JP4776259B2 - Pattern evaluation method, pattern alignment method, and program - Google Patents

Description

  The present invention relates to a pattern evaluation method, a pattern alignment method, and a program, for example, for product shape inspection.

  Methods for evaluating the shape of a pattern to be inspected (hereinafter referred to as an inspection pattern) by using the difference from the reference graphic as an index are widely used in various industrial fields. For example, in the evaluation of a semiconductor device pattern, an SEM (Scanning Electron Microscope) image and CAD (Computer Aided Design) data as design data are used as a pattern inspection image, and the processed shape of the device pattern and its design shape are used. The similarity is calculated and the quality of the pattern is determined. Hereinafter, an example of such a conventional evaluation method will be described.

  First, the outline of the inspection pattern is detected from the SEM image that is the inspection image. When a CAD figure is given as a design pattern in a binary file such as GDSII, the contour information of the design pattern is detected in the same manner by developing it as a binary image.

  Next, the position of both patterns is detected by image matching between the SEM image and the binary image. Specifically, the relative positions of both patterns are gradually changed in a predetermined manner, the degree of difference or similarity at that time is calculated each time, and the position where the value becomes maximum or minimum is detected. As the degree of difference, parameters such as cross correlation and normalized correlation obtained by normalizing image brightness are generally used. By the image matching described above, the contour of the SEM image and the CAD figure are superimposed.

  Subsequently, one or several ROIs (Region of interest) are set for the superimposed patterns, and the dimension between the contours of both patterns in the ROI is measured. The pass / fail of the inspection pattern is determined by comparing one or several dimension values obtained in this way with a separately defined standard.

  However, the above-described prior art has the following problems.

First, since the correctness of the overlay of the contour of the inspection pattern and the CAD figure depends on the pattern shape, it is difficult to compare the degree of coincidence between patterns having different shapes. In addition, since it is necessary to continuously execute two image processing operations such as image collation and pattern dissimilarity calculation, there is a problem in that the calculation takes time and eventually increases the cost of pattern shape evaluation. Furthermore, there is no clear standard for the threshold used for image matching, and its setting currently depends on the experience of the operator.
Japanese Patent Laid-Open No. 10-163283 JP 2002-183235 A JP 2000-232057 A Japanese Patent Laid-Open No. 10-270332 JP 2000-275010 A

  An object of the present invention is to perform pattern evaluation and alignment in a short time and with high accuracy.

According to a first aspect of the invention,
By processing contour data that is shape data including contour points detected from the image of the inspection target pattern, pointer addresses corresponding to pixel coordinates of the data of the image are respectively given, and gradation values of the image data are obtained. Correspondingly, a procedure for generating first array data composed of elements having a value of 1 at the contour point position and a value of 0 at a position other than the contour point;
By processing the contour data of the reference pattern including the contour point of the reference pattern serving as the inspection reference for the inspection target pattern, a pointer address is given to each, and a value of 1 is given to the contour point position and 0 is given to a position other than the contour point. Generating second array data composed of elements having values of:
The value of the component of the second array data is a distance value closest to the component and having a value of 1, and the value corresponding to the inside of the outline of the reference pattern and the outside A procedure for generating third array data by performing processing for converting each component of the second array data into a function value having a value different in sign from the value corresponding to
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern by calculating a value of a component of the fourth array data;
A pattern evaluation method is provided.

According to the second aspect of the present invention,
By acquiring a first image that is an image of an inspection target pattern, detecting contour data that is shape data including contour points of the inspection target pattern from the first image, and processing the detected contour data A pointer address corresponding to the pixel coordinate of the first image data is given, and a value of 1 is set at the contour point position and a value of 0 is set at a position other than the contour point corresponding to the gradation value of the image data. Generating a first array data composed of elements having;
By processing the contour data, which is the shape data including the contour point of the reference pattern serving as the inspection reference of the inspection target pattern, a pointer address is given, and a value of 1 is given to the contour point position and 0 is set to a position other than the contour point. Generating second array data composed of elements having values of:
The value of the component of the second array data is a distance value closest to the component and having a value of 1, and the value corresponding to the inside of the outline of the reference pattern and the outside A procedure for generating third array data by performing processing for converting each component of the second array data into a function value having a value different in sign from the value corresponding to
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A step of converting the fourth array data into a second image by mapping an index color with reference to a color table set in accordance with a value of a component value of the fourth array data;
A procedure for displaying the second image superimposed on the first image;
A pattern evaluation method is provided.

According to the third aspect of the present invention,
By processing contour data that is shape data including contour points detected from the image of the inspection target pattern, pointer addresses corresponding to pixel coordinates of the data of the image are respectively given, and gradation values of the image data are obtained. Correspondingly, a procedure for generating first array data composed of elements having a value of 1 at the contour point position and a value of 0 at a position other than the contour point;
By processing the contour data of the reference pattern including the contour point of the reference pattern serving as the inspection reference for the inspection target pattern, a pointer address is given to each, and a value of 1 is given to the contour point position and 0 is given to a position other than the contour point a step of generating a second sequence data consisting of elements having a value,
The component values of the second sequence data, closest from the elements, and a value of the distance in the configuration needed Motoma having a value 1, a value corresponding to the inner contour of the reference pattern A procedure for generating third array data by performing, for each component of the second array data, processing for converting the value corresponding to the outside and a value corresponding to the outside to a function value having a different sign.
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern by calculating a value of a component of the fourth array data;
A procedure for obtaining an optimal positional relationship between the inspection target pattern and the reference pattern based on the calculated numerical value and a set reference;
A pattern alignment method is provided.

Furthermore, according to the 4th aspect of this invention , the program which makes a computer perform the pattern evaluation method mentioned above is provided.

  According to the present invention, pattern shape evaluation and alignment can be performed in a short time and with high accuracy.

  Embodiments of the present invention will be described with reference to the drawings. In the following description, shape inspection of a fine pattern formed in a semiconductor manufacturing process such as a lithography process or an etching process will be taken up as appropriate. However, it should be noted that the present invention is not limited to these cases and can be applied to general pattern shape inspections in various other industrial fields.

(1) 1st Embodiment FIG. 1: is a block chart explaining the schematic procedure of the 1st Embodiment of this invention. FIG. 2 is a diagram showing an example of an inspection image of an inspection pattern which is an evaluation target of the pattern evaluation method shown in FIG. Further, FIG. 3 is a diagram showing an example of a pattern (hereinafter simply referred to as “design pattern”) based on design data as a reference pattern of the inspection pattern shown in FIG. In the present embodiment, design data is used as a reference pattern. However, the reference pattern is not limited to this, and for example, an image of a product pattern determined as a non-defective product may be used. In this case, as will be described later with respect to the inspection image, processing for detecting the contour is required (B14).

  Hereinafter, the case where the contour of the inspection image shown in FIG. 2 is collated with the design pattern shown in FIG. 3 and the pattern shape is evaluated will be described as an example. Here, it is assumed that the position of the inspection image shown in FIG. 2 is set with reference to the design data by some means. As this alignment method, various methods such as a method using correlation matching between an inspection pattern and a design pattern and a method of matching the centroids more simply can be considered. Whichever alignment method is employed, it is important to execute alignment suitable for the purpose of pattern evaluation.

  In the following, image data is treated as array data for general-purpose explanation. In the case of a gray scale image, the pixel coordinates of the image data correspond to the pointer address of the array data, and the gradation value of the image data corresponds to the element value of the array data. The difference between image data and array data is that only the normalized 0 and natural number values are allowed for the gradation value of the image data, whereas any numerical value including negative numbers and decimal numbers is allowed for the element value of the array data. It is a point. In this specification, the term “array” simply means a data structure of array data, and generally a two-dimensional array is used. However, an optimal data structure may be employed in accordance with a programming language or a computer system for performing calculations such as a bit plane type two-dimensional array or a one-dimensional array.

  First, as shown in FIG. 1, an inspection image and design data are read into a computer (B2, B12). Next, the ROI corresponding to the inspection pattern is specified in the inspection image, and the contour Es2 is detected over the entire periphery of the inspection pattern. The detection of the contour Es2 (see FIG. 2) may use a threshold method, a linear approximation method, or any other method. Contour points must also be detected with subpixel resolution. If the image size in FIG. 2 is 100 × 100 pixels, the contour data detected with a sampling resolution of 1/10 pixel has a value of 1 at the contour point position. Is stored in the storage device of the computer as a 1000 × 1000 array A having a value of 0.

  If design data (see design pattern Er2 in FIG. 3) is given as binary data, it is rendered into a figure, and the design pattern contour data is set to the contour point position with a value of 1 and other positions. Obtained as an array B having a value of 0 (FIG. 1, B16). At this time, the scale of the design data Er2 is determined based on the magnification of the inspection image. When performing pattern evaluation to be described later with subpixel accuracy, it is necessary to adopt sampling point detection sampling resolution matched to the array A, and store the array B in a storage device of a computer, for example, as a 1000 × 1000 array. is there. However, it is not always necessary to match the size of the array A. For example, the size of the array B may be 500 × 500. Rather, in order to improve the calculation speed, it is necessary to select an appropriate sampling resolution for both the arrays A and B.

  Next, an array conversion process is performed on the array B to generate an array C (FIG. 1, B18). In the present embodiment, the array conversion process refers to a conversion process in which the shortest distance from each element of the array B to the element having the value 1 is substituted into the value of that element as a distance value. In the present embodiment, the Euclidean distance is defined as a distance value and the array conversion process is executed. The distance value can be defined in various ways. In addition to the Euclidean distance, for example, a 4-neighbor distance (City Block), an 8-neighbor distance (Chessboard), an octagonal distance, and a quasi-Euclidean distance can be used. Since the distance value is not an integer, it cannot be handled as an image as it is, but for example, it is possible to generate an array conversion image by standardizing the distance value to an 8-bit value. FIG. 4 schematically shows an array C obtained by performing such an image conversion process on the design pattern Er2 shown in FIG. In the figure, a bright part indicates a distance value 0, that is, an outline. FIG. 4 shows a state in which the distance value increases as it moves away from the contour line as it moves to a dark gradation. When the accuracy of pattern evaluation is not required so much, the calculation time can be shortened by proceeding with the subsequent processing using this array conversion image. However, if high-precision pattern evaluation is required, the result of the array conversion process is handled as array data C at this stage.

  Returning to FIG. 1, when the sampling resolutions of the array A and the array C are different, it is necessary to divide the array with a large resolution into sub-arrays in accordance with the array with a small resolution (B22). The resolution is the same. When the size of the array A is larger than the size of the array B, it is necessary to align the size of the array A with the size of the array B (B22). In this embodiment, the array size is also the same. Under the above conditions, the product is obtained between the elements at the corresponding pointer positions of the arrays A and C, and the result is transferred to the same pointer position of the array D of the same size (B24). The array D generated in this way holds numerical values including positive and negative signs as array data.

  Subsequently, the total S of the elements of the array D is obtained (B26). The total value S is the sum of the distances from the design pattern of the inspection pattern shape and represents the degree of coincidence between the patterns. For example, if the shape of the inspection pattern is exactly the same as the design pattern, the value of S is 0, and the value of S increases as it deviates from the design pattern.

  By the above-described procedure, for example, which of two different inspection patterns is closer to the designed pattern can be quantitatively determined. It is also possible to set a threshold value of S and determine the pass / fail of the test pattern based on the degree of match with the design pattern of the test pattern (B28).

  According to the present embodiment, since the evaluation and alignment can be performed in a very short time, the resource efficiency of a CPU (Central Processing Unit) used for processing can be increased.

(2) Second Embodiment In the first embodiment described above, the degree of coincidence S between the inspection pattern and the reference pattern is calculated, but in this embodiment, in addition to calculating the degree of coincidence S, the inspection pattern Explains how to visually display how the pattern matches (and does not) match with the design pattern by converting the array D into an image and displaying it on the inspection image. To do.

  As described in the first embodiment, since the array data D includes a distance value as an element, it cannot be displayed as an image as it is. In addition, since the contour is detected with an accuracy of 1/10 pixel, it cannot be simply superimposed on the inspection image as it is. For this reason, an array is divided in units of 10 × 10 regions to define pixels, the total of the regions is further converted into an 8-bit integer value, and image conversion processing using the value as the pixel value is executed. In order to make it easier to grasp visually, an 8-bit gray scale is converted into an index color image with reference to an appropriate color table, and is displayed superimposed on the inspection image. FIG. 5 schematically shows an example of an image obtained by superposing the index color image and the inspection image on which such conversion processing has been performed. In the representation mode as shown in the figure, since the contour point is in units of one pixel, the exact contour position cannot be represented, but it is easy to grasp how much of the contour is deviated from the design data. be able to. In the superimposed image Img4 shown in FIG. 5, as the portion indicated by the symbol R (red) is close to the design pattern, and the symbol represents a color having a short wavelength such as Y (yellow) → G (yellowish green). It is expressed that it is away from.

  In the present embodiment, the array data D is finally converted into an image and displayed as an image. However, when high shape evaluation accuracy is not so required, the contour of the inspection pattern and the design pattern, and further, the array conversion processing result can be handled as an image. For example, the array data D is given as a distance contour image from the beginning.

(3) Third Embodiment Image data can be easily generalized as array data. In the following, the array data is treated as image data for the sake of simplicity. Therefore, the pointer coordinates of the array are described as pixel coordinates of the image, and the array values are described as the gradation values of the image. In addition, in the first and second embodiments described above, the relative position between the inspection pattern and the design pattern is determined by a normalized correlation method or the like. Although the degree of coincidence of the shapes in the positional relationship was evaluated, the optimal positional relationship may be obtained by collating one or a plurality of obtained numerical values with a certain standard, for example, the total S of elements of the array D. . For that purpose, the relative position between the inspection pattern and the design pattern is scanned and the procedure described in the first embodiment is repeated,
1) Total of element values 2) Variance of element values 3) A numerical value such as the number of elements having an element value of 0 is obtained for each relative position. Thereby, for example, if the relative position that minimizes the sum of the element values is obtained, it is possible to align the inspection pattern and the design pattern. This alignment result may be different from the result of normal correlation matching, and the alignment result by the method of this embodiment may be more preferable depending on the inspection purpose. By applying other rules, it is possible to respond to a wider variety of requirements.

(4) Fourth Embodiment In this embodiment, a case will be described in which a weight is incorporated in design data. This weight may be input by an operator at the time of inspection, or may be input in advance in design data and read by a computer when rendering it into a pattern. In any case, a different result can be obtained by taking in and calculating the influence of the array conversion process of a pattern component such as a side or a corner of a specified design pattern with a specified weight.

(5) Fifth Embodiment In this embodiment, a case where the minimum and maximum allowable dimensions are included in the design data will be described. Since the specific procedure is substantially the same as that of the first embodiment described above, the processing results are shown in FIGS. That is, FIG. 6 shows the same inspection image as the inspection image shown in FIG. 2, and FIG. 7 shows a design pattern including a minimum value (Er4a) and a maximum value (Er4b) of allowable dimensions. These minimum value (Er4a) and maximum value (Er4b) correspond to, for example, upper limit value correspondence array data and lower limit value correspondence array data, respectively. FIG. 8 is a diagram showing an array obtained by performing image conversion processing on the design pattern of FIG. Further, FIG. 9 schematically shows an example of an image obtained by superimposing the index color image and the inspection image in the arrangement shown in FIG.

(6) Sixth Embodiment A feature of this embodiment is that, in the above-described third embodiment, one or a plurality of numerical values (indicating the degree of coincidence between the shape of the inspection pattern and the design pattern in a certain positional relationship) For example, when an optimal positional relationship is obtained by collating the total S) of the elements of the array D with a certain reference, a value that can be calculated using a functional expression is adopted as a new distance value.

In addition to the case where the shape of the inspection pattern is complex, when the product pattern determined to be non-defective is used as the reference pattern and the magnification of the captured image of the product pattern is small, the density of the outline of the reference pattern increases. The elements of the array data generated by the array conversion process are almost uniform. As a result, there arises a problem that the alignment accuracy is lowered. In this case, in this case, in order to exaggerate and clarify the difference in shape within the reference pattern, the array data is generated after defining the value calculated by the following equation as a new distance value. To do.
A × exp (−C × D) (1)
Here, A is an arbitrary value, C is a constant as an attenuation parameter of the exponential function, and D is a distance from the design pattern of the inspection pattern shape. In this case, an accurate pattern alignment result can be obtained by selecting a value corresponding to the density of the inspection pattern as the constant C, for example, by setting a larger C value for a dense pattern.

(7) Program A series of procedures of the pattern evaluation method and the pattern alignment method described above may be incorporated into a program and read as a recipe file into a computer for execution. Thereby, the pattern evaluation method and the pattern alignment method described above can be realized using a general-purpose computer. Further, a series of procedures of the pattern evaluation method and the pattern alignment method described above may be stored in a recording medium such as a flexible disk or a CD-ROM as a program for causing a computer to execute and read and executed by a computer.

The recording medium is not limited to a portable medium such as a magnetic disk or an optical disk, but may be a fixed recording medium such as a hard disk device or a memory. Further, a program incorporating a series of procedures of the pattern evaluation method and the pattern alignment method described above may be distributed via a communication line (including wireless communication) such as the Internet. Furthermore, a program incorporating a series of procedures of the pattern evaluation method and the pattern alignment method described above is encrypted, modulated, or compressed, and is recorded via a wired line or a wireless line such as the Internet. It may be stored in a medium and distributed.
(8) Manufacturing Method of Semiconductor Device By applying the above-described pattern evaluation method and pattern alignment method to the manufacturing process of the semiconductor device, it becomes possible to manufacture the semiconductor device with high throughput and yield.

(9) Summary As described above, the present invention can be applied to the pattern evaluation method and the pattern alignment method, and is particularly characterized by the configurations (a) to (c).

(A) Pattern shape evaluation
(a1) In the pattern evaluation method,
A procedure for generating first array data from contour data which is shape data including contour points of the inspection target pattern;
Generating second array data from the contour data of the reference pattern including the contour points of the reference pattern serving as the inspection reference of the inspection target pattern;
For each component of the second array data, an array conversion process, which is a process of converting the value of the component of the second array data into a function value of the distance value from the component to the nearest contour point A procedure for generating third sequence data by performing;
A procedure for generating fourth array data by performing arithmetic processing between the first array data and the third array data;
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern using the constituent elements of the fourth array data;
Is provided.
(a2) In the pattern evaluation method,
A first image that is an image of an inspection target pattern is acquired, contour data that is shape data including contour points of the inspection target pattern is detected from the first image, and a first array is detected from the detected contour data The steps to generate data,
A procedure for generating second array data from contour data that is shape data including contour points of a reference pattern serving as an inspection reference of the inspection target pattern;
For each component of the second array data, an array conversion process, which is a process of converting the value of the component of the second array data into a function value of the distance value from the component to the nearest contour point A procedure for generating third sequence data by performing;
A procedure for generating fourth array data by performing arithmetic processing between the first array data and the third array data;
A step of converting the fourth array data into a second image by mapping an index color with reference to a color table set in accordance with a value of a component value of the fourth array data;
A procedure for displaying the second image superimposed on the first image;
Is provided.

Further, preferable embodiments of (a1) and (a2) include the following.
(1) The function value obtained by the array conversion process is characterized in that the sign corresponding to the value corresponding to the inside of the outline of the reference pattern is different from the value corresponding to the outside.
(2) The reference pattern is generated using design data.
(3) When A is an arbitrary value, C is a constant with an exponential function attenuation parameter, and D is a distance from the reference pattern of the inspection pattern shape, the function value is expressed by the following formula: A × exp (− | C × Distance value |)
It is calculated by the following.
(4) The design data includes a weight for the array conversion set for each component of the reference pattern,
The array conversion process includes a procedure of performing array conversion using the weights.
It is characterized by that.
(5) The reference pattern is design data including upper limit information and lower limit information of design tolerances,
The second array data includes upper limit value corresponding array data corresponding to the upper limit design allowable value, and lower limit value corresponding array data corresponding to the lower limit design allowable value,
The value of the third array data belongs only to a range defined by the upper limit value corresponding array data and the lower limit value corresponding array data.
(6) The reference pattern is design data including upper limit information and lower limit information of design tolerances,
The second array data includes upper limit value corresponding array data corresponding to the upper limit design allowable value, and lower limit value corresponding array data corresponding to the lower limit design allowable value,
The third array data is characterized by comprising positive and negative signs that differ within a range defined by the upper limit value-corresponding array data and the lower limit value-corresponding array data.

(B) Method for aligning patterns
(b1) In the pattern alignment method,
A procedure for generating first array data from contour data which is shape data including contour points of the inspection target pattern;
Generating second array data from the contour data of the reference pattern including the contour points of the reference pattern serving as the inspection reference of the inspection target pattern;
For each component of the second array data, an array conversion process, which is a process of converting the value of the component of the second array data into a function value of the distance value from the component to the nearest contour point A procedure for generating third sequence data by performing;
A procedure for generating fourth array data by performing arithmetic processing between the first array data and the third array data;
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern using the constituent elements of the fourth array data;
A procedure for performing alignment between the inspection target pattern and the reference pattern based on the calculated numerical value and a set reference;
Is provided.

Moreover, as a desirable embodiment of (b),
(1) The contour data is data detected from a first image that is an image of the inspection target pattern,
A procedure for mapping the index color by referring to a color table set according to the value of the component value of the fourth array data and converting the fourth array data to generate a second image When,
A procedure for displaying the second image superimposed on the first image;
Is further provided.
(2) The function value converted by the array conversion process has a positive / negative sign that differs between a value corresponding to the inside of the outline of the reference pattern and a value corresponding to the outside. The pattern alignment method described in 1.
(3) The reference pattern is generated using design data.
(4) When A is an arbitrary value, C is a constant as an attenuation parameter of an exponential function, and D is a distance from the reference pattern of the inspection pattern shape, the function value is expressed by the following formula: A × exp (− | C × Distance value |)
It is calculated by the following.
(5) The design data includes a weight for the array conversion set for each component of the reference pattern,
The array conversion process includes a procedure of performing array conversion using the weights.
It is characterized by that.
(6) The reference pattern is design data including upper limit information and lower limit information of design tolerances,
The second array data includes upper limit value corresponding array data corresponding to the upper limit design allowable value, and 2 lower limit value corresponding array data corresponding to the lower limit design allowable value,
The value of the third array data belongs only to a range defined by the upper limit value corresponding array data and the lower limit value corresponding array data.
(7) The reference pattern is design data including upper limit information and lower limit information of design tolerances,
The second array data includes upper limit value corresponding array data corresponding to the upper limit design allowable value, and lower limit value corresponding array data corresponding to the lower limit design allowable value,
The third array data is characterized by comprising positive and negative signs that differ within a range defined by the upper limit value-corresponding array data and the lower limit value-corresponding array data.

It is a block chart explaining the schematic procedure of the 1st Embodiment of this invention. It is a figure which shows an example of the outline of the test | inspection image of the test pattern which is the evaluation object of the pattern evaluation method shown in FIG. It is a figure which shows an example of the design pattern as a reference | standard pattern of the test | inspection pattern shown in FIG. It is an arrangement | sequence conversion image of the design pattern shown in FIG. It is a schematic diagram which shows an example of the index color image obtained by the 2nd Embodiment of this invention. It is a figure which shows an example of the outline of the test | inspection image to which the 5th Embodiment of this invention is applied. It is a figure which shows an example of the design pattern to which the 5th Embodiment of this invention is applied. 10 is an array conversion image of the design pattern shown in FIG. 9. It is a figure which shows an example of the distance outline image obtained by the 5th Embodiment of this invention.

Explanation of symbols

Dyg: Bonding part (defect part)
Es2, 4: Contours Er2, Er4a, Er4b of the inspection image: Design patterns Img2, Img6: Array conversion images Img4, Img8: Distance contour image G: Yellow green R: Red Y: Yellow

Claims (4)

By processing contour data that is shape data including contour points detected from the image of the inspection target pattern, pointer addresses corresponding to pixel coordinates of the data of the image are respectively given, and gradation values of the image data are obtained. Correspondingly, a procedure for generating first array data composed of elements having a value of 1 at the contour point position and a value of 0 at a position other than the contour point;
By processing the contour data of the reference pattern including the contour point of the reference pattern serving as the inspection reference for the inspection target pattern, a pointer address is given to each, and a value of 1 is given to the contour point position and 0 is given to a position other than the contour point. Generating second array data composed of elements having values of:
The value of the component of the second array data is a distance value closest to the component and having a value of 1, and the value corresponding to the inside of the outline of the reference pattern and the outside A procedure for generating third array data by performing processing for converting each component of the second array data into a function value having a value different in sign from the value corresponding to
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern by calculating a value of a component of the fourth array data;
A pattern evaluation method comprising: By acquiring a first image that is an image of an inspection target pattern, detecting contour data that is shape data including contour points of the inspection target pattern from the first image, and processing the detected contour data A pointer address corresponding to the pixel coordinate of the first image data is given, and a value of 1 is set at the contour point position and a value of 0 is set at a position other than the contour point corresponding to the gradation value of the image data. Generating a first array data composed of elements having;
By processing the contour data, which is the shape data including the contour point of the reference pattern serving as the inspection reference of the inspection target pattern, a pointer address is given, and a value of 1 is given to the contour point position and 0 is set to a position other than the contour point. Generating second array data composed of elements having values of:
The value of the component of the second array data is a distance value closest to the component and having a value of 1, and the value corresponding to the inside of the outline of the reference pattern and the outside A procedure for generating third array data by performing processing for converting each component of the second array data into a function value having a value different in sign from the value corresponding to
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A step of converting the fourth array data into a second image by mapping an index color with reference to a color table set in accordance with a value of a component value of the fourth array data;
A procedure for displaying the second image superimposed on the first image;
A pattern evaluation method comprising: By processing contour data that is shape data including contour points detected from the image of the inspection target pattern, pointer addresses corresponding to pixel coordinates of the data of the image are respectively given, and gradation values of the image data are obtained. Correspondingly, a procedure for generating first array data composed of elements having a value of 1 at the contour point position and a value of 0 at a position other than the contour point;
By processing the contour data of the reference pattern including the contour point of the reference pattern serving as the inspection reference for the inspection target pattern, a pointer address is given to each, and a value of 1 is given to the contour point position and 0 is given to a position other than the contour point. a step of generating a second sequence data consisting of elements having a value,
The component values of the second sequence data, closest from the elements, and a value of the distance in the configuration needed Motoma having a value 1, a value corresponding to the inner contour of the reference pattern A procedure for generating third array data by performing, for each component of the second array data, processing for converting the value corresponding to the outside and a value corresponding to the outside to a function value having a different sign.
An integration process is executed between the components of the pointer address corresponding to each other between the first array data and the third array data, and fourth array data having the result of the integration process as a component is generated. And the steps to
A procedure for calculating a numerical value that gives a degree of coincidence between the inspection target pattern and the reference pattern by calculating a value of a component of the fourth array data;
A procedure for obtaining an optimal positional relationship between the inspection target pattern and the reference pattern based on the calculated numerical value and a set reference;
A pattern alignment method comprising:   A program for causing a computer to execute the pattern evaluation method according to claim 1.

Images