JP2000294139A - Method and device for detecting defect of periodic pattern - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and a device for detecting defects in a periodic pattern realizing supersensitive defect detection without causing reduction in detectivity due to moire. SOLUTION: This method has a process S1 for photograpging so as to resolve a processed part, a process S2 for recording a multivalent image obtained by photograpging in a memory, a process S3 for threshold processing the recorded multivalent image by a predetermined slice level, a process S4 for labeling the threshold processed multivalent image according to binary data, a process S5 for seeking an address of a pixel constructing the same label according to a label result obtained by the labeling process, a process S6 for operating a pixel value of the recorded multivalent image corresponding to the address, a process S7 for expanding a representative value based on the operation result, a process S8 for differentiating the expanded representative value and a process S9 for extracting defects by comparison of the differentiated result with the slice level.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a work having a periodic pattern in which processing portions formed on a base material are periodically repeated, or a processing portion formed on a base material. The present invention relates to a method and an apparatus for inspecting a periodic pattern for inspecting a work having a pattern in which an arrangement pitch at which the pattern is gradually changed from the center of the work toward the outside.

[0002]

2. Description of the Related Art Conventionally, various inspection methods and apparatuses have been proposed in which a periodic pattern such as a shadow mask is photographed by a CCD camera and image processing is performed based on the photographed image to detect a defect of a work to be inspected. I have. However,
When a periodic pattern such as a shadow mask is photographed by a CCD camera, moire is generated due to interference between a pixel pitch of the CCD sensor and a pitch of a processing portion of a workpiece to be inspected (a pitch of a hole in the case of a shadow mask). This moiré causes a problem of lowering the detection accuracy in the periodic pattern inspection.

[0003] In recent years, along with the increase in quality requirements from customers required for shadow masks, it has been strongly desired to further improve the accuracy of defect detection.

[0004]

The present invention relates to a pixel pitch of a CCD sensor arranged in a CCD camera as a photographing means and a pitch of a processed portion of a work to be inspected (a pitch of a hole in the case of a shadow mask). ), Without lowering the detection sensitivity due to moiré caused by interference between
It is an object of the present invention to provide a pattern defect inspection method and apparatus that realize highly sensitive defect detection.

[0005]

According to the present invention, there is provided a method of inspecting a defect of a pattern, comprising: a work having a periodic pattern in which processed portions formed on a substrate are periodically repeated;
A pattern defect inspection method for inspecting a work having a pattern in which an arrangement pitch when arranging a processing portion formed on a base material gradually changes from a center portion of the work toward the outside, wherein the processing is performed. A step of photographing so as to resolve the part, a step of recording a multi-valued image obtained by the photographing in a memory, a step of performing binarization processing on the recorded multi-valued image at a predetermined slice level, Performing a labeling process based on the binarized binary data;
A step of obtaining an address of a pixel constituting the same label based on the label result obtained by the labeling process; a step of calculating a pixel value of the recorded multi-valued image corresponding to the address; Developing a representative value based on the on a memory, performing a differentiation process on the developed representative value, extracting a defect by comparing the result of the differentiation process with a slice level, It is characterized by having.

In the above-mentioned arithmetic processing step, it is further preferable that in the arithmetic processing step, each pixel value of the multi-valued image corresponding to the pixel address forming the same label is added to calculate the total sum for each label.

In the step of calculating, the sum of the pixel values of the multi-valued image corresponding to the contour address of the processed portion among the pixel values of the multi-valued image corresponding to the pixel addresses forming the same label is calculated. Is more desirable.

[0008] A pattern defect inspection apparatus according to the present invention comprises a workpiece having a periodic pattern in which processing portions formed on a base material are periodically repeated, or a processing portion formed on a base material. A pattern defect inspection apparatus for inspecting a work having a pattern in which an arrangement pitch at the time of arrangement is gradually changed from the center of the work toward the outside, wherein photographing is performed to resolve the processed portion. A first memory for recording a multi-valued image obtained by the photographing, and a binarization processing unit for performing a binarization process on the multi-valued image recorded in the first memory at a predetermined slice level A second memory for recording the binarized data obtained by the binarization processing unit, a labeling processing unit for performing a labeling process based on the binarized data recorded in the second memory, processing An arithmetic processing unit that obtains an address of a pixel constituting the same label based on the label result obtained by the above, performs an arithmetic operation on a pixel value of the multi-valued image corresponding to the address to obtain a representative value, and A third memory that expands a representative value based on the result in a memory, a differential processing unit that performs differential processing on the representative value expanded in the memory, and a comparison between the result of the differential processing and a slice level. A defect determining unit for extracting a defect.

