is displayed on the display device 8.

  As shown in FIGS. 2 (a) and 2 (b), the illuminating unit 3 is distributed in an arc shape on the inner peripheral portion of the inner wall of the dome-shaped housing 4, and is an indirect illumination that irradiates light on the upper inner wall surface. Become. The illuminating unit 3 is an artificial solar lamp, and employs LED illumination having a color rendering property of 92 or more, and is an indirect illumination type high color rendering white light source. Although it is compact and lightweight, it meets the conditions of a light source for color evaluation and overcomes the problem of chromatic aberration that occurs in general LEDs.

  The illumination source of the illumination unit 3 may employ a xenon lamp (pseudo sunlight), and includes a Fresnel lens assembly in addition to the xenon lamp. Xenon lamps shall illuminate evenly from the top of car J. In the present embodiment, since the position of the camera 2 is fixed, in the flip-flop measurement, there may occur a phenomenon that the brightness of the paint changes in the direction of light. For such flip-flop measurement, projection light from the left and right can be obtained by independently lighting a plurality of LED illumination units 3 arranged with respect to the camera 2. It is possible to inspect whether there is a door and a fender flip-flop. For example, the illumination area of the illumination unit 3 that is ring illumination can be divided into two or more areas. For example, in the example of two divisions, only the left side region (here, 6 or less, for example, 3) can be turned on and photographed (at this time, muting can be performed on a uniform surface for correction). Also, only the right side area (here, 6 or less, for example, 3) can be turned on and photographed (at this time, the right side light muting is taken and corrected on a uniform surface). The light can be applied to any of the ON / OFF and intensity differences because of the difference in the degree of coincidence. In addition, by attaching a left-right difference distribution to the illumination unit 3, it is possible to determine the occurrence of a flip-flop with a color distribution coincidence described later.

  The reason for adopting such a configuration is that the difference in brightness of the metallic coating by flip-flops occurs when the phenomenon that the bright material such as aluminum in electrodeposition coating is aligned in a certain direction occurs, the lighting is constant. However, the reflection effect of the glittering material is different depending on the viewing angle. In other words, even if the measurement angle is constant or the difference in the incident angle of the illumination, the same result is obtained. Therefore, in this case, the right and left balance of the illumination unit 3 is broken to capture such a phenomenon. Become. Furthermore, in this embodiment, since the difference in brightness due to the flip-flop is used as a statistical coincidence index, even if diffuse illumination is used, it can be quantified as a difference in human appearance.

  As shown in FIGS. 2A and 2B, the dome-shaped housing 4 is formed concentrically with the dome-shaped cover 41, the circular illumination opening 42, and the periphery of the cover 41 to form the illumination opening 42. A light shielding wall 43 having a predetermined shape, for example, an L-shaped cross section, and a mounting portion 44 for fixing the camera 2 at an angle with respect to the center from the center of the upper part of the interior. The focus and stop of the camera 2 can be adjusted from the illumination opening 42. The camera 2 is disposed on the mounting portion 44, and the illumination portion 3 is disposed on the inner side of the light shielding wall 43. The reason why the position of the mounting portion 44 is shifted from the center is to prevent the lens having a glossy object from directly moving in the image. As the angle, an angle inclined from the center, for example, an inclination angle of 10 ° to 30 °, for example, 20 ° is illustrated in FIGS.

  As shown in FIGS. 3 and 4, the grip portion 5 is formed in the shape of a grip that can be grasped by a hand, so that the automobile J can be manually imaged. An imaging button 6 of the camera 2 is provided in the vicinity thereof.

  Since the appearance differs depending on the angle due to the flip-flop, as shown in FIG. 4, the camera 2 is manually moved, and the camera 2 directly opens the illumination opening from multiple angles, for example, at least three different angles. 42 is used for imaging.

  The arithmetic processing unit 7 calculates the luminance, chromaticity, etc. in an arbitrary range based on the three-band visual sensitivity images S1i, S2i, S3i of the inspection object Q and the reference object R acquired by the camera 2, and paints them. The color and its color distribution data are measured, and the color distribution data of the paint colors are compared and indexed, for example, and the images of the reference object R and the inspection object Q are displayed in comparison.

  The RGB color system display device 8 is connected to the arithmetic processing device 7 so that the RGB image processed by the arithmetic processing device 7 is displayed on the screen. The arithmetic processing device 7 or the display device 8 includes an input unit (not shown) or the like as appropriate. The input means is a keyboard, a mouse, a touch panel provided in the image display device, or the like.

  The operation of the color inspection apparatus 1 will be described with a specific example. As shown in FIG. 1, the color inspection device 1 operates by connecting a camera 2, an arithmetic processing device 7, and a display device 8. The connection method can be selected regardless of wired or wireless. The flowchart in the camera 2 is shown in FIG. 7, and the flowchart in the arithmetic processing unit 7 is shown in FIGS.

  When the camera 2 is turned on, initialization is performed as shown in FIG. 7 (initialization S1).

  The reference object R and the vehicle J that is the inspection object Q are respectively imaged (imaging process S2). In the imaging process S2, the vehicle 2 is imaged by the camera 2 at a different angle. There are a plurality of imaging locations, and the number can be selected as appropriate. Here, the measurement is made from three directions of front (0 degree), left 45 degrees, and right 45 degrees. In addition, the 0-degree optical axis of the camera 2 is perpendicular to the body surface of the automobile J at the panel measurement location. The illumination part 3 is illumination from diagonally upward like sunlight.

  The imaging process S2 is a process of imaging the automobile J with the camera 2 having three normalized sensitivities (S1 (λ), S2 (λ), S3 (λ)). The characteristics of spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)) are given by FIG. Images are taken by the taking lens 21, the optical filters 22 a, 22 b, 22 c and the image sensor 23.

  The imaging light of the vehicle J that is the reference object R and the inspection object Q is converted into images A and B that are 3-band visual sensitivity images S1i, S2i, and S3i, respectively (input processing S3).

  The tristimulus values XYZ are calculated from the three-band visual sensitivity images S1i, S2i, and S3i according to Equation 1 by the arithmetic processing S4 in the arithmetic processing unit 24 of the camera 2 (arithmetic processing S4).

  The tristimulus values XYZ are transmitted to the arithmetic processing unit 7 (data transmission S5).

  It is determined whether or not to end the process (end process S6), and the process ends.

  Next, processing in the arithmetic processing unit 7 will be described with reference to FIG. First, when the power is turned on, initialization is performed (initialization S110).

  The tristimulus value XYZ transmitted from the camera 2 is received (S120).

  A chromaticity xyz coordinate histogram is calculated, and a color distribution matching index is calculated (S140).

  The calculation process S140 is a step of calculating and visualizing the color distribution matching index of the captured image, and converts the color information necessary for display on the display device 8 into RGB.

  The histogram distribution, color distribution coincidence index, and RGB data processed in S140 are transmitted to the display device 8, and the display device 8 performs display processing (display processing S150).

  It is determined whether or not to end the process (end process S160), and the process ends.

  The sub-flowchart of S140 in FIG. 9 will be described with reference to FIG. An image A of the reference object R is captured in advance, an image B of the inspection object Q to be compared next is captured, and a color distribution matching index is sequentially calculated as follows. The similarity of chromaticity is determined by this color distribution coincidence index.

  The inspection area K (see FIG. 10B) corresponding to the area T (see FIG. 10A) to be inspected for the captured image A is cut out and set (step S201). Size and place can be set freely.

  Next, similarly to the image A, the inspection area from the image B is cut out, specified, and set (S202).

  The chromaticity value xy is calculated from the tristimulus values XYZ of the images A and B (S203).

Expressions for converting the tristimulus values XYZ to the Y′xy color system are shown in Formulas 2 and 3, respectively. Here, a luminance meter (not shown) is used together with the camera 2, and Y is Y ′ calibrated by the luminance meter value (nt). Since color space conversion formulas are commonly used, other detailed formulas are omitted.

  Since x + y + z = 1, the calculation of z can be omitted. All of xyz may be calculated.

  An xy chromaticity histogram distribution or an XYZ chromaticity histogram distribution of the inspection object Q in the inspection region K cut out from the imaged image A of the inspection object Q is created (S204). This chromaticity histogram distribution is an integrated number obtained by counting pixels belonging to the overlapping region D of the two histogram distributions shown in FIG.

  As shown in FIG. 10C, the three-dimensional color distribution to be compared at the xyz coordinate position is projected onto a two-dimensional plane, the inspection region K is partitioned by a plane grid G, and pixels having xy values in the partition Create a planar histogram distribution for. A plane grid having a specific width, for example, a plane grid in which xy is cut by 1/1000 (1000 lines), and numerical values are scanned from end to end. When the numbers are scanned and integrated in the z direction, the histogram distribution is obtained. If the grid cells are made finer, the accuracy will be improved, but the calculation time will be longer.

  Similarly, an xy chromaticity histogram distribution of the image B of the reference object R is created (S204). The xy chromaticity histogram distribution is the cumulative number of pixels, and FIG.

  The minimum distribution of the xy chromaticity histogram distribution or the XYZ chromaticity histogram distribution is specified (S205).

FIGS. 10D and 10E are cross-sectional views of FIG. 10C taken along the line S-S. When viewed on the same line in the xy coordinates, there is an overlap. Instead of drawing a three-dimensional distribution, it is drawn in a plane for convenience. Since it is a histogram, it has a fine staircase distribution. Cumulative number _{H A} and the accumulated number _{H B} shown in FIG. 10 (d) are respectively the image A, corresponding to the image B. When the two histogram distributions are compared, an overlap region D exists. The xy chromaticity histogram distribution is the cumulative number of pixels, and FIG. 10 (d) shows the overlap region D and FIG. 10 (e) shows the minimum distribution.

As shown in FIG. 10E, H _A (x ₁ , y ₁ ) is the cumulative number of the xy chromaticity histogram distribution of the inspection object Q, and H _B (x ₂ , y ₂ ) is the xy chromaticity histogram of the reference object R. When the accumulated number of distribution, with H _a> H _B in overlapping left area, centrally H _a = H _B, and the on the right side is a H _a <H _B. H _A, of H _B, taking the smaller cumulative number (pixel frequency), the left H _A, becomes H _B, the minimum distribution is stepped histogram curve as a bold line can be identified on the right. By using this, the ratio of the overlapping region D to the entire region can be calculated.

The smaller integrated value is specified by this minimum distribution. Of H _A and H _B, if the addition operation the accumulated number of lesser overlap can be calculated the accumulated number of regions D, can be specified percentage of the total number of pixels. The total number of pixels in the inspection area K is determined, and the total number of pixels is the same for both the image A and the image B. This ratio calculation may be integrated three-dimensionally for all the lattices G. For example, as shown in FIG. 10C, the inspection region K is cut along the SS axis, and y is a predetermined value. The distribution of the cumulative number of pixels when x changes from end to end is two-dimensionally integrated. FIG. 10F is a two-dimensional map on the xy coordinates of the integration result. If there is no distribution in the inspection region K and the number of pixels is zero, it is excluded from the calculation.

The cumulative number of xy chromaticity histogram distributions in the overlapping region D is calculated (S206). As described above, the accumulated number in the overlapping area D of the H _A and H _B, adds calculates the cumulative number of lesser. Instead of the two-dimensional xy chromaticity histogram distribution, a three-dimensional XYZ chromaticity histogram distribution may be used.

  Next, a color distribution coincidence index is calculated (S207).

  This index is calculated by the following formula.

  Color distribution matching index = (total number of pixels belonging to overlapping region D / total number of pixels in inspection region K) × 100 (%)

  For example, the pixels belonging to the inspection region K are assumed to be 100 vertical pixels × 100 horizontal pixels = 10,000 pixels. Since the image is cut out in the same inspection region K, the total number of pixels of the image A and the image B is 10,000 pixels. From the xy chromaticity histogram, the number of pixels in the overlap region is integrated, and when the integration number is 5,000, the color distribution coincidence index is 50%. The degree of chromaticity difference increases as the color distribution matching index falls below 100%. If the distribution of the xy values is completely coincident, 100% is obtained. Thereby, when it determines with it being a numerical value more than fixed, it can determine with it being a conformity product.

  Finally, a histogram distribution and color distribution coincidence index storage process is performed, XYZ-RGB conversion of the image AB to be displayed is performed (S208), and the process returns.

  Through this process, the texture of the image can be grasped with a histogram distribution, so subtle differences in color tone are reflected by reflecting differences in the color texture (metallic feel, glitter, mottled pattern, color pattern, gritty feel, etc.) Can be determined.

  For example, as shown in FIGS. 11A to 11C, an example will be described in which three types are examined from a small metallic degree to a large metallic degree. An object having a small metallic degree is designated as a reference object R1, an intermediate degree of metallic degree is designated as an inspection object Q2, and an object having a large metallic degree is designated as an inspection object Q3. First, when a distribution on the xy chromaticity diagram after the above processing is performed on 1 to 3 is created, as shown in the xy chromaticity diagram of FIG. . The accumulated number is shown in light and dark. The brighter the color, the larger the accumulated number. FIG. 11C schematically shows the integrated number in three dimensions of the reference object R1 and the inspection objects Q2 and Q3. The xy axis is chromaticity, and the z axis is the integrated number of pixels. Basically, the stronger the metallic sensation, the lower the Yamagata, and the weaker the metallic sensation, the sharper the Yamagata. The glittering material (aluminum flakes), which is the source of the metallic feeling, gives a glittering feeling when exposed to illumination light. This glittering feeling is physically a light diffraction phenomenon. The color distribution matching index indicating the degree of overlap is calculated by comparing the two histogram distributions for the reference object 1 and the inspection object 2 or 3.

  As shown in Table 1, since the comparative example uses Lab in which ΔE is calculated based on the average value of the color that is the basis of the texture, the Lab and ΔE values are very small compared to the appearance, and inspection is difficult. there were. As the color distribution coincidence index of the present embodiment, since the integrated number within the range of the inspection region K is used as it is, the inspection objects Q2 and 3 are 58% and 27%, respectively, with respect to the reference object R1, and are clearly expressed numerically. In addition, the metallic degree can be easily identified.

  Although the first embodiment has been described above, the following effects are obtained.

  Since the image can be taken by directly applying the opening 42 to the surface of the automobile J, the accuracy is also improved. In addition, the vehicle J, which is a reference object R, at a site such as the head office is imaged by an XYZ color system camera such as the camera 2 or an XYZ system two-dimensional colorimeter, and an image A is obtained. To the manufacturing factory or the like. The manufacturing factory or the like receives the data, manufactures the automobile J of the inspection object Q, images the image B with the camera 2, calculates the color distribution histogram distribution and the color distribution coincidence index with the arithmetic processing unit 7, and displays the display unit 8 The calculation result is displayed in. It is also possible to display the images A and B themselves on the display device 8 in contrast. For example, by applying IT technology, inspection images and inspection data are displayed on a tablet or a smartphone, and transmitted and received between the base and the manufacturing factory. Finally, the image of the car, the color matching degree, etc. are comprehensively confirmed including the texture and judged. As a result, troubles during mass production in factories and the like can be prevented and time, cost and energy can be reduced.

  Other application examples will be described. Two images, A and B, obtained from the reference object R and the inspection object Q are overlapped, and each chromaticity histogram distribution is displayed on the display device 8, or each chromaticity histogram distribution is displayed as one chromaticity. A superimposed chromaticity diagram can be displayed on the diagram, and a color distribution matching index indicating the degree of overlapping can be displayed as a percentage. Thereby, the shift | offset | difference from the reference | standard object R of the chromaticity distribution of the test object Q can be confirmed by a numerical value. The inspection results are displayed numerically for each region K. The grid width of the grid G can be adjusted. An exponent threshold value can be arbitrarily set. Measurement results and captured images can be saved. It is possible to perform standardization of colors and stable color management by reducing problems of individual differences that cannot be avoided by visual inspection and troubles in judgment standards with customers.

  Since non-contact and wide-area shooting is possible, it is possible to digitize flip-flops by imaging the object to be measured from multiple angles with flat illumination, such as painting with aluminum flakes and pearl pigments. Evaluation that is close to the color and texture of the eyes is possible. Irregular pattern parts such as wood grain panels can also be matched. Since the captured images A and B can be displayed on the display device 8 (overlay function), alignment can be easily performed. The inspection object Q can be compared with the reference object R even if the size and material are different. Colors can also be used for fabrics with irregular patterns and textures such as leather. Resin parts inspection, color unevenness and color misregistration inspection are possible. For example, even an object with unevenness can be measured, such as inspection of a console box of an automobile. In addition, (1) Bumper and fender color shift inspection, (2) Fender and front door color shift inspection, and (3) Front door color unevenness flip-flop inspection are also possible. Colors can also be used for building materials with irregular patterns and textures such as flooring, irregular patterns such as wallpaper, and textures such as woodgrain, marble and geometric patterns. You can inspect the texture of teeth in the dental field.

