JP6371237B2 - Coloring evaluation apparatus and coloring evaluation method - Google Patents

Description

  The present invention relates to a coloring evaluation apparatus and method for inspecting the color distribution of an image in order to evaluate the coloring 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 imaging unit as a conventional means for obtaining 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 imaging unit approximates and expresses a 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, tristimulus values X, Y, and Z with high measurement accuracy are obtained. On the other hand, the tristimulus value direct reading method is a method of directly reading tristimulus values X, Y, and Z as 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 spectrum tristimulus values x (λ), y (λ), z (λ) ( The tristimulus values X, Y, and Z are obtained by calculation using 2).

  A colorimetric color difference meter directly measures tristimulus values X, Y, and Z, and its principle is similar to the process of human perception of color by looking at an object (sample). The integrating sphere, which corresponds to the eyeball, collects diffusely reflected light from the sample and guides it to the light receiver, and has the 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 determine the tristimulus values X, Y, and Z 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.

  In addition, as shown in FIG. 27, the conventional method of assembling and painting dissimilar materials is as follows. For example, an automobile is manufactured by manufacturing a door made of a steel plate at a parts factory 1, and 3 bumpers and fenders at a parts factory 2. After these are assembled and assembled at the assembly plant, the color-adjusted paint for each part is applied to the automobile. In other words, in the automobile painting process, parts are collected in an assembly factory, and after the finished product is assembled, the entire automobile is painted. In automobiles, everything except the bumpers is painted.

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 acquisition of color information, that is, the accuracy and method of the imaging unit that is the imaging means are the same as in the past, and problems remain in the color 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 is used. However, when painting is performed, the climatic conditions such as temperature and humidity, variations due to the condition of the base coat, the air gun scanning and painting 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. Different doors and fenders, bonnets and fenders, etc. may appear 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 XYZ value of this range is measured, and the pattern (texture and metallic feeling) is averaged. Yes, because the color cannot be photographed accurately, the color and texture appear slightly different from the appearance. 2. It is difficult to actually see the color image with numerical data alone. 3. Even if a standard color sample is taken with a conventional imaging unit, it appears in a slightly different color 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 imaging unit can only detect color data in a color gamut surrounded by RGB triangles, 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 the image captured by the conventional imaging unit 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.

  Also, depending on the brightness of the illumination, the color may appear different, but this was not taken into account.

  Further, for example, as shown in FIG. 27, when an automobile is composed of different materials in an assembly factory, the materials of the parts are different. However, since it is necessary to individually perform painting operations on parts with paints whose colors are adjusted every time, the problem that the painting operations become very troublesome remains. In addition, with metallic coating, it is difficult to adjust the metallic feeling by using a glitter material that is not only a color, and it has not been practically realized in mass production lines so far.

  Therefore, when the product is made of different materials, the present invention ensures the same color and texture coloring even if the parts are different, and improves the efficiency of the coloring work. Responding to requests for faithful color information acquisition, reproduction, analysis, etc., and obtaining accurate image color distribution that is faithful and accurate to the human eye, such as textures of products, human bodies, etc. It is an object of the present invention to accurately and easily inspect subtle shades such as irregularities, and to provide an appearance coloring evaluation apparatus and method.

In view of the above problems, the present invention provides an imaging unit 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. Coloring that is a part of an assembly, the parts are individually colored, and an image having three spectral sensitivities acquired by the imaging unit for the colored parts is converted into tristimulus values X, Y, and Z in the CIE XYZ color system An arithmetic processing unit that obtains and calculates data; and an illumination unit that irradiates the measurement object, wherein the arithmetic processing unit is identified from among the coloring data obtained by imaging the measurement object. An area is set, and the xy value or the XYZ value itself normalized from the XYZ value of each pixel of the inspection area is calculated for the inspection area for the inspection object and the reference object as the measurement object, and the xy chromaticity Inspection area of xy coordinates in the figure Was divided by the grating, by integrating the number of pixels of the inspected and reference belonging to each grid, to create a xy chromaticity histogram distribution or the inspection area of the XYZ coordinates of the XYZ chromaticity diagram in a grid By dividing and integrating the number of pixels of the inspection object and the reference object belonging to each grid, an XYZ chromaticity histogram three-dimensional distribution is created, and two xy chromaticity histogram distributions or XYZ colors of the inspection object and the reference object It is a coloring evaluation apparatus that compares three-dimensional distributions of degree histograms and selects a combination of the colored parts that approximate colors and textures based on the comparison.

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 a color distribution coincidence index that is a ratio of the accumulated number.

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

  It is preferable to image the inspection object flowing on the production line by attaching the imaging unit to a support member.

  The lattice may be a planar lattice or a three-dimensional lattice.

  In place of the two-dimensional lattice, if all XYZ are computed to form a three-dimensional lattice, the computation time increases. The overlapping ratio of the two xy chromaticity histograms is an index in a two-dimensional overlap or an index in a three-dimensional overlap.

  Another aspect of the invention is an imaging unit 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 the components of the assembly, the components are individually colored, and the images having the three spectral sensitivities acquired by the imaging unit for the colored components are converted into tristimulus values X, Y, and Z in the CIE XYZ color system. An arithmetic processing unit that acquires and calculates coloring data, and an illumination unit that irradiates the measurement object, wherein the arithmetic processing unit is identified from among the coloring data obtained by imaging the measurement object. Set the area, for the inspection object and the reference object as the measurement object, respectively, convert the XYZ value of each pixel of the inspection area into Lab, and divide the inspection area of the Lab coordinate in the Lab chromaticity diagram with a grid, Images of the inspection and reference objects belonging to each grid By integrating the prime numbers, a Lab chromaticity histogram distribution is created, the two Lab chromaticity histogram distributions of the inspection object and the reference object are compared, and based on the comparison, the color parts and the textures of which the color and texture are approximated are compared. It is a coloring evaluation apparatus characterized by selecting a combination.

