JPH10142158A - Method and apparatus for inspection of defect in axis-symmetric pattern - Google Patents

Abstract

PROBLEM TO BE SOLVED: To properly extract a defect in an axis-symmetric pattern such as a shadow mask. SOLUTION: A line-shaped high-luminance and high-uniformity light source 12 for transmitting illumination is arranged on one side of a shadow mask 11 as an object, to be inspected, which is attached to an X-Y stage, and a line-shaped CCD sensor 13 as an image-fetching device is arranged on the opposite side. An image which is fetched by the CCD sensor 13 is computed and processed by an image processing unit 14. First, a two-dimensional correlation coefficient between images in two inspection regions is positions which are symmetric with respect to the central axis is found, image data in one inspection region is moved, by means of software, to a position in which the two-dimensional correlation coefficient becomes maximum, and an image as the difference between the overlapped images is found. In addition, a smoothing treatment is executed, a sampling error is removed, and only a defect part is extracted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an axially symmetric pattern for detecting defects such as large holes, small holes, and clogged holes in an inspection object having an axially symmetric pattern such as a shadow mask used for a color picture tube. The present invention relates to a defect inspection method and apparatus.

[0002]

2. Description of the Related Art A color picture tube shadow mask is disposed inside a color picture tube so as to face a phosphor screen, and sorts an electron beam emitted from an electron gun through a large number of apertures provided in the shadow mask. In addition, it has a function of causing an electron beam to collide only with a phosphor layer having a geometrical one-to-one relationship with the opening of the shadow mask, and is also called a color selection electrode.

[0003] This shadow mask has a size of 0.1 mm to 0.2 mm.
Hundreds of thousands of minute circular or rectangular openings are provided in a predetermined arrangement in aluminum killed steel or invar material having a thickness of about mm. The large number of apertures are arranged symmetrically with respect to a central axis that coincides with the horizontal axis or the vertical axis of the screen through the tube axis, forming an axially symmetric pattern. Such a shadow mask is usually manufactured by a photo-etching method, but it is caused by inclusions present in the material, contamination of the material surface, adhesion of foreign matter and dust during the process, local progress of etching, and the like. Thus, the hole size may deviate from an expected value, or in some cases, a defect may occur in which the hole is completely clogged. If any one of such defects appears on the display screen of the color picture tube as an image defect such as color fringing, the manufactured shadow mask is inspected to exclude defective products.

Conventionally, this inspection work is usually carried out by visual inspection by a human. However, automation of the inspection process is demanded in view of an increase in labor costs and a reduction in hole diameter and hole pitch. FIG. 3 shows an example of a conventional shadow mask defect inspection apparatus, and FIG. 4 shows the principle of a conventional shadow mask defect inspection apparatus.

In FIG. 3, reference numeral 31 denotes a shadow mask to be inspected attached to an XY stage (not shown); 32, a light source for line-type transmitted illumination; 33, a line-type CCD sensor; Unit, 35 is X
A control computer for controlling the Y stage (not shown) and the image processing unit 34.

The principle of defect detection will be described with reference to FIG. 4 as an example of an image signal taken in here. FIG. 4A schematically shows a captured raw image signal. The white part of this image represents a base material part such as aluminum killed steel or amber material, and the hatched part represents an opening part. In this example image, the ordinates of the holes are a, b,
When c, d and abscissa are 1, 2, 3, 4, 5,
In (b, 3), there is a large hole defect in which the hole diameter is larger than an expected value, and in (d, 3), there is a small hole defect in which a part of the hole is missing. According to the conventional hole defect detection method, a detection algorithm for comparing images adjacent to a captured image is adopted. That is, for example, if we consider column b,
(B, 1)-(b, 2), (b, 2)-(b, 3),
(B, 3)-(b, 4), (b, 4)-(b, 5),
Subtraction is performed as (b, 5)-(b, 6). As a result, as shown in FIG. 4B, (b, 1)-(b, 2)
Are all normal apertures, the result of the calculation is 0,
(B, 2)-(b, 3)
3) is plus, (b, 2)-(b, 3) is minus,
(B, 3)-(b, 4) is positive. Similarly, considering row c, since a small hole defect exists at (c, 3),
(C, 2)-(c, 3) is plus, (c, 3)-(c,
4) is negative. By discriminating the hole size and the hole size based on the result, and by giving a threshold value indicating a failure in the area of the difference, it is possible to determine a defect if an abnormal value larger than this value is obtained.

[0007]

However, unlike a color filter used in a liquid crystal display device, which has the same pixel pitch and the same pixel shape, and has a peripheral shape different from a linear shape, the aperture pattern of the shadow mask is different from the center. The hole pitch and hole shape are gradually changed toward the periphery, and the periphery is also curved.

For this reason, in the conventional method of comparing the adjacent holes, a normal hole may be determined as a defect. In addition, if the first and fifth rows in FIG. 4 are both ends, there is no comparison object next to the first row and the fifth row, so that there is a problem that all of them are determined as defects as shown in FIG. 4C.

The present invention has been made in view of the above-described problems, and has a defect inspection method and a defect inspection apparatus capable of accurately and efficiently detecting a hole defect even in an axially symmetric pattern such as a shadow mask. The purpose is to provide.

[0010]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and has a defect of an axial symmetry pattern for detecting a defect of an inspection object having a pattern symmetric with respect to a central axis. In the inspection method, a step of setting a central axis of the inspection object, a step of capturing images of a pair of inspection areas symmetrical with respect to the central axis and storing the captured image data, and comparing the image data of the pair of inspection areas. And a method for inspecting a defect of an axially symmetric pattern.

The step of comparing the image data of the pair of inspection areas includes the step of calculating a two-dimensional correlation coefficient between the pair of image data, and the step of calculating the two-dimensional correlation coefficient between the pair of image data. It is characterized by comprising a step of superimposing data to obtain image data of a difference between image data, and a step of performing smoothing processing to extract only a defective portion.

Also, a light source for transmitted illumination provided on one surface side of a sheet-shaped inspection object having a pattern symmetrical with respect to a central axis, and an image provided on the other surface side of the inspection object. In a defect inspection device of an axially symmetric pattern having a capturing device and an image processing device that performs an arithmetic process on a captured image to determine a defect, the image processing device includes:
A defect inspection apparatus for an axially symmetric pattern, wherein a defect is determined by comparing images located at positions symmetrical with respect to a central axis.

Further, the image processing section calculates a two-dimensional correlation coefficient between the images of the two inspection areas symmetrical with respect to the central axis, and one inspection area at a position where the two-dimensional correlation coefficient becomes maximum. A processing unit that moves the image data, a processing unit that obtains a difference between the superimposed images, a processing unit that removes a sampling error, performs a smoothing process, and extracts only a defective portion, I have.

Further, the present invention provides a step of taking in image data consisting of light and dark, and inverting one of image data of a pair of image data sandwiching the axis of symmetry so as to be rotated by 180 ° around the axis of symmetry. And a step of superimposing the inverted image data and the other image data so that the axes of symmetry coincide with each other, and calculating a difference between the light and dark data thereof. It is a detection method.

As described above, in the present invention, since the image data on both sides of the symmetry axis are compared, a defect can be identified even when the shape and size of the pattern are changed.
In addition, since the end of the pattern is compared with the end of the opposite pattern, the entire end is not recognized as a defect.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. In addition, here, a shadow mask used for a color picture tube will be described as an example of the axial symmetry pattern.

FIG. 1 is a schematic diagram of a shadow mask defect inspection apparatus according to an embodiment of the present invention. In FIG. 1, reference numeral 11 denotes a shadow mask to be inspected mounted on an XY stage (not shown); 12, a line-type light source for transmitted illumination having high luminance and uniformity; and 13, a line-type light source, which is an image capturing device. A CCD sensor 14, an image processing unit that performs arithmetic processing on the captured image to determine a defect;
Reference numeral 15 denotes a control computer for controlling the XY stage (not shown) and the image processing unit 14.

The image processing unit includes a processing unit (not shown) for calculating a two-dimensional correlation coefficient between the images of the two inspection areas symmetrical with respect to the central axis. A processing unit that moves image data of one inspection region, a processing unit that calculates a difference between the superimposed images,
At least a processing unit that removes a sampling error and performs a smoothing process to extract only a defective portion.

Next, the principle of detecting a defect of a shadow mask according to the present invention will be described with reference to FIG. FIG. 2A is a raw image composed of light and dark captured by the line type CCD sensor 12. The white part of this image represents a base material part such as aluminum killed steel or amber material, and the hatched part represents an opening part. Also, let the ordinate of the hole be a,
Assuming that b,... g and the abscissas are 1, 2,..., in this image, a large hole defect having a hole diameter larger than the standard in (e, 7) and a hole missing in (c, 7) in FIG. There is a small hole defect. According to the conventional method, in addition to these two real defects, (b, 1), (b, 7), (c, 8), (d,
8), (e, 8) and the like were also detected as defects.

In the present invention, first, the central axis RR 'of the image is set by fetching the image data consisting of light and dark, obtaining the central axis at the position where the image is folded and overlapped.
Next, 2 at a position symmetrical with respect to the central axis RR ′.
Two inspection areas 10 and 20 are set, and these two images 1
A two-dimensional correlation coefficient between 0 and 20 is determined. Then, the image data of one inspection area is softly moved to the position where the two-dimensional correlation coefficient becomes the maximum, that is, the position where the overlap becomes the maximum when superimposed while comparing the hole size and the pitch. The difference image between the combined images is obtained. The result is shown in FIG. In this figure, 10 images are 2
The result of superimposing on 0 and subtracting 10 images from 20 images is shown. At this time, a sampling error 22 may occur due to the edge of the hole. In order to remove the sampling error 22, a smoothing process is performed to extract only a defective portion ((c)). By setting a threshold value for this image, if an abnormal value greater than this value appears, it can be determined as a defect.

In other words, the present invention takes in image data consisting of bright and dark areas such as apertures and non-apertures, and
One of the pair of image data is inverted so as to be rotated by 180 ° around the axis of symmetry, and the inverted image data and another image data are overlapped so that the axis of symmetry coincides. Then, by detecting the difference between the light and dark data, a defect of the axially symmetric pattern is detected.

In the above description, the defect detection of the shadow mask has been described as an example. However, the present invention is applicable to any pattern having a symmetric arrangement with respect to the central axis, such as a color filter or a plasma display used in a liquid crystal display device. be able to.

[0023]

As described above, according to the present invention, the defect size is determined by comparing the images of the two inspection areas symmetrical with respect to the central axis. Defects can be detected when the sheath shape has changed or even at both ends.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing an outline of a defect inspection apparatus used in the present invention.

2A is a diagram showing image data read by the apparatus of FIG. 1, and FIG. 2B is a diagram showing a difference between image data of an inspection area symmetrical with respect to a central axis. Yes, Figure 2
(C) is a diagram showing a result of removing a sampling error.

FIG. 3 is a schematic diagram showing an outline of a conventional defect inspection apparatus.

4 (a) is a diagram showing image data read by the apparatus of FIG. 3, FIG. 2 (b) is a diagram showing a result of determining a defect, and FIG. It is a figure explaining a fault by an apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... Shadow mask 12 ... Transmission illumination light source 13 ... CCD sensor 14 ... Image processing unit 15 ... Control computer

Claims (5)

[Claims] 1. An axially symmetric pattern defect inspection method for detecting a defect of an inspection object having a pattern symmetrical with respect to a central axis, comprising the steps of: setting a central axis of the inspection object; A method for inspecting a defect of an axially symmetric pattern, comprising: a step of taking in images of a pair of symmetric inspection regions and storing the image as image data; and a step of comparing the image data of the pair of inspection regions. 2. The step of comparing image data of the pair of inspection areas, the step of obtaining a two-dimensional correlation coefficient between the pair of image data, and the step of calculating the two-dimensional correlation coefficient at a position where the two-dimensional correlation coefficient is maximum. A defect inspection method for an axially symmetric pattern, comprising: a step of superimposing data to obtain image data of a difference between image data; and a step of performing a smoothing process to extract only a defective portion. 3. A light source for transmitted illumination provided on one surface of a sheet-shaped inspection object having a pattern symmetrical with respect to a central axis, and an image provided on the other surface of the inspection object. An axially symmetric pattern defect inspection device having a capturing device and an image processing device for performing arithmetic processing on the captured image to determine a defect, wherein the image processing device is located at a position symmetrical with respect to a central axis. A defect inspection apparatus for an axially symmetric pattern, characterized in that a defect is determined by comparing. 4. An image processing section for calculating a two-dimensional correlation coefficient between images of two inspection areas symmetrical with respect to a central axis, and one inspection area at a position where the two-dimensional correlation coefficient is maximum. A processing unit for moving the image data of the above, a processing unit for calculating a difference between the superimposed images, and a processing unit for removing a sampling error and performing a smoothing process to extract only a defective portion. 4. The defect inspection apparatus for an axially symmetric pattern according to claim 3, wherein 5. A step of capturing image data consisting of light and dark, and a step of inverting one of image data of a pair of image data sandwiching the axis of symmetry so as to rotate by 180 ° about the axis of symmetry. Superposing the inverted image data and another image data so that the symmetry axes thereof coincide with each other, and calculating a difference between the light and dark data, and a defect inspection method for an axially symmetric pattern.