JP2004144545A - Inspection method of color irregularity - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for performing quantitatively and quickly quality determination of color irregularity agreeing well with visual determination. <P>SOLUTION: This method is characterized as follows: reflected light acquired by irradiating the surface of an inspection object with light is photographed by a color camera; the output signal thereof is converted into chromaticity; the color difference between the determined chromaticity and a reference chromaticity and the frequency distribution thereof are determined relative to the whole inspection region; the frequency distribution is compared with a threshold curve wherein a frequency threshold is high in a small color difference set beforehand and the frequency threshold is suddenly decreased up to zero following increase in the color difference; and when the whole frequency distribution is below the threshold curve, it is determined that the color irregularity does not exist in the inspection object, and when the distribution has a part not below the threshold curve, it is determined that the color irregularity exists. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for inspecting color unevenness. More specifically, the present invention relates to a method for inspecting color unevenness of an object to be inspected such as a paint film, a film, a glass substrate, a multilayer thin film, etc., which quantitatively and quickly makes a determination that matches well with visual determination.
[0002]
[Prior art]
As a method for inspecting color unevenness, a method using image data obtained by photographing an inspection target is widely used. A primary image of the inspection target is captured by a color CCD camera, and three primary colors corresponding to three primary color light amounts are used. Data is generated for each pixel, the three primary color data is processed for each pixel to calculate uneven color distribution data, and a secondary image representing the uneven color distribution of the primary image is displayed on the monitor based on the uneven color distribution data, A method for making it visible (refer to Patent Document 1), calculating a color difference between the obtained three primary color data and a reference color for each pixel, and dividing the color difference data for each pixel into a grid composed of a predetermined number of pixel matrices. , Add the chrominance data of each pixel in each of the grids, and determine the change amount of the added value of the chrominance data between the respective grids in each of the horizontal direction and the vertical direction, and this change amount is larger than a predetermined amount. The case, the method determines that the color unevenness (see Patent Document 2) are proposed.
[0003]
[Patent Document 1]
JP-A-5-332838 (paragraph [0005])
[0004]
[Patent Document 2]
JP-A-8-138052 (Paragraph [0008])
[0005]
[Problems to be solved by the invention]
However, the method described in Patent Literature 1 can make the color unevenness distribution visible, but the determination of the color unevenness is based on visual observation. In other words, this is a method in which there is variation depending on the judge. Further, the method described in Patent Document 2 is a method for enhancing slight color unevenness that cannot be detected from a difference between pixels to facilitate detection, and is well suited to visual determination in consideration of the degree and frequency of color unevenness. It is not a method to make a judgment quantitatively and quickly.
[0006]
An object of the present invention is to provide a method for quantitatively and quickly performing a determination that is well suited to a visual determination in consideration of the degree and frequency of color unevenness of an inspection target.
The present inventors have intensively studied an inspection method of color unevenness in order to solve such a problem, and as a result, obtained a frequency distribution of a color difference between a chromaticity of an inspection object and a reference chromaticity, and obtained this frequency distribution and a small color difference. The present inventors have found that the object can be achieved by comparing and judging a threshold curve in which the frequency threshold is high and the threshold of the frequency rapidly decreases as the color difference becomes large and becomes zero, and the present invention has been achieved.
[0007]
[Means for Solving the Problems]
That is, the present invention captures reflected light obtained by irradiating light on the surface of the inspection object with a color camera, converts the output signal into chromaticity,
For the entire inspection area, the color difference between the obtained chromaticity and the reference chromaticity and the frequency distribution thereof are obtained,
This frequency distribution is compared with a threshold curve in which the threshold of the frequency is high in a small color difference set in advance and the threshold of the frequency rapidly decreases to zero as the color difference becomes large,
If the entire frequency distribution is equal to or less than the threshold curve, it is determined that there is no color unevenness in the inspection object, and if there is a portion that does not fit in the threshold curve, it is determined that there is color unevenness. Is the way.
With this method, it is possible to quantitatively and promptly determine color unevenness that matches well with visual determination.
[0008]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 is a diagram showing an example of an apparatus configuration used in the method for inspecting color unevenness of the present invention. FIG. 2 is a flowchart showing an example of a process for inspecting color unevenness of the present invention.
The inspection object 1 is irradiated with light 21 from the light source 2. As a light source, a halogen lamp, a fluorescent lamp, a xenon lamp or the like is usually used.
The reflected light 22 from the inspection object is photographed by the color camera 3 (S2). Usually, a color CCD camera is used as the camera. The reflected light captured by the color CCD camera is normally output as an electrical signal 23 of three primary colors (RGB) of red, green, and blue.
[0009]
The electric signal is sent to the image input unit 11 of the inspection operation unit 5. The check operation unit is usually constructed as software using a computer system. In the image input unit, the output signal 23 is digitized. Next, the output signal is converted into chromaticity by the chromaticity calculation unit 12 (S3). If the output signal from the color CCD camera is a digital signal, it is directly converted into chromaticity by the chromaticity calculation unit.
The chromaticity perceived at that time is uniquely determined from the spectral distribution of the reflected light (for example, Goro Kuwahara et al., "Optical Technology" published by Kyoritsu Shuppan).
The color system used for the chromaticity display is not particularly limited, and includes xy chromaticity, u'v 'chromaticity, L ^* a ^* b ^* color space, and the like.
[0010]
This chromaticity is usually obtained for each pixel in the entire inspection area. As the inspection area, the entire image captured by the camera, the entire image obtained by continuous imaging, or a part of the captured image may be set, and is appropriately selected depending on the purpose.
[0011]
In the reference chromaticity setting unit 14, a reference chromaticity is set in advance (S1). As the reference chromaticity, a chromaticity required for an inspection object, that is, a chromaticity set in advance or a chromaticity obtained for a standard sample is used.
Next, the color difference between the chromaticity obtained for each pixel of the entire inspection area and the reference chromaticity and its distribution are obtained (S4).
From the obtained color difference distribution, a color difference frequency distribution is obtained (S5).
[0012]
When visually evaluating color unevenness, it is determined that there is color unevenness if there is at least one place where the color difference is large, and it is determined that the color difference is relatively small, and if there are few such places, there is no color unevenness. If there are a large number of locations, it is determined that there is color unevenness.
In the present invention, in advance, the threshold value of the frequency is high for a small color difference, and the threshold value of the frequency is rapidly reduced and becomes zero as the color difference becomes large. Compare with the threshold curve, if the entire frequency distribution is equal to or less than the threshold curve, it is determined that there is no color unevenness, and if there is a portion of the frequency distribution that does not fall within the threshold curve, it is determined that there is color unevenness, The result is output (S6).
The inspection target is sorted based on this output.
The specific value of the threshold curve is appropriately set according to the criterion.
These calculations and determinations are quickly performed by a computer system, and are quantitatively determined by a threshold curve.
[0013]
【Example】
Hereinafter, the present invention will be described specifically with reference to Examples, but the present invention is not limited thereto.
[0014]
Example 1
On an acrylic film substrate (set thickness: 1 mm, refractive index: 1.49), a second layer (set thickness: 90 nm, refractive index: made of an ultraviolet-curable resin containing ZrO ₂ microparticles) sequentially from the substrate side. 1.72), a two-layer film formed of a first layer (set thickness: 80 nm, refractive index: 1.44) made of a thermosetting silane resin was inspected for color unevenness.
The surface of the two-layer film is irradiated using a fluorescent light source as a light source, and the reflected light is emitted from a color CCD camera (manufactured by SONY Corporation: XC-003) from an angle of 0 degrees from the normal to the surface of the two-layer film. Taken in. The digital signal from the color CCD camera is converted into chromaticity, and then the color difference from the chromaticity (reference chromaticity) previously obtained for the standard sample is obtained for each pixel of the entire photographed image (inspection area), and the distribution is obtained. Obtained. The result is shown in FIG.
Next, the frequency distribution of this color difference was determined. The result is shown in FIG.
[0015]
In FIG. 4, a dotted line indicates a threshold curve in which the threshold value of the frequency is high for a small color difference obtained in advance, and the threshold value of the frequency rapidly decreases to zero as the color difference increases. Specifically, if the color difference is 0.12 or less, the frequency threshold is 200; if the color difference is 0.14, the frequency threshold is 100; if the color difference is 0.16, the frequency threshold is 50; When the threshold is 25 and the color difference is 0.20, the frequency threshold is set to 12, when the color difference is 0.22, the frequency threshold is set to 6, and when the color difference is 0.24 or more, the frequency threshold is set to 0. In this case, since there is a portion in the frequency distribution of the color difference that does not fall below the threshold curve, it is determined that there is color unevenness.
[0016]
FIGS. 5 and 6 show the results obtained in the same manner for another two-layer film. FIG. 5 shows a color difference for each pixel, and FIG. 6 shows a frequency distribution of the color difference and a threshold curve. In this case, since the entire frequency distribution is equal to or smaller than the threshold curve, it is determined that there is no color unevenness.
[0017]
Reference Example 1
Using computer graphics, a total of 27 types of color shading images, including three types of color shading intensity, three types of color shading, and three types of color shading area width, and a reference non-defective image, for a total of 28 types of images. They were prepared and visually evaluated for each pair.
Two people felt that the color unevenness was strong, 1 point when they felt the color unevenness was 0, and 0 when they felt the color unevenness was weak. The evaluation score was added to quantify the color unevenness of each image. A non-defective evaluation score was 10 points, and an image in which color unevenness was most strongly felt was 49 points. An image with an evaluation point of 10 to 15 points is good with almost no color unevenness, and an image with an evaluation point of 40 or more has a strong color unevenness and is judged to be clearly bad.
[0018]
Next, the frequency distribution of the color difference between the chromaticity of 27 types of images having color unevenness and the chromaticity of a non-defective image (reference chromaticity) was determined. Some typical examples are shown in FIGS. In the figure, the horizontal axis represents the color difference, and the vertical axis represents the frequency (the number of pixels in the image). Also, for example, in the frequency distribution diagram of the color difference in the upper part of FIG. 7, M112 has one color unevenness (the first 1 indicates it), and its intensity is the weakest intensity ( The next 1 indicates that), and the color unevenness region width is that of an image having an intermediate width (2 indicates it) among the three types. It shows that the evaluation score by the above comparative evaluation method was 12 points. Similarly, in the frequency distribution diagram of the color difference in the lower part of FIG. 9, M332 indicates that the number of color unevenness is three, the intensity is the strongest of the three types, and the color unevenness region width is three. Of the images having an intermediate width, and evaluation / 45 indicates that the evaluation point of the image by the above-described comparative evaluation method was 45 points.
[0019]
If the frequency of the large color difference is small (there is a low mountain on the right side of the figure), the evaluation point is large, and it is determined visually that there is color unevenness. Even if the frequency of the small color difference is somewhat large (even if there is a slightly high mountain near the base of the graph), the evaluation score is small, and the color is visually judged to be small and the quality is good. Intermediate color differences and frequencies are generally moderately rated.
FIG. 10 is a diagram showing a difference between evaluators in the comparative evaluation. It shows that there is almost no difference between evaluators.
From these facts, the threshold value of the frequency is high in the small color difference, and the threshold value of the frequency decreases rapidly as the color difference becomes large. it can.
[0020]
【The invention's effect】
According to the method of the present invention, it is possible to quantitatively and quickly determine whether or not there is color unevenness in an inspection object, which is in good agreement with the visual determination, without variation among the determiners.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of the configuration of an apparatus used for a method for inspecting color unevenness according to the present invention.
FIG. 2 is a flowchart showing an example of a process for inspecting color unevenness according to the present invention.
FIG. 3 is a diagram illustrating a color difference obtained for each pixel of the entire inspection area for the two-layer film of the example.
FIG. 4 is a diagram showing a frequency distribution of color differences and a threshold curve for a two-layer film of an example.
FIG. 5 is a diagram showing a color difference obtained for each pixel in the entire inspection area for the other two-layer film of the example.
FIG. 6 is a diagram showing a frequency distribution of color differences and a threshold curve for the other two-layer film of the example.
FIG. 7 is a diagram showing a frequency distribution of color differences for an image of a reference example.
FIG. 8 is a diagram showing a frequency distribution of color differences for another image of the reference example.
FIG. 9 is a diagram illustrating a frequency distribution of color differences for still another image of the reference example.
FIG. 10 is a diagram showing a difference between evaluators in a comparative evaluation of a reference example.
[Explanation of symbols]
REFERENCE SIGNS LIST 1 inspection object 2 light source 3 camera 4 inspection operation unit 11 image input unit 12 chromaticity operation unit 13 color difference operation unit 14 reference chromaticity setting unit 15 frequency distribution operation unit 16 color unevenness determination unit 17 threshold curve setting unit 21 irradiation light 22 reflected light 23 output signal 24 judgment output signal

Claims (3)

The reflected light obtained by irradiating the surface of the inspection object with light is captured by a color camera, and the output signal is converted into chromaticity,
For the entire inspection area, the color difference between the obtained chromaticity and the reference chromaticity and the frequency distribution thereof are obtained,
This frequency distribution is compared with a threshold curve in which the threshold of the frequency is high at a small color difference set in advance and the threshold of the frequency rapidly decreases to zero as the color difference increases,
If the entire frequency distribution is equal to or less than the threshold curve, it is determined that there is no color unevenness in the inspection object, and if there is a portion that does not fit in the threshold curve, it is determined that there is color unevenness. Method. The method for inspecting color unevenness according to claim 1, wherein the reference chromaticity is a chromaticity determined for a standard sample. 2. The method for inspecting color unevenness according to claim 1, wherein the reference chromaticity is a chromaticity preset for the inspection object.

Images