JPH0620032A - Method and device for pattern check - Google Patents

Abstract

PURPOSE:To check a pattern with high accuracy and high reliability by calculating the weight coefficient of each selected similar dictionary pattern and reflecting this calculated weight coefficient onto the total unmatch value without fixing evenly the weight coefficient of the weight dictionary pattern. CONSTITUTION:An image acquiring means 11 is provided to acquire an image to be checked and outputs the image acquisition data DIN together with a dictionary information supply means 12 which outputs the dictionary data Dr serving as a deciding standard of the checked subject, the weight dictionary data DPt, and a weight coefficient K, and a comparing/deciding means 13 which compares both data Dr and DPt with the data DIN related to the checked area of the checked object. Then the means 12 is provided with a weight coefficient calculating means 12A which calculates the coefficient (K=S/N) based on the number of weight patterns N of the similar dictionary patterns PR1 and PR2 and the deciding reference value S with which the match or unmatch is decided between the number of both patterns N.

Description

Detailed Description of the Invention

[Table of Contents] Industrial Application Field of the Prior Art (FIG. 7) Problem to be Solved by the Invention Means for Solving the Problem (FIGS. 1 and 2) Action Example (FIGS. 3 to 6) Effect of the Invention

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspecting apparatus and a pattern inspecting method, and more particularly to an apparatus and a method for performing pattern matching processing on an object to be inspected based on a dictionary pattern.

In recent years, a pattern inspection apparatus has been used which compares a pattern to be inspected with a dictionary pattern and recognizes a difference between them when mounting inspection of a semiconductor integrated circuit device, a keyboard or the like.

According to this, in addition to the thick dictionary pattern and the thin dictionary pattern, a weight dictionary pattern in which the weight coefficient is fixed to a constant value is used, and the weight coefficient is fixed to a constant value in the pattern matching process. The weighted dictionary pattern and the inspected pattern are compared.

Therefore, in the case of identifying a pattern including a recognition target similar to each other in the pattern to be inspected, and when the dictionary patterns are similar to each other, an erroneous determination may be caused in the pattern matching process. There is.

Therefore, the weighting coefficient is calculated for each of the selected similar dictionary patterns without reflecting the weighting coefficient on the weighting dictionary pattern, and the weighting coefficient is reflected in the total amount of disagreement. An apparatus and a method capable of performing highly reliable pattern inspection are desired.

[0007]

2. Description of the Related Art FIG. 7 is a block diagram of a pattern inspection apparatus according to a conventional example. For example, an apparatus for inspecting the pattern of the inspection object 14 comprises an image processing system 1, a dictionary pattern supply system 2 and a pattern comparison / determination circuit 3 in FIG. The dictionary pattern supply system 2
Is composed of a dictionary memory circuit 2A, a similar dictionary selection circuit 2B, first and second selected dictionary memory circuits 2C and 2D, a dictionary pattern matching circuit 2E and a dictionary conversion circuit 2F.

The function of the apparatus is that when the image of the object 14 to be inspected is acquired by the image processing system 1, the image data DIN relating to the inspection area is output to the pattern comparison / determination circuit 3. Further, the dictionary data Dr and the weighted dictionary data Dt thereof are output from the dictionary pattern supply system 2 to the pattern comparison / determination circuit 3. As a result, the pattern comparison / determination circuit 3 compares the image data DIN relating to the inspection area of the inspection object 14 with the dictionary data Dr and the weight dictionary data Dt, and outputs the comparison result data DOUT.

The dictionary data Dr related to the dictionary pattern
Is read from the dictionary memory circuit 2A, and the weight dictionary data Dt related to the weight dictionary pattern is transferred from the dictionary conversion circuit 2F to the dictionary memory circuit 2A.

That is, similar dictionary data Dr1 and Dr2 from the dictionary memory circuit 2A through the similar dictionary selection circuit 2B.
Is selected, one of the dictionary data Dr1 is temporarily stored in the first selected dictionary memory circuit 2C, and the other dictionary data Dr2 is temporarily stored in the second selected dictionary memory circuit 2D.

Further, the dictionary pattern matching circuit 2E detects the number N of mismatches between the dictionary patterns based on the similar dictionary data Dr1 and Dr2, which is the dictionary conversion circuit 2
It is output to F. As a result, the dictionary conversion circuit 2F converts the number N of mismatches into weight data Dt related to the weight dictionary pattern, and the data Dt is output to the dictionary memory circuit 2A.

[0012]

By the way, according to the conventional example, the weighting coefficient K of the weighting dictionary pattern extracted based on the similar dictionary data Dr1 and Dr2 is fixed to a constant value.
In the pattern comparison / determination circuit 3, the image data DIN relating to the inspection area of the inspection object 14 is compared with the dictionary data Dr and the weight dictionary data Dt in which the weight coefficient K is fixed.

For this reason, in the case of identifying patterns that include a recognition target that is similar to each other among the inspection patterns, and the dictionary patterns are similar, there is a possibility that the pattern comparison and determination circuit 3 may make an erroneous determination. is there.

That is, in the pattern to be inspected, patterns to be recognized which are similar to each other, for example, Roman letters "O" and "C", "E" and "F", "I" and "L", etc. In the case where the dictionary pattern and the pattern to be inspected are matched, even if the amount of mismatch with the thick dictionary pattern or the thin dictionary pattern is small,
Since the amount of mismatch in the weight dictionary pattern becomes large, the total amount of mismatch u becomes large.

As a result, the pattern comparison / determination circuit 3 compares the determination level S with the total amount of non-coincidence u, and as a result, u> S, and "O", "C", and "E" of the pattern to be inspected. It is possible to accurately determine "mismatch" such as "F", "I" and "L".

However, in the case of performing a pattern inspection in which it is difficult to identify the difference between the recognized patterns similar to each other in the pattern to be inspected, for example, the unit symbol "Å" or the Roman letter "A", Since the mismatch amounts of the thick dictionary pattern, the thin dictionary pattern, and the weight dictionary pattern based on the matching process are all small, the total mismatch amount u is small, which causes an erroneous determination.

For example, in the pattern comparison / determination circuit 3, as a result of comparing the determination level S with the total amount of mismatch u, u
Since ≦ S, although the unit symbol “Å” related to the pattern to be inspected is originally different from the Roman letter “A”,
The unit symbol “Å” or the Roman letter “A” of the pattern to be inspected may be mistakenly determined as “match”.

As a result, there is a problem that accurate pattern inspection is hindered and the reliability of the pattern inspection is reduced. The present invention was created in view of the problems of such a conventional example, without uniformly fixing the weighting factor relating to the weighting dictionary pattern, calculating the weighting factor for each selected similar dictionary pattern, An object of the present invention is to provide a pattern inspection apparatus and a pattern inspection method that can perform pattern inspection with high accuracy and high reliability by reflecting it in the total amount of mismatch.

[0019]

FIG. 1 is a principle diagram of a pattern inspection apparatus according to the present invention, and FIG. 2 is a principle diagram of a pattern inspection method according to the present invention.

The pattern inspection apparatus of the present invention, as shown in FIG. 1, has an image acquisition means 11 for acquiring an image of an object 14 to be inspected and outputting image acquisition data DIN, and a criterion for judging the object 14 to be inspected. Dictionary information supplying means 12 for outputting the dictionary data Dr, the weight dictionary data DPt, and the weight coefficient K
And a comparison / determination unit 13 for making a comparison / determination of the dictionary data Dr, the weight dictionary data DPt, and the image acquisition data DIN relating to the inspection area of the inspection object 14, and the dictionary information supplying means 12 has a weighting coefficient. It is characterized in that a weighting factor calculation means 12A for calculating K is provided.

In the pattern inspecting apparatus of the present invention, the weighting factor computing means 12A matches the number N of weight patterns indicating the disagreement between the similar dictionary patterns PR1 and PR2 with the coincidence or disagreement between the similar dictionary patterns PR1 and PR2. The weighting coefficient K = S / N is calculated based on the determination reference value S for determining.

The pattern inspection method of the present invention is shown in FIG.
As shown in the processing flowchart of FIG.
In step 1, the similar dictionary patterns PR1 and PR2, which are the criteria for the subject 14 to be inspected, are selected, and then step P2 is performed.
In step P3, the weighting dictionary pattern Pt is extracted from the similar dictionary patterns PR1 and PR2. In step P3, the weighting factor K is calculated for each weighting dictionary pattern Pt. The above object is achieved by performing pattern matching processing of the inspection object 14 based on the patterns PR1 and PR2, the weight dictionary pattern Pt, and the weight coefficient K.

[0023]

[Operation] According to the pattern inspection apparatus of the present invention, as shown in FIG. 1, the image acquisition means 11, the dictionary information supply means 12 and the comparison / determination means 13 are provided.
2 is provided with a weight coefficient calculating means 12A for calculating the weight coefficient K.

For example, in connection with the pattern inspection of the object 14 to be inspected, the weighting factor calculation means 12A causes the number N of weighting patterns indicating the disagreement between similar dictionary patterns PR1 and PR2.
And the weighting factor K = S / N is calculated based on the determination reference value S for determining whether or not the dictionary patterns PR1 and PR2 match or do not match.

Therefore, when the image of the object 14 to be inspected is acquired by the image acquisition means 11 and the image acquisition data DIN relating to the inspection area is output to the comparison / judgment means 13, the dictionary data Dr, which becomes the judgment reference, The weight dictionary data DPt and the weight coefficient K are output from the dictionary information supply means 12 to the determination means 13.

Further, in the comparison / determination means 13, the dictionary data D
r, the weight dictionary data DPt and the image acquisition data DIN relating to the inspection area of the inspection object 14 are compared and determined. At this time, it is possible to correct the mismatch amount t related to the weight dictionary pattern Pt by the weight coefficient K.

As a result, the weighting coefficient K calculated for each similar dictionary pattern is reflected in the total disagreement amount u without uniformly fixing the weighting coefficient K related to the weighting dictionary pattern as in the conventional example. It is possible to perform highly accurate and highly reliable pattern inspection that matches the pattern to be inspected.

Further, according to the pattern inspection method of the present invention, as shown in the processing flow chart of FIG. 2, when the similar dictionary patterns PR1 and PR2 which are the judgment criteria of the object 14 to be inspected are selected in step P1, In step P2, the weighting dictionary pattern Pt is extracted from both dictionary patterns PR1 and PR2, and then in step P3, the weighting coefficient K is calculated for each weighting dictionary pattern Pt.

For this reason, even in the case of identifying a pattern including a recognition target object similar to each other in the inspection pattern and the dictionary patterns are similar, step P
4, it is possible to perform the comparison / determination process on the inspection object 14 with high accuracy based on the dictionary patterns PR1 and PR2, the weighting dictionary pattern Pt, and the weighting coefficient K that are similar to each other.

This makes it possible to reduce the erroneous determination as in the conventional example in the comparison and determination means 13 as much as possible.
As a result, it is possible to improve the reliability of the pattern inspection apparatus that performs the pattern matching process.

[0031]

Embodiments of the present invention will now be described with reference to the drawings. 3 to 6 are diagrams for explaining the pattern inspection apparatus and the pattern inspection method according to the embodiment of the present invention, and FIG. 3 is a configuration diagram of the pattern inspection apparatus according to the embodiment of the present invention.

For example, a pattern inspection apparatus applicable to mounting inspection of a semiconductor integrated circuit device or a keyboard comprises an image acquisition system 21, a dictionary information supply system 22 and a pattern comparison / determination circuit 23 in FIG.

That is, the image acquisition system 21 is an embodiment of the image acquisition means 11 and acquires the image of the object 14 to be inspected and outputs the image acquisition data DIN. For example, the image acquisition system 21 includes an imaging system 21A, an A / D conversion circuit 21B, a frame memory circuit 21C and a binarization conversion circuit 21.
Composed of D.

The image pickup system 21A acquires an image of a semiconductor integrated circuit device or a keyboard which is an example of the inspection object 14 and outputs an image acquisition signal Sg to the A / D conversion circuit 21B. The A / D conversion circuit 21B outputs image data DIN obtained by analog / digital conversion of the image acquisition signal Sg to the frame memory circuit 21C. The frame memory circuit 21C temporarily stores it, and the binarization conversion circuit 21D determines pixel data (hereinafter also referred to as "inspection data din") a in the image data DIN based on the threshold level L. The binary data b of "0" or "1" is converted. For example, if a <L, b = 0, and if a ≧ L, b = 1.

The dictionary information supply system 22 is one embodiment of the dictionary information supply means 12, and outputs the dictionary data Dr, the weight dictionary data DPt, and the weight coefficient K which are the criteria for the inspection object 14. For example, the dictionary information supply system 22 includes a dictionary memory circuit 22A and a similar dictionary selection circuit 22.
B, first and second selected dictionary memory circuits 22C and 22D, a dictionary pattern matching circuit 22E, a dictionary conversion circuit 22F and a weighting coefficient calculation circuit 22G.

The dictionary memory circuit 22A stores dictionary data Dr such as a thick dictionary pattern and a thin dictionary pattern, which are criteria for determining the object 14 to be inspected, a determination level S for determining matching / mismatching, data such as the size / area of the dictionary. Is stored.

The similar dictionary selection circuit 22B selects the similar dictionary data Dr1 and Dr2 related to the dictionary patterns PR1 and PR2 which are similar to each other from the dictionary patterns, and selects the similar dictionary data Dr1 and Dr2.
Of the selected dictionary memory circuits 22C and 22D. The first and second selection dictionary memory circuits 22C and 22 are for individually storing the similar dictionary data Dr1 and Dr2, respectively.

The dictionary pattern matching circuit 22E compares the similar dictionary data Dr1 and Dr2 and determines the dictionary pattern PR.
The number N of weight patterns indicating the disagreement between 1 and PR2 is extracted,
It is output to the dictionary conversion circuit 22F. Also,
The dictionary conversion circuit 22F converts the weight pattern number N into weight dictionary data DPt related to the weight dictionary pattern Pt and outputs it to the dictionary memory circuit 22A. The configuration so far is similar to that of the conventional example, but is different in the following points.

That is, the weighting coefficient calculating circuit 22G is an embodiment of the weighting coefficient calculating means 12A and calculates the weighting coefficient K. For example, the weighting factor calculation circuit 22A determines the weighting factor K based on the number N of weighting patterns indicating the disagreement between the similar dictionary patterns PR1 and PR2 and the determination reference value S for determining the coincidence or disagreement of the similar dictionary patterns PR1 and PR2. = S
Calculate / N.

The pattern comparison / determination circuit 23 is an embodiment of the comparison / determination means 13, and includes the dictionary data Dr, the weight dictionary data DPt, and the inspection data din for the inspection object 14.
It is a comparison judgment with. For example, the pattern comparison / determination circuit 23 is the pattern matching circuit 23A, the determination circuit
23B and a result output circuit 23C.

The pattern matching circuit 23A compares the inspected data din with the dictionary data d1 related to the thick dictionary pattern, and the large dictionary pattern mismatch amount W, and the thin dictionary comparing the data din with the dictionary data d2 related to the thin dictionary pattern. The pattern mismatch amount h, the data din and the weight dictionary pattern Pt
Is extracted as a weight pattern mismatch amount t obtained by comparing the weight dictionary data DPt according to the above with u = (W
+ H + t × K) is output to the determination circuit 23B.
Here, K is a weighting coefficient, and the weighting coefficient K = S / N calculated by the weighting coefficient calculation circuit 22G.

The judgment circuit 23B is for judging a match / mismatch based on the judgment level S. For example, the total disagreement amount u and the determination level S are compared and evaluated. The result output circuit 23C outputs the evaluation result. That is, if the relationship between the total mismatch amount u and the determination level S is u ≦ S, a match is output, and conversely, if the relationship between the total mismatch amount u and the determination level S is u> S, a mismatch is determined. Output.

As described above, according to the pattern inspection apparatus of the embodiment of the present invention, as shown in FIG. 3, the image acquisition system 21, the dictionary information supply system 22, and the pattern comparison / determination circuit 23 are provided, and the dictionary is provided. Information supply system 22
Is provided with a weighting coefficient calculation circuit 22A for calculating the weighting coefficient K.

For example, in connection with the pattern inspection of the object 14 to be inspected, the weighting factor calculation circuit 12A causes the number N of weighting patterns indicating a disagreement between similar dictionary patterns PR1 and PR2.
And the weighting factor K = S / N is calculated based on the determination reference value S for determining whether or not the dictionary patterns PR1 and PR2 match or do not match.

Therefore, the image of the inspection object 14 is acquired by the image acquisition system 21 and the inspection data din is acquired.
Is output to the pattern comparison / determination circuit 23, the dictionary data Dr, the weight dictionary data DPt, and the weight coefficient K, which are the determination criteria, are output from the dictionary information supply system 22 to the comparison / determination circuit 23.

In the pattern comparison / determination circuit 23, the dictionary data Dr, the weight dictionary data DPt and the inspected data din of the inspected object 14 are compared and determined. At this time, it is possible to correct the mismatch amount t related to the weight dictionary pattern Pt by the weight coefficient K. For example, the weight coefficient K is added to the weight pattern mismatch amount t obtained by comparing the inspection data din with the weight dictionary data DPt related to the weight dictionary pattern Pt.

As a result, the weighting coefficient K calculated for each similar dictionary pattern is reflected in the total disagreement amount u without uniformly fixing the weighting coefficient K related to the weighting dictionary pattern Pt as in the conventional example. It is possible to perform a highly accurate and highly reliable pattern inspection that matches the pattern to be inspected.

Next, the pattern inspection method according to the embodiment of the present invention will be described while supplementing the operation of the apparatus. FIG. 4 is a flow chart of the pattern inspection process according to the second embodiment of the present invention, and FIGS. 5 and 6 show supplementary explanatory diagrams thereof. For example, in the case of inspecting a DUT pattern PG1 having a roman letter "C" as shown in FIG. 5 or a DUT pattern PG2 having a unit symbol "Å" as shown in FIG. 6, based on the dictionary patterns Pr1 and Pr2, In FIG.
First, in step P1, an image acquisition process for the DUT patterns PG1 and PG2 is performed.

At this time, the image of the DUT pattern PG1 such as the Roman character "C" or the unit symbol "Å" is acquired by the imaging system 21A, and the image acquisition signal Sg is analog / analogized by the A / D conversion circuit 21B. Digitally converted. Further, the image data DIN is temporarily stored in the frame memory circuit 21C, and the inspected data din is binarized by the binarization conversion circuit 21D. For example, the pixel data a related to the region to be inspected is converted into the binarized data b of “0” or “1” based on the threshold level L.

Here, if the pattern matching process of the pattern under test PG1 is performed first, step P2
Then, selection processing of similar dictionary patterns PR1 and PR2, which are reference criteria for the DUT pattern PG1, is performed. At this time,
From the dictionary memory circuit 22A of the dictionary information supply system 22, the similar dictionary selection circuit 22B selects the similar dictionary data Dr1 and Dr2 related to the dictionary patterns PR1 and PR2 which are similar to each other from the dictionary pattern PR.

For example, Roman letters "C" and "O", which are criteria for judging the pattern under test PG1 as shown in FIG.
The dictionary data Dr such as the thick dictionary pattern Pd11 and the thin dictionary pattern Pd21, the determination level S for determining the matching / mismatching of the dictionary data, and the data such as the size / area of the dictionary are read out, which are the first and second selections. The dictionary memory circuits 22C and 22D are individually stored.

Next, in step P3, a weight dictionary pattern Pt is extracted from the similar dictionary patterns PR1 and PR2 such as Roman letters "C" and "O". Here, the dictionary pattern matching circuit 22E causes the similar dictionary data Dr1, D1.
r2 is compared, the number N of weight patterns indicating the disagreement between the dictionary patterns PR1 and PR2 is extracted, and this is output to the dictionary conversion circuit 22F.

In the dictionary conversion circuit 22F, the weight pattern number N is the weight dictionary data DPt related to the weight dictionary pattern Pt.
Is converted to the dictionary memory circuit 22A. Further, in step P4, the weight coefficient K is calculated for each weight dictionary pattern Pt. At this time, the weighting coefficient calculation circuit 22G calculates the weighting coefficient K. For example, in the weighting factor calculation circuit 22A, similar dictionary patterns PR1 and PR2 are used.
Judgment reference value S for judging whether or not the dictionary patterns PR1 and PR2 similar to the number N of weight patterns indicating the mismatch between
Based on and, the weighting factor K1 = S / N is calculated.

Then, in step P5, similar dictionary patterns PR1 and PR2, a weight dictionary pattern Pt, and a weight coefficient K are used.
Based on the above, the pattern matching processing of the inspection object 14 is performed. Here, the pattern matching circuit 23A of the pattern comparison / determination circuit 23 compares the inspected data din with the dictionary data d1 related to the thick dictionary pattern, and the pattern data shown in FIG.
As shown in, the thick dictionary pattern mismatch amount W1 is extracted. Further, the data din is compared with the dictionary data d2 related to the fine dictionary pattern, and the fine dictionary pattern mismatch amount h is extracted. Further, the data din and the weight dictionary pattern P
The weight dictionary data DPt relating to t is compared, the weight pattern mismatch amount t1 is extracted, and as the total mismatch amount u,
u = (W1 + h + t1 × K1) is output to the determination circuit 23B.

Further, the decision circuit 23B decides on the basis of the decision level S whether there is a match or a mismatch. For example, the total disagreement amount u is compared with the determination level S, and the evaluation result is output from the result output circuit 23C. Here, if the relationship between the total amount of mismatch u and the determination level S is u ≦ S, a match between the Roman character “C” related to the DUT pattern PG1 and the Roman character “C” related to the dictionary pattern Pr is detected. On the contrary, if the relationship between the total amount of mismatch u and the determination level S is u> S, the mismatch between the Roman character “C” related to the DUT pattern PG1 and the Roman character “O” related to the dictionary pattern Pr is detected. To be done.

Then, in step P6, the inspection target 1
It is determined whether or not the pattern matching processing of 4 has been completed. At this time, if it is all finished (YES)
Then, the control relating to the pattern inspection is ended. If all of them are not completed (NO), the process returns to step P2 and the similar dictionary patterns PR1 and PR2, which are the determination criteria for the DUT pattern PG2, are selected.

In the embodiment of the present invention, in order to inspect the pattern of the object pattern PG2 having the unit symbol "Å", the dictionary pattern is supplied from the dictionary memory circuit 22A of the dictionary information supply system 22 to the similar dictionary selection circuit 22B. Similar dictionary data Dr1 and Dr2 relating to dictionary patterns PR1 and PR2 which are similar to each other are selected from PR.

For example, as shown in FIG. 6 (a), the unit symbol "Å" or the Roman alphabet "A" that serves as the criterion for the test object pattern PG2, the thick dictionary pattern Pd12 and the thin dictionary pattern P are used.
The dictionary data Dr such as d22, the determination level S for determining the match / mismatch, and the data such as the size / area of the dictionary are read out, which are the first and second selected dictionary memory circuits 22C,
The weighting coefficient K2 is calculated in step P4, and the unit symbol “Å” related to the DUT pattern PG2 is calculated based on steps P5 and P6.
Pattern inspection is performed.

As described above, according to the pattern inspection method of the embodiment of the present invention, as shown in the flow chart of FIG. 4, similar dictionary patterns serving as the judgment criteria of the DUT patterns PG1 and PG2 in step P2. When PR1 and PR2 are selected, the weighting dictionary pattern Pt is extracted from both dictionary patterns PR1 and PR2 in step P3, and then the weighting coefficient K1 is calculated for each weighting dictionary pattern Pt in step P4.

For this reason, the recognition target objects similar to each other in the test object patterns PG1 and PG2, for example, patterns including Roman letters “O” and “C”, unit symbol “Å” and Roman letters “A” are identified. Even if the dictionary patterns PR1 and PR2 are similar, the DUT pattern PG is determined based on the similar dictionary patterns PR1 and PR2, the weighting dictionary pattern Pt, and the weighting factors K1 and K2 in step P5.
1 and PG2 can be accurately compared and judged.

That is, in the object patterns PG1 and PG2, objects to be recognized which are similar to each other, for example, Roman letters "O" and "C", "E" and "F", "I" and "L" are recognized. In the case of identifying a pattern including
When R1 and PR2 and the DUT patterns PG1 and PG2 are matched, even if the amount of mismatch between the thick dictionary pattern Pd11 and the thin dictionary pattern Pd21 is small, the weight dictionary pattern Pt is calculated based on the weight coefficient K1. Disagreement amount t related to
Since 1 becomes large, the total mismatch amount u becomes large.

As a result, the pattern comparison / determination circuit 23 compares the determination level S with the total amount of non-coincidence u, and as a result, u> S, and "O", "C", and "E" of the DUT pattern. It is possible to accurately determine "mismatch" such as "F" and "I" and "L".

Further, in the case of pattern inspection in which it is difficult to identify differences between the test object patterns PG1 and PG2 which are similar to each other, for example, the unit symbol "Å" or the Roman letter "A". In addition, as shown in FIG. 6B, the thick dictionary pattern Pd1 based on the matching process.
2, even if the amount of mismatch h1 of the fine dictionary pattern Pd22 becomes small, the weight dictionary pattern P is calculated based on the weight coefficient K2.
Since the disparity amount t2 related to t is large unlike the conventional example, the total disparity amount u is large.

For example, in the pattern comparison / determination circuit 23, the result of comparison between the determination level S and the total amount of mismatch u is not u ≦ S as in the conventional example, but u> S. Even when the unit symbol "Å" related to the pattern is different from the Roman letter "A", the unit symbol "Å" or Roman letter "A" of the DUT pattern can be determined as "mismatch". .

As a result, it is possible to reduce the erroneous determination as in the conventional example in the pattern comparison / determination circuit 23 as much as possible. As a result, it is possible to improve the reliability of the pattern inspection apparatus that performs the pattern matching process.

[0066]

As described above, according to the pattern inspection apparatus of the present invention, the image acquisition means, the dictionary information supply means and the comparison / determination means are provided, and the weight coefficient calculation for calculating the weight coefficient in the dictionary information supply means. Means are provided.

For this reason, when the weight coefficient is calculated by the weight coefficient calculating means based on the number of weight patterns between similar dictionary patterns and the judgment reference value of both dictionary patterns, the weight dictionary is used in the pattern matching process. It is possible to correct the amount of disagreement related to the pattern by the weighting coefficient.
As a result, the weighting coefficient calculated for each similar dictionary pattern can be reflected in the total amount of disagreement without uniformly fixing the weighting coefficient related to the weighting dictionary pattern as in the conventional example.

Further, according to the pattern inspection method of the present invention, when a similar dictionary pattern is selected as a criterion for the object to be inspected, the weight dictionary patterns are extracted from both dictionary patterns, and thereafter, each weight dictionary pattern is extracted. The weighting factor calculation process is performed.

Therefore, since different weighting factors are obtained for each dictionary pattern, it is necessary to identify patterns in the pattern to be inspected that include objects to be recognized that are similar to each other, and if the dictionary patterns are similar to each other. Also, based on the similar dictionary pattern, weight dictionary pattern, and weight coefficient, it is possible to perform the comparison / determination process on the inspection object with high accuracy. This makes it possible to reduce erroneous determination as in the conventional example as much as possible.

This makes it possible to perform highly accurate and highly reliable pattern matching processing that matches the pattern to be inspected. This greatly contributes to the provision of a highly reliable pattern inspection device.

[Brief description of drawings]

FIG. 1 is a principle diagram of a pattern inspection apparatus according to the present invention.

FIG. 2 is a principle diagram of a pattern inspection method according to the present invention.

FIG. 3 is a configuration diagram of a pattern inspection apparatus according to an embodiment of the present invention.

FIG. 4 is a flowchart of a pattern inspection process according to the embodiment of the present invention.

FIG. 5 is a supplementary explanatory diagram (part 1) of the flowchart according to the embodiment of the present invention.

FIG. 6 is a supplementary explanatory diagram (part 2) of the flowchart according to the embodiment of the present invention.

FIG. 7 is a configuration diagram of a pattern inspection apparatus according to a conventional example.

[Explanation of symbols]

 11 ... Image acquisition means, 12 ... Dictionary information supply means, 13 ... Comparison determination means, 12A ... Weight coefficient calculation means, DIN ... Image acquisition data, Dr ... Dictionary data, DPt ... Weight dictionary data, K ... Weight coefficient, PR1, PR2 ... Similar dictionary pattern, Pt ... Weight dictionary pattern.

Claims (3)

[Claims] 1. An image acquisition means (11) for acquiring an image of an inspection object (14) and outputting image acquisition data (DIN).
And dictionary information supply means (12) for outputting the dictionary data (Dr), the weight dictionary data (DPt), and the weighting coefficient (K) which are the criteria for the inspection object (14), and the dictionary data (Dr). , Weighting dictionary data (DPt) and image acquisition data (DIN) related to the inspection area of the inspection object (14)
Comparing and judging means (13) for making a comparative judgment with
The pattern inspection apparatus, wherein the dictionary information supply means (12) is provided with a weighting coefficient calculation means (12A) for calculating a weighting coefficient (K). 2. The pattern inspection apparatus according to claim 1, wherein the weighting factor computing means (12A) indicates the number of weighting patterns (N) indicating a disagreement between similar dictionary patterns (PR1, PR2), and the similar dictionary. A pattern inspection apparatus characterized in that a weighting coefficient (K = S / N) is calculated on the basis of a judgment reference value (S) for judging whether patterns (PR1, PR2) match or do not match. 3. A process of selecting a similar dictionary pattern (PR1, PR2) serving as a criterion for the subject (14) to be inspected, and a weight dictionary pattern (Pt) is extracted from the similar dictionary pattern (PR1, PR2). Then, the weighting coefficient (K) is calculated for each weighting dictionary pattern (Pt), and based on the similar dictionary patterns (PR1, PR2), the weighting dictionary pattern (Pt) and the weighting coefficient (K). A pattern inspection method comprising performing pattern matching processing on the inspection object (14).