JPH06105603B2 - Image correction device for scanning electron microscope - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image correction apparatus for a scanning electron microscope, and more particularly to an image correction for a scanning electron microscope, which is suitable for correcting the image disturbance due to the vibration of scanning. Regarding the device.

[Background of the Invention]

FIG. 2 schematically shows how an electron microscope image is disturbed when vibration is superimposed on electron beam scanning. Part (a) of the same figure shows the pattern and the scanning position of the imaging target. Each scan has a random position deviation in the horizontal direction, and also has a random position deviation in the vertical direction. The scanning positions may also change in the vertical direction. The maximum frequency of scanning vibration in the figure is
This is a case where the scanning frequency is lower than that of the scanning frequency, and linear scanning is performed although the scanning line moves, expands, contracts, and tilts. FIG. 2B is a scan image, and shows a state in which the periphery of the pattern is disturbed in the horizontal direction in an oscillating manner and the pattern is discontinuous in the vertical direction.

Conventionally, the image disturbance caused by the scanning vibration of the scanning image pickup system has also occurred in the surface observation by the artificial satellite, and the "satellite image distortion correction processing method" (Japanese Patent Application No. 57-16).
No. 8354) has been implemented. However,
In this method, the satellite attitude vibration, which is the cause of the scanning vibration, is measured by a precise attitude angle detector, the scanning positional deviation is determined by a geometric model, and the image data is rearranged and corrected by the determined scanning positional deviation. I was going. However, in the scanning electron microscope, the factors of the scanning vibration are not only mechanical vibration but also vibration of the magnetic field, fluctuation of the power supply voltage, etc. Therefore, all the factors are measured with sufficient accuracy, and from the model of the scanning deviation,
It has been difficult to accurately obtain the amount of displacement of the image.

[Object of the Invention]

The present invention has been made to eliminate the above-mentioned drawbacks, and the purpose thereof is to correct the image disturbance, and it is not necessary to accurately measure mechanical vibration, magnetic field vibration, etc. that cause the disturbance. An object is to provide an image correction device for a scanning electron microscope.

[Outline of Invention]

In order to achieve the above object, in the present invention, the amount of relative positional deviation between lines in the scanning direction is detected from the cross-correlation function calculation regarding the scanned image line piece. The detection of the inter-line position replacement in the scanning vertical direction is performed by using the relative correlation value calculation between the scan image line pieces spanning a plurality of lines and the relaxation method calculation for the correlation value.

The outline of the present invention will be described below. FIG. 3 (a) shows the image intensity distribution of the scan images of consecutive numbers. Although the scanning speed and position fluctuate, since both scanning images scan a position close to a horizontal straight line, the image intensity pattern is close except for the position shift. The shift amount of the image intensity pattern is obtained by the peak position of the cross-correlation function between the image line pieces. This is shown in FIG. 3 (b). The peak position of the cross-correlation function, that is, the relative shift amount of the image between the two lines changes depending on the position on the line.

Now, the digital image data is represented by x (i, j). i is the pixel number in the scanning direction, and j is the scanning line number. An image line piece of length N from the i-th pixel on the j-th line, (j + i)
If the cross-correlation function with the image piece on the line is expressed as R (i, j, l), Determined by. l is the relative position of the image line piece from line j to line (j + 1). The peak position l _m (i, j) of the cross-correlation function When a _{R (i, j, l m} (i, j)) R (i, j, l) (2). For each scan line number j, the immediately following line (j
The relative displacement function l _m (i, j) with respect to + i) generally contains noise.
It is advisable to perform a second-order polynomial approximation.

Overall position shift amount L (i, j) of each scanning line j (2)
Is the absolute displacement of the first line, _Lo (i), And the polynomial approximation coefficient A _t (j) is also Can be expressed as here, Is. The image distortion in the scanning direction is corrected for each scanning image line j by the pixel position on the next scanning image corresponding to the correct pixel position i, Then, the image data at that position or the closest position may be arranged.

