JPH03209137A - Color difference inspection device - Google Patents

Abstract

PURPOSE:To improve the efficiency of online color difference discrimination by adjusting the inspection speed in accordance with binary color difference inspection and fuzzy color difference inspection. CONSTITUTION:With respect to a color difference value (hue value), data inputted from a color camera or a color sensor is fetched into a computer (CPU) through interface. At this time, the input voltage value of the color difference value is converted to a color space display value quantitatively expressing the color. The CPU controls an online driving part and also the carrying system power or the like of a sheet-shaped object. A fuzzy set which numerically expresses the human discrimination reference for delicate and fuzzy general color difference inspection is used and membership functions are generated from data required for color difference inspection and are used for the inspection in a boundary area where the ambiguity for online color difference check of the object is especially remarkable. Consequently, fuzzy color difference inspection is performed and the line speed of the inspection object is reduced only when the measured color difference value is between set values epsilon1 and epsilon2.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an apparatus for automatically inspecting color differences in objects by using an industrial camera, a color sensor, or the like.

(Prior Art) Color difference inspection of objects is performed using industrial cameras, color sensors, and the like. In the conventional method, the difference between the reference color (color difference value) and the color difference value of the object is judged, and the judgment value is set in advance and the color difference value (Δ
If E) is larger than the judgment value, there is a color difference (NO),
If the difference is smaller, a binary color difference determination is made in which there is no color difference (OK).

In the color difference test (color difference judgment), the results are always OK/NO.
There are cases where the value cannot be determined. In particular, the current situation is that visual inspections are performed by humans when inspecting color differences in sheet-like color distributions, such as uneven dyeing of textiles and uneven coloring of carpets.

However, visual inspection has large chronological variations depending on the conditions at the time of inspection (environment, human health, etc.), which poses a quality problem.

The present invention attempts to perform color difference inspection while maintaining ambiguity in cases where such delicate color difference inspection is required, and aims to further improve efficiency compared to conventional color difference inspection. .

(Means for Solving the Problems) The present invention solves the problems of the prior art and provides an apparatus with excellent uniformity and reliability that can discriminate within an acceptable range for inspecting subtle color differences. This is what we intend to provide. That is, the present invention is an apparatus for continuously inspecting the color difference of an object, including means (a) for detecting the color difference of the object, and a binary color difference for comparing the detected color difference value with a reference value to perform a binary color difference determination. a fuzzy color difference test means (c) for performing a fuzzy color difference judgment on the detected color difference value when the detected color difference value is between the specified color difference value ε, and ε2; The color difference inspection apparatus is characterized in that it also includes an inspection frequency control means (d) that adjusts the inspection speed according to the ratio of the inspection according to (c) and the inspection according to (c).

The present invention uses a color difference (color difference value) to judge the color of an object, and is particularly preferably applied to continuous online color inspection, and the uniformity (no smearing over time) of the equipment. This is a color difference inspection device for sheet-like objects that has the ability to judge comprehensively and within an acceptable range of ambiguity, which humans are capable of. In other words, we distinguish between the range where OK or No can be clearly judged, and the range between OK and NO that requires delicate judgment. Set the set value according to the color and material of the object to be measured, and if there is a measured color difference between the set values, perform a color difference test using fuzzy theory, and perform a color difference test for the subsequent α times (α is a number of 2 or more). When the measured color difference is between the set values and fuzzy judgment is performed, the line speed is slowed down to perform precise measurements, or in other words, to increase the number of times the object is measured (measurement frequency). .

More specifically, the measured color difference value of a sheet-like object from an industrial color camera or color sensor is input into a calculator (computer) as a quantitative value. On the other hand, using a fuzzy set that numerically expresses the judgment criteria of humans who conduct comprehensive color difference inspections, which are subtle and ambiguous, membership functions are created from the data necessary for color difference inspection, and color difference inspection of objects is performed online. Used for inspection in boundary areas where ambiguity is particularly noticeable. Therefore, the measured color difference value (Δ
E) Perform fuzzy color difference inspection only when the value is between a certain set value ε1 and ε2, and when the measured color difference value of the subsequent α times is between ε, and ε2, the line speed of the inspection object is The purpose of this is to slow down the number of color measurements and substantially increase the number of color measurements to improve accuracy.

On the other hand, if the measured color difference value is outside the range of ε1 and ε2, that is, ΔE<ε1 or ΔE>ε2, a binary color difference test is performed at a regular test speed.

(Example) FIG. 1 shows the system configuration of the color difference inspection apparatus for sheet-like objects of the present invention. Color difference values (color values) are data input from a color camera or color sensor and are taken into the CPU through an interface. At this point, the input voltage value of the color difference value is converted into a color space display value that quantitatively represents the color.

Here, the description will be made using the CIE Lab color system. The CPU also controls the online drive section, and also controls the power of the conveyance system for the sheet-like object. Furthermore, this system is comprised of a notification device (for example, a buzzer, abnormal lighting, lamp, etc.) when a color difference abnormality is detected, a recording device (hard disk, floppy disk device, etc.), and peripheral devices such as a printer, a display, and a keyboard.

Figure 2 shows an operational flowchart of this case. This will be explained in detail below. When the system is powered on and becomes operational, initial value settings are performed. (Here, initial value setting means setting the color difference meter measurement as L% a N b color system, inputting reference color data values L., a0, b0, etc.) In the example, data of a reference color is input to confirm that the mode is for measuring color difference and to determine (inspect) the color difference test of a sheet-like object online. It is assumed that the value (Lo, ao11)o) is set by inputting this reference color data from the keyboard or by measuring the color of the reference object.

At this time, the setting of the color difference determination value differs depending on the input value. This shows that the color difference value (ΔE) is not uniform throughout the Lab color space because the degree to which humans perceive color differences differs depending on, for example, red-based colors or blue-based colors. In other words, it is necessary to change the color difference determination value εr according to the input value of the reference color data. The setting of this color difference judgment value varies depending on the object whose color is to be measured, the purpose, etc., and may also vary depending on the person making the final judgment.

Therefore, we applied the fuzzy set concept to this part, created a membership function based on the inspector's know-how and experience, and used this to make the same judgments as conventional human inspections. A reference value for color difference determination is calculated by fuzzy inference.

(The membership function and inference method will be discussed later.

) The color difference determination value (εr) mentioned above is the color value L1 of the object to be determined.
The color difference value Δ is calculated from the color difference LO, ao, and bo of a1b and the reference color.
This value is used to determine whether there is a color difference compared to E. For example, if we take conventional binary determination as an example, ΔE>ε
When r, there is a color difference, and when NO1ΔE<εr, it is OK without a color difference. In addition, the standard values for color difference judgment are, for example, ΔE=2.01 for red and ΔE=2.01 for dark blue.
E = 0. A value close to 7 indicates that there is or is not a color difference. Although there are values based on conventional human judgment, the acceptable values for color difference judgment differ depending on the color and color system. Here, the standard color is L. , a0, b0,
It is necessary to set reference values for color difference determination in advance, and the values for determination differ depending on different reference colors.

Since it is derived from Losaozbo, here, the reference value for color difference determination is expressed as εI'o. Therefore, in the present invention, εr has set values ε1 and ε2 set on the tenth side with εr0 as the center, and if ΔE is smaller than ε1, there is no color difference, and if ΔE is larger than ε2, there is a color difference. When there is ΔE between ε1 and ε2, a color difference test using fuzzy theory is performed, and the colorimetric value of the sheet-like object continues to be α times ε1.
When the state of ΔE<ε2 is reached, the inspection speed is controlled to be slower than the normal inspection speed, and the color difference inspection is performed in more detail.

Note that the sensor that measures color values is operated at regular intervals. (In the example described later, color measurement is performed every 3 seconds.) Now, from the membership function, a rule for obtaining a reference value for color difference determination is created. This is represented by the product {IF-THEN~}, 81. As an example, as shown in Figure 3, we create a rule that IF L is dark (AND) a is red (AND) b is yellowish THEN Color difference judgment reference value εl'o takes a value that seems to be 1.0. . In creating this membership function and rules, we bring in the know-how and experience of inspection companies that perform actual color inspections. Next, also in the case of color difference determination fuzzy as the second step, membership functions are configured as shown in FIG. Here, membership functions such as those shown in FIGS. 4(a), (b), and (c) can be considered in order to set the input value to (ΔL1Δa1Δb) in order to perform color difference determination. Also, the final color difference judgment is an example: There seems to be no color difference/There is some color difference, but it is almost O.
It shows three types of color ship functions: K, color difference, and color difference.

Similarly to the previous section, rules are created such that, for example, IF ΔL is slightly brighter < (AND) Δa is approximately the same (AND) Δb is approximately the same THEN There is a slight color difference, but it is approximately OK. FIG. 5 shows an inference method for determining the fuzzy color difference test.

This method is a direct method that is often used in fuzzy control. To explain this diagram, first, when the input value ΔL1Δa1Δb for judgment is obtained, a rule is selected. Now let's call this the rule ■, ■. And the actual ΔL
The consequent proposition is truncated from the grade value which is multiplied by χ1 to the value of 1Δa1Δb, and a membership function is obtained by combining them.

Furthermore, from this, the center of gravity value ε. is calculated.

This fuzzy set in which ε0 belongs to the highest grade of the membership function, that is, in this example, it is determined that "there seems to be a color difference". (The fuzzy inference method is described in detail in "Fuzzy Control" by Michio Kanno (Nikkan Kogyo Shimbun), pages 78 to 83.) The system with the above configuration was experimentally verified, and the conventional 2
We confirmed that this method has superior judgment efficiency compared to the value-based color difference judgment method. This will be explained below.

The color sensor used in the experiment was a Minolta color difference meter CR-210, which was connected to a personal computer to perform operations such as data input, processing, and discrimination.

In addition, the colorimetric object selected was a dark blue color used in fabrics that require delicate color difference determination, especially men's suit fabrics. At the same time, we collected data on color difference limit samples that were actually judged, and created a membership function for color judgment based on this data. Figure 6 shows the limit color difference sample data for the navy blue system, (a) the a-b plane, (b) the a-L plane.
It is represented as a plan view. Figure 7 shows the membership function created from the speed in Figure 6, and rules were created based on this. On the other hand, men for color discrimination?
-ship function is shown in Figure 8. The platform set (variable definition range) of each membership function of ΔL1Δa1Δb was set using the color difference determination reference value εr0.

We realized the above fuzzy inference color judgment operation as software and conducted experiments. First, the method that calculates the reference value for color difference determination from fuzzy estimation obtained results that were almost 100% as expected. For this purpose, we investigated the limit color difference sample data as input values. Furthermore, the fuzzy inference for color difference judgment yielded satisfactory results, almost the same as those made by humans.

Under these conditions, we prototyped an experimental machine that can transport actual fabrics with a length of approximately 7 meters to a factory dresser, and attempted online inspections.

With CR-210, the recent repeated color measurement is 3 seconds, so 1
It was decided to measure at a rate of times/3 seconds.

Regular line inspection is 20m/min. This is a speed at which the color measurement interval for the fabric is approximately 1 m, and is considered a practical test. On the other hand, when the state of ε■×ΔE×ε2 continues α times, the inspection speed is set to be slow.As a basic condition, if α>1, there is no problem, but due to the limitations of the experimental machine, α=2. The inspection speed at that time was 10 m/min, which was 1/2 of the regular inspection speed. In other words, this means that the colorimetric interval is 5
If the speed is 0 cm and there is ambiguity in color difference judgment, that is, if α = 2 times of ε, <ΔE < ε2 state continues, set the steady inspection speed to 1/2 and increase the number of color measurements. This means that the accuracy of color difference inspection of the color distribution of a sheet-like object is improved. (Measurement at intervals of 1 m becomes measurement at intervals of 50 cm) It will be explained that the fuzzy color difference test with such a configuration is more reliable than the binary color difference test.

(FIG. 9) (a) shows the color difference value when performing binary color difference determination, and (b) shows the color difference value when performing fuzzy color difference determination on the horizontal axis. (a) is OK if it is smaller than the judgment guideline value εI'o; if it is larger, it is No.
shows. However, considering the measurement system of the colorimeter used for this judgment, that is, the variation, when the measured value takes a value very close to εro, it is possible to make an OK/NO judgment with a confidence level of 0.5 (50%). This means that you are doing the following. On the other hand, in fuzzy color judgment, the final question is whether there is a color difference or not, so the fuzzy set also says, ``There is some color difference, but it is almost OK,'' and it seems that there are four color differences.'' If the value obtained by fuzzy inference is a point on the line segment BC in figure (b), the value that belongs to either set and cannot be determined is ΔPBC. (indicated by diagonal lines).In other words, if it belongs to ΔAPB or ΔDPC, it can be determined whether there is a color difference or not, so when a point on BC is given, the reliability of the determination is that the area ratio = (ΔAPB+ΔDPC)/QAPDCB=area where color difference can be determined accurately/given by the entire grade area in which points on BC can exist.

In this experiment, the reliability is 0.88 because all membership functions are expressed as straight lines (for example, w, ■1) and the intersection point P is set to a grade value of 0.5.
7 (86.7%).

Although this is a calculation based on the assumption that binary color difference determination is ultimately performed, fuzzy color discrimination is superior in terms of reliability. Reliability can be further improved by creating a membership function that reduces this ΔPBC.

As described above, even though the reliability is affected by the shape of the membership function, color difference determination could theoretically be achieved with higher reliability than the binary color difference determination method.

(Effect of the invention) When performing color difference judgment of a sheet-like object, especially when performing color difference judgment online, because the judgment method is ambiguous, subtle color differences can be judged in the same way as humans, which is better than conventional methods. The reliability is higher than the OK/No binary color difference determination method.

When a more ambiguous color difference determination is required, the inspection speed is controlled to be slow, thereby substantially increasing the number of measurements and improving the colorimetric accuracy.

Therefore, the efficiency of online color difference determination can be improved.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of a system configuration in an embodiment of the present invention. FIG. 2 is a schematic diagram of an operation flowchart of the embodiment. FIG. 3 is a schematic diagram of an example of a membership function in the embodiment. , the same example 〃 Fig. 5 is a schematic diagram of the inference method for fuzzy color difference judgment in the same example. Fig. 6 shows the navy blue sample data in the same example. A schematic diagram of an example of a membership function. Figure 8 is an explanatory diagram for calculating efficiency from the membership function in the same example. Figure 9 is an illustration of the difference in reliability between the conventional binary judgment method and the method of the present invention. Schematic diagram for

Claims (1)

[Claims] (1) An apparatus for continuously inspecting the color difference of an object, comprising means (a) for detecting the color difference of the object, and a binary color difference testing means for comparing the detected color difference value with a reference value to make a binary color difference determination. (b)
, the detected color difference values are predefined color difference values ε_1 and ε
fuzzy color difference testing means (c) that performs a fuzzy color difference judgment on the detected color difference value when the detected color difference value is between A color difference inspection device characterized in that it also includes inspection frequency control means (d).