  In another aspect of the invention, 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, It is preferable to calculate the color distribution coincidence index by calculating a ratio of the accumulated number to the total number of pixels in the inspection area.

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

  It is preferable that the imaging unit is attached to a support member to image an inspection object flowing on the production line.

Another aspect of the invention is a color evaluation using an imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function. In the method, under illumination, an object to be measured is an assembly part, the parts are individually colored, and an image having three spectral sensitivities acquired by imaging with the imaging unit for the colored part is represented by the CIE XYZ color system. Generating the coloring data converted into the tristimulus values XYZ in the above, and for the inspection object and the reference object as the measurement object, the specified inspection area among the coloring data obtained by imaging the measurement object A step of setting, for the inspection object and the reference object as measurement objects, a step of calculating an xy value normalized from an XYZ value of each pixel of the inspection area for the inspection area, and x An inspection area of the xy coordinates of chromaticity diagram is partitioned in a grid, by integrating the number of pixels of the inspected and reference belonging to each grid, to create a xy chromaticity histogram distribution, or, XYZ chromaticity diagram By dividing the inspection area of the XYZ coordinates by a grid and adding up the number of pixels of the inspection object and the reference object belonging to each lattice, an XYZ chromaticity histogram three-dimensional distribution is created, and 2 of the inspection object and the reference object Comparing the two xy chromaticity histogram distributions or the three-dimensional distribution of the XYZ chromaticity histograms , and selecting a combination of the colored parts whose colors and textures are approximated based on the comparison. It is an evaluation method.

  Another aspect of the invention is a color evaluation using an imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function. In the method, an object to be measured is a part of an assembly, the parts are individually colored, and an image having three spectral sensitivities obtained by imaging by the imaging unit under illumination with respect to the colored part is represented by the CIE XYZ color specification. A step of generating coloring data converted into tristimulus values X, Y, and Z in the system, a step of setting a specified inspection region among the coloring data obtained by imaging the measurement target, and a measurement target For each of the inspection object and the reference object, a step of converting the XYZ value of each pixel of the inspection area into Lab, and an inspection area of Lab coordinates in the Lab chromaticity diagram are partitioned by a grid, and the inspection belonging to each grid A step of creating a Lab chromaticity histogram distribution by integrating the number of pixels of the object and the reference object, comparing the two Lab chromaticity histogram distributions of the inspection object and the reference object, and based on the comparison, Selecting a combination of the colored parts that approximates the texture, and a coloring evaluation method.

  There may be a single imaging unit or a plurality of imaging units. The imaging unit may be fixed or imaged, or moved and imaged. When there are a plurality of imaging units, they can be installed at locations corresponding to imaging angles.

  In the present invention, the image pickup unit preferably picks up an image of the observation object by dividing it into three channels with three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)). Can be used regardless of whether it is an optical filter, a dichroic mirror, a dichroic prism, or the like set to obtain these spectral sensitivities.

  The spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) of the imaging unit are mountain shapes having a single peak that has no negative value from the CIE XYZ spectral characteristics. It is equivalently converted from the condition that the peak values are equal and the overlap of spectral sensitivity curves is minimized. The curve of the spectral characteristic S1 has a peak wavelength of 582 nm and a half-value width of 523 to 629 nm. 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.

  Therefore, as shown in FIG. 1, if the same color and texture can be colored, the inspection process at the assembly plant is further shortened.

  In the present invention, even when the object to be measured is made of different materials, the same color and texture can be colored, so the coloring process is further shortened. Moreover, the coloring matching function regarding the material of a different color is excellent. In addition, in response to the demand for faithful color information acquisition, reproduction and analysis in the coloring field, it obtains accurate image color distribution faithful to the human eye, such as lamé, pearl, metallic, etc. It is possible to accurately and easily inspect subtle shades such as texture and unevenness. The calculation time can be shortened, the application examples are wide, and the industrial utility value is great.

It is explanatory drawing which shows the outline | summary of the coloring evaluation apparatus 1 of this invention Embodiment 1. FIG. It is a flowchart which shows the process of the coloring evaluation apparatus 1 of this Embodiment 1. FIG. It is a front view of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a perspective view of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a perspective view 1 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a perspective view 2 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a perspective view 3 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is the top view 1 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is the top view 2 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a top view 3 of the upper part of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a block diagram of the coloring evaluation apparatus 1 of Embodiment 1 of this invention. It is a function which shows the spectral sensitivity of the XYZ color system imaging part 2 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 imaging part 2 of this invention Embodiment 1. FIG. It is a flowchart in the arithmetic processing unit 3 of Embodiment 1 of this invention. It is a sub chart in the arithmetic processing unit 3 of Embodiment 1 of this invention. (A) is explanatory drawing which shows the test | inspection area | region T in the arithmetic processing unit 3 of Embodiment 1 of this invention, (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 how chromaticity overlaps 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 an example of a display of OK goods of Embodiment 1 of the present invention. It is an example of a display of the NG product of Embodiment 1 of the present invention. It is a block diagram which shows the structure of the coloring evaluation apparatus 101 of Embodiment 2 of this invention. It is a flowchart in the arithmetic processing unit 103 of the coloring evaluation apparatus 101 of Embodiment 2 of the present invention. It is a block diagram which shows the structure of the coloring evaluation apparatus 201 of this Embodiment 3. FIG. It is a figure which shows the structure and operation | movement of the coloring evaluation apparatus 301 of Embodiment 4 of this invention. It is a block diagram which shows the structure of the coloring evaluation apparatus 301 of this invention Embodiment 4. FIG. It is a block diagram which shows the structure of the coloring evaluation apparatus 401 of this Embodiment 5. FIG. It is explanatory drawing which shows the conventional painting process.