Next, FIG. 4 shows the principle of detecting the order change of the images in the vertical scanning direction. The image data after the scanning direction correction is x ′ (i,
j), the cross-correlation value rj, j + _m between the image line pieces at the pixel position i spanning multiple lines after the scanning direction correction
(I) is Can be obtained with. FIG. 4A shows the actual position on the target pattern to which each scanning line after correction in the scanning direction corresponds. For the line pieces from pixel number i to (i + N-1), the scanning positions of the third line and the second line are reversed in the scanning vertical direction. By the way, generally, the correlation of intensity values between two points on an image decreases monotonically with respect to the distance between the two points. Similarly, the cross-correlation value between line pieces also decreases monotonically according to the distance on the actual pattern. Therefore, the line piece cross-correlation value rj, j + across multiple line pieces
By examining _m (i), it is possible to detect permutation in the scanning vertical direction. FIG. 4B shows the line piece cross-correlation value at the pixel position i on the directed graph between the line numbers. Arcs with small correlation values are omitted. Although the correlation value originally satisfies the symmetry r _pg = r _gp ,
The following modifications have been made to strengthen the forward relationship.

p> g r _pg = r _{g, p} (8) p <g r _pg = r _{p, g} + d What we want to find is to start a directed graph starting from node # 1 in FIG. It is a path that passes through a node once, and the sum of related correlation values is maximum. FIG. 4 (c) is a graph in which nodes, that is, line numbers are rearranged in accordance with such a path. Conventionally, such a path has been found by solving a combination problem, but it takes time, so here, FIG. 4 (d) is used.
The optimum route is obtained by the relaxation method calculation for the table of the correlation value r _pg shown in. That is, the following calculation is repeated.

In the iterative calculation, a relatively large correlation value r _pg suppresses the correlation values in the same row and the same column. On the contrary, it has the effect of increasing the correlation values that are not in the same row and the same column. Therefore, after the iterative calculation of the equation (a), there is one larger correlation value in every row and column, and only one outstanding correlation value. If the correlation value is found by the threshold value processing, that is, the row and column values at the position, the correct line order shown in FIG. 4 (c) can be known. After that, the correction may be performed according to this.

Example of Invention

An embodiment of the present invention will be described below with reference to FIG. The scanning image signal from the scanning electron microscope 1 is converted into a digital signal by the A / D converter 2 and stored in the image data buffer 3 anytime. The image data of two lines before and after in the image data buffer 3 are input to the correlation calculation device 4,
The cross-correlation function between the image line pieces represented by the equation (1) is calculated. FIG. 5 shows the processing contents of the correlation calculation device 4. Two lines of image data x (i, j) and x (i, j + 1),
One of (i = 1, ..., Number of pixels) is read out by the address control device 18 with a delay of 1 pixel, multiplied by the multiplier 19, and stored in the l-th position of the multiplied image buffer 20.
The shift read amount l is changed to complete the read multiplication. From multiplication image x (i, j) · x (i, j + 1) in buffer (1)
The cross-correlation function R (i, j, l) of the equation is calculated by the following sequential equation in order to significantly reduce the amount of calculation.

That is, from the multiplication image buffer 20, two pixel positions i,
The data of (i + N) is read by the adders 21 and 22 (1
The result R (i + 1, j, l) of expression (0) is obtained. The correlation value as the result is stored in the correlation value register 23 at any time, and is used in the sequential calculation (10) at the next position (i + 1). As for the correlation function in the register 23 for the first pixel, the maximum value detector 5 detects the peak position l _m (i, j), and the approximate polynomial detector 6 using the least square method scans the line j. It is converted into a coefficient of a polynomial representing the relative line shift in the direction. The adder 7 sequentially calculates the equation (5). That is, the content of the absolute deviation polynomial coefficient register 8 in the scanning direction is read and added to the relative deviation polynomial coefficient to obtain the absolute deviation polynomial coefficient for the line j. When outputting the corrected image of the i-th pixel, the scanning-direction image correction address calculation device 9 calculates the equation (6) and obtains the read pixel number in the image data buffer 3.

The image data corrected in the scanning direction is stored in the scanning vertical direction correction image data buffer 10. The image data buffer stores the number of lines m corresponding to the disturbance of the image scanning in the vertical direction. The correlation calculator 11 obtains the image-line-piece cross-correlation value rj, j + _m (i) of the equation (7) at several representative pixel positions with respect to the current line j. The correlation value buffer 12 stores cross-correlation values for m lines, and the correlation value is input from the correlation calculator 11 by the first-in first-out method. An example of the correlation value table in the correlation value buffer is shown in FIG. 4 (d). The relaxation order calculation of the equation (9) by the optimum order replacement calculation means 13 is added to the optimum order at the representative pixel position. Replacement is required. The interpolation processing device 14 approximates the optimum order permutation with a low-order expression regarding the pixel position, and outputs the coefficient to the scanning vertical direction image correction device 15. Based on the coefficient, the correction device 15 calculates an address so that the image data closest to the output image line (shown by the broken line) is selected from the image buffer 10 and outputs the corrected image, as shown in FIG. To do. The image data corrected in the vertical scanning direction is converted into a normal video signal by the D / A converter 16 and displayed on the image display 17. According to this embodiment, when the scanning electron microscope 1 starts scanning and the correction image buffer 10 in the scanning vertical direction is filled, the display of the corrected image starts on the image display. Therefore, it can be said that the image correction according to the present embodiment is real-time processing.

〔The invention's effect〕

As described above, according to the present invention, for the output image having the disturbance caused by the scanning vibration of the scanning electron microscope, in the scanning direction, the peak position of the cross-correlation function between the scanning image line pieces is calculated and the line spacing is calculated. Relative deviation amount is detected and disturbance is corrected,
Scanning image in the vertical direction of scanning Scanning lines across multiple lines The relaxation method calculation for the correlation value between line pieces can detect line position switching and correct the disturbance, so mechanical vibration that causes scanning disturbance as in the past By accurately measuring the magnetic field vibrations and accurately estimating the amount of the scanning turbulence by the scanning turbulence model, it is possible to provide an image correction apparatus for a scanning electron microscope that does not require any means.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of a scanning electron microscope image correction apparatus according to the present invention, FIG. 2 is an explanatory diagram showing how an image is disturbed when vibration is superimposed on scanning, and FIG. 3 is between adjacent scanning image data. FIG. 4 is a diagram showing the principle of detection of relative scanning deviation in the scanning direction by the line piece cross-correlation function and its peak position, and FIG. 4 shows the principle of scanning vertical direction permutation detection by the relaxation method based on the correlation value between line pieces. FIG. 5 is a configuration diagram of the scanning direction correlator, and FIG. 6 is a diagram showing an example of scanning vertical direction image correction.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Tetsuo Yokoyama 1099, Ozenji, Aso-ku, Kawasaki-shi, Kanagawa Inside the Hitachi, Ltd. System Development Laboratory (72) Inventor, Koichi Haruna 1099, Ozen-ji, Aso-ku, Kawasaki, Kanagawa Ceremony company Hitachi Systems Development Laboratory

Claims (2)

[Claims] 1. An A / D converter for A / D converting scanned image data by a scanning electron microscope apparatus, an image data buffer for temporarily storing the A / D converted scanned image data, and an image in the image data buffer. Correlation calculation device for obtaining the cross-correlation function between line pieces with respect to the relative position in the scanning direction and the image line piece relative position shift amount from the peak position, and adding the relative position shift amount with the absolute position shift amount obtained in the previous step The absolute position deviation amount calculation device that uses the calculated value as the absolute position deviation amount of the current step, and the image position deviation in the scanning direction is corrected by controlling the read address of the image data in the image data buffer based on the absolute position deviation amount. An image correcting apparatus for a scanning electron microscope, comprising: 2. A scanning image data obtained by a scanning electron microscope apparatus.
A / D converter that converts the data into A / D and the scanned image that has been A / D converted
Image data buffer that temporarily stores data and image data
Scan direction cross-correlation function between image line pieces in buffer
Image line from the peak position obtained with respect to the relative position
One-sided relative displacement amount is calculated, and the image at the peak position
The correlation value between line pieces is calculated over a plurality of lines, and
The correlation value buffer that temporarily stores the correlation value and the buffer in the buffer
Based on the correlation value, the sum of correlation values between adjacent lines is the maximum.
Optimal ordering to find the ordering of image lines
The replacement calculation device and the image according to the optimum order replacement
Vertical scan direction to read out the image data in the image data buffer
Scanning electron characterized by being provided with an image correction device
Image correction device for microscope.