JP2002259951A - Repetitive pattern erasing method, defect inspection method and device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To resolve a problem in application of a conventional pattern erasing process where a linear defect portion which should be left as a defect is erased by pattern erasure or a pattern is left when the pattern has dispersion of pattern pitches. SOLUTION: A concentration difference nearest to 0 is determined from respective concentration differences between a target picture element and a comparison picture element group A arranged on the same column as the target picture element and between the target picture element and a comparison picture element group B arranged in a direction perpendicular to the comparison picture element group A, a concentration difference farthest from 0 is determined by comparing a particular concentration difference to the comparison picture element group A and a particular concentration difference to the comparison picture element group B, and a pattern erased image of adding a maximum concentration difference to a reference concentration in a pattern erasure target image is generated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention mainly relates to a method of erasing a repetitive pattern in an image and a defect inspection for judging a defect in a repetitive pattern included in an electronic device such as a liquid crystal panel, a plasma display, and a semiconductor wafer. The present invention relates to a defect inspection method and apparatus to be performed.

[0002]

2. Description of the Related Art In many cases, electronic devices such as liquid crystal panels, plasma displays, and semiconductor wafers have the same pattern as a partial pattern formed repeatedly as a whole, and detect defects in the repeated pattern as a target. In this case, a defect is detected mainly by the following processing.

[0003] First, for an original image including a repetition pattern that is input as an image,

[0004]

(Equation 1)

[0005] The process of (Equation 1) is performed to perform pattern erasure. Here, V is the density of each pixel of the processed image, I is the density of each pixel of the original image, I1, I2,..., And In are the density of the comparison pixels. Selected pixel. By using a plurality of comparison pixels, it is possible to be robust against disturbance such as noise. F selects the one with the smallest absolute value of each element,
A function that returns that element. C is added as a reference density in the processed image, and in the case of 256 gradations of 8 bits, it is often 128 gradations at the center. This process is called a pattern erasing process, and the image obtained here is called an image after the pattern erasing process or a background image.

Next, a mass which deviates from the background density of the background image is detected and determined as a defect. This process is called a defect detection process.

FIG. 9 is a view for explaining a conventional repeated pattern erasing method. FIG. 9A shows an image before pattern erasing processing in which a rectangular shape is repeated, and FIG. 9B shows an image after processing. It can be seen that the pattern has been erased.

The lower part of FIGS. 9A and 9B shows density profiles 13 and 14 on check lines 11 and 12 in an image before pattern processing and an image after pattern processing. Brighter is closer to 255 gradations, darker is 0
Close to gradation.

In the defect detection processing, an image satisfying a certain density condition is detected as a defect in the image after the processing of FIG. Here, taking the density profile 14 on the check line 12 as an example, white defects are defined above the specified density gradation 135 and black defects are defined below the specified density gradation 120.
Here, the black defect 15 is detected.

[0010]

However, the conventional pattern erasing process has the following problems.

In a liquid crystal panel or a plasma display,
Due to the characteristics of equipment, linear defects often appear as defects. FIG. 10A shows a state in which a linear defect exists on the repeated pattern. FIG.
14A is a processed image obtained by applying a conventional pattern erasing process to FIG. Originally, a linear defect portion that should remain as a defect has been erased by pattern erasure.

This is a phenomenon that has occurred because a density difference between a plurality of comparison pixels and the density difference closest to 0 is selected. If only one comparison pixel is used, the linear defect is not erased, but the robustness is lost. If the plurality of comparison pixels are set in a direction perpendicular to the linear defect, the linear defect does not need to be erased, but only a linear defect in one direction can be dealt with.

When the pattern pitch is not constant due to the manufacturing process and is slightly varied, the pattern pitch deviates from a specified amount because there is only one comparison distance between the target pixel and the comparison pixel. Since the patterns cannot be compared with each other in the portion where the pattern is present, there is a problem that the pattern is not completely erased. It goes without saying that the pattern remaining without being erased is erroneously detected as a defect.