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

  As shown in FIGS. 1 and 2, the coloring evaluation apparatus 1 has three spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) that are linearly converted equivalently to the CIE XYZ color matching functions. The CIE XYZ color image of the part 2 and the image having three spectral sensitivities acquired by the imaging unit 2 for the painted part 5 when the object to be measured is an assembly part, the parts are individually painted at the parts factories 1 and 2 An arithmetic processing unit 3 that acquires and calculates coloring data converted into tristimulus values X, Y, and Z in the system, and an illumination unit 6 that irradiates a painted part 5 that is a measurement object. The arithmetic processing unit 3 sets the specified inspection area among the coloring data obtained by imaging the painted part 5, and the XYZ of each pixel in the inspection area for the inspection object and the reference object as the measurement objects, respectively. By calculating the xy value normalized from the value for the inspection area, dividing the inspection area of the xy coordinates of the xy chromaticity diagram by 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 is created, the two xy chromaticity histogram distributions of the inspection object and the reference object are compared, and a combination of painted parts 5 whose colors and texture are approximated is selected based on the comparison. .

  As shown in FIG. 1 and FIG. 2, a steel plate door is manufactured at the parts factory 1 and painted in the same color, while a carbon fiber or plastic bumper and fender are manufactured at the parts factory 2, and these are the same color. Paint each. In addition, for each component 5, an IC tag storing unique information, for example, a component number, a manufacturing factory name, a manufacturing date, and the like is attached to the component by built-in or mounting (S11).

  In the parts factories 1 and 2, color information such as a color distribution matching index and a color difference ΔE is measured by the coloring evaluation apparatus 1 for the parts after painting, and stored in the computer (S12). In addition to the color distribution information indicating the color and metallic feeling, the color distribution spread information may be stored as the paint color distribution information.

  Before carrying the parts to the assembly factory and assembling them, the head office or the parts factory pre-matches the color distribution coincidence index information with the ones with similar numerical values based on the color difference ΔE, etc., and associates the color information with the IC tag to the part number The combination table is created (S13). Matching may be performed by computer software. In order to combine the most optimal part number and color data, the data is shuffled, the best matching state is calculated, and a part combination table is created. As a result, it is possible to avoid color mismatch due to different coating conditions and to increase manufacturing efficiency.

  The combination table data is uploaded to the cloud using a computer and a network (S14).

  A combination table is downloaded from the cloud (S15).

  When the parts are brought into the assembly factory, the unique information is read from the IC tag, the painted parts are flowed on the line, and the automobile is assembled with the optimum combination of parts based on the table (S16). As a result, even when the automobile is made of different materials, the same color and texture can be colored, so the coloring process is further shortened. Moreover, the coloring matching function regarding the material of a different color is excellent.

In addition to automobiles, the present invention can also be applied to products manufactured by assembling parts, for example, home appliances such as refrigerators, houses, airplanes, and trains. The application range of the present invention is not limited to the above field as long as the same color tone is integrated.

  In the parts factory, the appearance of the paint color varies depending on the angle due to the flip-flop. Here, the imaging unit 2 is manually moved to capture images from at least three different angles. There is an illuminating unit 6, the imaging unit 2 is installed below, and the imaging unit 2 can change its angle manually. The imaging unit 2 can measure the coating color of the coating component 5 and its color distribution data from multiple angles.

  Moreover, in this embodiment, the guide part 4 which attaches the imaging part 2 so that a movement is possible is provided.

The spectral sensitivity of the imaging unit 2 satisfies the router condition, and the spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) are 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 spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) have 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.

The three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) are obtained using the following Equation 1. For details about the spectral characteristics themselves, refer to JP-A-2005-257827.

  The specifications of the imaging unit 2 are, for example, an effective frequency value of about 5 million pixels, an effective area of 9.93 mm × 8.7 mm, an image size of 3.45 μm × 3.45 μm, a video output of 12 bits, an imaging unit interface GigE, and the number of frames (during focus adjustment) 3 ~ 7 frames / sec, shutter speed 1 / 15,600sec ~ 1 / 15sec, integration time up to 3 seconds, S / N ratio 60dB or more, lens mount F mount, operating temperature 0 ℃ -40 ℃, operating humidity 20% -80% It is.

  As shown in FIG. 11, the imaging unit 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. And an image pickup device 23 (CCD, CMOS, etc.). The three spectral sensitivities (S1 (λ), S2 (λ), and S3 (λ)) of the imaging unit 2 are given by the product of the spectral transmittances of the optical filters 22a, 22b, and 22c and the spectral sensitivity of the imaging element 23. Is. The arrangement relationship between the optical filters 22a, 22b, and 22c and the image sensor 23 in FIG. 11 is merely schematically shown. Specific examples of the method of acquiring image information according to the three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) will be given below. In the first embodiment, any of these may be adopted. It is also possible to adopt other methods. The imaging unit 2 includes an arithmetic processing unit 24 and a display device 25.

FIG. 13A 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 transmitted, and the remaining light according to the spectral sensitivity S2 is transmitted . The light transmitted through the dichroic mirror 22a ′ is imaged by the image sensor 23b to obtain the spectral sensitivity S2. The light reflected by the dichroic mirror 22a ′ is reflected by the reflecting mirror 29, and the spectral sensitivity S1 is obtained by the imaging device 23a . 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. 13B 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. S1, S2, and S3 are obtained.

  FIG. 13C 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.

  The imaging unit 2 transmits image information acquired by spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)) to the arithmetic processing device 3, and the arithmetic processing device 3 uses the tristimulus values in the XYZ color system. The image data converted to X, Y, and Z is normalized by the conversion process and converted to xy by the conversion process, and the visualized image is displayed on the display device 7. .

