JPH1114323A - Pattern inspection method and pattern inspection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a pattern inspection method and a pattern inspection device by which elements composed of different kinds of metals and circuit patterns can be inspected with comparatively simple constitution. SOLUTION: After being converted into a black-and-white variable density image video signal Ey by adjusting respective luminances by a luminance adjusting computing element 15 on color video signals EE, EG and EB from a color camera 7 picking up an image of an IC element 5a on a wafer 5 to avoid the excessive approach to a plurality of respective elements or the mutually adjacent distribution of the circuit pattern concentration distribution, the respective elements or circuit patterns are separated by an image processing processor 16, and are converted into a black-and-white variable density image. These method and device are also constituted so as to judge the existence of a defect part 18 by processing this black-and-white variable density image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspection method and a pattern inspection apparatus for inspecting an element or a circuit pattern formed on a semiconductor wafer for defects.
The present invention relates to a pattern inspection method and a pattern inspection apparatus for inspecting a plurality of elements or circuit patterns formed of dissimilar metals for defects.

[0002]

2. Description of the Related Art Conventionally, the most frequently used method for inspecting a defect of an element or a circuit pattern formed on a semiconductor wafer is to prepare a reference image of a normal element or a circuit pattern and input a pattern to be inspected. In general, an inspection method is used in which the reference image is subtracted from the input image after the registration with the image, and a difference is obtained, and a pattern abnormality portion is revealed as a dark portion or a bright portion.

For example, when a pattern to be inspected is photographed by a monochrome (monochrome) camera, Japanese Patent Application Laid-Open No.
And Japanese Patent Application Laid-Open Nos. 8-94537 and 8-94537. In the case where a pattern to be inspected is photographed by a color camera, Japanese Patent Application Laid-Open Nos. As disclosed in Japanese Unexamined Patent Publication No. 7-12742, etc., for each image of red (R), green (G), and blue (B) components, an input image of a pattern to be inspected and a reference image of a normal element or circuit pattern. In general, a method using a so-called difference method of extracting and inspecting a difference between the two is used.

[0004]

In the case of inspecting a plurality of elements or a circuit pattern formed of a dissimilar metal on a semiconductor wafer, first, it is necessary to distinguish a specific pattern from another pattern. In the black-and-white grayscale image information, when the densities of different kinds of metals are similar, even if the density range is appropriately extracted to extract a specific different kind of metal pattern and binarized to make a black-and-white pattern, the other densities similar to each other are obtained. There is a possibility of overlapping with a different kind of metal pattern, and it is substantially impossible to extract a specific pattern when the density of the different kind of metal is similar based on the black and white image information.

[0005] The above problem can be solved by a color image processing system.
Image processing of each of the green (G) and blue (B) components,
Since a specific pattern can be extracted only after performing a complicated operation such as synthesis, a process and an apparatus for performing these processes are complicated, and an inspection time is increased and an inspection cost is increased.

[0006] The present invention is to solve the above-mentioned problems,
An object of the present invention is to provide a pattern inspection method and a pattern inspection apparatus which enable an element or a circuit pattern made of dissimilar metal to be inspected with a relatively simple configuration.

[0007]

In order to achieve the above object, a pattern inspection method according to the present invention is directed to a method of inspecting a device or a circuit surface in which a plurality of elements or circuit patterns made of a plurality of different materials are substantially on the same plane. In a pattern inspection method in which the reflected light is photographed by a color camera and each pattern is inspected by image processing, the elements or circuit patterns are separated by a density distribution corresponding to each material forming the plurality of elements or circuit patterns. In image processing,
First, after adjusting the luminance of each of red, green, and blue components on a color video signal photographed by the color camera in order to avoid excessive proximity to a plurality of adjacent elements or circuit pattern density distributions, It is characterized in that the element or circuit pattern is separated and converted into a black and white gray image, and the black and white gray image is subjected to image processing for determination.

Further, the pattern inspection apparatus according to the present invention comprises:
In a pattern inspection apparatus in which a plurality of elements or circuit patterns made of a plurality of different materials are exposed to light on an element or a circuit surface on substantially the same plane, the reflected light is photographed by a color camera, and each pattern is inspected by image processing. A luminance adjustment computing unit that performs each luminance adjustment of red, green, and blue components on a color video signal photographed by the color camera to avoid excessive proximity to a plurality of adjacent elements or circuit pattern density distributions; An image processing device that separates each element or circuit pattern as an input of the image processing, converts the output from the luminance adjustment arithmetic unit into a black and white gray image, and performs image processing on the black and white gray image for determination. Features.

Since the present invention is constructed as described above, the red (R), green (G), and blue (B) components of the color video signal are adjusted in brightness using a color camera and converted into a black-and-white shading video signal. By providing a luminance adjustment arithmetic unit that performs the above processing and using the output of the luminance adjustment arithmetic unit as an input signal for image processing, it is possible to inspect patterns of elements and circuits made of dissimilar metals substantially only by binarization processing.

[0010] That is, the luminance adjustment arithmetic unit is provided with red (R),
The video signals of the green (G) and blue (B) components are represented by E _R ,
E _G and E _B , the coefficients of the video signals E _R , E _G and E _B are α, β, and γ, and the black-and-white shading video signal of the calculation result output from the _luminance adjustment calculator is E _Y , ｛E _Y = α
It has a function of performing an operation of E _R + βE _G + γ E _B } and outputting a black-and-white shading video signal.

Then, several sample images to be inspected are sampled from an IC (semiconductor integrated circuit) wafer to be inspected, and arithmetic processing by the luminance adjustment arithmetic unit and averaging processing by an image processing apparatus are performed. To create a standard image in black and white shaded images (standard image creation function).

Further, the standard image is subjected to image processing by an image processing apparatus, binarized in a predetermined density range and converted into a black-and-white pattern to extract a dissimilar metal pattern (dissimilar metal pattern extraction function). The metal pattern to be extracted is expressed in white for convenience.

Then, from the IC wafer, red (R), green (G), and blue (B) color images of a portion to be inspected are sequentially collected, and the brightness adjustment arithmetic processing unit performs the brightness adjustment arithmetic processing. To create a black and white inspection image (inspection image generation function).

Then, the difference between the inspection image and the standard image is obtained, the image of the defective portion is made obvious, the difference result image is binarized in a predetermined density range, converted into a black and white pattern, and the defective portion is extracted (defective portion). Extraction function). In addition, a defective part is represented by white for convenience.

The logical product (A) of the white portion of the defective portion extracted black and white pattern and the dissimilar metal pattern extracted black and white pattern
ND) to determine which metal pattern has a defect (pattern position determining function of a defective portion).

With the above arrangement, when the density of different metal patterns is similar in both color image information and black and white gray image information, the density adjustment of the different metal patterns having similar densities is performed by the luminance adjustment arithmetic unit using the color image information. Can be converted to a black-and-white grayscale image by separating them from each other, so that a dissimilar metal pattern can be easily extracted.

Then, by comparing and comparing the defective portion and the metal pattern portion, it is possible to know which metal pattern portion the defect is mixed in. By the above-mentioned comparison and collation, it is possible to rank up to whether a defective portion straddles between different metal patterns, stops within one metal pattern, or determines whether or not a defect should be made by judging the size of an abnormal portion on a different metal. Become.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a pattern inspection method and a pattern inspection apparatus using the same according to the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an overall configuration of a pattern inspection apparatus according to the present invention, FIG. 2 is a diagram showing a configuration of a standard image creation function in the pattern inspection apparatus according to the present invention,
FIG. 3 is a diagram showing a configuration of a function of a positioning process in the pattern inspection apparatus according to the present invention, FIG. 4A is a plan view showing a configuration of an IC element in a wafer to be inspected, and FIG. 5A is a plan view showing a configuration of an IC element in a wafer to be inspected having a defect, FIG. 5A is a diagram showing an image of a defect revealed by a difference method, FIG.
6A is a frequency distribution diagram of density data of an image captured by a monochrome camera, FIG. 6B is a frequency distribution diagram of density data of a red (R) component image captured by a color camera, and FIG.
FIG. 6D is a frequency distribution diagram of density data of a green (G) component image captured by a color camera, FIG. 6D is a frequency distribution diagram of density data of a blue (B) component image captured by a color camera, and FIG.
(A) shows a red (R) component and a blue (B) component by a brightness adjustment arithmetic unit.
FIG. 7B is a frequency distribution diagram of density data of a black-and-white grayscale image obtained by synthesizing the video signal of the component, and FIG. FIG. 8A is a diagram showing a pattern of pattern 1 extracted with the density threshold a shown in FIG. 7B, and FIG.
7B is a diagram showing a pattern of pattern 2 extracted at the density thresholds b and c shown in FIG. 7B, and FIG.
FIG. 8D shows a pattern 3 extracted with the density thresholds d and e shown in FIG. 7B, and FIG. 8D shows a pattern 4 extracted with the density threshold f shown in FIG. FIGS. 9A and 9B are diagrams showing the collation of the positions of the defective portion and the metal pattern.

In FIG. 1, a pattern inspection method according to the present invention and an inspection object by a pattern inspection apparatus using the same include, for example, a plurality of different materials formed on a wafer 5 of a semiconductor element such as a Hall element wafer. It is possible to inspect for defects such as pinholes, scratches, plating defects, pattern formation defects, etc., occurring on the element surface or circuit surface where the element or circuit pattern is substantially on the same plane.

The semiconductor IC (integrated circuit) element 5a has several
It is formed on a flat wafer 5 having a size of about 1 cm to several tens of cm, and has a size of about 1 mm square.

In order to inspect the IC element 5a, as shown in FIG. 1, a color camera 7 is set on a microscope 6 and an enlarged photograph is taken. The microscope 6 has a cylinder 6a for setting a color camera 7 and a revolver 6c for setting an objective lens 6b.
And a microscope stage 6d. An XY movement stage 6e is provided above the microscope stage 6d to move in the XY coordinate direction with respect to the microscope stage 6d and to place the wafer 5 to be inspected. ing.

An X axis moving motor 8a for moving the XY moving stage 6e in the X coordinate direction in order to sequentially photograph the IC elements 5a on the wafer 5 mounted on the XY moving stage 6e, A Y-axis moving motor 8b for moving the XY moving stage 6e in the Y-coordinate direction, and a stage controller 9 for controlling rotation of the X-axis moving motor 8a and the Y-axis moving motor 8b are provided. The stage controller 9 outputs and controls a motor drive signal to each of the X-axis movement motor 8a and the Y-axis movement motor 8b.

A ring-shaped halogen illumination lamp 10 is arranged between the objective lens 6b and the wafer 5 to irradiate light on the surface of the IC element 5a on the wafer 5. The light is guided to the halogen illumination lamp 10 by the bundle cable 12 and the I
Light is applied to the surface of the C element 5a, and the reflected light is photographed by the color camera 7.

The color camera 7 is driven by a control signal from the color camera controller 13 and outputs color video signals E _R , E _G , and E _B. This color video signal E
_R, E _G, luminance adjustment computing device 15 of the inspection control panel 14 E _B
The image signal is converted to a black-and-white gray-scale image video signal E _Y and input to an image processor 16 serving as an image processing device.

The image processing processor 16 is a personal computer for driving operation (hereinafter simply referred to as "driving personal computer").
Inspection conditions are received from 17) by data communication, inspection processing is performed, and the result is returned to the operation personal computer 17 again by data communication. Driving operation PC 17 is keyboard 17
a and the operation monitor 17b, the operation is performed, and the inspection result is output to the form printer 17c.

Next, the function of creating a standard image in detecting a defect of the IC element 5a will be described with reference to FIG. Since the standard image must not include a defective portion, a method is employed in which, for example, 20 images are photographed while aligning the image patterns, and the average is taken. The alignment of each image pattern is performed by controlling the image processor 16, the stage controller 9, and the operation personal computer 17 so that the image pattern position is perfectly matched for each photographing.

In photographing an image, the color video signals E _R , E _G , and E _B of the red (R), green (G), and blue (B) components output from the color camera 7 are all subjected to luminance. The adjustment arithmetic unit 15 converts the image signal into a monochrome grayscale image video signal _EY , converts it into a monochrome grayscale image, and records it in the image memory of the image processor 16.

In the image memory, the density is divided by the number of times of photographing and recorded.
And store it. If all of the 20 1/20 density images are added, an averaged standard image that does not include a defective portion can be obtained.

Next, a method of aligning each sample image in creating the standard image will be described with reference to FIG. The first image capture follows the basic method above,
One IC elements 5a portion of the predetermined position in the wafer 5 captured by the color camera 7, a color video signal from the color camera 7 E _R, E _G, black-and-white grayscale image video signal E by the luminance adjustment computing device 15 E _{B The} image data is converted into _Y , the density is reduced to 1/20 by the image processor 16, converted into image data, and stored in the image memory in the image processor 16.

FIG. 3 shows the stored monochrome grayscale image data.
As shown in the figure, an X-axis projected density cumulative frequency distribution and a Y-axis projected density cumulative frequency distribution are obtained by differentiating in each of the X-axis and Y-axis directions and projecting the cumulative number of density data with respect to each of the coordinate axes X and Y of this differential image. Get.

Next, a certain distance away from the first photographing position
The second image is taken by the color camera 7 in the same manner as described above.
Shooting and a color video signal from the color camera 7
E_R, E _G, E_BBlack and white shading with the brightness adjustment calculator 15
Image video signal E_YTo 1/20 density
Data is converted to the image memory in the image processor 16
Remember.

An X-axis projected density is obtained by differentiating the stored monochrome gray-scale image data in the X-axis and Y-axis directions in the same manner as described above, and projecting the cumulative number of density data for each coordinate axis X and Y of this differential image. A cumulative frequency distribution and a Y-axis projected density cumulative frequency distribution are obtained.

Next, the alignment between the first 1/20 density image and the second 1/20 density image is performed by adjusting the density of each X-axis and Y-axis projected density cumulative frequency distribution data with respect to each coordinate axis X, Y. The position of the wafer 5 is finely moved by the stage controller 9 so as to minimize the cumulative frequency difference, and the photographing of the second photographing point and the comparison of the X-axis and Y-axis projected density cumulative frequency distribution data are repeated.

Next, while the third image is photographed in the same manner as described above, the first 1/20 density image and the third 1
/ 20 density image is aligned. The same applies to the fourth and subsequent image positions in the same manner.

1/20 of the 20 sheets thus obtained
When all the density images are added together, a standard image of the IC element 5a corresponding to the original density and not including the averaged defect is obtained as shown in FIG.

Next, an embodiment of an inspection image creation function and a defect extraction function by a difference method will be described. FIG. 4 shows an image when the IC element 5a to be inspected has a defective portion.
(B). The inspection image is obtained by controlling the X-axis moving motor 8a and the Y-axis moving motor 8b by the stage controller 9 based on the command from the operation personal computer 17, and the I-image on the wafer 5
The C element 5a is sequentially photographed.

When taking an image, as described above,
The color video signals E _R , E _G , E of the red (R), green (G), and blue (B) components output from the color camera 7.
_B is converted into a black-and-white gray-scale image video signal E _Y by a luminance adjustment arithmetic unit 15 and converted into a black-and-white gray-scale image.

The alignment method between the standard image and the inspection image is the same as the first and Nth alignment methods for creating the standard image. For example, the X-axis projection density cumulative frequency distribution and the Y-axis projection are obtained by differentiating the monochrome image data of the standard image in each of the X-axis and Y-axis directions, and projecting the cumulative number of density data for each coordinate axis of the differentiated image. Obtain the concentration cumulative frequency distribution.

