JPH0968497A - Method and system for inspecting unevenness - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine the unevenness of gray pattern quantitatively by deter mining a blurred image of gray pattern and a power spectral distribution in a partial region thereof, adding the power spectrums in a specified spatial frequency region thereof and comparing the sum with a threshold value. SOLUTION: An image processor 54 executes a software through a CPU 66 to perform various operations. At first, a CCD camera 49 picks up the image of shadow mask SM of an object for which the unevenness is inspected at a position shorter by a specified length than the focal length thus inputting a blurred image. A plurality of square processing regions are then clipped from the image and two-dimensional Fourier transform is performed for each region thus calculating a power spectrum. Subsequently, the power spectrum values are added in a specified frequency region for each region to obtain a sum of power spectrums. Furthermore, a maximum sum is determined for each region and compared with a predetermined threshold thus deciding whether the mask SM is acceptable or not. According to the method, acceptability of the mask SM can be determined quantitatively.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern of through holes in a shadow mask and an original plate for producing a shadow mask (generally,
The present invention relates to a method for inspecting unevenness in a substantially periodic light and dark pattern such as a light and shade pattern formed on a glass plate) and an apparatus therefor.

[0002]

2. Description of the Related Art A cathode ray tube shadow mask is a through-hole plate in which minute through-holes are formed in a metal plate at a substantially periodic pitch. A black-and-white image is formed on the original plate used in the photoetching process of the shadow mask and the color filter for the liquid crystal display panel to expose a fine pattern. In this specification, the pattern of through holes formed in the shadow mask and the pattern of the black and white image formed on the original plate are referred to as "bright and dark patterns".

The quality of the shadow mask and its original plate depends on the error in the bright and dark pattern (error in hole diameter and pitch). Since the diameter of the through hole is as small as about 100 μm, the through hole itself cannot be seen by observing the shadow mask with the naked eye. However, when there is an error in the light-dark pattern of the shadow mask or the original plate, and in particular when the error is locally deviated, unevenness in shade is recognized when observing the shadow mask or the original plate with the naked eye. Therefore, conventionally, in the quality inspection of the shadow mask or the original plate, the inspector visually inspects the optical density unevenness to determine the quality of the error in the light-dark pattern.

[0004]

However, since the inspection result of the unevenness by the naked eye depends on the skill and physical condition of the inspector, the stable inspection result may not be obtained in some cases. Further, since the inspection requires concentration, it is difficult to carry out the inspection continuously for a long time, and the inspection amount per day is also limited. Therefore, there has been a long-felt need for a method and apparatus for quantitatively inspecting unevenness without relying on the naked eye.

The present invention has been made to solve the above problems in the prior art, and an object of the present invention is to provide a method and apparatus for quantitatively inspecting unevenness of a bright and dark pattern formed almost periodically. And

[0006]

Means for Solving the Problem and Its Action / Effect To solve at least a part of the above-mentioned problem, the first invention is a method for inspecting unevenness of a bright and dark pattern formed almost periodically. , (A) obtaining a blurred image of the light-dark pattern, (b) obtaining a power spectrum distribution for at least a part of the blurred image, and (c) a predetermined spatial frequency region of the power spectrum distribution. And (d) comparing the power spectrum addition value with a predetermined threshold,
And a step of judging whether the unevenness of the bright and dark pattern is good or bad.

Since unevenness of the bright and dark pattern appears in the blurred image, the power spectrum value at the spatial frequency of the unevenness is large. Therefore, the power spectrum addition value obtained by adding the power spectrum values in the predetermined spatial frequency region has a large value when the unevenness is large and a small value when the unevenness is small. By comparing the power spectrum addition value with the threshold value, it is possible to determine the quality of the unevenness. Therefore, it is possible to quantitatively inspect the unevenness of the light-dark pattern formed almost periodically.

In the first aspect, the step (a)
(1) obtaining a focus position between the image pickup means for picking up the blurred image and the light-dark pattern;
(2) shifting the focal position of the imaging means from the focused position by a predetermined amount, and (3) capturing the image of the light-dark pattern at the focal position shifted from the focused position to obtain the blurred image. And a step of obtaining.

By doing so, it is possible to easily generate a blurred image suitable for the unevenness inspection.

Alternatively, in the first aspect of the invention, the step (a) includes (1) an aperture stop of the image pickup means for determining a focus position between the image pickup means for picking up the blurred image and the light-dark pattern. And the process of obtaining with open
(2) a step of narrowing down the aperture stop, and (3) a state in which the aperture stop is narrowed down, at the in-focus position,
The step of obtaining the blurred image by capturing the image of the light-dark pattern may be included.

In this way, it is possible to easily create a blurred image suitable for unevenness inspection without changing the imaging magnification.

A second aspect of the present invention is an apparatus for inspecting unevenness of a bright and dark pattern formed almost periodically, and a blurred image generating means for obtaining a blurred image of the bright and dark pattern, and at least a part of the blurred image. Power spectrum calculating means for obtaining the power spectrum distribution for the area of, the addition means for adding the power spectrum values in the predetermined spatial frequency area of the power spectrum distribution to obtain the power spectrum addition value, the power spectrum addition value and the predetermined value. A determination unit that determines whether the unevenness of the light-dark pattern is good or bad by comparing with a threshold value.

The second invention also has the same operation and effect as the first invention, and it is possible to quantitatively inspect the unevenness of the bright and dark patterns formed almost periodically.

In the second aspect of the invention, the blurred image generating means includes a focus position calculating means for obtaining a focus position between the image pickup means for picking up the blurred image and the light-dark pattern, and Focus position adjusting means for shifting the focus position of the image pickup means from the in-focus position by a predetermined amount, and image pickup means for obtaining the blurred image by capturing the image of the light-dark pattern at the focus position displaced from the in-focus position. , Are preferably provided.

In the second aspect of the invention, the blurred image generating means includes an image pickup means having an aperture stop,
A focus position calculating unit that obtains a focus position between the image pickup unit and the light-dark pattern in a state where the aperture stop is opened; and the image pickup unit is a state where the aperture stop is narrowed down. The blurred image may be obtained by capturing an image of the light / dark pattern at the in-focus position.

[0016]

Next, embodiments of the present invention will be described based on examples. FIG. 1 is a conceptual diagram showing a shadow mask inspection apparatus 30 to which an embodiment of the present invention is applied. This shadow mask inspection device 30 is composed of an optical measuring device 40 for taking an image of the shadow mask SM and obtaining image data thereof, and a data processing device 50 for performing various data processing based on the image data.

The optical measuring device 40 is provided with a surface plate 41, and on the surface plate 41, a measurement table 42 having a transmission hole for passing illumination light is formed at the substantially center. Further, a glass plate (for example, frosted glass) 43 that diffuses and transmits light is placed on the upper surface of the measurement table 42 in order to make the light amount distribution uniform, and below the measurement table 42, a light source 44. Are arranged. A diffusion plate may be adopted instead of the glass plate 43. As the light source 44, for example, a high frequency lighting type fluorescent lamp is used. A shadow mask SM is attached and fixed to the upper surface of the glass plate 43 with an adhesive tape or the like.

A stand supporting arm 46 extending upward is provided upright from the surface plate 41, and the stand supporting arm 4 is provided.
6, a camera holding beam 47 for holding a CCD camera 49 as an image pickup means is provided so-called cantilever. A Z-axis stage 48 is fixed to the tip of the camera holding beam 47. Camera holding beam 47
Is connected to a ball screw unit (not shown) built in the stand support arm 46 along the Z axis in the figure,
It is driven by a Z-axis drive motor (not shown) and moves (moves up and down) in the Z direction. Therefore, the CCD camera 49 moves up and down together with the camera holding beam 47. Therefore, the CCD so that the perforation areas of the through holes in the shadow masks SM of various sizes and the imaging area of the CCD camera 49 coincide with each other.
The shadow mask SM can be imaged by moving the camera 49 up and down.

The CCD camera 49 is a CCD camera in which CCD elements are two-dimensionally arranged, and outputs the image of the light and dark pattern of the shadow mask SM as digital image data of 10 bits / pixel. From the CCD camera 49,
An image of an area of 1534 pixels × 1024 pixels can be acquired.

The data processing device 50 includes a display 52 for displaying images, which will be described later, an image processing device 54 for performing various image processes, a Z-axis stage controller 56 for vertically moving the CCD camera 49, and a CCD camera 49. An auxiliary display 58 for displaying the captured image for confirming the image capturing area of the device, an external storage device 60 for storing the image data, and a camera control device 62 for adjusting the focus and the like of the CCD camera 49 are provided. The camera control device 62 also has a function of capturing image data captured by the CCD camera 49 and sending the image data to another device.

FIG. 2 is a block diagram showing the electrical construction of the shadow mask inspection apparatus 30. The light emitted from the light source 44 sequentially passes through the measurement table 42, the glass plate 43, and the through hole of the shadow mask SM, and then the CCD camera 4
9 is incident. The CCD camera 49 is a shadow mask S
An image of M is obtained as image data of a grayscale image (multivalued image) arranged two-dimensionally. The image data Di is captured by the camera control device 62 and output to the auxiliary display 58. Therefore, the auxiliary display 58 has a CCD
The raw image obtained by the camera 49 is displayed.

The image processing device 54 is a computer system that executes various processes by executing a software program. This image processing device 54 is a CP
U66 and RAM68 as a temporary storage means for data
A keyboard 70 as a data input means, a flexible disk device 72 for driving a flexible disk (FD), and a printer 74 for issuing inspection results and the like, which are connected to each other via a bus line 64. ing.

A camera control device 62 and a Z-axis stage controller 5 are provided on a bus line 64 of the image processing device 54.
6 and the external storage device 60, a display 52,
A Z-axis rotary encoder 76 for detecting the Z-axis position of the CCD camera 49 is connected. The Z-axis rotary encoder 76 is attached to a drive shaft of a Z-axis drive motor (not shown) for vertically moving the CCD camera 49 together with the camera holding beam 47, and outputs a signal of the Z-axis position of the CCD camera 49. To do. It should be noted that this signal is input to the image processing device 54, the camera control device 62, and the Z-axis stage controller 56. Then, based on this signal and the control signal from the image processing device 54, the focus adjustment of the CCD camera 49 by the camera control device 62 and the CCD camera 4 by the Z-axis stage controller 56 are performed.
9, the Z-axis position adjustment is performed.

The CPU 66 executes a software program to obtain a focus position of the CCD camera 49, a focus position calculating means, and the CCD camera 4.
Focus position adjusting means for shifting the focus position of 9 from the in-focus position by a predetermined amount, power spectrum calculating means for obtaining the power spectrum distribution of the blurred image obtained by the CCD camera 49,
The functions of various means such as addition means for obtaining the power spectrum addition value by adding the power spectrum values in a predetermined spatial frequency area, and determination means for determining the quality of unevenness of the light-dark pattern from the power spectrum addition value are realized. ing.

FIG. 3 is a flow chart showing the procedure of the unevenness inspection method in the embodiment. In step T1, the CCD mask 49 captures an image of the shadow mask SM to be inspected for unevenness, and the image is input. At this time, in order to capture a blurred image, the distance L (see FIG. 1) between the CCD camera 49 and the shadow mask SM is set to be shorter than the in-focus position by a predetermined amount. For example, when the focus position of the CCD camera 49 is 1 m from the CCD camera 49, the distance between the CCD camera 49 and the shadow mask SM is set to a value 20 cm shorter (80 cm) from the focus position. By doing so, a blurred image can be intentionally obtained. In this way, the blurred image is intentionally picked up by the pixel array of the CCD camera 49 and the shadow mask SM.
This is to prevent an interference pattern called a beat from being generated in the image due to the interference with the bright and dark patterns. That is, in step T1, a blurred image having no beat may be generated. In a blurred image without a beat, individual through holes cannot be observed, but unevenness of light and shade due to an error in a through hole pattern appears in the image.

In order to obtain a blurred image without a beat, the CCD camera 49 is added to or subtracted from a value of about 10% to about 30% of the distance between the CCD camera 49 and the shadow mask SM at the in-focus position. It is desirable to set. Alternatively, it is desirable to set the CCD camera 49 to a distance obtained by adding or subtracting a value about 1.5 times to about 4 times the depth of focus of the CCD camera 49 in the photographing state.

In step T2 in FIG. 3, a processing region is cut out from the blurred image of the shadow mask SM that has been picked up and is processed. FIG. 4 is an explanatory diagram showing the six processing regions R1 to R6 cut out in step T2. Each of the six processing regions R1 to R6 is a square region, and the six processing regions R1 to R6 are arranged so as to partially overlap each other and cover almost the entire surface of the blurred image of the shadow mask. As will be described later, since the two-dimensional Fourier transform is performed in this embodiment, each processing area needs to be square. On the other hand, since the shadow mask SM has a substantially rectangular shape, as shown in FIG. 4, a plurality of square processing regions are set to inspect the entire substantially rectangular shadow mask SM. The processing areas do not necessarily have to overlap each other. At least one processing area
You only need to set one.

In step T3 of FIG. 3, each processing region R1
Two-dimensional Fourier transform (2d-FFT) is performed for R6 to R6, and the power spectrum is calculated from the transform coefficient in step T4. The power spectrum is the sum of squares of the real and imaginary parts of the Fourier transform coefficient. The power spectrum thus obtained is mathematically equivalent to that obtained by Fourier transforming the autocorrelation of the image data of each processing region. Therefore, in steps T3 and T4, the cross-correlation of the image data of each processing area may be obtained and Fourier-transformed. That is, in steps T3 and T4, the power spectra of the processing regions R1 to R6 may be obtained by some method.

FIG. 5 is a graph showing an example of power spectrum distributions obtained for non-defective products and defective products. The horizontal axis of the graph is the spatial frequency fsp, and the vertical axis is the power spectrum value. Here, the spatial frequency fsp is defined as the frequency in the length W of one side of each processing region, as shown in FIG.

As shown in FIG. 5, a good product has a relatively flat power spectrum distribution, but a defective product (having a large deviation in hole diameter or pitch error) has a specific spatial frequency. The power spectrum value in 1 tends to be high. This is because a defective product has an error deviation, and the power spectrum value at the deviation frequency becomes large.

FIG. 7 is a conceptual diagram showing the power spectrum distribution obtained for each of the processing regions R1 to R6. FIG.
In step T5, the power spectrum addition value is obtained by adding the power spectrum values of the specific frequency region FR for each processing region. As shown in FIG. 7, since the processing regions R1 to R6 have different power spectrum distributions, the power spectrum addition values S1 to S6 also have different values. In addition,
The power spectrum addition value is a value obtained by integrating the power spectrum distribution in the specific frequency region FR.

Specific frequency region FR for obtaining the added value
For example, the value of the spatial frequency fsp is preferably in the range of 5 to 19. The power spectrum value of which the spatial frequency fsp is less than 5 is influenced by the grade of the shadow mask SM. Here, the "grade" means the shadow mask SM
It means that the design values of the hole diameter and pitch of the through holes are changed near the periphery of the. For example, the pitch of the through holes in the peripheral portion of the shadow mask SM is designed to be larger than the pitch in the central portion. Such grades are
Appears in regions of relatively low spatial frequency. Therefore, in order to remove the influence of the grade in the unevenness inspection, the relatively low spatial frequency region of less than about 5 is removed and the power spectrum value is added. On the other hand, the power spectrum values in the relatively high spatial frequency region do not show a large difference between the non-defective product and the defective product, as can be seen from FIG. Therefore, in the unevenness inspection, the power spectrum value is added after removing the relatively high spatial frequency region of more than about 20.

As can be seen from the above description, step T
The specific spatial frequency region FR indicating the range of addition in 5 does not include a relatively low spatial frequency region affected by the grade of the shadow mask SM, and does not include a relatively high spatial frequency region equal to or higher than a predetermined frequency. Area.
The range of the spatial frequency domain FR to be added is
It is set to an appropriate range that differs depending on the design of the shadow mask, the size of the processing area, and the like.

In step T6 of FIG. 3, the power spectrum addition values S1 to S6 for the six processing regions R1 to R6 are added.
Of the shadow mask SM is compared with a predetermined threshold value to determine whether the shadow mask SM is uneven. Then, in step T7, the determination result is displayed on the display 52. The larger the unevenness, the larger the power spectrum addition value. The reason why the maximum value of the power spectrum addition values S1 to S6 is compared with the threshold value is to determine the quality of the uneven state in the processing area having the largest unevenness among the six processing areas R1 to R6.

FIG. 8 is a graph showing the correlation between the power spectrum addition value of the shadow mask and the rank of the quality of the unevenness of the shadow mask by the human. The graph in Figure 8 is
6 is a graph in which an inspector visually inspects an unevenness and gives a rank of goodness of unevenness and a power spectrum addition value obtained by the unevenness inspection apparatus of the embodiment. According to the inspector's ranking, those having a rank of 3.5 or higher are defective products, and those having a rank of 3 or lower are good products. FIG.
As can be seen from the above, there is an extremely high correlation between the power spectrum addition value obtained by the unevenness inspection apparatus of this embodiment and the rank given by the inspector. For example, in the graph of FIG. 8, the power spectrum addition value f3 corresponding to the lowest rank 3 of the good product and the power spectrum addition value f corresponding to the highest rank 3.5 of the defective product in the visual inspection by the inspector.
If 3.5 is used as the threshold value, it is possible to automatically judge the quality of the unevenness of the shadow mask. In addition,
Considering the error of the unevenness inspection, a rank between 3 and 3.5 may be defined as a gray area, and the inspector may visually confirm the quality of the shadow mask included in this gray area. . This has the advantage that the inspection efficiency is improved because the inspector only needs to inspect the shadow mask in the gray region with the naked eye. Moreover, since most of the shadow masks are automatically judged to be good or bad, it is possible to ensure the stability of the inspection result.

As described above, according to this embodiment, since a blurred image which is not affected by the beat of the shadow mask SM is picked up, the influence of the beat of the shadow mask SM on the result of the unevenness inspection should be reduced. You can In addition, the power spectrum of the blurred image is obtained, and the shadow mask SM
The power spectrum value in a specific spatial frequency region excluding the frequency of the grade is added, and the quality of the unevenness is determined according to the added value of the power spectrum. Therefore, the influence of the grade of the shadow mask SM on the result of the unevenness inspection is affected. Can be made smaller. That is, according to this embodiment, it is possible to quantitatively inspect the unevenness while suppressing the influence of the beat and grade of the shadow mask SM.

The present invention is not limited to the above embodiment, but can be implemented in various modes without departing from the gist of the present invention. For example, the following modifications are possible.

(1) Although a plurality of processing regions R1 to R6 are set in the shadow mask in the above embodiment, at least one processing region may be set. However, as shown in FIG. 4, if a plurality of processing regions are set so as to cover almost the entire surface of the shadow mask, it is possible to detect a region having a large unevenness in the plurality of processing regions. There is an advantage that unevenness can be judged better.

(2) In the above embodiment, the blurred image that is not influenced by the beat is directly captured by the CCD camera 49, but the CCD camera 49 captures a sharp original image and processes the original image. It is also possible to obtain a blurred image that is not affected by the beat. For example, CC
It is possible to create a blurred image by performing smoothing processing such as a median filter on the sharp original image captured by the D camera 49. Alternatively, by dividing the image data of the original image by the image data of the blurred image after the smoothing process, an image in which the influence of the grade is removed can be obtained. The unevenness inspection may be performed using such a grade-removed image as a processing target.

(3) In the above embodiment, the CCD camera 49 is set at a position deviated from the in-focus position in order to obtain a blurred image which is not affected by the beat, but the CCD camera 49 is set at the in-focus position. A blurred image may be obtained by reducing the resolution (optical MTF).
FIG. 9 is a conceptual diagram showing the configuration of the lens system 80 of the CCD camera 49. The lens system 80 has a lens barrel 82, lenses 84, 85, 86 and an aperture stop 88 provided in the lens barrel 82. Using such a lens system 80, the in-focus position is obtained in a state where the aperture stop is opened (FIG. 9A), and the aperture stop is stopped until the influence of the beat disappears (FIG. 9B). If you take an image with
A blurred image can be obtained. In this case, there is an advantage that a blurred image can be generated without changing the imaging magnification.

[Brief description of drawings]

FIG. 1 is a conceptual diagram showing a shadow mask inspection apparatus 30 to which an embodiment of the present invention is applied.

FIG. 2 is a block diagram showing an electrical configuration of a shadow mask inspection device 30.

FIG. 3 is a flowchart showing the procedure of the unevenness inspection method in the embodiment.

FIG. 4 is an explanatory diagram showing six processing regions R1 to R6 cut out in step T2.

FIG. 5 is a graph showing an example of power spectrum distributions obtained for non-defective products and defective products.

FIG. 6 is an explanatory diagram showing the definition of the spatial frequency fsp in the embodiment.

FIG. 7 is a conceptual diagram showing a power spectrum distribution obtained for each of the processing regions R1 to R6.

FIG. 8 is a graph showing the relationship between the power spectrum added value of the shadow mask and the rank of the quality of the unevenness of the shadow mask by a person.

FIG. 9 is a conceptual diagram showing a configuration of a lens system 80 of a CCD camera 49.

[Explanation of symbols]

 30 ... Shadow mask inspection device 40 ... Optical measurement device 41 ... Surface plate 42 ... Measurement table 43 ... Glass plate 44 ... Light source 46 ... Stand support arm 47 ... Camera holding beam 48 ... Z-axis stage 49 ... CCD camera 50 ... Data processing device 52 ... Display 54 ... Image processing device 56 ... Z-axis stage controller 58 ... Auxiliary display 60 ... External storage device 62 ... Camera control device 64 ... Bus line 66 ... CPU 68 ... RAM 70 ... Keyboard 72 ... Flexible disk device 74 ... Printer 76 ... Z-axis rotary encoder 80 ... Lens system 82 ... Lens barrel 84, 85, 86 ... Lens 88 ... Aperture stop

Claims (6)

[Claims] 1. A method for inspecting unevenness of a light-dark pattern formed almost periodically, the method comprising: (a) obtaining a blurred image of the light-dark pattern; and (b) at least a part of the blurred image. A step of obtaining a power spectrum distribution for a region; (c) a step of adding power spectrum values in a predetermined spatial frequency region of the power spectrum distribution to obtain a power spectrum addition value; and (d) the power spectrum addition value and a predetermined value. A non-uniformity inspection method, which comprises determining whether the non-uniformity of the light-dark pattern is good or bad by comparing the non-uniformity of the bright and dark patterns. 2. The unevenness inspection method according to claim 1, wherein in the step (a), (1) an in-focus position between an image pickup unit for picking up the blurred image and the bright / dark pattern is obtained. The steps of: (2) shifting the focus position of the imaging means from the in-focus position by a predetermined amount; and (3) capturing the image of the light-dark pattern at the focus position deviated from the in-focus position. An unevenness inspection method comprising: a step of obtaining a blurred image. 3. The unevenness inspection method according to claim 1, wherein in the step (a), (1) the focus position between the image pickup means for picking up the blurred image and the bright / dark pattern is set to the focus position. A step of obtaining with the aperture stop of the image pickup means opened,
(2) a step of narrowing down the aperture stop, and (3) a state in which the aperture stop is narrowed down, at the in-focus position,
And a step of obtaining the blurred image by capturing the image of the light-dark pattern. 4. An apparatus for inspecting unevenness of a light-dark pattern formed almost periodically, wherein a blurred image generating means for obtaining a blurred image of the bright-dark pattern, and a power for at least a part of the area in the blurred image. A power spectrum calculating means for obtaining a spectrum distribution, an adding means for adding a power spectrum value in a predetermined spatial frequency region of the power spectrum distribution to obtain a power spectrum addition value, and comparing the power spectrum addition value with a predetermined threshold value. By doing so, a nonuniformity inspection device comprising: a determination unit that determines whether the lightness / darkness pattern is nonuniform. 5. The unevenness inspection apparatus according to claim 4, wherein the blurred image generation unit obtains a focused position between an imaging unit for capturing the blurred image and the bright-dark pattern. A blurring image by calculating an image of the light-dark pattern at a focus position that is displaced from the in-focus position by a calculation unit, a focus-position adjusting unit that shifts the focus position of the imaging unit from the in-focus position by a predetermined amount. A nonuniformity inspection apparatus comprising: 6. The unevenness inspection apparatus according to claim 4, wherein the blurred image generation means includes an image pickup means having an aperture stop, and the image pickup means and the light-dark pattern in a state where the aperture stop is opened. Focusing position calculating means for determining a focusing position between the, the imaging means, in a state in which the aperture stop is narrowed,
An unevenness inspection apparatus that obtains the blurred image by capturing an image of the bright and dark pattern at the in-focus position.