FIG. 11A shows a pattern having a pattern pitch p and a pattern pitch p + α due to variation. FIG. 11B shows a case where the conventional repeated pattern erasing process is applied at the comparison distance p. It can be seen that the pattern is successfully erased at the pattern pitch p, but the pattern is not completely erased but remains at the pattern pitch p + α.

The reasons why the pattern pitch is not constant may be that the pitch is not actually constant due to the manufacturing process, or that moire occurs as a characteristic of the optical system when an image is input and the pitch is not constant. . Further, the pitch may not be constant due to vibration or the like during image capture.

An object of the present invention is to solve the above-mentioned problem and to provide a repetitive pattern erasing method capable of inspecting a repetitive pattern having a linear defect.

Further, the pattern pitch is not constant,
It is an object of the present invention to provide a method in which pattern erasure is completely performed even when there is variation.

[0018]

SUMMARY OF THE INVENTION In order to solve this problem, the present invention provides a pixel of interest, a group of comparative pixels A arranged on the same column as the pixel of interest, and the group of comparative pixels A arranged vertically. A density difference closest to 0 is obtained from the respective density differences with the comparison pixel group B, and the specific density difference for the comparison pixel group A and the specific density difference for the comparison pixel group B are compared with each other.
A pattern erasure image is generated by determining a density difference far from the reference density and adding the maximum density difference to a reference density in the pattern erasure image.

Thus, the above problem can be solved and only the pattern can be properly erased without erasing the linear defect of the repeated pattern having the linear defect.

In addition, since the method of comparing the specific density difference for the comparative pixel group A and the specific density difference for the comparative pixel group B can be selected from those closer to 0, there is no linear defect. In the case of a test object for which is known, the detection method and the hardware of the apparatus can be largely shared.

Further, a comparative pixel group D is provided for the comparative pixel groups A and B in consideration of the shift amount of the pattern pitch, and a specific density difference with respect to the comparative pixel group D is determined. By determining the density difference closest to 0 from each specific density difference between B and the comparative pixel group D and setting the density difference as the specific density difference, even for a pattern having a variation in pattern pitch, without leaving a pattern, The above problem can be solved and the pattern can be properly erased.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which a repetitive pattern erasing method capable of detecting a line defect is applied to a defect inspection of a liquid crystal display for an object to be inspected will be described below with reference to FIGS.

In FIG. 1, which shows a schematic configuration of a pattern defect inspection apparatus, an inspection object 1 is installed, illuminated by an epi-illumination 2, and an imaging device 3 such as a CCD area sensor is provided.
Is imaged. The image data in the image sensor 3 is transferred to the image memory 5 in the computer 4 as a processing device in one-to-one correspondence. The computer 4 stores a processing program 6 for reading the image data in the image memory 5 and performing predetermined processing. Here, the image density is handled in 256 gradations from 0 to 255.

FIG. 2 shows a processing flow of a method for detecting a defect by erasing a repetitive pattern. In FIG.
First, in the image input step of step # 1, the captured image data from the image sensor 3 is stored in the image memory 5 of the computer 4.
Is stored in Next, in the multiple density difference detection step of step # 2, the density difference between the target pixel and the comparison pixel group is obtained. Further, in the specific density difference detecting step of step # 3, a specific density difference is selected from the density differences detected in step # 2. Here, the plurality of comparison pixel groups are as shown in FIG.
A group set in one repetition direction and a B group set in the direction perpendicular thereto are divided into groups, and the processing of step # 2 and step # 3 is performed on each group. Since they are the same, a processing method for the group A will be described here.

The size p previously determined from the pattern pitch
Then, the following processing is performed. The plurality of comparison pixel groups A have a size p
The density difference between the pixel of interest and the plurality of comparison pixel groups A is determined. FIG. 4 shows a relationship between a certain pixel of interest and a comparative pixel group. Comparative pixels 11 and 12 having a size p are set for the target pixel 10.

[0026]

(Equation 2)

In (Equation 2), x is the pixel of interest, and n is an integer multiple of the above. Here, two density differences are obtained, V1 and V
Let it be 2.

Next, in a specific density difference determination step of step # 3, a final output density is determined from the two density differences V1 and V2. Here, the one closest to 0 is selected. For example, when V1 = 3 and V2 = -2, V2 is selected as the specific density difference. This is defined as a specific density difference VA for the comparison pixel group A.

The same operation is performed for the comparison pixel group B which is selected in the vertical direction with respect to the comparison pixel group A. This is defined as a specific density difference VB for the comparison pixel group B.

Then, in the maximum density difference determination step of step # 4, the final output density is determined from VA and VB obtained for each of the groups A and B in the specific density difference determination step. Here, the one farthest from 0 is selected. For example, V
If A = -2 and VB = 5, VB is selected as the maximum density difference.

Next, in the erased image creation process of step # 5,
The reference density in the pattern erased image is added to the maximum density difference obtained in the maximum density difference determination step. The concept of the reference density is the same as that of C described in the conventional example, and the reference density is 0 to 2
In the case of 8 bits of 55 gradations and 256 gradations, 128 gradations are often used.

FIG. 5A is a diagram before the pattern erasing process of a repeated pattern having a linear defect, and FIG. 5B is a diagram after the process. It can be seen that even when the pattern erasing process is performed, the linear defect remains without being erased.

In the present invention, in the maximum density difference determination step of step # 4, the density difference obtained from each of the groups A and B is 0%.
By selecting an object close to 0 without selecting an object far from the same, the same operation as the conventional pattern erasing process can be realized. In this case, the robustness with respect to disturbance such as noise is increased.

The steps other than the step of determining the maximum density difference in step # 4 are shared with the conventional method, so that the hardware and the like can be largely shared with the conventional method.

Next, a method of erasing a repetitive pattern when there is a pattern shift will be described with reference to FIGS. This is an example in which the object to be inspected is applied to a defect inspection in a liquid crystal display. The schematic configuration of the pattern defect inspection apparatus is the same as that in FIG.

The method of erasing a repetitive pattern having a pattern shift is realized by adding an optimum density difference determining step to the processing flow of FIG. 2, as shown in FIG.

As shown in FIG. 7, a group C is provided at a position separated from the target pixel by a specified pattern pitch, and a group C is provided at a position separated from the group C by a variation in pattern pitch (here, α). Two of the D groups are provided as comparison pixel groups of the pattern.

The density difference between the target pixel and the comparison pixel group C is determined by the plural density difference detection step and the specific density difference detection step described above. As a result, the determined density difference is expressed as V
c. Similarly, the density difference between the target pixel and the comparative pixel group D is determined by the multiple density difference detection step and the specific density difference detection step. Thereby, the determined density difference is set to Vd.

Next, one of Vc and Vd which is close to 0 is set as the optimum density difference. At this time, Vd may be appropriately weighted. For example, when Vc = 4 and Vd = 3 with a weighting factor of 1.5, Vd after weighting is 4.5, so Vc = 4 is set as the optimum density difference.

Thus, pattern erasing can be performed even when the pattern pitch varies. FIG. 8B shows the result of applying this method to FIG. 8A. It can be seen that even when the pattern pitch varies, the pattern is completely erased.

Further, by performing a defect inspection after these processes, the inspection object can be inspected with high accuracy.

[0042]

As described above, according to the present invention, only a pattern can be properly erased without erasing a linear defect of a repeated pattern having a linear defect.

Also, in the case of an object to be inspected which is known not to have a linear defect, the detection method and the hardware of the apparatus can be largely shared.

Further, even for a pattern having a variation in pattern pitch, the above problem can be solved and the pattern can be properly erased without leaving the pattern.

The advantageous effect that the defect inspection can be performed accurately can be obtained.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a defect inspection apparatus according to an embodiment of the present invention.

FIG. 2 is a processing flowchart of a repeated pattern erasing method according to the embodiment;

FIG. 3 is an exemplary view showing the arrangement of comparative pixel groups in the repeated pattern erasing method according to the embodiment;

FIG. 4 is a view showing a positional relationship of a comparison pixel group in the repeated pattern erasing method according to the embodiment;

FIG. 5 is an explanatory diagram of an image before and after processing when applied to a case having a linear defect in the repeated pattern erasing method according to the embodiment.

FIG. 6 is a processing flowchart in the case where there is a pattern shift in the repeated pattern erasing method according to the embodiment;

FIG. 7 is a view showing the arrangement of comparative pixel groups in the repeated pattern erasing method according to the embodiment;

FIG. 8 is an explanatory diagram of an image before and after processing when applied to a case where there is a pattern shift in the repeated pattern erasing method according to the embodiment.

FIG. 9 is an explanatory view of a conventional repeated pattern erasing method.

FIG. 10 is an explanatory diagram of an image before and after processing by a conventional repeated pattern erasing method when a linear defect is present.

FIG. 11 is an explanatory diagram of an image before and after processing by a conventional repeated pattern erasing method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Inspection object 3 Image sensor 4 Computer

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Hajime Kawano 1006 Kadoma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (72) Koji Yoshii 1006 Kadoma Kadoma, Kadoma City, Osaka Matsushita Electric Industrial F Terms (reference) 5B057 AA03 BA02 CA02 CA08 CA12 CA16 DA03 DB02 DB05 DB09 DC22 5L096 AA03 AA06 BA03 CA14 DA01 EA37 FA17 GA07 GA51

Claims (5)

[Claims] 1. A method for erasing a repetitive pattern in a grayscale image obtained by imaging an object to be inspected, comprising: a pixel of interest; a comparative pixel group A arranged on the same column as the pixel of interest; A specific density difference determining step of obtaining a density difference closest to 0 from each density difference between the pixel group A and the comparative pixel group B arranged in the vertical direction; a specific density difference for the comparative pixel group A; A maximum density difference determining step of determining a density difference far from 0 by comparing a specific density difference with respect to the pattern density, and a pattern erasing image generating step of adding the maximum density difference to a reference density in the pattern erasing image. Repeated pattern erasing method. 2. The method according to claim 1, wherein the maximum density difference determining step includes a switching step of selecting a method of comparing the specific density difference with respect to the comparison pixel group A and the specific density difference with respect to the comparison pixel group B from 0. 2. The repeated pattern erasing method according to claim 1, wherein 3. A comparison pixel group D in which a shift amount of a pattern pitch is taken into consideration with respect to the comparison pixel groups A and B, and a specific density difference with respect to the comparison pixel group D is determined. 3. The method according to claim 1, further comprising the step of: determining a density difference closest to 0 from each specific density difference between B and the comparative pixel group D and setting the density difference as a specific density difference. The repeated pattern erasing method described above. 4. A defect inspection method, comprising: performing a defect inspection of an object to be inspected after performing the repeated pattern erasure according to claim 1; 5. An image pickup means for picking up an image of an object to be inspected, a pixel of interest in a grayscale image obtained by the image pickup means, a comparative pixel group A arranged on the same column as the pixel of interest, and the comparative pixel group A Means a specific density difference determining means for obtaining a density difference closest to 0 from each density difference with a comparative pixel group B arranged vertically, a specific density difference for the comparative pixel group A and a specific density difference for the comparative pixel group B. A defect inspection apparatus comprising: a maximum density difference determining means for comparing a difference to determine a density difference far from 0; and a pattern erased image generating means for adding the maximum density difference to a reference density in the pattern erased image. .