It is further desirable that the arithmetic processing section performs an addition process on each pixel value of the multi-valued image corresponding to a pixel address forming the same label to calculate a total sum for each label.

[0010] The arithmetic processing unit may further calculate a sum of pixel values of the multivalued image corresponding to the contour address of the processing unit among the pixel values of the multivalued image corresponding to the pixel addresses forming the same label. desirable.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A case where a shadow mask is used as a periodic pattern will be described below.

First, the entire apparatus will be described. FIG.
1 is a diagram for explaining an overall configuration of a periodic pattern defect inspection apparatus according to an embodiment of the present invention. This is an inspection apparatus that uses a shadow mask as an inspection target and performs transmission illumination.

In FIG. 4, a shadow mask 44 to be inspected is placed on a transport device 43, and the transport controller 43 moves the transport device 43 in the direction of the arrow. Then, the transport device 4 on which the shadow mask 44 is placed
3 is passed below the CCD line sensor camera 41. At this time, the shadow mask 44 is illuminated from below the transfer device 43 by the linear light source 42, and the transmitted light passing through the hole of the shadow mask 44 placed on the transfer device 43 is converted into C light.
An image is taken by the CD line sensor camera 41. Of course, the transport device has a structure in which the light of the linear light source is passed and the shadow mask 44 placed on the device is illuminated.

The shadow mask 4 obtained by the above photographing
An image processing unit 45 performs image processing of a signal based on light transmitted through the hole 4 to perform a defect inspection. This inspection result is displayed on the man-machine interface 47. In this embodiment, the man-machine interface 47 also serves as a transport control device.

In the apparatus according to the embodiment of the present invention, the image captured by the CCD line sensor camera uses a transmission image, but the image captured may be a transmission image or a reflection image.

That is, an image obtained by photographing a pattern has a C side on one side with respect to a workpiece such as a shadow mask.
A photographing unit such as a CD camera is placed, and an irradiation light source is placed on the other side. The transmitted light is a transmitted image of a penetrated processed part (a hole in a shadow mask) formed on a work such as a shadow mask. It may be. Alternatively, a CCD camera or the like may be arranged on one side of a work such as a shadow mask, and an irradiation light source may be arranged on the same side, and a reflected image in which irradiated light is reflected by the work such as a shadow mask may be used.

Next, the inspection method of the present invention implemented in the image processing section of the inspection apparatus shown in FIG. 4 will be described with reference to FIG. FIG. 1 used here is a diagram for explaining a pattern defect inspection method according to the embodiment of the present invention.

The inspection method shown in FIG. 1 will be described. In step 1 (S1), the shadow mask to be inspected is photographed with a CCD camera at a resolution enough to resolve a hole as a processing portion. In this embodiment, one hole is formed by 5 × 5 sensor pixels of a CCD line sensor camera.
Is taken so as to correspond to the 25 pixels. Of course, this is not a limitation. However, if the number of pixels is roughly 4 pixels (2 × 2) or about 1 pixel, moiré is generated, and high-precision inspection cannot be performed. Conversely, if the number of pixels is too large, if the number of pixels is 10,000 or more per hole, image processing time such as inspection takes a long time, so that high-precision inspection can be performed, but it is not very practical.

In step 2 (S2), the multi-valued image obtained by the photographing is recorded in the first image memory.
In step 3 (S3), a binarization process is performed at a predetermined slice level at which the multi-valued image can be set as needed. Here, a second image memory different from the first image memory of S2 is used.
In the image memory. However, it is sufficient that these multi-valued image data and binary image data can be stored so that the multi-valued image data to be used later is not lost. In addition, the area to be binarized changes around the outline of the hole depending on the slice level to be set. This slice level is obtained in advance by a test.