Next, the monochrome density image data of the inspection image is differentiated in each of the X-axis and Y-axis directions, and an X-axis projected density cumulative frequency distribution is obtained by projecting the cumulative number of density data for each coordinate axis of the differential image. And the Y-axis projection density cumulative frequency distribution are obtained.

The wafer 5 is finely moved by the stage controller 9 so that the frequency difference between the respective X-axis and Y-axis projected density cumulative frequency distribution data obtained from the standard image and the inspection image with respect to each coordinate axis is minimized. Is repeated by repeating the photographing of the photographing point and the comparison of the X-axis and Y-axis projected density cumulative frequency distribution data.

After aligning the standard image and the inspection image, and subtracting the standard image from the inspection image to obtain a difference,
An exposed image of the defective portion 18 as shown in FIG. It should be noted that if the visualized image of the defective portion 18 as shown in FIG. 5A is binarized with an appropriate density and the defective portion 18 is expressed in pure white, the defective portion 18 as shown in FIG. An extracted image is obtained.

Next, the function of the brightness adjustment arithmetic unit 15 and the function of extracting a dissimilar metal pattern will be described. In order to determine which metal pattern the defective portion 18 is in, it is necessary to recognize the metal pattern.

Therefore, according to the present invention, even when the densities of different kinds of metal patterns are similar in the color image information and the black-and-white gradation image information, it is possible to clearly extract the different kinds of metal patterns by the function of the luminance adjustment arithmetic unit 15. You can do it.

The case of an IC element 5a on a wafer 5 made of a plurality of metals having different patterns 1 to 4 will be described with reference to FIGS. IC element 5 without defect
The standard image of a is shown in FIG. FIG. 6A shows an image of this IC element 5a taken by a monochrome (black and white) camera, and the distribution of two-dimensional density data is expressed.

In FIG. 6A, the abscissa represents the density expressed in digital values, black represents 0, white represents 255 digits, and the vertical axis represents the frequency with respect to the density. In this case, as shown in FIG. 6A, the density distributions of the pattern 1 and the pattern 2 overlap, and the brightness becomes the same, making it difficult to determine. Further, the density distributions of the pattern 3 and the pattern 4 partially overlap, and complete brightness separation is impossible.

Therefore, when the IC element 5a is photographed by the color camera 7, and the distribution of the two-dimensional density data is expressed for each of the red (R), green (G), and blue (B) components, the red (R) )
6B, the component of green (G) is shown in FIG. 6C, and the component of blue (B) is shown in FIG. 6D. still,
6 (b) to 6 (d), similarly to FIG. 6 (a), the horizontal axis represents the density, black represents 0, pure white represents 255 digits, and the vertical axis represents the frequency with respect to the density.

As shown in FIG. 6B, in the distribution of the density data of the red (R) component, the density distributions of pattern 1 and pattern 2 completely overlap, and the density distributions of pattern 3 and pattern 4 also partially overlap. . Further, as shown in FIG. 6C, in the distribution of the density data of the green (G) component, the density distributions of the pattern 1 and the pattern 2 completely overlap, and the pattern 3 overlaps the density distributions of the patterns 1, 2, and 4. Some overlap. Further, as shown in FIG. 6D, in the distribution of the density data of the blue (B) component, the density distributions of the pattern 2 and the pattern 3 completely overlap.

As described above, even if an image is taken by a monochrome camera, red (R), green (G),
Even in the image of the blue (B) component, the density of any one of the metal patterns overlaps, and it is impossible to completely extract the different metal patterns.

In the present invention, these disadvantages are solved by using the luminance adjustment arithmetic unit 15. First, an image of the IC element 5a on the wafer 5 is taken by the color camera 7,
Each of the color video signals E _R , E _G , E _B is multiplied by a coefficient α, β, γ, respectively, and ｛α · E _R
A luminance adjustment calculator 15 is provided to output a composite video signal of + β · E _G + γ · E _B }.

Here, for example, in this embodiment, α = 1,
By setting β = 0 and γ = 1, the output of the luminance adjustment calculator 15 with respect to the red (R), green (G), and blue (B) component signals input to the brightness adjustment calculator 15 It will be set to red (R) + blue (B), whereby, apparently output luminance adjustment computing device 15, becomes converted into black-and-white grayscale image video signal E _Y.

The video signal obtained as a result is imaged to show the distribution of density data of a two-dimensional image.
become that way.

Next, a method for extracting the pattern of each metal element from the result calculated by the brightness adjustment calculator 15 will be described. FIG. 7B shows threshold values a, b, c, d, e, and f for density separation in FIG. 7A.

In FIG. 7B, when pattern 1 is cut out, binarization is performed by image processing at the density of the threshold value a, and the portion having the density equal to or less than the threshold value a is completely white (density 255).
Digit). Similarly, when the pattern 2 is cut out, the threshold b
And the part of the density between the threshold value c and pure white (density 255)
Digit).

When the pattern 3 is cut out, a portion having a density between the threshold value d and the threshold value e may be extracted by image processing so as to be completely white (a density of 255 digits). When the pattern 4 is cut out, binarization may be performed at a density of the threshold value f by image processing, and a portion having a density of f or more may be extracted so as to be completely white (density of 255 digits).

In this embodiment, threshold values for concentration separation are a = 30 digits, b = 50 digits, and c = 30 digits.
The measurement was performed by setting 120 digits, d = 140 digits, e = 200 digits, and f = 220 digits. The result of extracting the pattern of each metal element in this manner is shown in FIG.
(A) to (d). 8A shows a pattern 1 extracted at the density threshold a, FIG. 8B shows a pattern 2 extracted at the density thresholds b and c, and FIG. FIG. 8D shows a pattern 3 extracted with the threshold values d and e, and FIG. 8D shows a pattern 4 extracted with the density threshold value f.

The principle of the operation is as follows. In the distribution of the density data of one of the red (R), green (G), and blue (B) components, the metal pattern that is too close or overlaps the other metal pattern. This utilizes the effect of separation based on the concentration data of the components.

As shown in FIG. 4B, the distribution of each metal element composed of a plurality of different materials can be clearly separated in FIG.

Next, the function of determining the pattern position of the defective portion 18 will be described. In order to know in which metal pattern portion the defective portion 18 is mixed, an extracted image of the defective portion 18 as shown in FIG. 5B in which the defective portion 18 is expressed in white, and the pattern of each metal element is expressed in pure white 8 (a) to 8 (d) and the logical product (AN
By taking D), it is possible to determine on which metal pattern the defective portion 18 is located.

The defect 18 shown in FIG.
9 (a) and 9 (b) show an example on the pattern 4 and FIG.
In this case, if two defective portions 18 are present over the two patterns 4 and the pattern 2 and the material of the defective portion 18 is not an insulating foreign material, that is, if the defective foreign material is a conductive foreign material, the two patterns 4 It is possible to judge the possibility of an important defect that short-circuits between four.

[0060]

Since the present invention has the above-described configuration and operation, it can be used for color image information or black-and-white shaded image information.
When the densities of different types of metal patterns are similar, it is possible to convert the densities of different types of metal patterns having similar densities to a black and white grayscale image by separating them using a color adjustment information using a luminance adjustment arithmetic unit. Can be easily extracted.

By comparing and comparing the defective portion with the metal pattern portion, it is possible to know in which metal pattern portion the defective portion is mixed. By the above-described comparative check, the defective portion straddles between different metal patterns. It is possible to determine whether or not a defect should be determined by judging the size of a defective portion, whether it is within one metal pattern, or the size of a defective portion on a dissimilar metal.

[Brief description of the drawings]

FIG. 1 is a diagram showing an overall configuration of a pattern inspection apparatus according to the present invention.

FIG. 2 is a diagram showing a configuration of a standard image creation function in the pattern inspection apparatus according to the present invention.

FIG. 3 is a diagram showing a configuration of a function of a positioning process in the pattern inspection apparatus according to the present invention.

FIG. 4A is a plan view showing a configuration of an IC element in a wafer to be inspected, and FIG. 4B is a plan view showing a configuration of an IC element in a wafer to be inspected having a defective portion.

FIG. 5A is a diagram illustrating an image of a defective portion that is revealed by a difference method, and FIG. 5B is a diagram illustrating an image of a defective portion extracted.

6A is a frequency distribution diagram of density data of an image captured by a monochrome camera, FIG. 6B is a frequency distribution diagram of density data of a red (R) component image captured by a color camera, and FIG. FIG. 4 is a frequency distribution diagram of density data of a green (G) component image captured by a camera, and FIG. 4D is a frequency distribution diagram of density data of a blue (B) component image captured by a color camera.

7A is a frequency distribution diagram of density data of a black-and-white grayscale image in which a video signal of a red (R) component and a blue (B) component are synthesized by a brightness adjustment calculator, and FIG. 7B is a brightness adjustment calculation. FIG. 7 is a diagram showing density thresholds for extracting respective metal patterns in a frequency distribution diagram of density data of a monochrome grayscale image synthesized by a filter.

8A is a diagram showing a pattern 1 extracted at the density threshold a shown in FIG. 7B, and FIG.
FIG. 7B shows a pattern of pattern 2 extracted with density thresholds b and c shown in FIG. 7B, and FIG. 7C shows a pattern of pattern 3 extracted with density thresholds d and e shown in FIG. FIG. 9D is a diagram showing a pattern 4 extracted at the density threshold f shown in FIG. 7B.

FIGS. 9A and 9B are diagrams showing the collation of the position between a defective portion and a metal pattern.

[Explanation of symbols]

1-4 Pattern 5 Wafer 5a IC element 6 Microscope 6a Cylindrical body 6b Objective lens 6c Revolver 6d Microscope base 6e XY moving stage 7 Color camera 8a X-axis moving motor 8b Y-axis moving Motor 9 ... Stage controller 10 ... Halogen lamp 11 ... Halogen lamp light source 12 ... Light guide fiber bundle cable 13 ... Color camera controller 14 ... Inspection control panel 15 ... Brightness adjustment calculator 16 ... Image processing processor 17 ... Operation personal computer 17a ... keyboard 17b ... driving operation monitor 17c ... form printer 18 ... defect portion R ... red G ... green B ... blue _{E R, E G, E B} ... color video signal E _Y ... black-and-white grayscale video signal a to f ... threshold

Claims (2)

[Claims] 1. A method in which a plurality of elements or circuit patterns made of a plurality of different materials are exposed to light on an element or a circuit surface having substantially the same surface, the reflected light is photographed by a color camera, and each pattern is inspected by image processing. In the pattern inspection method, in separating an element or a circuit pattern by a density distribution corresponding to each material forming the plurality of elements or circuit patterns and performing image processing, first, a plurality of adjacent elements or circuit pattern density distributions are adjacent to each other. After adjusting the red, green, and blue components of the color video signal captured by the color camera in order to avoid excessive proximity to the distribution, each element or circuit pattern is separated and converted to a black and white grayscale image. A pattern inspection method, wherein image processing is performed on the black-and-white grayscale image to make a determination. 2. A method in which a plurality of elements or circuit patterns made of a plurality of different materials are exposed to light on an element or a circuit surface having substantially the same surface, the reflected light is photographed by a color camera, and each pattern is inspected by image processing. In the pattern inspection apparatus, a luminance for performing each luminance adjustment of red, green, and blue components on a color video signal photographed by the color camera in order to avoid excessive proximity to a plurality of adjacent distributions of each element or circuit pattern density distribution. An image processing device that separates each element or circuit pattern into a black-and-white gray image by using an output from the luminance adjustment calculator as an input of image processing, and performs image processing on the black-and-white gray image to determine A pattern inspection apparatus characterized by having: