JPH01274280A - Method and device for inspecting pattern - Google Patents

Abstract

PURPOSE:To accurately perform the inspection of defect by comparing an inspection pattern with a master pattern after matching them corresponding to the inclination of the inspection pattern. CONSTITUTION:Image data from a CCD camera 3 are stored in shift register groups (9a-9d) sequentially. A controller 4 read outs the image data stored in the shift register groups (9a-9d), and stores them in a bit map memory 10a as the read data of the inspection pattern. Also, the readout of a table 1 and the register groups (9a-9d) are controlled corresponding to the inclination of the inspection pattern 3. The existence of the defect in a pattern formed in an object 2 to be inspected can be inspected by comparing the contour image data of the inspection pattern stored in the bit map memory 10a with that of the master pattern stored in a bit map memory 10c.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a pattern defect inspection method and apparatus for inspecting the presence or absence of defects in patterns formed on masks, reticles, printed circuit boards, etc.

[Conventional technology]

For example, as a conventional pattern inspection method,
Publication No. 21523 (first conventional example), JP-A-59-126
There is one described in Publication No. 830 (second conventional example).

A first conventional example is a photomask defect inspection method, which includes:
A signal for detecting a scanning signal corresponding to a pattern to be inspected obtained by providing at least three marks as alignment marks on a mask and scanning the center position of each of these marks in one direction with a light irradiation unit. Detected by the detection unit, the amount of expansion/contraction in the X direction, the amount of expansion/contraction in the Y direction, and the deviation in orthogonality of the mask are determined from each coordinate of these center positions, and based on these, the reference signal generating unit (mask pattern) is scanned for comparison. By correcting the position ↑n information when the signal is generated, erroneous judgments due to positional deviation between the pattern to be inspected and the mask pattern are prevented.

In the second conventional example, scanning is performed with a one-dimensional line sensor used for scanning tilted at a fixed angle with respect to a scanning area in a subject to prevent deviation in the line direction with respect to the image.

[Problem to be solved by the invention]

However, in the first conventional example, it is necessary to provide at least three alignment marks on the member to be inspected separately from the pattern to be inspected, and these alignment patterns are read to detect the center position coordinates. There was an unresolved problem that the device would be complicated and large, such as the need for a pattern alignment mark reading means and the need for a pet.

Furthermore, the second conventional example described above is effective for correcting the read pattern when the pattern to be inspected is continuously moved in the Y direction while scanning the one-dimensional line sensor in the X direction. In order to correct the deviation between the inspection pattern and the mask pattern, there is an unresolved problem that the tilt of the one-dimensional line sensor cannot be controlled without separately detecting the deviation angle between the pattern to be inspected and the mask pattern. there were.

Therefore, the present invention has been made by focusing on the unresolved problems of the conventional example, and it is possible to detect the problem even when the pattern to be inspected has an angular deviation without providing a special alignment mark. To provide a pattern inspection method and apparatus capable of solving the problems of the conventional example by reading the pattern to be inspected while scanning it in the X and Y directions in this state and correcting angular errors during image processing. It is an object.

[Means for Solving the Problem] In order to solve the above problem, the invention of claim (1) scans an object to be inspected on which a pattern to be inspected is formed and an imaging device relatively one-dimensionally or two-dimensionally. In the pattern inspection method, defects in the inspection pattern are inspected by obtaining inspection pattern data and comparing the inspection pattern data with master pattern data stored in advance. The pattern data is divided into a plurality of pixel groups and stored, and the slope of the test pattern is calculated by comparing the stored test pattern data with the master pattern data, and the test pattern is adjusted according to the slope of the test pattern. The present invention is characterized in that the reading order of either of the master pattern data and the master pattern data is changed to match the two pattern data, and then the two are compared. Here, it is preferable to perform exclusive OR with the inspection pattern data and the master pattern data.

Furthermore, the invention of claim (3) obtains inspection pattern data by relatively scanning the inspection object on which the inspection target pattern is formed and an imaging device in one or two dimensions, and stores the inspection pattern data and the inspection pattern data in advance. In the pattern inspection apparatus, the inspection pattern data is inspected for defects by comparing the inspection pattern data with the master pattern data obtained by the imaging device, and the inspection pattern data for each scanning line from the image pickup device is divided into a plurality of pixel groups and stored therein. a tilt detecting means for detecting the tilt of the test pattern by comparing the test pattern data stored in the divided storage means with the master pattern data; A data matching means for changing the reading order of any of the master pattern data to match both pattern data, and a comparison means for comparing the inspection pattern data and the master pattern data matched by the data matching means. It is characterized by Here, it is preferable that the comparison means extracts the defective portion by performing an exclusive OR of the inspection pattern data and the master pattern data.

[Effect]

In this invention, inspection pattern data for one scan obtained by scanning an object to be inspected is divided into a plurality of pixel groups and stored, and the divided and stored inspection pattern data and master pattern data are compared. By detecting the tilt error between the two and changing the data reading order of either the inspection pattern data or the master pattern data according to this tilt error, the consistency of the pattern data between the two is ensured. By comparing both pattern data in a matched state, accurate pattern inspection can be performed with tilt errors corrected without providing any special alignment marks or means for reading them.

〔Example] Embodiments of the present invention will be described below based on the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention.

In the figure, reference numeral 1 denotes an R mounting table on which a test object 2 having a predetermined inspection pattern P formed on the surface of a mask, a reticle, a printed circuit board, etc. is placed and is movable at least in the XY force direction. X-axis drive motor 1a and Y-axis drive motor 1b
is moved in the XY force direction by

A CCD camera 3 as an imaging device is fixedly arranged on the upper surface side of the mounting table 1 at a position facing the object to be inspected 2.
A CD camera 3 images an inspection pattern P of an object to be inspected 2 and outputs the image data.

The 'R mounting table and the CCD camera 3 are controlled by a control device 4 including, for example, a computer, and the image data of the CCD camera 3 is processed by the control device 4.

The control device 4 includes a motor controller 7 that controls the X-axis motor 1a and the Y-axis motor 1b, a camera controller 8 that controls the CCD camera 3, and eight shift registers R1 for one scanning line of image signals from the CCD camera 3.
A plurality of shift register groups 9a to 9d connected in series, for example, four shift register groups 9a to 9d, which are divided and stored in ~R6, and four bitmap memories 10a, 10b, 10c as image memories.
and 10d, an ifffp device 1), a magnetic disk device 12, a CRT display 13, a printer 14, and a defect inspection circuit 20 are connected. Magnetic tape device 1
) includes input data (
A magnetic tape on which CAD data (hereinafter referred to as CAD data) is recorded is attached. Here, the CAD data is data necessary for forming the mask pattern M, as shown in FIG. 2, and has a format similar to that of a numerical control device, etc. Selection data D of aperture A1 and aperture A2 scanning in the Y-axis direction
IO, Dll, aperture A1 and A2 movement coordinate data (XXi, Y7,) and shutter open data D
01. For example, in order to scan the upper half of the mask pattern M, 'G'OOX Il
l Y□Do 2D10'' to close the shutter and move aperture A to the initial position (X+, Y+), then r G 01 XxzYy□DOIJ to open the shutter and move aperture A1 to the end position (xz, yz).

The control device 4 executes a predetermined process according to a program written in a built-in ROM in advance, and first, the image data for four scanning lines output from the CCD camera 3 are sequentially transferred to shift register groups 9a to 9d as buffer memories. The left and right edge coordinates (X p + , Y
p I) and (X F t , Y P Z ), and these edge coordinates (X p + , Y P I
) and (XP2. Yrt) and the edge coordinates of the same pattern of mask pattern M (X Mr , Y , 4
I) and (XM2. YMZ) to create a mask pattern M
The slope tan θ of the test pattern P is calculated, and the successive address order for each of the shift registers R1 to R8 of the shift register groups 9a to 9d is determined from this slope tanθ, thereby performing coordinate conversion, and according to this successive address order, the shift register The image data stored in the groups 9a to 9d is read out and stored in the video node map memory 1.
0a as test pattern read data. In addition,
The deviation of the CCD camera 3 in the line direction due to the inclination of the inspection pattern P is corrected by moving the table in the line direction during scanning by a predetermined amount set in advance according to the inclination.

The control device 4 also generates coordinate data (edge X) indicating the outer shape of the mask pattern M from the CAD data.

y coordinate as numerical data) and store this as master pattern data in the magnetic disk device 12,
Outline image data representing the outline is formed from the master pattern data stored in the magnetic disk device 12, and this is stored in the bitmap memory 10b, and the inspection data stored in the bitmap memory 10a in a similar manner is The outline image data of the pattern data is formed and stored in the bitmap memory IOC, and the outline image data stored in both memories 10b and 10c is transferred to the defect inspection device 20.
By comparing the two, the pattern formed on the object to be inspected 2 is inspected for defects. If there is a defect, the data of the defective part is stored in the bitmap memory 10.
d, and outputs this stored data to the CRT display 14 and printer 15 as necessary.

As shown in FIG. 3, the defect inspection circuit 20 selects the input signal supplied to one input terminal 1 when the selection input signal supplied to the select input terminal S has a logical value of "0°". When the logical value is "1'", the other input terminal I
two selectors 2 for selecting the input signals supplied to 2;
0A and 20B, and an exclusive OR circuit 21 whose selected output is supplied to the input side, respectively. One input terminal 1) of the selector 20A receives a readout signal from the center pixel W9 of the inspection window WHO corresponding to a mask pattern M to be described later. The readout signal of each pixel WMI to WM of the inspection window W14, which will be described later, is sent to the select terminal S via the I10 interface 22 and the OR circuit 23, and the center pixel wps of the inspection window W corresponding to the inspection pattern P, which will be described later. is the above I10
The center pixel WP5 of the inspection window WP corresponding to the inspection pattern P, which will be described later, is inputted through the interface 22, and one input terminal (1) of the selector 20B is inputted via the interface 22.
A readout signal of each pixel W, to WP9 of the inspection window WP, which will be described later, is sent to the other input terminal 2 through the I10 interface 22.
and the center pixel W of the inspection window W corresponding to the mask pattern M is connected to the select terminal S via the OR circuit 24.
,4. are respectively input via the I10 interface 22. Further, the output signal of the exclusive OR circuit 21 is outputted to the control device 4 via the I10 interface 22 as a defect inspection signal.

Next, the operation of the above embodiment will be explained with reference to flowcharts of FIGS. 4 to 8 showing the processing procedure of the control device 4.

That is, as shown in FIG. 4, in step (2), the magnetic tape device 1) is driven to read the CAD data of the mask pattern M, and the contour data is stored in the bitmap memory 10b.
Executes mask pattern contour data registration processing to be written to.

Next, the process proceeds to step (3), where image data from the CCD camera 3 is read, and based on this, when the inspection pattern is tilted with respect to the mask pattern, the tilt is corrected and the inspection is performed to register it in the bitmap memory 10a. Execute pattern coordinate conversion processing.

Next, move to step ■ and focus map memory 1.
A test pattern contour data registration process is executed to form contour data of the test pattern stored in 0a and register it in the bitmap memory 10C.

Next, proceed to step ■, and bitmap memory 1
The contour data of the mask pattern and inspection pattern stored in bitmap memory 10d are read out and compared and judged to form defect data.
Execute the defect inspection process registered in .

Next, the process proceeds to step ■, where it is determined whether or not to end the inspection. If the inspection is to be continued, the process proceeds to step ■, where it is determined whether or not an inspection start command to start the next inspection has been input. If the test start command is not input, the process remains on standby; if the test start command is input, the process returns to step (2), and if the test is to be completed, the process is terminated.

Then, in the master pattern contour data registration process of step (2), as shown in FIG. In this state, drive the magnetic tape device 1) to read the CAD data in step (a), then proceed to step (b) to determine the starting point of the outer shape of the mask pattern M from the aperture type, starting point coordinates, and ending point coordinates. The coordinates (X+,)'+) and the end point coordinates (xz, yz) are calculated. In other words, the size of the aperture can be determined depending on the type of aperture.
Since the extending direction of the mask pattern M can be determined from the starting point coordinates (x+, y+) and the ending point coordinates (X2°yz), for example, if the mask pattern M is an inverted pattern shown in FIG. 2 for simplicity, then , aperture A, from the starting point coordinates (x+, y+) and ending point coordinates (x,,'yz), the starting point of the outline of the mask pattern M is located at the upper left corner with the aperture AI at the starting point coordinates (x+, y+). Coordinates (x
+, y+) are recognized as the edge coordinates of the outline,
Similarly, the edge coordinates of the other end of the outline are aperture A,
is at the end point coordinates (Xz, Yz), it is recognized as the lower right corner coordinates (x2゜yz), and there are no end point coordinates of other apertures connected to the lower side at the end point coordinate position of aperture A1. By doing so, the coordinates of the lower right corner (X3
．． V3) can also be recognized as edge coordinates, and in the same way, the starting point coordinates of aperture A2 (X!, Y3)
) and the end point coordinates (X4.Y4) to the edge coordinates of the outline (
Xn, )'t), (Xs,)'s) and (xb,y
6), and when the calculation of these edge coordinates is completed, proceed to step ■C, connect each edge coordinate, and calculate
The contour pattern data shown in 1 is formed and stored in the focus map memory 10b in step d, and the contour data registration process of the mask pattern M is completed.

In the inspection pattern coordinate conversion process, as shown in FIG. Occupies ~9d. Since each shift register group 9a to 9d is connected in series, writing of data to the shift register group 9a to 9d is performed by sequentially inputting data output from the CCD camera in time order to the shift register R9 of the shift register group 9a. By writing, the first write data is sequentially shifted and written to the final stage of the shift register R1) of the shift register group 9d.

Next, the process moves to step (2)b and the shift register group 9a
It is determined whether writing of the read data to ~9d has been completed, and if the writing is not completed, the process returns to step (a) to continue writing the read data, and when the writing is completed, step (2) is performed. Move to C.

In this step ■C, the inspection pattern read data written in the shift register groups 9a to 9d is searched to determine the edge coordinates (XPl, Y□) for the preset pattern P.
and (X p z , Y p z ), and these edge coordinates and the edge coordinates of the mask pattern M (X s
+ , Y , + ) and (X N Z , Y , 4
The inclination tan θ of the inspection pattern P with respect to the mask pattern M is calculated from □). The calculation of this slope tanθ is
First, the edge coordinates (X p r , Y p r ) and (X□,
YMI), calculate the deviation ΔX in the X direction and the deviation ΔY in the Y direction according to the following formulas (1) and (2),
Next, the following equation (3) is calculated based on both to calculate the slope tanθ.

ΔX= l XPI X, 4+ l...
・・・・・・・・・(1) ΔY= l Y,, -Y□1
・・・・・・・・・・・・(2) tanθ, = ΔX/Δ
Y ・・・・・・・・・・・・(3) Similarly, perform the same calculation for the other edge coordinate to obtain the slope tanθ2
Calculate the slope tanθ and tanθ2, whichever is larger, or when both are equal, set either one as the slope ta
Determine as nθ. This determination may be performed when the inspection pattern P is tilted around one of the edge coordinates, and in this case, the tilt for the central edge coordinate is zero, and for the other edge coordinate, the tilt is determined according to the tilt. Since tanθ is obtained, the slope tanθ is determined as the slope of the inspection pattern.

Next, the process moves to step ■d, and the slope t stored in advance is calculated based on the slope tanθ calculated in step ■C.
The order of read addresses in shift register groups 9a to 9d (for example, shift register R[l, R? of shift register group 9d, shift register Rb of shift register group 9C) is determined by referring to a memory table showing the correspondence between anθ and successive addresses. , Rs, shift registers R4, R3 of the shift register group 9b and shift registers R2°R1) of the shift register group 9a are selected. In this case, as the slope tan θ becomes smaller, the data stored in the shift registers 9a, 9b, and 9C are not read out in this order.
When the slope tanθ is zero, the shift register 9d
Read only the stored data.

Next, the process moves to step ■e, and the shift register groups 93 to 93 are sequentially selected in the order of successive addresses selected in step ■d.
The write data is read from each of the shift registers R3 to R8 of 9d and sequentially stored in the bitmap memory 10a.

Next, the process moves to step f, and it is determined whether or not writing of all the image data to the bitmap memory 10a is completed, and if the writing is not completed, the process proceeds to step f.
Then, the image data for the next four scan lines from the CCD camera 3 is stored in the shift register group 9a to 9d, and then the process returns to step e, and when all the image data has been written, the process ends. Then move on to step (2) in FIG.

By this inspection pattern coordinate conversion processing, the positions of the shift registers R+-Re of the shift register groups 9a to 9d are changed according to the inclination of the inspection pattern P with respect to the mask pattern M.
The inspection pattern data stored in the mask pattern has its tilt corrected and becomes consistent with the mask pattern.

Furthermore, as shown in FIG. 7, in the inspection pattern contour data registration process of step (2), in step (a), the image data of the inspection pattern P stored in the bitmap memory 10a is input pixel by pixel in the row and column directions. Scan and find the sign “
Search for a bit position that is inverted from "0" to "1" or vice versa, store this bit position sequentially as edge coordinates, and connect each edge coordinate to create outline data of the inspection pattern P. Then, The process moves to step (2) b, where the contour data is stored in the bitmap memory 10c, and then the process moves to step (4) in FIG.

Furthermore, in the defect inspection process of step (2), as shown in FIG.
For example, the contour data of the mask pattern M shown in FIG. 9 (fat) stored in 10c and the contour data of the inspection pattern P shown in FIG.
In a) and (C), the center pixel WM5. W
PS is set to be pixel D0°, and inspection windows (reading ranges) WH and W are set to, for example, 3×3 pixels, and then, proceeding to step b, inspection windows WH and inspection windows W are set. Serially reads out each pixel of , outputs it to the defect inspection circuit 20, and
The defect data of this defect inspection circuit 20 is read and stored in the bitmap memory 10d, and then step
It is determined whether defect inspection has been performed on all the pixels stored in the bitmap memories 10b and 10c. If the defect inspection is not completed, the process proceeds to step d and the inspection is performed. Windows WM and W, 1 to the right
After shifting by the number of pixels, the process returns to step (2) a, and when all defect inspections are completed, the process is terminated and the process proceeds to step (4) in FIG. Incidentally, in step d, the center pixels W, 4. and W2. When reaches the right edge, it is shifted to the left edge of the row below.

Therefore, the pixel D0 in FIGS. 9(a) and (C1
゜.

I) o+... The central pixels W i and WP of the inspection windows WM and We both have a logical value II O
II, selectors 20B and 20A select input terminal I, so center pixels WMS and Wl,
Since the bit signal with the logical value "O" is input to the exclusive OR circuit 21, its output becomes the logical value "0°", and the corresponding pixel position in the bitmap memory 10d has the logical value "0" indicating that there is no defect. °゛ is written.

Similarly, the inspection window W corresponding to the mask pattern M
When one of the central pixels WH5 and WP of the inspection window W corresponding to H and the inspection pattern P has a logical value of "1" and the other has a logical value of "°0°", the logical value is "1".
The central pixel WMS (or W P S ) that has become "°,
Each pixel of the other side W p I” W P 9 (or W, 4
The OR output of ゜ ~ W s q ) is the exclusive OR circuit 2
1, each pixel WP1 to WP9 (or W□
~ W W2. (or W141-WHl) is the logical value “0′”, the output of the exclusive OR circuit 21 is the logical value “1”°
Therefore, it can be determined that there is a defect.

Then, the contents of the bitmap memory 9c are read out and displayed on the CRT display 13 or printer 14 as necessary.
By outputting the image to the image data, the defective portion of the inspection pattern P can be displayed or printed.

In addition, if the inspection pattern P of the inspected object 2 captured by the COD camera 3 has an internal defect as shown in FIG. As shown in Figure (b), contour data representing the outer peripheral side contour and the contour around the internal defect is obtained, and this data is stored in the map memory 10c.

Then, when the defect inspection process shown in FIG. Although defect data with a logical value of '°1' is not written in the center pixel W of the inspection window W2 corresponding to the inspection pattern P,
When Ps reaches the position of the pixel Dl13, the bit of the pixel D1)3 of the mask pattern M has the logical value ""O", and the bit of the pixel Dm3 of the inspection pattern P has the logical value "1'
degree, and each pixel WM1 to WM of the inspection window WM corresponding to the mask pattern M has the logical value "O1), so it is determined that there is a defect, and the bitmap memory
While writing the logic value "'1" into the pixel 0d, 3, the inspection windows WH and W2 are shifted to the right by one pixel, and the process returns to step (2) a.

Similarly, the center pixel WPS of the inspection window W corresponding to the inspection pattern P is the pixel D84 of the inspection pattern P.
DI+6・D 9:l+D Q&・DA3・D
MA・D1)3・DI&+ It is determined that DC3 to DC& are defective and a logical value ``1'' is written in the corresponding pixel position of the bitmap memory 10d.

Therefore, as shown in FIG. 10 (C1), the defective portion of the inspection pattern P can be recognized from the stored contents of the bitmap memory 10d.

According to this embodiment, a 3×3 bit inspection window W
, and W2 are used to perform defect inspection, so even if there is a quantization error or positional misalignment of ±1 bit in the mask pattern M and inspection pattern P, it can be determined that there is no defect, and no errors can be detected. Detection can be prevented.

In the above embodiment, a case has been described in which the inspection windows W, 4, and W are selected to be 3×3 bits, but it goes without saying that the present invention is not limited to this, and inspection windows of any size can be applied.

The inspection windows W, 4, and W can also be selected to have a minimum pixel unit size, and in this case, the exclusive OR of the bits of corresponding pixels of the mask pattern M and the inspection pattern P is performed. Therefore, if the two match, it can be determined that it is normal, and if they do not match, it can be determined that there is a defect.

Further, in the above embodiment, a case has been described in which the order in which the test pattern data is successively produced is changed to obtain consistency between the test pattern and the master pattern, but the invention is not limited to this. The order may be changed.

Furthermore, in the above embodiment, the case where the inclination is calculated based on the edge coordinates of both ends of the pattern was explained, but in the case of a printed circuit board, etc., the inclination is calculated based on the position coordinates of a minute pattern such as a predetermined through hole. may be calculated.

Furthermore, in the above embodiment, C is used as an imaging device.
Although the case where the CD camera 3 is applied has been described, the present invention is not limited to this, and a combination of a line scanning type image sensor and a mounting table that moves in synchronization can also be applied.

Furthermore, in the above embodiment, a case has been described in which input data to a photoplotter or a pattern generator is applied as CAD data, but the invention is not limited to this, and any graphic data that can extract the outline of a pattern may be used. Furthermore, as the data recording medium, any recording medium such as a floppy disk network other than magnetic tape can be used.

[Effects of the Invention] As explained above, according to the present invention, inspection pattern data for each scanning line from the imaging device is divided into a plurality of pixel groups and stored, and the stored inspection pattern data and Calculating the slope of the test pattern by comparing it with the master pattern data, and changing the reading order of either the test pattern or the master pattern data according to the slope of the test pattern to match both pattern data. Since the two are compared from the beginning, it is possible to reliably prevent misjudgment of defective patterns due to angular deviation between the inspection pattern and the mask pattern, and also to simplify the configuration. It will be done.

Moreover, by calculating the exclusive OR of the contour data of the inspection pattern and the contour data of the mask pattern, accurate defect inspection can be performed even when there is some coordinate shift between the two data.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is an explanatory diagram showing a mask pattern, Fig. 3 is a block diagram showing an example of a defect inspection circuit, and Fig. 4 shows a processing procedure of a control device. 5 is a flowchart showing the mask pattern contour data processing procedure, FIG. 6 is a flowchart showing the inspection pattern coordinate conversion processing procedure of the control device, and FIG. 7 is a flowchart showing the inspection pattern contour data processing procedure of the control device. Flowchart, FIG. 8 is a flowchart showing the defect inspection processing procedure of the control device, and FIG. 9 (al to (d)
10(a) to 10(C) are explanatory diagrams of patterns for explaining the operation of the embodiment of the present invention, respectively. In the figure, ■ is! 1 device, 2 is an object to be inspected, 3 is a COD camera (imaging device), 4 is a control device, 9a to 9d are shift register groups, R3-R4 are shift registers, 10a to 10d
is a bit map memory, 1) is a magnetic tape device, 12 is a magnetic disk, 13 is a CRT display, 14 is a printer, M is a mask pattern, P is a test pattern, WH,
W, is the inspection window, WMS, WF2 is the center pixel,
20 is a defect inspection circuit. Figure 1 Figure 2 Figure 3 Figure 6 Figure 7 Figure 8 Cause Figure 9 CG) Figure 9 (b) Figure 9 (C) Figure 9 (d) Figure 10 (G) Figure 10 (b)

Claims (4)

[Claims] (1) Obtain inspection pattern data by relatively scanning the inspection object on which the inspection target pattern is formed and an imaging device in one or two dimensions, and compare the inspection pattern data with pre-stored master pattern data. In the pattern inspection method, inspection pattern data for each scanning line from the imaging device is divided into a plurality of pixel groups and stored, and the stored inspection pattern data and the Calculate the slope of the test pattern by comparing it with the master pattern data, change the reading order of either the test pattern or the master pattern data according to the slope of the test pattern, and then match the data of both patterns. A pattern inspection method characterized by comparing the two. (2) The pattern inspection method according to claim (1), wherein the inspection pattern data and the master pattern data are compared by performing an exclusive OR of the two. (3) Obtain inspection pattern data by relatively scanning the inspection object on which the inspection target pattern has been formed and the imaging device one-dimensionally or two-dimensionally, and compare the inspection pattern data with pre-stored master pattern data. The pattern inspection apparatus is configured to inspect defects in an inspection pattern using a plurality of pixel groups; a tilt detecting means for detecting the tilt of the test pattern by comparing the stored test pattern data and the master pattern data; and reading either the test pattern or the master pattern data based on the detection result of the tilt detecting means. A pattern inspection device comprising: data matching means for matching both pattern data by changing the order; and comparison means for comparing the inspection pattern data and master pattern data matched by the data matching means. . (4) The pattern defect inspection apparatus according to claim (3), wherein the comparison means extracts the defective portion by calculating an exclusive OR of the inspection pattern data and the master pattern data.