  The arithmetic processing device 3 calculates the luminance, chromaticity, etc. at an arbitrary position of the image acquired by the imaging unit 2 and performs a visualization process. Illumination is applied to the painted part 5 from an oblique direction, and the support part 4 is manually operated to compare the color distribution data of the painted colors and index them.

  As shown in FIGS. 3 to 10, the support unit 4 fixes the head 40, the arm 41 composed of a plurality of beams that move the head 40, the support column 42 that rotatably mounts the arm 41, and the support column 42. And a reinforcing member 42a for reinforcing the support of the column 42.

  The head 40 includes a linear guide 44 and an arcuate member 45 connected to the linear guide 44, and the illumination unit 6 is fixed to both ends of the head 40 so that the position and angle of the illumination unit 6 can be adjusted by a lamp adjusting tool 46. The imaging unit 2 is attached to 44 so as to be movable. An imaging unit angle adjuster 47 that can adjust the imaging angle of the imaging unit 2 is attached. The head 40 is provided with an imaging unit / shift adjustment motor 48 a and an imaging unit / pan adjustment motor 48 b, and the imaging unit 2 is slidable along the linear guide 44. The head 40 is attached to the arm 41 so as to be tiltable by a head / tilt adjusting tool 49. A weight 41 a is fixed to the lower end of the arm 41. The arm 41 is connected to the support 42 by a height adjuster 41b.

  The painted part 5 is imaged at one location by the imaging unit 2, and the imaging unit 2 moves and images at another location. Here, for example, it is possible to shoot at three locations (an appropriate number of locations may be used) at 45 degrees on the front and left and right.

  The coloring evaluation device 1 is installed for the painted part 5 to be measured, and it is determined whether or not the painted part 5 is properly painted, and the paint is checked for sampling.

  The illumination source of the illumination unit 6 employs a xenon lamp (pseudo sunlight). The illumination unit 6 includes a Fresnel lens assembly in addition to the xenon lamp. The xenon lamp is assumed to illuminate uniformly from diagonally above the painted part 5. In addition to xenon lamps, LED artificial solar lights may be used. When attached to a robot arm or the like, it is very convenient because it is compact and lightweight, satisfies the conditions of a light source for color evaluation, and overcomes the problem of chromatic aberration that occurs in general LEDs.

  The display device 7 is connected to the arithmetic processing device 3, receives an image signal processed by the arithmetic processing device 3, and displays an image on a screen. The arithmetic processing device 3 or the display device 7 includes input means (not shown) and the like as appropriate. The input means is a keyboard, a mouse, a touch panel provided in the image display device, or the like.

  In addition, although illustration is abbreviate | omitted, depending on the condition of a test | inspection, you may attach a light-shielding cover to the coloring evaluation apparatus 1. FIG.

  The operation of the coloring evaluation apparatus 1 will be described with specific examples. As shown in FIG. 11, the coloring evaluation device 1 operates by connecting the imaging unit 2, the arithmetic processing device 3, and the display device 7. The connection method can be selected regardless of wired or wireless. A flowchart in the imaging unit 2 is shown in FIG. 14, and a flowchart in the arithmetic processing unit 3 is shown in FIG.

  When the power of the imaging unit 2 is turned on, initialization is performed as shown in FIG. 14 (initialization S1). Next, the painted part 5 is imaged by spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)) (imaging process S2), and then the imaged image data is input by the image sensor 23 (input). Processing S3), the processing unit 3 converts the tristimulus values X, Y, and Z (conversion processing S4). Spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) are transmitted to the display device 7 (data transmission S5). When the image is a moving image, a series of processing from imaging processing S2 to data transmission S5 is continuously performed. The image is displayed on the image display device 7.

  In the imaging process S2, the painted part 5 is imaged by the imaging unit 2 at different angles with respect to the specific region. 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. Further, the panel measurement location is such that the 0-degree optical axis of the imaging unit 2 is perpendicular to the body surface of the painted part 5. In addition, the illumination is characterized by illumination from obliquely above as with sunlight.

Since the relationship between numerical values and colors is difficult to understand in the XYZ color system, an xyY color system for expressing an absolute hue from the XYZ color system is used, and the chromaticity value xy of the chromaticity coordinates is obtained from the following equation.

  Y uses Y of XYZ as it is. All colors can be represented by a two-dimensional plane with chromaticity values xy and Y indicating brightness. Y uses Y of XYZ as it is. It is also possible to convert from xyY to XYZ. Conversion to Lab can be performed on the XYZ values from the XYZ values of the maximum brightness white point of Xn, Yn, and Zn.

  The imaging process S2 is a process of imaging the painted component 5 by the imaging unit 2 having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) (see FIG. 14). Spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) are given according to the above formula 1. Input processing S3 is continuously performed at the same time as imaging is performed by the imaging lens 21, the optical filters 22a, 22b, and 22c, and the imaging device 23.

  Since the input image data is a value according to the spectral sensitivity (S1 (λ), S2 (λ), S3 (λ)), the image captured by the conversion processing S4 in the arithmetic processing unit 3 of the imaging unit 2 is used. Are converted into tristimulus values X, Y, and Z. This conversion is performed according to Equation 1. That is, the tristimulus values X, Y, and Z are obtained by multiplying the inverse matrix of the coefficients in Equation 1. The imaging unit 2 transmits the values in accordance with the spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) to the arithmetic processing unit 3 (S5).

  When the processing unit 3 is turned on, initialization is performed as shown in FIG. 15 (initialization S110). The display device 7 receives the spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) transmitted from the imaging unit 2 in a state where it is connected to the imaging unit 2 (data reception S120). Thereafter, the spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) are converted into tristimulus values X, Y, Z, a chromaticity xy coordinate histogram is calculated, and a color distribution matching index is calculated ( S140). The contents are transmitted to the display device 7 (display process S150). In accordance with data reception S120 from the imaging unit 2, a series of processing from conversion processing S130 to display processing S150 is continuously performed.

  The arithmetic processing S140 is a step of calculating and visualizing the color distribution matching index of the captured image, and converting the color information into RGB or the like when necessary for display on the display device 7.

  The display process S150 is a process of displaying the visualized color distribution coincidence index on the image display device, and the process returns.

  A sub-flowchart of S140 in FIG. 16 will be described. A first image of the reference object is captured, a second image of the inspection object 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.

  An inspection region K (see FIG. 17B) corresponding to a region T (see FIG. 17A) to be inspected with respect to the captured image AB is set (step S141). Size and place can be set freely.

  The chromaticity value xy is calculated to obtain the chromaticity value xy (S142).

  An xy chromaticity histogram distribution of the reference object Q in the region K cut out from the imaged image A of the inspection object is created (S143). This xy 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. 17C, the color distribution to be compared at the position of the xy coordinates is described, and the three-dimensional color distribution is projected onto the two-dimensional color distribution. The inspection area K is partitioned by a plane grid G, and a histogram distribution of pixels having the xy values of the section is created. Let xy coordinates be a grid (solid grid) having a specific width, for example, a grid obtained by cutting xy by 1/1000 (1000 lines). The histogram is scanned from end to end, the number of pixels belonging to the area divided into the grid G is scanned on the same xy plane, and the count number is accumulated in the z direction. In addition, if the xy coordinates are specified in the inspection region K, the calculation time can be shortened. If the grid cells are made finer, the accuracy will be improved, but the calculation time will be longer. Note that if you want to see only a certain color in detail, you can inspect only that part.

  Similarly to S143, an xy chromaticity histogram distribution of the image B of the reference object R is created (S144). The xy chromaticity histogram distribution is the cumulative number of pixels, and FIG.

  Instead of the xy chromaticity value, an Lab space Lab value converted from the XYZ value in the XYZ space or the XYZ value before normalization by the following Equation 5 may be used. The Lab color space is a type 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 color distribution coincidence index in consideration of the brightness direction can be obtained in the Lab color space in contrast to the color distribution coincidence index in the XYZ color space. .

In the case of the calculation of the chromaticity histogram distribution in the Lab space corresponding to the inspection region K, the calculation of the color distribution coincidence index is the same in both the three-dimensional and the plane. First, the XYZ value is converted to Lab, and mapped to the three-dimensional Lab space. The xy chromaticity is determined by a distribution in a three-dimensional space of L axis, a axis, and b axis. 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 Lab three-dimensional space corresponding to the inspection area K in space is partitioned by a three-dimensional lattice called a voxel, and the number of pixels belonging to each three-dimensional lattice is counted up. In the chromaticity histogram distribution of U ₁ (L, a, b) and U ₂ (L, a, b) in three dimensions, a place where there is no pixel in the Lab space is not counted, and there is a mutual count. The minimum value is obtained by comparing the two values and performing an operation that leaves only the smaller count, and the count number is added. Since the count value where the distribution overlaps in the Lab three-dimensional space remains, the color matching index is calculated in three dimensions by performing a comparison operation with the count number corresponding to 25000 pixels cut out first. In this case, it is a feature that a difference in L value (lightness) also appears in the color matching index.

In the case of xy chromaticity, since the brightness information is lost, the color distribution matching index of only the color information is shown. However, in the Lab space, when the brightness of the image changes, the L value changes, The distributions D ₁ and D _{2 of the} color distribution coincidence index are displaced in the Lab space. For example, even if the xy chromaticity is the same, L is high under illumination with high lightness. The position shifts. In the three-dimensional case, if the same object is illuminated with different brightness, the entire three-dimensional distribution appears to shift, and the difference in brightness can be evaluated by the degree of shift, so both color and brightness Can be evaluated. For example, the Lab chromaticity histogram distribution can be determined in consideration of the contrast of the image, such as shifting downward when dark and shifting upward when bright. If you just want to compare colors in a place where normal illumination light is not stable, the two-dimensional color distribution coincidence index is enough because the same value comes out, but even if the color is the same, the brightness is different If it is desired to detect this, a three-dimensional color distribution matching index is used.

The color distribution coincidence index is calculated (S145) and returned. This index is calculated by the following formula. The xy chromaticity histogram distribution is the cumulative number of pixels. FIG. 17D shows the overlap region D, and FIG. 17E shows the minimum distribution.
Color distribution matching index = total number of pixels belonging to overlapping region D / total number of pixels in inspection region K × 100 (%)

17D and 17E are cross-sectional views taken along the line S-S of FIG. 17C. When viewed on the same line in the xy coordinates, there is an overlap. Instead of drawing in three dimensions, it is drawn in a plane for convenience. Moreover, since it is a histogram, it has a minute staircase distribution. Cumulative number _{H B} and the accumulated number _{H A} of FIG. 17 (d) are respectively the image A, corresponding to the image B. When the two histogram distributions are compared, an overlap region D exists.

As shown in FIG. 17E, H _A (x ₁ , y ₁ ) is the integrated 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 inspection 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 region K is determined, and the total number of pixels of the inspection object Q and the reference object R is the same. The calculation of this ratio may be integrated three-dimensionally for all the lattices G. As shown in FIG. 17 (c), the distribution of the cumulative number of pixels when the inspection area K is cut along SS and y is a predetermined value and x changes from end to end in a two-dimensional manner. It is the one that was integrated. FIG. 17F is a two-dimensional map on the xy coordinates of the integration result. When there is no distribution of pixels in the grid G of the inspection region K, the number of pixels is zero, and is excluded from the calculation. Return processing.

  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.

  For the image, the color information obtained primarily is the three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) by a function equivalent to the XYZ color matching function, and is acquired by RGB. Compared to the case, it is more accurate and sensitive to the sensitivity of the human eye. Since the overlap of spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) is small, the S / N ratio is sufficient, and the curve in the spectral sensitivity curve changes naturally, so the error in colorimetry is Minimized.

  Since the texture of the image can be grasped by the histogram distribution, even a subtle difference in hue can be determined by reflecting the difference in color texture (metallic feeling, glittering feeling, mottled pattern, color pattern, feeling of roughness, etc.).

  For example, as shown in FIGS. 18A to 18C, an example will be described in which three types are inspected from a small metallic degree to a large metallic degree. An object having a small metallic degree is referred to as a reference object 1, an intermediate degree of metallic degree is referred to as an inspection object 2, and an object having a large metallic degree is referred to as an inspection object 3. 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. 18C schematically shows the integrated number in three dimensions of the reference object and the inspection object. The xy axis is chromaticity, and the z axis is the cumulative number. 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. By comparing the two histogram distributions of the reference object 1 and the inspection object 2 or 3, a color distribution matching index indicating the degree of overlap is calculated.

  As shown in Table 1, since the comparative example uses Lab in which ΔE is calculated based on the average color of the texture, the Lab value and the value of ΔE are very small compared to the appearance. In the comparative example, It was difficult to inspect the metal feeling. On the other hand, the color distribution coincidence index of the present embodiment uses the integrated number within the range of the inspection region K as it is, so that the inspection objects 2 and 3 are 58% and 27% with respect to the reference object 1, respectively. The metallicity can be identified clearly and easily with numerical values.

Here, the calculation formula for obtaining the color difference ΔE in the L ^* a ^* b ^* color system is shown below. The color difference ΔE is given by the Euclidean distance in the color space, and may be calculated by another more accurate ΔE2000. Further, what is shown in the graph is not limited to the color difference ΔE. For example, there is a method of displaying the L ^* value, a ^* value, and b ^* value in the L ^* a ^* b ^* color system in a graph in parallel.

  Note that the relationship between ΔE and the color matching index is that the metallic feeling cannot be evaluated only with the color difference ΔE, and the color matching index alone is a relative comparison. Close to eye evaluation.

  As an example, the threshold value of the distribution coincidence index is 75%, and the threshold value of the color difference ΔE is 2.5, and the distribution coincidence index is 75% or more and the color difference ΔE is 2.5 or less.

  19 and 20 are display examples on the screen of the display device 7 (S150 in FIG. 15). As shown in FIG. 19, the matching index is 89% and ΔE0.111 is displayed together with the threshold value, OK is displayed in the center area to indicate that it is within the threshold value, the reference image and the inspection image are displayed side by side, and the right area is displayed. The spectral sensitivity distribution and xy chromaticity distribution are displayed. On the other hand, as shown in FIG. 20, the color distribution coincidence index is 81%, ΔE5.546 is displayed together with the threshold value, and NG indicating that the threshold value is outside the threshold value is displayed in the center area, and the reference image and the inspection image are displayed side by side. The spectral sensitivity distribution and the xy chromaticity distribution are displayed in the right area. In the case of an RGB display device, an image converted from XYZ to RGB is displayed.

  Next, a coloring evaluation apparatus 101 according to the second embodiment will be described with reference to FIGS. For the corresponding similar elements, the explanation is cited as 100 series, and the difference is mainly explained.

  An image processing unit 103 that images the reference object / inspection vehicle 105, an image processing unit 102 connected to the image capturing unit 102 via the switch 106, receives a signal, and calculates a color distribution matching index; And a display device 107 for displaying an index.

  As illustrated in FIG. 21, the arithmetic processing device 103 is acquired by imaging a vehicle serving as a test object and a calculation unit 103A that calculates a stimulus value XYZ1 acquired by imaging the vehicle 105 serving as a reference object R. The calculation unit 103B for calculating the stimulation value XYZ2, the calculation unit 103A and the calculation unit 103B are connected, the calculation unit 103C for calculating the color distribution matching index of the vehicle, and the OK signal or the NG signal from the calculation unit 103C are displayed. 107 or to the outside. Note that the switch 106 selectively inputs the stimulus value XYZ1 and the stimulus value XYZ2. The target vehicle is an example.

FIG. 22 is a flowchart for calculating a color distribution coincidence index by comparing chromaticity histogram distributions from two images A and B. As shown in FIG. 24, when the program is started, the inspection area K is cut out from the image A, specified, and set (S201). Next, an inspection area similar to that of the image A is cut out from the image B, specified, and set (S202). The chromaticity value xy or Lab value is calculated from the images A and B (S203). In the inspection region K, the xy chromaticity histogram distribution or the Lab chromaticity histogram distribution of the inspection object Q and the reference object R is calculated and created (S204). The minimum distribution of the xy chromaticity histogram distribution or the Lab chromaticity histogram distribution is specified (S205). The integrated number of the xy chromaticity histogram distribution or the Lab chromaticity histogram distribution in the overlapping region D is calculated (S206). In place of the Lab chromaticity histogram distribution, an XYZ chromaticity histogram distribution may be used. Next, the color distribution coincidence index is calculated (S207), and the process returns. Color distribution matching index = (integrated number of pixels belonging to overlapping region D / total number of pixels in inspection region K) × 100 (%). Cumulative number of in the overlapping area D of the H _A and H _B, adds calculates the cumulative number of lesser.

  Next, the coloring evaluation apparatus 201 of Embodiment 3 will be described with reference to FIG. About the corresponding similar element, description is used as 200 series, and a difference is mainly demonstrated.

  As shown in FIG. 23, the color determination target is a partial region of the painted part 5, and the imaging unit 202 images the target region of the painted part 5. The arithmetic processing unit 203 is connected to the arithmetic unit 203A that calculates the reference stimulus value XYZ1, the arithmetic unit 203B that calculates the stimulus value XYZ2 that is the determination target, and the arithmetic unit 203A and the arithmetic unit 203B. The calculation unit 203C that performs the calculation and the exponent value from the calculation unit 203C are transmitted to the coloring device 257. While looking at the screen of the painted part 5 according to the index value, it is determined whether the paint color is appropriate by looking at the screen, and further a coloring process is performed. Note that the switch 206 selectively inputs the reference XYZ and the target XYZ. Since the main processing is generally the same as the flowcharts of the first and second embodiments, the description is incorporated.

  Next, the coloring evaluation apparatus 301 of Embodiment 4 will be described with reference to FIGS. Regarding the corresponding similar elements, the above description is cited as the 300 series, and the differences are mainly described.

  Here, the support unit 4 is replaced with a robot arm 341, and the angle of the imaging unit 302 can be automatically changed by a motor (not shown). In this embodiment, the image pickup unit 302 is moved to a place where an automobile 305 moving on the automobile production line 309 is desired to be observed, and images are taken at a plurality of locations at different angles, and the color distribution is measured. If appropriate, the automobile 305 is stopped and imaged.

  In the first embodiment, a structure that can manually change the imaging angle and evaluate the coating of the painted part 5 is configured. In the fourth embodiment, the change of the imaging angle is automated. 340 is attached to the tip of the robot arm 341 and the measurement is made online. With respect to the automobile 305, the illumination unit 306 is set to illuminate from diagonally upper portions on both sides, and the imaging unit 302 can move around the automobile 305 by a motor drive. The head 340 includes an arc-shaped housing, and an illumination unit 306 is fixed to the housing, and the imaging unit 302 can move along the arc-shaped locus along the housing.

  In the in-line processing of inspection, it is preferable to cover the illumination and the imaging unit with a light shielding cover. In Embodiments 1 to 3 described above, such a light shielding cover may not be provided. However, when the robot arm 341 performs inspection in an in-line flow in the factory, the illumination unit 306 and the imaging unit 302 itself are provided. It is preferable to cover with a light shielding hood.

  Since the body of the automobile 305 is usually determined where to see, a plurality of angles are determined. The appearance of the automobile 305 can be imaged from various angles by the imaging unit 302 in the automobile production line 309. The illumination unit 306 is an illumination that irradiates the automobile 305 from the upper side and drops the light reflected directly downward. Although the single imaging unit 302 is moved, a plurality of imaging units 302 may be installed and moved. A plurality of imaging units may be fixed to the head 340 so that the position thereof can be adjusted.

  In the case of 100% inspection in-line, when the measurement is performed more precisely, the left and right angles can be made finer. It is also possible to adopt a method in which the imaging part 5 is scanned and inspected by the imaging unit.

  As an in-line application, it can be attached to the robot arm of the production line. All products can be inspected if they are built into the production line. By combining with the robot arm machine control, a wide range of products can be inspected.

  Next, the coloring evaluation apparatus 401 according to the fifth embodiment will be described with reference to FIG. For the corresponding similar elements, the above description is cited as the 400 series, and the differences are mainly described. The coloring evaluation apparatus 401 is generally the same as that of the fourth embodiment, and includes a plurality of imaging units 402 and artificial sun lamps 406 arranged and fixed in the vicinity of the production line 409, and further, measurement PC1 to PC3 424. Connected to the main PC 410 for measurement control via the hub 408, and connected to the display device 407 directly or via the KVM 411 for measurement PC1 to PC3 424. Is. The main PC for measurement control communicates with the inline host and can send and receive data such as vehicle type and color. Instead of the imaging unit 402, a two-dimensional colorimeter may be used. Since the color of each pixel within the specified inspection range is mapped to the xy chromaticity diagram, the Lab chromaticity diagram, or the XYZ chromaticity diagram in the same manner as in the first embodiment, and the spread and density are obtained as a histogram distribution, It can be quantified including the glitter of lamé and pearl.

  Other application examples will be described. Two images of the acquired AB images of the reference product and the inspection product are overlapped, and the respective chromaticity histogram distributions are displayed on the display device 7, or the respective chromaticity histogram distributions are overlapped on one chromaticity diagram. A combined chromaticity diagram can be displayed, and a color distribution matching index indicating the degree of overlap can be displayed as a percentage. Thereby, the deviation from the reference product of the chromaticity distribution of the inspection product can be confirmed numerically. 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 7 (overlay function), alignment can be performed easily. The inspection object can be compared with the reference object 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 be displayed. In the case of color inspection of different materials such as the front door and fender of an automobile, the color difference (ΔE) from the start point to the end point can be seen to verify and inspect the color difference between different materials.

  In addition to the tristimulus values XYZ, it is possible to display the gain of each parameter of CMYK and Lab. The imaging unit uses an imaging unit 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.

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

  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 clearly and easily quantified, and the comparative inspection between the inspection object and the reference object 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 comparison by chromaticity histogram distribution, (3) Comparison by quantified numerical value by color distribution matching index (%) of chromaticity histogram is there. 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 in a non-contact manner by the imaging unit 2, there is no fear 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.

  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 coloring evaluation apparatus of the present invention has great applicability in color inspection and other industrial applications at the production site of electrical appliances, vehicles, and housing building materials. 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 of cosmetics, sportswear, shoes, golf balls, etc.

1, 101, 201, 301, 401, 501, 601, 701 ... Coloring evaluation device 2, 102, 202, 302, 402, 502, 602 ... Camera (imaging unit)
3, 103, 203, 303, 403, 503, 603 ... arithmetic processing unit 4 ... support parts 5, 105, 205, 305, 405 ... painted parts 6, 106, 206, 306, 406,. Illumination units 7, 107, 207, 307, 407 ... display device 21 ... photographic lenses 22a, 22b, 22c ... optical filter 23 ... imaging elements 22a ', 22c' ... dichroic mirror 23a , 23b, 23c ... image sensor 24 ... arithmetic processing unit 25 ... display device 26 ... reflecting mirror 27 ... filter turret 40 ... head 41 ... arm 42 ... post 43 ... Stand 42a ... Reinforcing member 44 ... Linear guide 45 ... Old member 46 ... Ramp adjuster 47 ... Imaging unit angle adjuster 48a ... Imaging unit / shift adjustment motor 48 ... Imaging unit / pan adjustment motor 49 ... Head / tilt adjustment tool 41a ... Weight 41b ... Height adjustment tool 106, 206 ... Switches 103A, 103B, 103C, 203A, 203B, 203C ..Calculation unit 207 ... Coloring device 309, 409 ... Automobile production line

Claims (7)

An imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function;
An object to be measured is a part of an assembly, the parts are individually colored, and images having three spectral sensitivities acquired by the imaging unit with respect to the colored parts are tristimulus values X, Y, Z in the CIE XYZ color system. An arithmetic processing unit that acquires and calculates the coloring data converted into
An illumination unit for irradiating the measurement object;
With
The arithmetic processing unit is
Of the coloring data obtained by imaging the measurement object, set the specified inspection area,
For the inspection object and the reference object as the measurement object, the xy value or the XYZ value itself normalized from the XYZ value of each pixel of the inspection area is calculated for the inspection area,
Create an xy chromaticity histogram distribution by dividing the inspection area of the xy coordinates of the xy chromaticity diagram with a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid , or XYZ chromaticity By dividing the inspection area of the XYZ coordinates of the figure with a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid, an XYZ chromaticity histogram three-dimensional distribution is created,
Compare the two xy chromaticity histogram distributions or the XYZ chromaticity histogram three-dimensional distributions of the inspection object and the reference object,
A coloring evaluation apparatus that selects a combination of the colored parts that approximates color and texture based on the comparison. An imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function;
An object to be measured is a part of an assembly, the parts are individually colored, and images having three spectral sensitivities acquired by the imaging unit with respect to the colored parts are tristimulus values X, Y, Z in the CIE XYZ color system. An arithmetic processing unit that acquires and calculates the coloring data converted into
An illumination unit for irradiating the measurement object;
With
The arithmetic processing unit is
Of the coloring data obtained by imaging the measurement object, set the specified inspection area,
For the inspection object and the reference object as measurement objects, the XYZ values of each pixel in the inspection area are converted to Lab,
A Lab chromaticity histogram distribution is created by dividing the inspection region of the Lab coordinate of the Lab chromaticity diagram by a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid,
Comparing the two Lab chromaticity histogram distributions of the test object and the reference object,
A coloring evaluation apparatus that selects a combination of the colored parts that approximates color and texture based on the comparison. 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 coloring evaluation apparatus according to claim 1 or 2 , wherein a color distribution coincidence index that is a ratio of the accumulated number is calculated.   The colored part includes unique information, and the unique information can be electromagnetically erased or read, and the unique information is associated with the distribution unique to the colored part. 3. Any one of the coloring evaluation apparatuses.   The coloring evaluation apparatus according to claim 1, wherein the imaging unit is attached to a support member and images an inspection object flowing on a production line. In a coloring evaluation method using an imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function,
Under illumination, the object to be measured is a component of an assembly, the components are individually colored, and an image having three spectral sensitivities obtained by imaging the colored component by the imaging unit is tristimulus in the CIE XYZ color system. Generating coloring data converted into values XYZ;
For the inspection object and the reference object as the measurement object, the step of setting the specified inspection region among the coloring data obtained by imaging the measurement object;
Calculating the xy value or the XYZ value itself normalized from the XYZ value of each pixel of the inspection area for the inspection area as the measurement object,
Create an xy chromaticity histogram distribution by dividing the inspection area of the xy coordinates of the xy chromaticity diagram with a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid , or XYZ chromaticity By dividing the inspection area of the XYZ coordinates of the figure with a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid, an XYZ chromaticity histogram three-dimensional distribution is created,
Compare the two xy chromaticity histogram distributions or the XYZ chromaticity histogram three-dimensional distributions of the inspection object and the reference object,
Selecting a combination of colored parts that approximates color and texture based on the comparison; and
A coloring evaluation method comprising: In a coloring evaluation method using an imaging unit having three spectral sensitivities (S1 (λ), S2 (λ), S3 (λ)) linearly converted equivalently to the CIE XYZ color matching function,
An object to be measured is a part of an assembly, the parts are individually colored, and an image having three spectral sensitivities acquired by imaging by the imaging unit under illumination is displayed on the colored part in the CIE XYZ color system. Generating coloring data converted into stimulus values X, Y, Z;
Of the coloring data obtained by imaging the measurement object, setting the specified inspection region,
Converting the XYZ value of each pixel in the inspection area into Lab for the inspection object and the reference object, respectively, as measurement objects;
Creating a Lab chromaticity histogram distribution by dividing an inspection region of Lab coordinates of the Lab chromaticity diagram by a grid and integrating the number of pixels of the inspection object and the reference object belonging to each grid;
Comparing the two Lab chromaticity histogram distributions of the test object and the reference object,
Selecting a combination of colored parts that approximates color and texture based on the comparison; and
A coloring evaluation method comprising: