Color difference inspection method, color difference inspection device, and color difference inspection program

Abstract

To provide a color difference inspection method which can estimate a range of a value of a color difference that may be taken by the entire lot from a small number of samples in a hue inspection and properly evaluate a color difference.SOLUTION: A color difference inspection method by a color difference inspection device comprises: a measurement step of acquiring an observed value of each component of one or a plurality of color differences by measuring one or a plurality of inspection object items or the inspection object item and one or a plurality of standard items corresponding to the inspection object items; a storage step of storing the observed value of each component of the color difference; a calculation step of calculating an evaluation value from the observed value of each component of the color difference; and an output step of outputting the evaluation value. The calculation step sets a probability density function followed by the observed value of each component to the observed value of each component of the color difference, approximately derives the probability density function followed by a color difference evaluation expression calculated from the observed value of each component, and calculates a characteristic value of the probability density function followed by the evaluation expression as the evaluation value.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a data processing technique for estimating, from a small number of samples, the range of color difference values ​​that can be taken by a large number of other samples in a color test. [Background technology]

[0002] Conventionally, color testing has generally been performed by calculating the color difference between the color of a product manufactured sequentially and the color of a standard product. Color difference is a numerical representation of the perceived distance between two colors, and the larger the color difference, the greater the difference in color.

[0003] The color difference evaluation formula is, for example, L * a * b * Evaluation formula ΔE defined as the Euclidean distance for each component of the color space (ISO 11664-4) * ab (Z 8781-4 Equation (19)) (hereinafter sometimes referred to as ΔE) is widely used. In color quality control using a color difference evaluation formula, a predetermined threshold is set as a quality control standard for the evaluation value obtained from the evaluation formula, so that the color of the product can be controlled to fall within a certain standard value. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2009-236785 Summary of the Invention [Problem to be solved by the invention]

[0005] However, color quality control using such color difference evaluation formulas has the problem that it is not possible to estimate the range of values ​​that the entire lot can take from a small number of samples, and there are also problems with evaluations based only on the mean and variance not providing an appropriate estimate. That is, the commonly used sampling inspection using the average and standard deviation is sometimes inappropriate, especially in the range of small values, because the color difference evaluation formula is an index distributed on a half line.

[0006] On the other hand, when measuring color using a color difference meter or colorimeter to perform accurate evaluation, the measurement takes time, which limits the feasibility of conducting 100% inspection. Furthermore, when performing a simple evaluation using a normal color (RGB) camera, which can relatively quickly measure, the variability in the measurements is large, and the variability in the color information differs from that when measuring color using a color difference meter or the like. Furthermore, it was not known how to estimate the evaluation formula ΔE based on a small number of samples obtained by sampling inspection or the like.

[0007] Therefore, the inventors conducted extensive research and succeeded in developing a color difference inspection method that can calculate the probability distribution of the amount calculated as a color difference evaluation formula by treating the measured values ​​of each color difference component as stochastically sampled random numbers, i.e., as random variables. Furthermore, by finding, for example, a cumulative distribution function from the probability distribution of the color difference evaluation formula, it became possible to numerically calculate the probability of exceeding a certain color difference evaluation value.

[0008] According to the color difference inspection method of the present invention, even in a sampling inspection, it is possible to estimate the variation of the entire lot and to grasp the probability of the observable color difference including the variation of the measurement system. Moreover, since the number of samples used for calculating the probability can be small, it is possible to appropriately evaluate the risk of the color difference exceeding a certain range even when 100% inspection is not possible.

[0009] Here, Patent Document 1 discloses a color accuracy verification prediction method for a control device that verifies the color accuracy of the print colors of a printer. This method is said to enable efficient color accuracy verification and reduce unnecessary materials and labor. However, when using the color difference evaluation formula, there is a problem that the estimation of the variation obtained from the standard deviation of the measured data does not necessarily provide an appropriate evaluation. Furthermore, this color accuracy verification and prediction method does not achieve the above-mentioned effects obtained by the color difference inspection method of the present invention.