  The color profile, color unevenness and texture can be verified with graphs and numerical values. The deviation of the chromaticity histogram within the designated inspection range can be confirmed, and a graph showing each change of ΔE and the Lab value can also be displayed. In the case of color inspection of different materials such as the front door and fender of an automobile, the color difference between different materials can be verified and inspected by looking at the color difference (ΔE) from the start point to the end point.

  In addition to the tristimulus values XYZ, it is possible to display the gain of each parameter of CMYK and Lab. The camera uses a camera filter equivalent to a color matching function and can detect all color data in a color gamut that can be recognized by human eyes. The detection accuracy can be measured with high accuracy such that the color difference ΔE is 1.0 or less.

  Since it feels like a human being, such as metallic feeling and glittering feeling of glittering pearl pigments, which were difficult to quantify in the past, it has data using the cumulative number of each grid within the inspection area K. The texture such as the metallic degree can be quantified clearly and easily, and the comparative inspection between the inspection object Q and the reference object R can be rationalized.

  It is possible to inspect the three data of the color of the measurement object together. (1) sensory comparison by image, (2) objective and visually easy-to-understand comparison by chromaticity histogram distribution, and (3) comparison by quantified numerical value by degree of coincidence (%) of chromaticity histogram. By using these three types of data together, it can be used as a common language for colors inside and outside the organization, and communication between sales, product management, manufacturing, business partners, etc. will be smooth. The color management according to the present invention eliminates the conventional problems, can quantify the difference between the objective color management standard and the texture and brightness of the appearance, and can deviate from the conventional visual inspection and limit sample inspection. Become. The present invention has a feature that an actual color can be accurately seen from an image as well as accurate digitization of color data, which is not found in a conventional system. It is possible to reduce problems such as individual differences that are unavoidable by visual inspection and troubles caused by discriminating criteria from customers.

  In addition to the above three data, various data can be viewed from one image according to the purpose, which is useful for troubleshooting. In addition, storing images and data can contribute to the accumulation of color data and the realization of traceability, and can also be used as product history data and restoration materials.

  In addition to objective data that has been quantified when developing new colors for color research and experimenting with new paints, new materials, and new groundwork, faithful image data can be saved together with color data. Contribute to the accumulation of traceability and the realization of traceability.

  By using communication technologies such as the Internet and cloud, products made in factories around the world can be managed at the headquarters offices in Japan based on unified standards, saving time and money for color management. In addition, the witness inspection process can be carried out on PC monitors such as the head office and clients.

  Since the measurement object can be photographed without contact with the camera, there is no worry of damaging the product. By presenting an image converted to a color close to the real in a situation where the real cannot be seen, troubles such as feeling a gap when the real is seen later can be avoided. Since it is possible to measure over a wide range on the chromaticity diagram, it is possible to see a wide range of chromaticity distribution, and it is possible to measure not only colors but also complex patterns and textures. Since the color measurement range is wide, the color shift is small depending on the measurement location, and the measurement conditions can be adjusted every time.

  Next, the coloring inspection apparatus according to the second embodiment will be described with reference to FIG. In the second embodiment, instead of the xy value or the XYZ value of the first embodiment, the Lab space Lab value converted from the XYZ value before normalization to xy by the following Equation 4 is calculated, and the xy chromaticity histogram distribution is calculated. Alternatively, a Lab chromaticity histogram distribution is used instead of the XYZ chromaticity histogram distribution.

  This Lab color space is a kind of complementary color space, and has a dimension L indicating lightness and complementary color dimensions A and B, and is based on nonlinearly compressed coordinates of the CIEXYZ color space. By converting from an XYZ value before normalization to Lab using Equation 4, a matching degree that takes the brightness direction into consideration in the Lab color space can be obtained as compared to the matching degree in the XYZ color space.

In the case of calculation of the chromaticity histogram distribution in the Lab space corresponding to the inspection area K, the XYZ value is converted to Lab. The degree of coincidence is calculated based on a distribution in a three-dimensional space of the L axis, a axis, and b axis, instead of the xy chromaticity or the XYZ chromaticity. Lab values in the Lab space coordinates of the inspection object Q and the reference object R are U ₁ (L, a, b) and U ₂ (L, a, b), respectively. In the Lab color space, the histogram distribution has a globe-like shape, and the two histogram distributions overlap three-dimensionally. The inspection region K in the three-dimensional space is partitioned by a lattice, and the chromaticity histogram distribution and the minimum distribution of U ₁ (L, a, b) and U ₂ (L, a, b) in three dimensions are obtained, and the same agreement is obtained. Calculate the degree. The accumulated number of grids is projected on a plane, and the accumulated number of overlapping areas on the grid is calculated by the same accumulation within the plane. In the case of xyz chromaticity, the brightness information is lost, so the matching degree of only the color information is shown. However, in the Lab space, when the brightness of the image changes, the L value changes, and the matching degree Since the positions of the distributions D ₁ and D ₂ are shifted in the Lab space, it is possible to make a determination in consideration of light and dark. This is because, even if the xyz chromaticity is the same, if the brightness of the image is different, the position of the distribution is shifted. For example, the Lab chromaticity histogram distribution shifts downward when dark and shifts upward when bright.

  Next, the coloring inspection apparatus 301 of Embodiment 3 will be described with reference to FIG. In the first embodiment, the object to be imaged is the automobile J, but in the third embodiment, the wig (wig) made to order and the hair W of the person are compared. Regarding the corresponding similar elements, the above description is cited as the 300 series, and the differences are mainly described. The coloring inspection apparatus 301 obtains an image by capturing the hair W of the person with the camera 302 under illumination, and obtains the xy chromaticity in the inspection region K in the arithmetic processing apparatus 324. The xy chromaticity is transmitted from the arithmetic processing unit 324 to the arithmetic processing unit 307 in the manufacturing factory using the communication line 326 (also via the management server 327 if necessary). The processing unit 307 in the manufacturing factory receives the data, manufactures the wig that is the inspection object Q, images it with the RGB camera 302A of the tablet, obtains the xy chromaticity, and obtains the color distribution histogram distribution and color distribution with the processing unit 307. The coincidence index is calculated and the calculation result is displayed on the display device 308. When the wig is completed, in order to confirm the completed state with the person, the inspection image and the inspection data are displayed on the tablet or smartphone which is a portable information terminal by applying IT technology, and transmitted / received between the bases. Finally, the three-band visual sensitivity images S1i, S2i, and S3i of the person's hair and wig are compared with each other, and the colors and textures including the xy chromaticity histogram distribution and color distribution matching index are comprehensively compared. Check and determine whether the product is acceptable. As a result, it is possible to prevent troubles in order-made ordering by the conventional photo print or electronic file method of FIG. Save time, money and energy.

  The embodiments of the present invention are not limited to the above-described embodiments, and modifications and the like can be made without departing from the technical idea of the present invention. Needless to say, objects and the like are also included in the technical scope of the present invention and can take various forms as long as they belong to the technical scope. For example, with respect to a method of acquiring image information according to three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)), the methods given in the present embodiment are only specific examples. However, the technical idea of the present invention is not limited to the above and can be implemented by other methods.

  The color inspection apparatus of the present invention has a great color applicability in manufacturing sites for electrical appliances, vehicles, housing building materials, and other industrial applicability. Inspection of color deviation and color unevenness of printed materials, inspection of cosmetics including lame and pearl pigments, automotive coating inspection (color shift, color unevenness, color shift between different materials are quantified, flip by light intensity graphing ), Car grain inspection, automotive leather seat inspection (inspecting complex texture materials for color unevenness and color misregistration), flooring color matching, tile color matching (accurately to colors and textures) Photography), color matching inspection for cosmetics, sportswear, shoes, golf balls, and artificial teeth in the dental field.

1, 201, 301 ... Coloring inspection device 2 ... Camera 3 ... Illumination unit 4 ... Dome-shaped housing 5 ... Grasping unit 6 ... Switch 7 ... Arithmetic processing device 8 ..Display devices
J ... Automobile 21 ... Shooting lens 22a, 22b, 22c ... Optical filter 23 ... Imaging element 22a ', 22c' ... Dichroic mirror 23a, 23b, 23c ... Imaging element 24 ... -Arithmetic processing unit 25 ... display unit 27 ... filter turret

Claims (8)

A camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalent to the CIE XYZ color matching function;
An illumination unit that indirectly illuminates the measurement object;
A dome-shaped housing having an illumination opening, having the camera at the center, and the illumination unit around, and an illumination surface for illuminating illumination light toward the inner surface of the dome;
An arithmetic processing unit for converting the three-band visual sensitivity image acquired by the camera into tristimulus values XYZ in the CIE XYZ color system;
With
The arithmetic processing unit is
Of the three-band visual sensitivity image obtained by imaging the measurement object, set the specified inspection region,
As the measurement object, for the inspection object and the reference object, respectively, the XYZ value of each pixel in the inspection area or the xy value normalized from this XYZ value is calculated,
By dividing the inspection area of the xy coordinates of the xy chromaticity diagram or the XYZ coordinates of the XYZ chromaticity diagram with a grid, and integrating the number of pixels of the inspection object and the reference object belonging to each grid, an xy chromaticity histogram distribution or XYZ Create a chromaticity histogram distribution,
A color inspection apparatus for inspecting a color by calculating a color distribution coincidence index indicating an overlapping ratio of two xy chromaticity histogram distributions or XYZ chromaticity histogram distributions of the inspection object and the reference object.   The color inspection apparatus according to claim 1, wherein a Lab value calculated from an XYZ value is used instead of the xy value or XYZ value, and a Lab chromaticity histogram distribution is used instead of the xy chromaticity histogram distribution or the XYZ chromaticity histogram distribution.   In the chromaticity histogram distribution, the pixel integration number of the two pixels is compared in a grid unit, the smaller pixel integration number is specified, the pixel integration number is integrated, and the total number of pixels in the inspection region is The color inspection apparatus according to claim 1 or 2, wherein the color distribution coincidence index is calculated by calculating a ratio of the accumulated number.   The illumination unit is divided into a plurality of regions, and the illumination units belonging to the respective regions are independently turned on, so that a three-band visual sensitivity image is calculated by projection light from different directions, and a flip-flop is applied to the measurement object. 4. The color inspection apparatus according to claim 1, wherein the color inspection apparatus inspects whether or not there is any color.   The lighting unit is divided into left and right regions, and lighting units belonging to the left and right regions are individually turned on to calculate a three-band visual sensitivity image using the projection light from the left and right directions, and flip the measurement object. 5. The color inspection apparatus according to claim 4, wherein the color inspection apparatus inspects whether or not there is a color.   The color inspection apparatus according to claim 1, wherein data indicating the color distribution coincidence index is displayed on a display device and transmitted / received between computers via a communication line. In a color inspection method using a camera having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to a CIE XYZ color matching function,
Illuminating the imaging object with indirect illumination light from the opening of the dome-shaped housing; and
Converting a three-band visual sensitivity image having three spectral sensitivities acquired by imaging with the camera into tristimulus values XYZ in the CIE XYZ color system;
Setting a specified examination region among the three-band visual sensitivity images obtained by imaging the measurement object;
Calculating the XYZ value of each pixel in the inspection region or the xy value normalized from the XYZ value for each of the inspection object and the reference object as the measurement object;
By dividing the inspection area of the xy coordinates of the xy chromaticity diagram or the XYZ coordinates of the XYZ chromaticity diagram with a grid, and integrating the number of pixels of the inspection object and the reference object belonging to each grid, an xy chromaticity histogram distribution or XYZ Create a chromaticity histogram distribution,
Inspecting the color by calculating a color distribution coincidence index indicating the overlapping ratio of the two xy chromaticity histogram distributions or the XYZ chromaticity histogram distributions of the inspection object and the reference object;
A coloring inspection method comprising:   8. The color inspection method according to claim 7, wherein a Lab value calculated from an XYZ value is used instead of the xy value or XYZ value, and a Lab chromaticity histogram distribution is used instead of the xy chromaticity histogram distribution or the XYZ chromaticity histogram distribution.