In step 4 (S4), a labeling process is performed based on the binarized binary data. The labeling process refers to a process of sequentially numbering connected regions on a binary image. The number is stored in the memory as image data. The processing result is called a label image. In step 5 (S5), an address of a pixel having the same label is obtained based on the processing result (here, referred to as a label result) shown in the label image obtained by the labeling processing.

In step 6 (S6), a pixel value of the multi-level image corresponding to the address is calculated. More specifically, there are two types here, and they will be described below.

In the first arithmetic processing, the arithmetic processing in S6 calculates the total sum for each label by adding pixel values of the multi-valued image corresponding to the address, and calculates the total sum of the holes of the shadow mask. Determine as a representative value. However, the average value obtained from the sum may be used as the representative value.

In the second arithmetic processing, the arithmetic processing in S6 calculates the total sum of the contours of the holes of the shadow mask by adding pixel values of the multi-valued image corresponding to the address and calculates the sum based on the sum. It is determined as a representative value of the hole of the shadow mask. Of course, an average value obtained by summing the contours of the hole may be used as the representative value.

In step 7 (S7), a representative value based on the calculation result is two-dimensionally developed on the image memory. That is, based on the label result, the representative value of the hole is developed on the third image memory in the order of the pattern of the hole of the shadow mask.

Further, S6 and S7 will be described with reference to FIG. FIG. 2A is an image in which an image of a pattern (part) of the hole 21 of the shadow mask is superimposed on the memory 22. Here, there are 14 holes. An area indicated by a rule drawn vertically and horizontally so as to overlap the image of the hole 21 is a memory 22, and “□” surrounded by the rule is a pixel 23 indicating one pixel of memory.

When the labeling process is performed, one label 24 (for example, "1") is provided for each hole 21 which is a connected area.
And the label is recorded in the pixel on the memory 22 corresponding to the hole. Here, since there are 14 holes, the labels “1” to “14” are given to each of the holes 21. Then, based on the addresses of the pixels to which the same label has been given, the sum is obtained by adding the pixel values of the multi-valued image that has already been photographed and recorded in the first image memory. This is a representative value of the hole.

For each hole 21 shown in FIG.
The storage area 2 on the third image memory 25 shown in FIG.
6 is prepared in two dimensions. That is, the storage area 2
6 are used according to the arrangement order of the holes 21. The sum (representative value of the hole 21) is stored in the third image memory 25 in the hole 2
1 and are recorded in the storage area 26 on the third image memory 25 and developed two-dimensionally.

In step 8 (S8), a differentiation process is performed on the two-dimensionally expanded representative value. The differentiation processing is not particularly limited, but here, an operation for obtaining a difference from an adjacent image is performed. In step 9 (S9), a slice process is performed on the result of the differential process using a preset slice level to extract a defect.

By performing such steps, it becomes possible to inspect defects occurring in a periodic pattern such as a shadow mask with high accuracy.

Next, a block diagram of the present inspection apparatus will be described with reference to FIG. FIG. 3 is a diagram for explaining a block configuration of the periodic pattern defect inspection apparatus according to the embodiment of the present invention.

A transmission image of a shadow mask is taken by a CCD camera which is an input unit 31 of the inspection apparatus. After the photographing signal is converted into digital image data (multi-valued image data) by the A / D converter 32, it is stored in the image memory 300 as multi-valued image data. The image memory is prepared as a plurality of image memory boards. Here, the single image memory board is used.

The binarization processing section 33 performs binarization processing on the multi-valued image at a predetermined slice level set in advance. Based on the binarized data thus obtained, the labeling processing unit 34 assigns a number to the connected area of the image and records the number in the memory.

In the addition processing unit 35, based on the label result obtained by the labeling processing, the address of the pixel of the connected area forming the same label is obtained, and the pixel value of the multi-valued image corresponding to the address is added. Then, the sum of each connected region is calculated as a representative value. Here, as described above, the value to be calculated may be the sum of the values of the addresses constituting the contour of the hole to which the label result is attached.

The sum of the pixel values calculated by the addition processing unit 35 is developed two-dimensionally on one image memory substrate on the image memory 300 as an image as a representative value of each connected region (hole). At the time when the image is obtained, an image from which the influence of moire has been eliminated is obtained. In the pass / fail determination unit 37,
Differential processing is performed on the image to emphasize a defective portion, and a defect is extracted based on a preset slice level.

[0035]

As described above, by implementing the present invention, the sensitivity is reduced by the moire caused by the interference between the pixel pitch of the CCD sensor for photographing the periodic pattern and the pitch of the holes of the shadow mask. Thus, highly accurate defect detection can be realized.

[Brief description of the drawings]

FIG. 1 is a view for explaining a periodic pattern defect inspection method according to an embodiment of the present invention;

FIG. 2A is a diagram for explaining a defect inspection method of a periodic pattern according to an embodiment of the present invention; FIG. 2B is a diagram for explaining an image after a labeling process; FIG. Diagram for explaining the image

FIG. 3 is a diagram for explaining a block configuration of a periodic pattern defect inspection apparatus according to the embodiment of the present invention;

FIG. 4 is a view for explaining the overall configuration of a periodic pattern defect inspection apparatus according to an embodiment of the present invention;

[Explanation of symbols]

 Reference Signs List 30 image processing CPU 31 input unit 32 A / D conversion unit 33 binarization processing unit 34 labeling processing unit 35 addition processing unit 36 differential processing unit 37 image quality determination unit 38 image processing device 39 Output unit 300: Image memory

Claims (6)

[Claims] 1. A work having a periodic pattern in which processed parts formed on a base material are periodically repeated, or
A pattern defect inspection method for inspecting a workpiece having a pattern in which an arrangement pitch when arranging a processing portion formed on a base material gradually changes outward from a center of the workpiece, wherein the processing is performed. A step of photographing so as to resolve the part, a step of recording a multi-valued image obtained by the photographing in a memory, a step of performing binarization processing on the recorded multi-valued image at a predetermined slice level, Performing a labeling process based on the binarized binary data;
A step of obtaining an address of a pixel constituting the same label based on the label result obtained by the labeling process; a step of calculating a pixel value of the recorded multi-valued image corresponding to the address; Developing a representative value based on the on a memory, performing a differentiation process on the developed representative value, extracting a defect by comparing the result of the differentiation process with a slice level, A defect inspection method for a periodic pattern, comprising: 2. The arithmetic processing step according to claim 1, wherein each pixel value of a multi-valued image corresponding to a pixel address forming the same label is added to calculate a sum for each label. Defect inspection method for periodic patterns. 3. The multivalued image pixel value corresponding to the contour address of the processed portion, among the pixel values of the multivalued image corresponding to the pixel addresses forming the same label, in the arithmetic processing step according to claim 1. A defect inspection method for a periodic pattern, comprising: calculating the sum of 4. A work having a periodic pattern in which processed parts formed on a base material are periodically repeated, or
A pattern defect inspection apparatus for inspecting a workpiece having a pattern in which an arrangement pitch when arranging a processing portion formed on a base material gradually changes outward from a center of the workpiece, wherein the processing is performed. A photographing unit for photographing a part so as to resolve the part, a first memory for recording a multi-valued image obtained by the photographing, and a binarization of the multi-valued image recorded in the first memory at a predetermined slice level. Binarization processing unit for performing binarization processing, a second memory for recording binarized data obtained by the binarization processing unit, and labeling based on the binarized data recorded in the second memory A labeling processing unit that performs processing, and obtains an address of a pixel configuring the same label based on a label result obtained by the labeling processing unit, and performs an arithmetic processing on a pixel value of the multi-valued image corresponding to the address. An arithmetic processing unit as a representative value, a third memory that expands a representative value based on a result of the arithmetic processing in a memory, a differential processing unit that performs a differentiation process on the representative value expanded in the memory, A defect discriminating unit for extracting a defect by comparing a result of the differential processing with a slice level. 5. The cycle according to claim 4, wherein the arithmetic processing unit adds each pixel value of the multi-valued image corresponding to the pixel address forming the same label to calculate a total sum for each label. Inspection device for characteristic patterns. 6. The sum of the pixel values of the multi-valued image corresponding to the contour address of the processing unit among the pixel values of the multi-valued image corresponding to the pixel addresses forming the same label. The periodic pattern defect inspection apparatus characterized by calculating: