JP2008116207A - Image measuring apparatus, image measuring method and program - Google Patents

Abstract

An object of the present invention is to provide a configuration in which an optimum edge extraction method is automatically applied when measuring a measurement object based on image data of the measurement object.
Edges corresponding to the contour of a measurement object in order to perform measurement of the measurement object based on the image data by attaching predetermined imaging condition data relating to imaging of the image to image data Is extracted, the edge extraction method to be applied to the image data is determined based on the imaging condition data attached to the image data.
[Selection] Figure 2

Description

  The present invention relates to an image measurement device, an image measurement method, and a program, and more particularly to an image measurement device, an image measurement method, and a program for measuring a measurement object based on an image obtained by imaging.

  The measurement object is manufactured by imaging the measurement object using a scanning electron microscope for length measurement, so-called CD-SEM, and measuring the length of a specific portion of the measurement object based on the image data. Process management in the process is performed. Examples of measurement objects in this case include various semiconductor devices, recording / reproducing heads of hard disk devices, and the like.

  Here, assuming that a specific graphic exists in the captured image, processing for extracting edge points constituting the graphic is performed. In this case, an edge extraction algorithm using a curve obtained by differentiating the luminance value profile in one direction is used.

  However, since the differential curve does not become steep for an image in which the brightness of the edge portion is not so bright, an edge extraction algorithm for determining an intersection with the brightness profile curve by setting a threshold value may be suitable.

Even if the image taken by the CD-SEM is for the same part, the luminance distribution and profile shape may differ depending on the stage in the manufacturing process of the measurement object. For this reason, there is a possibility that an image suitable for the edge extraction method based on the differentiation and an image suitable for the threshold value processing method are mixed. In this case, it is necessary to switch the edge extraction algorithm by some method. This edge extraction algorithm will be described later with reference to FIG.
JP 2006-217082 A

  The present invention has been made in view of the above problems, and provides a configuration in which an optimal edge extraction method is automatically applied when measuring a measurement target based on image data of the measurement target. With the goal.

  In the present invention, imaging condition data relating to imaging of the image is attached to the image data of the measurement object, and the measurement object is provided to measure the measurement object based on the image data. In addition, when extracting an edge corresponding to an outline of a predetermined shape, an edge extraction method to be applied to the image data is determined based on imaging condition data attached to the image data.

  The attachment of the imaging condition data can be realized by storing, for example, an image data file and a corresponding imaging condition data file in the same directory in a storage device such as a hard disk device of a computer.

  According to the present invention, the operator attaches corresponding imaging condition data to each image data of the measurement object in advance, and stores the image data with the imaging condition data attached to the storage device of the image measuring apparatus. By storing the data, it is possible to cause the image measurement device to automatically and sequentially perform measurement operations on these image data.

  Conventionally, even if the measurement process for the image data of the measurement object is automatically executed, in the process of each image data, the operator himself / herself visually checks the image data in the step of determining the edge extraction method. It was necessary to judge the edge extraction method. For this reason, the execution of the measurement process is interrupted each time.

  On the other hand, in the present invention, the imaging condition data for determining the optimum edge extraction method is attached to the image data set in the image measuring apparatus in advance, so that the effective efficiency of the measurement process can be improved.

  Here, the imaging condition data of the image data is data indicating at which stage in the manufacturing process of the measurement object the image data is obtained by imaging. Therefore, the data can be acquired when the operator obtains the image data by imaging the measurement object, and can be attached at that time.

  Then, by referring to this imaging condition data indicating at which stage in the manufacturing process of the measurement object the image data to be processed was obtained, the image measurement apparatus automatically extracts the corresponding edge extraction. What is necessary is just to provide the structure which can determine a law. For example, the relationship between the imaging condition data and the optimum edge extraction method for this is tabulated in advance, and the image measuring apparatus automatically refers to this table and automatically selects the optimum edge extraction method for the imaging condition data of the given image data. It is possible to have a configuration that can be determined in an automated manner.

  ADVANTAGE OF THE INVENTION According to this invention, the efficiency of the measurement process of the said measurement target based on the image data of a measurement target can be improved effectively.

  The configuration of the embodiment of the present invention will be described in detail below.

  In the embodiment of the present invention, a file of photographing condition data corresponding to photographing an image of a measurement object with a CD-SEM is attached. At the time of actual measurement processing, an image extraction condition data is read from the file to identify the type of the image, and an edge extraction algorithm, that is, an edge extraction method is determined based on the identification result.

  In other words, the stage at which the given image was captured in the manufacturing process of the measurement object is recorded in advance in a file of imaging condition data, and the imaging condition data is obtained during the edge extraction process based on the image. The optimal edge extraction algorithm is automatically selected based on the strings in the file.

  According to the image measuring apparatus according to the embodiment of the present invention, when imaging the measurement object, the imaging condition data is attached as a separate file to the image of the measurement object, and a figure is extracted from the image, and measurement is performed based on the extracted figure. Processing is done. The attached imaging condition data is read out before the actual measurement process, and an optimum edge extraction algorithm is selected according to the read imaging condition data.

  Further, a one-dimensional luminance profile is extracted from the image of the measurement object, and a process for searching for the position of the symmetry axis of the figure related to the extraction is performed on the luminance profile. Then, when extracting the edge points constituting the figure from the luminance profile, the threshold points are separately applied to the right edge and the left edge of the symmetry axis, thereby performing the process of extracting the edge points. The If the shape of the luminance profile is asymmetric, the edge is re-extracted by appropriately correcting the threshold value on one side.

  Further, when the threshold value processing is performed once to extract the edge point and one edge is re-extracted when it is determined as asymmetric from the result, the following operation is performed.

  That is, in the luminance profile, the correction target is a longer distance between the edge extraction point and the maximum luminance point. When the distance between the edge extraction point to be corrected and the maximum luminance point used for threshold determination is n times the distance between the other edge extraction point and the corresponding maximum luminance point, the threshold determination is performed. The threshold value of the correction target is corrected so that the difference between the maximum brightness value used for the correction and the threshold value to be corrected is 1 / n times the difference between the maximum brightness value and the corresponding threshold value. To do. Then, the threshold value processing is performed again for the correction target with the corrected threshold value, and the edge is re-extracted.

  Further, instead of using the difference between the maximum luminance value and the threshold value to correct the threshold value, the threshold value ratio to the difference between the maximum luminance value and the minimum luminance value is applied. The value may be corrected.

  When the threshold value is corrected, an upper limit of the corrected threshold value is set in advance.

  When determining the position of the symmetric axis, an image near the symmetric axis is set in advance as teacher image data, and the position of the symmetric axis in the processing target image is determined by pattern matching with the teacher image.

  Alternatively, in this case, instead of using pattern matching, an operation in which a candidate point is a midpoint between the maximum point obtained when searching from the right side of the luminance profile and the maximum point obtained when searching from the left side of the luminance profile is performed. The search position may be changed at least twice, and the position obtained by taking the average of the plurality of candidate points thus obtained may be obtained as the position of the true symmetry axis.

  The function of the image measuring apparatus according to the embodiment of the present invention, that is, the image measuring method according to the embodiment of the present invention will be described with reference to FIG.

  An image measuring apparatus or an image measuring method according to an embodiment of the present invention includes a preprocessing unit 1 (that is, a preprocessing stage), an edge extracting unit 2 (that is, an edge extracting stage), and a length measuring unit 3 (that is, a length measuring stage). Including.

  The pre-processing unit 1 performs pre-processing such as removal of information unnecessary for subsequent measurement processing, such as removal of white lines included in a given image. The edge extraction unit 2 extracts edges included in the image, and the length measurement unit 3 determines a measurement position for measuring the extracted edge, and measures the distance between the edges at the measurement position. . Based on the measurement values obtained in this way, it is checked whether or not the measurement object that is the basis of the image is manufactured to a prescribed size.

  Hereinafter, the function of each part of the configuration shown in FIG. 1 will be described in detail.

  FIG. 2 is a processing flowchart for explaining the function of the preprocessing unit 1.

  In step S11, white lines unnecessary for the subsequent processing are removed from the image data to be processed, and in step S12, the type of image data, that is, the measurement object, from the imaging condition data file attached to the image data to be processed. Information indicating at which stage in the manufacturing process the image related to imaging, and dimensional information for calibration, that is, information indicating the relationship between the image data and the actual dimensions of the measurement object on which the image data is based. read.

  In step S13, calibration is performed.

  That is, the calculation for obtaining the correspondence between the number of pixels in the image data and the actual size of the measurement object that is the basis of the image data, that is, the number of pixels of the image data corresponding to the actual unit length of the measurement object I do.

  In step S14, the type of image data is determined. Here, the type of image data is selected in accordance with information obtained from the file of the imaging condition data, that is, information indicating at which stage the image data relates to imaging in the manufacturing process of the measurement object.

  In step S15, pattern matching is performed in step S16 according to any of the three types of images prepared in advance according to the selection result. By the alignment by this pattern matching, the position of the symmetry axis of the figure used in the edge extraction process in the next edge extraction unit 2 is extracted.

  That is, a template image that matches the imaging condition is selected based on information regarding which stage of image data is related to imaging in the manufacturing process of the measurement object, and template matching with the template image is performed. Then, the position of the symmetry axis of the template image in the relative positional relationship between the template image and the image to be processed having the maximum similarity is acquired as the position of the symmetry axis of the process target image. As a result, the symmetry axis of the figure included in the image can be extracted with high accuracy.

  This is because alignment by pattern matching between images obtained under different imaging conditions does not provide a sufficiently high similarity between them, so it is difficult to improve the alignment accuracy, but the other is equivalent imaging. In the alignment by pattern matching between the images obtained under the conditions, a sufficiently high similarity can be obtained, so that the alignment accuracy can be increased. Therefore, by using an image obtained under an imaging condition that matches the imaging condition of the processing target image as the template image, it is possible to improve the accuracy of pattern matching alignment, and thus increase the symmetry axis of the figure included in the image. It becomes possible to obtain the accuracy.

  FIG. 3A is a process flowchart for explaining the process executed by the edge extraction unit 2.

  First, in step S21, a variable y for acquiring the luminance profile of the processing target image is set. Here, the variable y corresponds to the y coordinate, that is, the vertical coordinate in the example of the image shown in FIG.

  In step S22, the left edge El (y) and the right Er (y) with respect to the symmetry axis are extracted.

  In step S23, it is determined whether or not these left and right edges are in a symmetrical position with respect to the symmetry axis. If the result of determination is that it is not symmetrical (NO), step S22 is repeated.

  Here, the case where it is determined in step S23 that it is not symmetrical is a case where it is determined that the edge extracted in step S22 is due to erroneous extraction. Therefore, in this case, the edge search is further continued in step S22, and a correct edge is searched for, and this is repeated until the determination result in step S23 becomes YES. However, even when the edge search is performed for the entire corresponding search range, if the determination result in step S23 is not YES, this process ends in error, and a result indicating that measurement is impossible for the corresponding variable y is output.

  In step S24, the next value is set to the variable y. In step S25, it is determined whether or not the processing for the predetermined range of the variable y is completed.

  Details of the edge extraction processing in step S22 will be described in detail with reference to FIG.

  In step S31, the imaging condition data is read from the imaging condition data file attached to the image data to be processed by the same processing as the image type determination in step S14 in FIG. It is determined at which stage in the manufacturing process the image data related to imaging, and an optimum edge extraction algorithm is selected according to the determination result. That is, it is determined whether to perform edge extraction (steps S32 to S34 or steps S35 to S37) using the differential value, or to perform edge extraction (steps S38 to S42) by threshold processing.

  The method using the differential value further includes the processes in steps S32 to S34 and the processes in steps S35 to S37, and any one of them is selected in step S31.

  In step S32 and step S35, different smoothing processing, that is, filter processing is performed. Thereafter, a differential value of the luminance value on the luminance profile is calculated in step S33 or S36. In step S34, the maximum value of the differential value is searched for inside the edge from the symmetric axis toward both sides. In step S37, the maximum value of the differential value is set outside the edge from the outside of the edge toward the symmetric axis. Explore.

  At the edge portion of the graphic included in the image, for example, as shown in FIG. 5, FIG. 6, or FIG. Therefore, the method of obtaining the position of the edge by obtaining the position where the differential value of the luminance value on the luminance profile is the maximum, that is, the processing of steps S32 to S34 or the processing of steps S35 to S37 is applied.

  If an edge extraction algorithm by threshold processing is selected, the maximum value of luminance on the luminance profile is searched for in step S38, the minimum value is searched for in step S39, and the threshold value is set in step S40. .

  Here, for example, the threshold value is obtained as a predetermined ratio with respect to the difference between the maximum value and the minimum value of the luminance obtained in steps S38 and S39, for example, 70%.

  Then, in step S41, the brightness below the threshold value obtained here is searched to obtain the position where the brightness profile crosses the threshold value, that is, the edge position. In step S42, left / right asymmetric correction processing, which will be described later with reference to FIG. 7, is performed as necessary.

  This threshold value processing, that is, the processing in steps S38 to S42, is based on the threshold method shown in FIG. At the edge portion of the graphic included in the image, for example, as shown in FIG. 5, FIG. 6, or FIG. Therefore, a method is adopted in which a certain threshold value is provided and a point crossing the threshold value is determined as the edge position, that is, the threshold method.

  The reason for changing the edge extraction algorithm depending on the type of image in this way is as follows.

  That is, since the shape or material of the measurement object varies depending on the stage in the manufacturing process of the measurement object, for example, the contour of a predetermined shape such as a protrusion provided on the measurement object in the manufacturing process, that is, the light at the edge The degree of reflection is different. For this reason, the optimum edge extraction algorithm differs depending on the stage of the image in the manufacturing process.

  FIG. 4A is a process flowchart for explaining in detail the process executed by the length measuring unit 3.

  In step S50, the contour of the figure included in the image data is linearly approximated, and a reference point is extracted by a predetermined method using the approximate line thus obtained.

  This obtains a position in the image corresponding to a specific position in the measurement target object to be imaged from a graphic included in the image data as a reference point, and determines a position to be measured in the image with reference to the reference point. This is to identify.

  In step S51, a number corresponding to the first position among a plurality of positions to be measured in the image is set in a variable t. In step S52, the reference point and the unit obtained by the calibration in step S13 in FIG. The position to be measured in the image is specified from the number of pixels corresponding to the dimension.

  In step S53, an edge is obtained by threshold processing using a predetermined threshold from the luminance profile of the image data at the position to be measured.

  In step S54, a desired length W (t) is calculated from the width between the edges thus obtained.

  In step S55, the number corresponding to the next position to be measured is set in the variable t, and in step S56, it is determined whether or not the W (t) values for all of the predetermined positions to be measured have been acquired. .

  As a result of the determination, if there remains a position to be measured that does not yet live in the measurement process, the process returns to step S52 and the above process is repeated.

  Next, the edge re-extraction process in step S53 will be described with reference to FIG.

  In steps S61 and S62 in the figure, the maximum luminance and the minimum luminance are obtained in the vicinity of the edge once obtained in step S22 on the luminance profile. In step S63, the threshold value is obtained again, and in step S64, the edge is re-extracted by threshold processing again. In step S65, the left / right asymmetry correction process is performed by the same process as the process of step S42 in FIG.

  Here, the processing of steps S61 to S65 is basically the same as the processing of steps S38 to S42 in FIG. 3B, but the processing accuracy is improved. In particular, in the edge re-extraction in step S64, subpixel conversion is performed by linear interpolation unlike the case of step S41. As a result, when the length measurement process in step S64 is performed, measurement can be performed with high accuracy.

  Next, the left / right asymmetry correction performed in step S42 in FIG. 3 and step S65 will be described.

  In the image obtained by the CD-SEM imaging of the measurement object to be subjected to the threshold value processing, for example, as shown in FIG. 5B, an image having a different luminance profile inclination between the left half surface and the right half surface is displayed. In such a case, if the same threshold value is applied to the image on the left and right sides, as shown in FIG. 6B, the edge of the right half surface is farther from the maximum luminance point than that of the left half surface. Detected by position.

  As a result, the positional relationship between the left and right maximum luminance points and the detected edges becomes unbalanced. That is, as shown in FIG. 5A, when the luminance profile is substantially symmetrical, the maximum luminance points on both sides can be obtained as shown in FIG. The point where the luminance profile crosses the threshold outside of (the left half maximum point or the right half maximum value in the figure), that is, the extracted left edge and right edge (in the figure, “left edge extraction” ”And“ Right edge extraction ”) are present at approximately equal distances from the nearest luminance maximum point.

  On the other hand, when the luminance profile is asymmetrical as shown in FIG. 5B, if the same threshold value is used on the left and right, the luminance maximum points on both sides (shown in FIG. 6B) are obtained. , The point where the luminance profile crosses the threshold outside the “left half maximum point” or “right half maximum value”), that is, the extracted left edge and right edge (“left edge extraction”, “ "Right edge extraction" is present at different distances from the most recent luminance maximum point. For this reason, the said imbalance arises.

  In order to correct this imbalance, threshold correction described later with reference to FIGS. 7 and 8 is performed, and in this example, the threshold for extracting the right half face is set high.

  Specifically, in steps S71 and S72 in FIG. 7, left and right edges are first extracted by the threshold processing. As a result, in the case of the example of FIG. 5B, for example, as shown in FIG. 8, the edge extracted due to the slope of the luminance profile being gentle on the right half is compared to the case of the left half, n times outside. That is, the distance between the maximum luminance point on the right side and the extracted edge point is Δxr = nΔxl with respect to the distance Δxl between the maximum luminance point on the left side and the extracted edge point.

  In such a case, in step S73, the right threshold value is corrected so that the difference between the maximum luminance value and the threshold value becomes 1 / n. That is, as shown in FIG. 8, the difference Δyl between the maximum luminance value and the threshold value is set to 1 / n for the right half surface. That is, the right threshold value is corrected so that Δyr = 1 / n × Δyr.

  As a result, as shown in FIG. 9B, the right edge is extracted on the inner side as compared with the case where there is no correction, that is, in the case of FIG. In both cases, edge extraction can be performed at a position that is approximately the same distance from the maximum luminance point.

  This is because human vision has non-linearity and a function that takes into account factors such as changes in luminance, so that the edges of figures determined by humans simply do not match the position defined by the threshold value. . Therefore, in order to correspond to such a property of human vision, in the above example, correction is performed so that the extracted edge position approaches the maximum luminance point on the right side.

  In the correction by such a method, it is assumed that the threshold value is very close to the maximum luminance value. When the threshold value is close to the maximum luminance point in this way, a point where the luminance suddenly increases due to noise is likely to be erroneously detected as an edge.

  In order to prevent this harmful effect, an upper limit is set to the corrected threshold value, and it is determined whether or not the upper limit is exceeded in step S74.

  If the upper limit is exceeded, the upper limit value is set as a threshold value in step S75.

  In step S76, an edge is extracted again by threshold processing on the side on which the above correction is performed, that is, on the right side in the above example.

  In this embodiment, it is necessary to obtain the position of the symmetry axis of the extracted figure. In this embodiment, the symmetry axis is obtained by performing pattern matching in step S16 of FIG.

  In addition to this method, for example, as shown in FIG. 10, a process of searching for the maximum point from both the left and right ends on the luminance profile and obtaining the midpoint of the obtained maximum point can be obtained with several different y coordinates. The brightness profiles, that is, the brightness profiles of (b) and (c) obtained from the image of FIG. 10 (a) are averaged, and the average of the x coordinates of the plurality of midpoints thus obtained is taken. Thus, the symmetry axis may be obtained.

  The present embodiment is an application example of the present invention to image measurement processing related to imaging by a CD-SEM. However, in principle, the same method can be applied to an image captured with an optical microscope.

  According to the image measurement apparatus of the present embodiment, by selecting an optimum edge extraction algorithm, the luminance distribution and the luminance profile on the image obtained by imaging the measurement object at each stage in the manufacturing process of the measurement object Even in measurement systems with different values, a large amount of images can be automatically processed, and the efficiency of product management during mass production can be effectively improved. Further, by performing left / right asymmetry correction processing even on an image in which the inclination of the luminance profile is asymmetrical, measurement processing can be performed while obtaining the right / left balance by removing the influence. Therefore, measurement accuracy can be improved effectively.

  FIG. 18 is a block diagram showing the configuration of the computer 50 or 110 constituting each embodiment of the present invention.

  As shown in FIG. 18, the computer 50 or 110 includes a CPU 51 for executing various operations by executing instructions constituting a given program, a keyboard, a mouse, etc. An operation unit 52 for inputting, a display unit 53 composed of a CRT, a liquid crystal display, etc., which displays processing progress and processing results by the CPU 51 to the user, a program executed by the CPU 54, data, etc. composed of ROM, RAM, etc. A memory 54 used for storage or as a work area, a hard disk device 55 for storing programs, data, and the like; a CD-ROM drive 56 for loading programs and data from outside via a CD-ROM 57; Or an external server via a communication network 59 such as the Internet or a LAN. And a modem 58 for download, such as a program.

  The computers 50 and 110 cause the CPU 51 to execute the measurement process executed by the image measurement apparatus of the above-described embodiment, that is, the above-described process together with FIGS. 1 to 10 through the CD-ROM 57 or the communication network 59 as a medium. Load or download a program consisting of instructions for Then, this is installed in the hard disk device 55, loaded into the memory 54 as appropriate, and executed by the CPU 51. As a result, an image measuring device is realized by the computers 50 and 110.

  Next, the Hough transform for obtaining the approximate line of the contour of the figure included in the image obtained by the CD-SEM imaging in the process for obtaining the approximate line in step S50 of FIG. 4 will be described.

  Hough conversion is described, for example, on pages 204 to 206 of “Computer Image Processing” edited by Hideyuki Tamura (Ohm Corp., 2002).

  That is, if the shape of the line to be detected is predetermined and the shape can be expressed by an algebraic equation, the Hough transform (which maps the feature points (edge points, etc.) in the image to the parameter space representing the shape of the line ( Hough Transform) is said to be effective. Here, the most standard Hough transform for line detection will be described.

First, as an algebraic equation representing a straight line,

ρ = x cos θ + ysin θ. . . (B-1)

Is used.

In the above equation, ρ is a length of a perpendicular line drawn straight from the coordinate origin, and θ is a parameter representing an angle between the perpendicular line and the x axis. Using this algebraic equation, a straight line ρ ₀ = x ₀ cos θ + y ₀ sin θ in the xy image space shown in FIG. 13A is expressed as one point in the ρ-θ parameter space shown in FIG. Is done.

An arbitrary straight line passing through the feature points (x ₀ , y ₀ ) in the image is

ρ = x ₀ cos θ + y ₀ sin θ. . . (B-2)

A straight line group diagram 13c satisfying this formula forms a locus as shown in FIG. 13d in the parameter space.

In other words, this means that the point (x ₀ , y ₀ ) in the image space is mapped to the locus in the parameter space represented by the equation (B-2).

  Based on the relationship between the image space and the parameter space, the Hough transform performs straight line detection as follows.

  1) Prepare a two-dimensional array representing the parameter space and initialize all of its values to zero.

  2) An expression obtained by substituting the coordinate values of each feature point in the image into x and y in the expression (B-1) is regarded as an equation relating to ρ and θ, and a locus represented by the equation is drawn in the parameter space (FIG. 13). (See (e)).

  To draw a trajectory, the array element through which the trajectory passes is calculated by calculating the value of ρ that satisfies the equation while increasing θ by a constant interval Δθ, and the trajectory is overwritten by incrementing the value by 1 (voting). .

  3) After drawing trajectories corresponding to all feature points, a position where a large number of trajectories are concentrated, that is, an element having a large maximum value in an array representing the parameter space is obtained.

Let the parameters represented by such elements be (ρ ^* , θ ^* ). This parameter represents a linear ρ ^* = xcosθ ^* + ysinθ ^* in the image, the straight line passing through a large number of feature points will have been detected (see FIG. 13 (e), (f) ).

  Hough transform is said to be able to detect a straight line well even if feature points in the image are not continuous, and to be more resistant to noise than edge tracking. As described above, the Hough transform has excellent characteristics, and since the first proposed 1962, many improvements and extensions have been made.

  The Hough transform can be used to detect circles and ellipses in addition to straight lines. However, three parameters are required for a circle and five parameters for an ellipse, and the simple method requires the preparation of 3D and 5D arrays, respectively, which significantly increases both storage capacity and calculation time. It is said that there is a problem.

  Various ideas have been considered for this problem. In addition, there is a generalized Hough transform as a method for detecting a two-dimensional figure that cannot be expressed by a simple algebraic equation such as a polygon. Furthermore, the Hough transform concept can also be used to detect and recognize 3D planes and objects.

  Next, an edge extraction method in the edge re-extraction process (step S53 in FIG. 4) of the edge extraction unit 2 or the length measurement unit 3 will be described.

  For the edge detection method, for example, the roadmap on the homepage of the Japan Semiconductor Manufacturing Equipment Association (2003) (URL: http://www.seaj.or.jp/rdmp/indx_f.htm, October 2006) (As of the 12th) is shown in the fifth edition “Measurement”.

  That is, there are various systems shown in FIG. 14 that are actually applied to the CD-SEM as using the luminance value profile of the pixel of the microscope image. In general, there are many other methods, and as described with reference to FIG. 3B, there is also a method using a differential value profile of luminance values. In FIG. 14, L indicates a length measurement value, and A indicates an actual pattern width.

The present invention can take the configurations described in the following supplementary notes.
(Appendix 1)
An image measurement device that measures the measurement object based on image data obtained by attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object,
Edge extraction means for extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
An image measurement apparatus comprising an edge extraction method determination unit that determines an edge extraction method applied in an edge extraction unit based on imaging condition data attached to the image data.
(Appendix 2)
The image measurement apparatus according to appendix 1, wherein the imaging condition data is configured to include data indicating a stage related to the imaging in the manufacturing process of the measurement object.
(Appendix 3)
The edge extraction method determination means selects one method from a plurality of edge extraction methods, and the plurality of edge extraction methods obtain a differential value by differentiating the luminance value of the image data, and the differential value is A first method for extracting the maximum position as an edge, and a predetermined threshold value for the luminance value of the image data, and a position where the luminance value of the image data crosses the threshold value is extracted as an edge. The image measurement apparatus according to appendix 1 or 2, wherein the image measurement apparatus is configured to include two methods.
(Appendix 4)
Furthermore, it has a means for acquiring the symmetry axis of the figure indicated by the image data,
The image measuring apparatus according to appendix 3, further comprising means for setting different threshold values on both sides of the symmetry axis in the second method.
(Appendix 5)
The means for setting different threshold values on both sides of the symmetry axis is arranged such that the distance between the position where the luminance value of the image data has a peak and the position having the threshold value is equal on both sides of the symmetry axis. Item 5. The image measurement device according to appendix 4, wherein the image measurement device is configured to set.
(Appendix 6)
When setting the threshold value so that the distance between the position where the profile of the luminance value of the image data has a peak and the position having the threshold value is equal on both sides of the symmetry axis, the maximum value and the minimum value of the luminance value are set. The image measurement apparatus according to appendix 5, wherein the image measurement apparatus is configured to use a ratio of a threshold value to a value difference.
(Appendix 7)
The image measurement device according to any one of appendices 3 to 6, wherein a predetermined upper limit is provided for the threshold value.
(Appendix 8)
The means for obtaining the symmetry axis is configured to obtain a reference image in advance and obtain a position of the symmetry axis by performing pattern matching with the reference image. The image measurement device according to any one of the above.
(Appendix 9)
The means for obtaining the symmetry axis performs the operation of obtaining the maximum value of the luminance value profile a plurality of times at different measurement positions, and determines the midpoint position of the position having the maximum value on the luminance value profile for each measurement position. The image measuring device according to any one of appendices 4 to 7, which is configured to obtain the symmetry axis by obtaining an average of the plurality of measurements for the midpoint position.
(Appendix 10)
An imaging condition attaching step of attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object;
A measurement step of measuring the measurement object based on the image data,
The measurement step includes an edge extraction step of extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
An image measurement method comprising an edge extraction method determination step of determining an edge extraction method to be applied in an edge extraction step based on imaging condition data attached to the image data in the imaging condition attachment step.
(Appendix 11)
The method according to appendix 10, wherein the imaging condition data is configured to include data indicating a stage related to imaging in a manufacturing process of a measurement object.
(Appendix 12)
In the step of determining the edge extraction method, one method is selected from a plurality of edge extraction methods. The plurality of edge extraction methods obtain a differential value by differentiating the luminance value of the image data, and the differential value is maximum. And a second method of setting a predetermined threshold value for the luminance value of the image data and extracting a position where the luminance value of the image data crosses the threshold value as an edge. The method according to appendix 10 or 11, wherein the method is configured to include:
(Appendix 13)
And further comprising obtaining a symmetry axis of the figure indicated by the image data,
The method according to claim 12, wherein the measuring step includes the step of setting different threshold values on both sides of the symmetry axis in the second method.
(Appendix 14)
In the step of setting different threshold values on both sides of the symmetry axis, the threshold value is set so that the distance between the position where the luminance value of the image data has a peak and the position having the threshold value is equal on both sides of the symmetry axis. 14. The method according to appendix 13, wherein the method is configured to set.
(Appendix 15)
When setting the threshold value so that the distance between the position where the profile of the luminance value of the image data has a peak and the position having the threshold value is equal on both sides of the symmetry axis, the maximum value and the minimum value of the luminance value are set. 15. The method according to appendix 14, wherein the method is configured using a ratio of a threshold value to a value difference.
(Appendix 16)
16. The method according to any one of appendices 12 to 15, wherein a predetermined upper limit is provided for the threshold value.
(Appendix 17)
In the step of acquiring the symmetry axis, the reference image is provided in advance, and pattern matching is performed with the reference image to obtain the position of the symmetry axis. The method according to any one.
(Appendix 18)
In the step of obtaining the symmetry axis, the operation for obtaining the maximum value of the luminance value profile is performed a plurality of times at different measurement positions, and the midpoint position of the position having the maximum value on the luminance value profile is determined for each measurement position. The method according to any one of supplementary notes 13 to 16, wherein the symmetry axis is obtained by obtaining and calculating an average of the plurality of measurements for the midpoint position.
(Appendix 19)
A program for causing a computer to execute a measurement step of measuring the measurement object based on image data obtained by attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object. And
The measurement step includes an edge extraction step of extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
A program comprising: an edge extraction method determination step for determining an edge extraction method to be applied in an edge extraction step based on imaging condition data attached to the image data in the imaging condition attachment step.
(Appendix 20)
The program according to appendix 19, wherein the imaging condition data includes data indicating a stage related to imaging in a manufacturing process of a measurement object.

  The application example of the present invention is not limited to the measurement process for the image by the scanning electron microscope in the manufacturing process of the semiconductor device and the hard disk device as described above. For example, the present invention can be applied to all industries that handle microscope images. The imaging apparatus for obtaining the measurement target image is not limited to the scanning electron microscope, but may be an optical microscope, a CCD camera, a transmission electron microscope, or the like. Is also applicable.

It is a functional block diagram of the image measuring device by the Example of this invention. 3 is an operation flowchart for explaining an operation of a preprocessing unit shown in FIG. 1. 3 is an operation flowchart for explaining an operation of an edge extraction unit shown in FIG. 1. It is an operation | movement flowchart for demonstrating operation | movement of the length measurement part shown in FIG. It is a figure for demonstrating an example of the image data handled with the image measuring device by the Example of this invention. It is a figure for demonstrating an example of the processing result with respect to the image shown in FIG. FIG. 5 is an operation flowchart for explaining a specific method of left / right asymmetric correction shown in FIGS. 3 and 4. FIG. It is a figure for demonstrating an example of the specific method of the left-right asymmetric correction shown in FIG. It is a figure for demonstrating the effect by the image measuring device by the Example of this invention. FIG. 5 is a diagram for explaining a method of determining a symmetry axis performed when extracting the edges El (y) and Er (y) in FIG. 3 and re-extracting the edge in FIG. 4. It is a block diagram for demonstrating the image measuring device by the Example of this invention. It is a block diagram for demonstrating the structure of the said computer in the case of implement | achieving the image measuring device by the Example of this invention with a computer. It is a figure for demonstrating Hough conversion. It is a figure for demonstrating the edge detection method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Pre-processing part 2 Edge extraction part 3 Length measurement part 50 Image measurement apparatus 100 CD-SEM

Claims (7)

An image measurement device that measures the measurement object based on image data obtained by attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object,
Edge extraction means for extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
An image measurement apparatus comprising an edge extraction method determination unit that determines an edge extraction method applied in an edge extraction unit based on imaging condition data attached to the image data.   The image measurement apparatus according to claim 1, wherein the imaging condition data includes data indicating a stage related to the imaging in a manufacturing process of the measurement object.   The edge extraction method determination means selects one method from a plurality of edge extraction methods, and the plurality of edge extraction methods obtain a differential value by differentiating the luminance value of the image data, and the differential value is A first method for extracting the maximum position as an edge, and a predetermined threshold value for the luminance value of the image data, and a position where the luminance value of the image data crosses the threshold value is extracted as an edge. The image measurement device according to claim 1, wherein the image measurement device is configured to include two methods. Furthermore, it has a means for acquiring the symmetry axis of the figure indicated by the image data,
The image measuring apparatus according to claim 3, further comprising: a threshold value that is different on both sides of the axis of symmetry in the second method. An imaging condition attaching step of attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object;
A measurement step of measuring the measurement object based on the image data,
The measurement step includes an edge extraction step of extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
An image measurement method comprising an edge extraction method determination step of determining an edge extraction method to be applied in an edge extraction step based on imaging condition data attached to the image data in the imaging condition attachment step.   The method according to claim 5, wherein the imaging condition data includes data indicating a stage related to imaging in a manufacturing process of a measurement object. A program for causing a computer to execute a measurement step of measuring the measurement object based on image data obtained by attaching imaging condition data related to the imaging to image data obtained by imaging the measurement object. And
The measurement step includes an edge extraction step of extracting an edge corresponding to a contour of a predetermined shape on the measurement object based on the image data;
A program comprising: an edge extraction method determination step for determining an edge extraction method to be applied in an edge extraction step based on imaging condition data attached to the image data in the imaging condition attachment step.