[0010] The present invention has been made in consideration of the above circumstances, and aims to provide a color difference inspection method, a color difference inspection device, and a color difference inspection program that are capable of estimating the range of color difference values ​​that can be taken by an entire lot from a small number of samples in a color inspection, and are capable of appropriately evaluating color differences. [Means for solving the problem]

[0011] In order to achieve the above object, the color difference inspection method of the present invention is a color difference inspection method using a color difference inspection device, and includes a measurement step of obtaining observed values ​​of one or more color difference components by measuring one or more inspection target products, or the inspection target products and one or more standard products corresponding to the inspection target products, a storage step of storing the observed values ​​of each of the color difference components, a calculation step of calculating an evaluation value from the observed values ​​of each of the color difference components, and an output step of outputting the evaluation value, in which the calculation step sets a probability density function that the observed values ​​of each of the color difference components follow, approximately derives a probability density function that follows a color difference evaluation formula calculated from the observed values ​​of each of the components, and calculates a characteristic value of the probability density function that the evaluation formula follows as the evaluation value.

[0012] Furthermore, it is also preferable that the color difference inspection method of the present invention is a method in which, in the calculation step, a multivariate normal distribution is set as the probability density function, a ΔE function is set as the color difference evaluation formula, the probability density function that the evaluation formula follows is approximated as a quadratic form of a random variable vector that follows the multivariate normal distribution, and a characteristic value of the probability density function that the evaluation formula follows is calculated as the evaluation value.

[0013] Furthermore, it is also preferable that the color difference inspection method of the present invention be a method in which, in the measuring step, a region of similar colors within an inspection target is measured at a plurality of locations to obtain observed values ​​of one or more color difference components, and in the calculating step, a characteristic value of a probability density function that the evaluation formula follows is calculated as the evaluation value, thereby evaluating a color dispersion range that can be taken by the entire region of similar colors. It is also preferable that the color difference inspection method of the present invention is a method for calculating a cumulative distribution function as the characteristic value of the probability density function to which the evaluation formula conforms.

[0014] Furthermore, a color difference inspection device of the present invention comprises a memory unit that stores observed values ​​of one or more color difference components obtained by measuring an inspection target product and one or more standard products corresponding to the inspection target product, a calculation unit that calculates an evaluation value from the observed values ​​of each of the color difference components, and an output unit that outputs the evaluation value, wherein the calculation unit sets a probability density function that the observed values ​​of each of the color difference components follow, approximately derives a probability density function that the color difference evaluation formula calculated from the observed values ​​of each of the components follows, and calculates a characteristic value of the probability density function that the evaluation formula follows as the evaluation value.

[0015] It is also preferable that the color difference inspection device of the present invention further comprises a measurement unit that acquires observed values ​​of one or more color difference components by measuring one or more inspection target products, or the inspection target products and one or more standard products corresponding to the inspection target products.

[0016] Furthermore, the color difference inspection program of the present invention is configured to cause a color difference inspection device to set a probability density function to which the observed values ​​of one or more color difference components conform for the observed values ​​of the components obtained by measuring one or more inspection target products, or the inspection target products and one or more standard products corresponding to the inspection target products, approximately derive a probability density function to which a color difference evaluation formula calculated from the observed values ​​of the components conforms, calculate a characteristic value of the probability density function to which the evaluation formula conforms, and output the characteristic value.

[0017] Furthermore, it is also preferable that the color difference inspection program of the present invention is configured to cause a color difference inspection device to acquire observed values ​​of one or more color difference components based on measurements of one or more inspection target products, or the inspection target products and one or more standard products corresponding to the inspection target products, and to store the observed values ​​of each of the color difference components. [Effects of the Invention]

[0018] According to the present invention, it is possible to provide a color difference inspection method, a color difference inspection device, and a color difference inspection program that are capable of estimating the range of color difference values ​​that can be taken by an entire lot from a small number of samples in a color inspection and appropriately evaluating color differences. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a block diagram showing a configuration of a color difference inspection device according to an embodiment of the present invention. [Figure 2] 4 is a flowchart showing a processing procedure performed by the color difference inspection device according to the embodiment of the present invention. [Figure 3] FIG. 1 is a diagram showing a simulation result of Example 1 using the color difference test device according to the embodiment of the present invention. [Figure 4] FIG. 10 is a diagram showing a simulation result of Example 2 using the color difference test device according to the embodiment of the present invention. [Figure 5] FIG. 10 is a diagram showing a simulation result of Example 3 using the color difference test device according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0020] Hereinafter, embodiments of the color difference inspection method, color difference inspection device, and color difference inspection program of the present invention will be described in detail. However, the present invention is not limited to the following embodiments and the specific contents of the examples described later.

[0021] First, the color space and color difference evaluation formula in the color difference inspection method of this embodiment will be described. L * a * b * A color space is a three-dimensional space standardized by the International Commission on Illumination (CIE) in 1976 to represent the color of an object numerically. * a * b * In color space, lightness is L * chromaticity, which represents hue and chroma, is a * and b * Displayed by.

[0022] Color difference is the difference between two colors. The following formula, which is defined as the Euclidean distance for each component of color difference, is widely used to evaluate this color difference. The subscripts [.] for lightness and chromaticity in the formula are: (q) (q=1, 2) represent the values ​​of the first and second colors, respectively. Note that there are various other color difference evaluation formulas, and the calculation operations in the embodiments described below can be applied to any evaluation formula expressed in a quadratic form, similar to ΔE, but are not limited to this.

number

[0023] The color difference inspection method of this embodiment is a color difference inspection method using a color difference inspection device, and is characterized by having the following steps: Note that the details of the processing in each step will be described in detail in the explanation of the color difference inspection device. (A) A measurement step of measuring one or more test samples, or the test samples and one or more standards corresponding to the test samples, to obtain observed values ​​of one or more color difference components. (B) A storage step for storing the observed values ​​of each color difference component. (C) A calculation step for calculating an evaluation value from the observed values ​​of each color difference component. (D) Output process to output the evaluation value

[0024] In the color difference inspection method of this embodiment, it is preferable to perform the following processing in the calculation step. (C1) For the observed values ​​of each component of color difference, a probability density function to which the observed values ​​of each component follow is set. (C2) Approximately derive the probability density function that the color difference evaluation formula calculated from the observed values ​​of each component follows. (C3) A characteristic value of the probability density function that the evaluation formula follows is calculated as an evaluation value.

[0025] Furthermore, in the color difference inspection method of this embodiment, it is more preferable to perform the following processing in the calculation step. (C11) Set the multivariate normal distribution as the probability density function. (C21) A ΔE function is set as a color difference evaluation formula. (C22) The probability density function to which the evaluation formula conforms is approximated as a quadratic form of a vector of random variables that conform to a multivariate normal distribution. (C31) A characteristic value of the probability density function that the evaluation formula follows is calculated as an evaluation value. (C311) Calculate the cumulative distribution function as a characteristic value of the probability density function that the evaluation formula follows.

[0026] Furthermore, in the color difference inspection method of this embodiment, it is also preferable that in the measurement step, areas of similar colors within the product to be inspected are measured at multiple locations to obtain observed values ​​of one or more color difference components, and in the calculation step, a characteristic value of the probability density function that the evaluation formula follows is calculated as an evaluation value, thereby evaluating the color dispersion range that the entire area of ​​similar colors can take. At this time, it is more preferable to calculate a cumulative distribution function as a characteristic value and evaluate the color dispersion range that can be taken by the entire area of ​​similar colors.

[0027] Here, the color difference inspection method of this embodiment is a method that can calculate the color difference exceedance risk of the entire lot (product accumulation unit) including products that have not been selected from products that have been randomly (thinned out) selected. However, due to the limitations of color measurement devices such as color difference meters, the measurement range is limited (for example, the measurement range is a circle with a diameter of a few millimeters), making it difficult to evaluate every part of the product. In such cases, measurable locations must be selected as "representative points" and measurement data must be obtained.

[0028] In such a situation, according to the color difference inspection method of the present embodiment, it is possible to perform a risk assessment of the entire area (of the same color) of a product based on color values ​​measured at multiple locations, by regarding the in-plane color variation, including locations that were not measured, as the lot.

[0029] As shown in FIG. 1, the color difference inspection device of this embodiment includes an inspection object product measurement unit 10, a standard product measurement unit 11, a color information storage unit 12, a calculation unit 13, a probability distribution setting unit 14, an evaluation formula setting unit 15, and an output unit 16. The inspection object measurement unit 10 and the standard object measurement unit 11 can be configured, for example, with a color difference meter, a colorimeter, or a normal color camera, and can measure the colors of the inspection object and the standard object, respectively, and store each color information in the color information storage unit 12.

[0030] Here, the difference between evaluation using a color difference meter or the like and simple evaluation using a normal color camera will be explained. Simple evaluation is an evaluation system that does not have equipment for collecting diffused light, such as an integrating sphere, and uses a regular color camera as a measuring device. In simple evaluation, even if the lighting angle is fixed, the amount of reflected light that reaches the light receiver (camera) increases or decreases depending on the surface characteristics (unevenness and tilt) of the object. In addition, regular color cameras convert transmitted light through a filter into RGB brightness values, making it difficult to correct for spectral changes in the light source. Furthermore, evaluations using colorimeters that perform accurate color measurements use the average of 5 or 10 measurements as the measurement value, whereas inspection lines using cameras typically only take a single image, which means that the amount of random error from trials is relatively large. Therefore, when using color information from a color difference meter or other device designed to determine accurate color, there will be a difference in the variability of the calculated color information compared to when using color information from a camera commonly used in product inspection.

[0031] In this embodiment, both a color difference meter, a colorimeter, or the like, and an ordinary color camera can be used as the inspection target product measurement unit 10 and the standard product measurement unit 11. If data from a color difference meter is used, an approximate distribution with a narrowed prediction range and little variance can be obtained, whereas if data acquired using a camera is used, the data obtained will be in a wide range, and therefore the range of color differences estimated from the approximate distribution will also be wide. According to this embodiment, even when data from a color difference meter or colorimeter is used to determine accurate colors, the range of color difference values ​​that the entire lot can take can be estimated from a small number of samples, making it possible to appropriately evaluate color differences.

[0032] In the color difference inspection device 1 of this embodiment, the inspection target product measurement unit 10 and the standard product measurement unit 11 do not need to be physically separate units, and the same device can be used for both. The color information storage unit 12, calculation unit 13, probability distribution setting unit 14, evaluation formula setting unit 15, and output unit 16 can be provided in, for example, a microcomputer in a color difference meter. Furthermore, these components may be provided in an information processing device such as a computer, tablet, or smartphone.

[0033] The color information storage unit 12 stores the observed values ​​of each component of color difference input from the inspection target product measurement unit 10 and the standard product measurement unit 11. A memory, a hard disk, an SSD, etc. may be used as the color information storage unit 12. Also, online storage such as the cloud may be used.

[0034] The calculation unit 13 includes a color information distribution parameter estimation unit 131 , a color difference evaluation formula approximate distribution calculation unit 132 , and a characteristic value calculation unit 133 . The color information distribution parameter estimation unit 131 receives the type of probability distribution function for obtaining the color information distribution from the probability distribution setting unit 14. It also receives color information from the color information storage unit 12. Then, based on this information, it estimates the parameters (population variables) used in the probability density function.

[0035] Specifically, when a multivariate normal distribution is input as the probability density function, the color information distribution parameter estimation unit 131 estimates a mean vector μ and a covariance matrix Σ as parameters. In this case, the mean vector μ can be estimated by averaging each component of the color information, and the covariance matrix Σ can be estimated by first calculating the mean vector μ and then averaging the tensor products of the difference vectors between each component and the mean vector μ.

[0036] The color difference evaluation formula approximate distribution calculation unit 132 receives the color difference evaluation formula from the evaluation formula setting unit 15. It also receives the parameters estimated from the color information distribution parameter estimation unit 131. Based on this information, it calculates an approximate distribution of the probability density function to which the color difference evaluation formula calculated from the observed values ​​of each component conforms. Specifically, when a ΔE function is input as the color difference evaluation formula, the color difference evaluation formula approximate distribution calculation unit 132 can approximate the probability density function that the evaluation formula follows as a quadratic form of a random variable vector that follows a multivariate normal distribution.

[0037] The characteristic value calculation unit 133 calculates the characteristic value of the probability density function to which the color difference evaluation formula conforms as the evaluation value. Specifically, for example, a cumulative distribution function can be calculated as a characteristic value of the probability density function that the evaluation formula follows. This makes it possible to obtain information such as "there is a risk that the color difference will be greater than the reference value with a probability of ○%" corresponding to the value of the color difference evaluation formula.

[0038] The above processing by the color difference evaluation formula approximate distribution calculation unit 132 and the characteristic value calculation unit 133 can be performed by, for example, the following calculation. First, let us assume that color 1 and color 2 are random variable vectors that follow a multivariate normal distribution. (q) This means that μ is expressed by the following formula: (q) and Σ (q) are x (q) are the mean vector and covariance matrix for

number

[0039] x (q) Assuming that is a multivariate normal distribution, the difference vector δ between two colors follows the multivariate normal distribution shown below due to the reproductive property of normal distribution.

number

[0040] Therefore, the square of the color difference evaluation formula is expressed as the following quadratic form with respect to the difference vector:

number

[0041] The exact probability distribution in the above equation must be solved approximately because the probability density function cannot be found in a closed form. Mathai, Provost, et al. have derived the exact cumulants and moments for a general quadratic form of a random variable that follows a multivariate normal distribution (Mathai, AM & Provost, SB (1992). Quadratic Forms in Random Variables, Theory and Applications, Marcel Dekker Inc., New York.), which are called ΔE 2 When applied to the following equations, (0) =1.

[0042] The sth-order cumulant κ(s) is

number

number

[0043] Hyung, Provost, et al. have proposed a method for creating an approximate distribution based on the results of the moments calculated by the above formula (Hyung TAEHA & Provost, SB (2013). An accurate approximation to the distribution of a linear combination of non-central chi-square random variables, REVSTAT - Statistical Journal Volume 11, Number 3, November 2013, 231-254). In the proposed method, δ T Note that δ is positive definite, and the probability density function of the approximate distribution f Yd(y) can be written as follows using d-term Laguerre polynomials:

number

[0044] Now, the following calculation can be performed:

number

number

number

[0045] Also, the cumulative distribution function F Yd (c0) is expressed as follows: In addition, Γ (x) is the gamma function, Γ (x,α) is the lower incomplete gamma function of the second kind.

number

[0046] From the above equation, the probability density function f of the approximate distribution that the random variable Z of the color difference evaluation formula ΔE follows is z (z) and cumulative distribution function F z (z) is derived using the following formula:

number

[0047] The probability distribution setting unit 14 sets the type of probability distribution function for obtaining the color information distribution, and outputs the probability distribution function to the color information distribution parameter estimation unit 131. This probability distribution function is set by the probability distribution setting unit 14 based on selection information designated by a person. At this time, the probability distribution setting unit 14 sets a probability density function to which the observed value of each component of color difference follows, for the observed value of each component.

[0048] The evaluation formula setting unit 15 sets a color difference evaluation formula and outputs it to the color difference evaluation formula approximate distribution calculation unit 132. This color difference evaluation formula is set by the evaluation formula setting unit 15 based on selection information designated by a person.

[0049] The output unit 16 outputs the evaluation value. That is, the output unit 16 outputs the characteristic value calculated by the characteristic value calculation unit 133 as the evaluation value. The output unit 16 can also output the approximate distribution calculated by the color difference evaluation formula approximate distribution calculation unit 132. Of course, the output unit 16 may output the parameters estimated by the color information distribution parameter estimation unit 131, and the information output by the output unit 16 is not particularly limited. The output unit 16 can output the output information to a display or monitor of a color difference meter, an information processing device, etc. The output information may also be output to a printer or the like.

[0050] Next, the processing procedure performed by the color difference inspection device of this embodiment will be described with reference to FIG. First, the inspection target measuring unit 10 of the color difference inspection device 1 measures the color of the inspection target, and stores the color information in the color information storage unit 12 (step 10). Furthermore, the standard product measuring unit 11 measures the color of the standard product and stores the color information in the color information storage unit 12 (step 11).

[0051] The order of steps 10 and 11 may be reversed. Alternatively, step 11 may be performed in advance to store the color information of the standard product in color information storage unit 12, and thereafter only step 10 may be performed without performing step 11.

[0052] Next, the probability distribution setting unit 14 sets the type of probability distribution function for obtaining the color information distribution, and outputs the probability distribution function to the color information distribution parameter estimation unit 131 (step 12). Then, the color information distribution parameter estimation unit 131 estimates parameters to be used in the probability density function based on the probability distribution function and the color information (step 13).

[0053] Furthermore, the evaluation formula setting unit 15 sets a color difference evaluation formula and outputs it to the color difference evaluation formula approximate distribution calculation unit 132 (step 14). Then, the color difference evaluation formula approximate distribution calculation unit 132 calculates the approximate distribution of the probability density function that the color difference evaluation formula follows (step 15).

[0054] Furthermore, the characteristic value calculation unit 133 calculates the characteristic value of the probability density function to which the color difference evaluation formula conforms (step 16). Then, the output unit 16 outputs these calculation results (step 17). It should be noted that the setting of the probability density function in step 12 and the setting of the color difference evaluation formula in step 14 do not need to be performed every time the characteristic values ​​are calculated unless there is an intention to change them.

[0055] The color difference test device of the above embodiment can be realized by using a computer controlled by the color difference test program of the present invention. The CPU of the computer sends commands to each component of the computer based on the color difference test program, causing the components to perform predetermined processes required for the operation of the color difference test device, such as estimating color information distribution parameters, calculating an approximate distribution of a color difference evaluation formula, and calculating characteristic values. In this way, each process and operation in the color difference test device of the present invention can be realized by specific means in which the program and the computer work together.

[0056] The program is stored in advance on a recording medium such as ROM or RAM, and is executed by having the computer read the program from the recording medium installed in the computer, but it can also be read into the computer via a communication line, for example. The recording medium for storing the program can be configured as any recording means readable by any computer, such as a semiconductor memory, a magnetic disk, an optical disk, or the like.

[0057] As described above, this embodiment makes it possible to probabilistically evaluate the range in which color differences can be measured for a target product group from actual measurements of a small number of color spaces, making it suitable for use in sampling inspections of products that are produced in large quantities at high speed. Here, the commonly used sampling inspection using the mean and standard deviation may not be appropriate, particularly in a small value range, because the color difference evaluation formula is an index that is distributed on a half line. According to this embodiment, even in such a case, it is possible to estimate the variation of the entire lot, including the products that were not selected, and to appropriately calculate the color difference excess risk.

[0058] Furthermore, according to this embodiment, it is possible to grasp the probability of observable color differences, including variations in the measurement system. Variations due to the measurement system generally occur due to instability of the light source (illumination), optical path, sensor, etc., used to obtain the spectrum of reflected light from the target. The reason why this embodiment can calculate the observable color difference, including the variability of the measurement system, is that μ and Σ calculated from the measured data include bias and variability of the data, and a multivariate normal distribution is assumed based on these, and the distribution of color differences is a probability distribution of norms (color differences) that can be calculated from a probability vector based on the multivariate normal distribution, so it can be said to be a distribution of observable color differences that includes the variability of the measurement system.

[0059] Furthermore, according to this embodiment, the number of samples used to calculate the probability can be small, so even if full inspection is not possible, it is possible to appropriately evaluate the risk that the color difference will exceed a certain range. In other words, in this embodiment, a model expressed with a small number of parameters, such as a multivariate normal distribution with a mean vector μ and a covariance matrix Σ, can be used, making it possible to perform probabilistic evaluation using a small number of samples.

[0060] Furthermore, this embodiment has a high affinity with the output of statistical machine learning. That is, if there is already a system that predicts color information using some kind of machine learning algorithm, the predicted value output by this system will have an expected value (mean) and a prediction error (variance). By replacing the mean and variance calculated from actual measurement data with these values, it becomes possible to calculate the probability distribution of the color difference that can be taken between the predicted value output by statistical machine learning and the actual value. Specifically, it is possible to suitably apply a method that can output both the mean and variance of the predicted value, such as Gaussian process regression. [Example]

[0061] A detailed description will now be given of a simulation performed to confirm the effects of the color difference inspection method, color difference inspection device, and color difference inspection program according to the embodiment of the present invention. In the following examples, the test product and the standard product were not actually measured, and raw color information data was not input. Instead, μ and Σ were set as examples of multivariate normal distributions that would be calculated from the data, and the processing from step 15 onwards was carried out.

[0062] [Example 1] A multivariate normal distribution was set as the probability distribution, and a ΔE function was set as the color difference evaluation formula. In addition, μ and Σ were set as follows:

number

[0063] The probability density function to which the evaluation formula conforms is then approximated by the above-mentioned method as a quadratic form of a random variable vector conforming to a multivariate normal distribution, thereby obtaining a probability density function. In addition, we obtained a cumulative distribution function as a characteristic value of the probability density function that the evaluation formula follows. These were then output to a graph, and the results are shown in Figure 3. The cumulative distribution function values ​​for ΔE=2.5 were as follows: F Z (2.5)=0.998041..

[0064] [Example 2] A multivariate normal distribution was set as the probability distribution, and a ΔE function was set as the color difference evaluation formula. In addition, μ and Σ were set as follows:

number

[0065] The probability density function to which the evaluation formula conforms is then approximated by the above-mentioned method as a quadratic form of a random variable vector conforming to a multivariate normal distribution, thereby obtaining a probability density function. In addition, we obtained a cumulative distribution function as a characteristic value of the probability density function that the evaluation formula follows. These were then output to a graph, and the results are shown in Figure 4. The cumulative distribution function values ​​for ΔE=2.5 were as follows: F Z (2.5)=0.999823..

[0066] [Example 3] A multivariate normal distribution was set as the probability distribution, and a ΔE function was set as the color difference evaluation formula. In addition, μ and Σ were set as follows:

number

[0067] The probability density function to which the evaluation formula conforms is then approximated by the above-mentioned method as a quadratic form of a random variable vector conforming to a multivariate normal distribution, thereby obtaining a probability density function. In addition, we obtained a cumulative distribution function as a characteristic value of the probability density function that the evaluation formula follows. These were then output to a graph, and the results are shown in Figure 5. The cumulative distribution function values ​​for ΔE=2.5 were as follows: F Z (2.5)=0.998259..

[0068] As shown in Figures 3(A), 4(A), and 5(A), according to this embodiment, it is possible to plot a probability density function equivalent to a histogram created from a large amount of data. By creating a histogram, it becomes possible to determine that "there is a risk that the color difference will be greater than the reference value with a probability of x%, but to obtain an accurate value for this determination, it is necessary to collect hundreds of thousands of data points.

[0069] In other words, as is clear from these figures, observing large color differences is a rare event, and so in order to provide a well-founded numerical value for a rare event (obtain multiple samples), it is necessary to observe other samples that occur frequently as well. Therefore, if there is not enough data to collect for the histogram, it is usually not possible to create the upper graph. In this embodiment, by estimating μ and Σ from finite data, it is possible to know the distribution of ΔE without collecting hundreds of thousands of actual data to create a histogram.

[0070] In the example of Figure 3-5, when an approximate distribution is created by gradually increasing the number of samples to estimate μ,Σ from 4, 8, 16, 32, 64,..., 1024, μ,Σ close to the true value is obtained from 128 samples onwards. Therefore, in these cases, the number of samples can be said to be reduced to about 1 / 1000. As described above, according to this embodiment, it is possible to significantly reduce the amount of work required to collect actual data, and it is also possible to reduce the cost of color testing.

[0071] The present invention is not limited to the above-described embodiment, and various modifications are possible within the scope of the present invention. For example, it is possible to use a color difference evaluation formula other than ΔE as the color difference evaluation formula. [Industrial Applicability]

[0072] The present invention can be suitably used when calculating the color difference exceedance risk for the entire lot, including products that were not selected from randomly selected products. [Explanation of symbols]

[0073] 1. Color difference inspection device 10. Inspection object measurement section 11 Standard product measurement section 12 Color information storage section 13 Calculation section 131 Color information distribution parameter estimation unit 132 Color difference evaluation formula approximate distribution calculation part 133 Characteristic Value Calculation Unit 14 Probability distribution setting section 15 Evaluation formula setting section 16 Output section

Claims

1. A color difference inspection method using a color difference inspection device, comprising: a measuring step of measuring one or more test objects, or the test objects and one or more standard objects corresponding to the test objects, to obtain observed values ​​of one or more color difference components; a storage step of storing the observed values ​​of each component of the color difference; a calculation step of calculating an evaluation value from the observed values ​​of each component of the color difference; an output step of outputting the evaluation value, In the calculation step, setting a probability density function to which the observed value of each component of the color difference follows for the observed value of each component; Approximately deriving a probability density function according to a color difference evaluation formula calculated from the observed values ​​of each of the components; A characteristic value of a probability density function according to the evaluation formula is calculated as the evaluation value. A color difference inspection method characterized by the above.

2. In the calculation step, A multivariate normal distribution is set as the probability density function, A ΔE function is set as the color difference evaluation formula, Approximating the probability density function that the evaluation formula follows as a quadratic form of a random variable vector that follows a multivariate normal distribution; A characteristic value of a probability density function according to the evaluation formula is calculated as the evaluation value.

2. The color difference inspection method according to claim 1.

3. In the measuring step, a plurality of areas of similar colors within the inspection object are measured to obtain observed values ​​of one or more color difference components; In the calculation step, a characteristic value of a probability density function according to the evaluation formula is calculated as the evaluation value, thereby evaluating the color dispersion range that can be taken by the entire area of ​​similar colors.

3. The color difference inspection method according to claim 1 or 2.

4. Calculating a cumulative distribution function as the characteristic value of the probability density function according to the evaluation formula 4. The color difference inspection method according to claim 1, wherein the color difference is detected by the color detecting means.

5. A color difference inspection device, a storage unit that stores observed values ​​of one or more color difference components obtained by measuring the test object and one or more standard products corresponding to the test object; a calculation unit that calculates an evaluation value from the observed values ​​of each component of the color difference; an output unit that outputs the evaluation value, The calculation unit: setting a probability density function to which the observed value of each component of the color difference follows for the observed value of each component; Approximately deriving a probability density function according to a color difference evaluation formula calculated from the observed values ​​of each of the components; A characteristic value of a probability density function according to the evaluation formula is calculated as the evaluation value. A color difference inspection device characterized by the above.

6. The apparatus further includes a measurement unit that acquires observed values ​​of one or more color difference components by measuring one or more test items, or the test items and one or more standard items corresponding to the test items.

6. The color difference inspection device according to claim 5.

7. A color difference inspection program, Color difference inspection equipment, setting a probability density function to be followed by the observed value of each of one or more color difference components obtained by measuring one or more test items, or the test items and one or more standard items corresponding to the test items; Approximately deriving a probability density function according to a color difference evaluation formula calculated from the observed values ​​of each of the components; Calculating a characteristic value of a probability density function that the evaluation formula follows, and outputting the characteristic value A color difference inspection program characterized by:

8. Color difference inspection equipment, Obtaining observed values ​​of one or more components of color difference based on measurements of one or more test items, or the test items and one or more standard items corresponding to the test items, and storing the observed values ​​of the components of color difference 8. The color difference inspection program according to claim 7.

Images