JPS62102376A - Edge position detecting method for pattern - Google Patents

Abstract

PURPOSE:To find an edge position stably by subtracting a predetermined threshold value from plural density difference values nearby the edge and finding different values, finding the position of the center of gravity from positive values among those difference values, and using it as the edge position. CONSTITUTION:Light information on an image pattern is converted by an image pickup device 31 into an image signal, which is digitized by an A/D converting circuit 33 and stored temporarily in an image memory 35, and then differentiated by a differentiating circuit 39 and supplied to an edge position detecting circuit 14. This edge position detecting circuit 41 subtracts the predetermined threshold value from plural density differential values nearby the edge to find difference values, finds the position of the center of gravity from positive difference values among those found difference values, and uses the position of the center of gravity as the edge position of the image pattern.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to the position and detection method of grayscale images, and particularly to a method for detecting edge positions of patterns.

(Prior Art) Various image pattern edge position detection methods have been proposed for detecting edge positions of patterns in grayscale images with high precision.

As an edge position detection method of such an image pattern, for example, the literature (Transactions of the Telecommunications Society 0.85 [2] (
1982) P, 274-281).

In this method, the position where the density changes in the grayscale image input by the imaging device is taken as the edge position of the pattern, multiple edge positions of this pattern are detected, and the midpoint between these edge positions is determined as the edge position of the pattern. It is a method of estimating the center position, and specifically, it is a method of measuring the eye movement. It has been reported as one method. Here it is.

This method will be explained in detail below.

A kinematic meter for a test object that moves at high speeds like eye movements+
1111, it is necessary to sample images at high speed. Therefore, it is necessary to speed up the image sampling period by using a low-resolution imaging element with a small number of pixels. Furthermore, in order to detect pupil movement (movement amount) from this low-resolution eyeball image with a resolution of a fraction of a pixel, the position of the pupil outline and the detection resolution must be adjusted as much as possible. It is necessary to increase it.

In this document, in order to know the edge position of the pupil contour, the spatial differentiation of the density of the contour (see Figure 6 (A)) is obtained, and the centroid position of the differential waveform (see Figure 6 (1)) is calculated. The position of the center of gravity is taken as the edge position difference.

On the other hand, the amount of movement of a test object such as a pupil can be determined from the center of gravity G1 at a certain time t1 and the center of gravity G2 at time t2 after the test object has moved. Furthermore, center of gravity G
l and G2 can be determined from the following equations (1) and (2), respectively.

G+=(f″x f' (x)dx l/(J knee
f'(x)di)...(1)G2=(f!
r' (x-Δx)dx)/(f f' (X-
Δx) dXl=Gl +ΔX・・・・・・・・・・・・
(2) However, Fig. 6 (A), (B) and formulas (1), (2
), f(ri indicates the density of the image, and f'(x) indicates the one-time differential of f(ri).

In order to accurately determine the amount of movement of the object under test from the centers of gravity G1 and G2 mentioned above, it is important to accurately detect the edge position of each image in both the images at time tl and t2. .

Next, a method for detecting edge positions will be explained.

In the above-mentioned equations (1) and (2), the density of one image is expressed by a continuous function, but in a digital image, the differential is a difference. First, let the density value of point (X, Y) in one image be g(X, Y), and the difference value cl(X, Y) at this point is
Y) is determined from the following equation (3).

a (X, Y)-1, g (X+t, Y) -g (X
-1,Y)! ...-(3) There is usually little change in density in the areas on both sides of the edge of the image (the difference value is small)
However, the density distribution is not necessarily flat (the difference value is zero). Therefore, the following operation is performed to limit the range in which calculations for determining the center of gravity are performed to the vicinity of the edge portion.

In the method of this document, when the difference value d (X, Y) goes around a certain threshold (fi T Difference I[d(X,Y) >Tt (However.

A range of pixels where TL <TH) is determined, this range is set as a target area for calculating the center of gravity, and the center of gravity is determined from the following equation (4).

However, "G" indicates the edge position of the j-th row when the grayscale image is divided into a matrix and measured. According to this method, the edge position of a pattern can be read with a resolution of a fraction of one pixel.

This operation is performed for each row (X direction), and a similar operation is performed for the vertical force direction (Y direction) to detect the edge position of the pupil.

FIG. 7 is a diagram for explaining this edge position detection method. The horizontal axis represents the j-th pixel position, the vertical axis represents the difference value, and the difference value for each pixel is plotted and shown. be. In the figure, TL is a threshold value, and a difference value region 11 (indicated by a black line in the figure) with a value larger than this TI,
, and the pixel position x1 corresponding to these difference values (4
) It is possible to find the center of gravity G from the formula. However, even if the pattern of the same C light image is read by the input means, it may be difficult to read due to noise or quantization errors in the A/D conversion circuit. There is no guarantee that the concentration will be the same value every time, so the difference value obtained by calculating with formula (3) based on this concentration may also vary, albeit slightly.
In some cases, as shown in Figure 8 (A) and CB), the distribution of difference values obtained by reading the pattern of grayscale images at exactly the same time may differ only by the difference value at the pixel position Xp. For example, the difference value d (Xp) corresponding to this pixel Xp is smaller than the threshold value Tt that defines the range for calculating the center of gravity (Fig. 758 (A)), and the other value is d (Xp
) becomes a larger value than TL (FIG. 8(B)).

By the way, in the conventional edge position detection method, as shown in FIG.
), the difference value ψbIi13 (in the figure) is larger than the threshold ratio Tt.

The center of gravity was determined from equation (4) using the area indicated by the diagonal line downward to the left) as the target area for the center of gravity calculation.Also, in the case of the difference value distribution as shown in Figure 8(B), the area 13 and the area 15 ( The center of gravity was determined from equation (4) using the area indicated by the downward diagonal line in the figure as the target area for calculating the center of gravity.

(Problem to be solved by the invention) However, in the pixel position direction of the target area for single center of gravity calculation,
Particularly, since the parts at both ends are far from the center of gravity, they have a large influence when determining the center of gravity.Therefore, whether pixel 1 at the end of the target area, for example, the aforementioned pixel There is a problem in that the position of the center of gravity fluctuates greatly and the edge position of the pattern cannot be accurately determined.

This problem arises not only from the quality of reproducibility in reading grayscale images, but also from the threshold (fi T
This also occurs because the target area for gravity center calculation changes depending on the setting of L. Therefore, in the conventional edge position detection method, even a small change in the threshold (Ei T L ) causes a large change in the edge position, so there is a problem in that it is difficult to set the threshold.

The purpose of this invention is to solve the problems mentioned above.

Even if the threshold value T1 is changed, the edge position of the grayscale image pattern can be stably determined. An object of the present invention is to provide a method for detecting edge positions of a pattern.

(Means for Solving the Problem) In order to achieve this object, according to the present invention, when detecting the edge position of a pattern of a grayscale image input by an imaging device, the density of a plurality of pixels of interest is detected. Find the difference value.

A predetermined threshold value is subtracted from a plurality of density difference values near the edge to obtain a difference value.

The center of gravity position is determined from the positive difference values among these difference values.

The feature is that this center of gravity position is taken as the edge position of the pattern described above.

(Function) According to this method, the difference value used in the calculation for determining the center of gravity is calculated as the difference value (d (Xi) - TL) obtained by subtracting a predetermined threshold value from this difference value. There is.

Therefore, in the case of a difference value distribution as shown in FIG. 1 (A) corresponding to FIG. 8 (A), the target area for 1 center of gravity calculation is 2 in the figure
This is the part indicated by 1. In addition, in the case of a difference value distribution as shown in FIG. 1(B) corresponding to FIG. 8(B) described above, the target area for gravity center calculation is the portions indicated by 21 and 23 in the figure.

Therefore, whether or not a pixel located at the edge of the target area, such as the aforementioned pixel Xp, is targeted for centroid calculation, the difference value of pixel Xp is a difference smaller than the original difference value. (Since it is converted into [(d(Xp)-T[), the influence of this difference value on the result of center of gravity calculation can be reduced.

(Example) An example of the present invention will now be described with reference to the drawings. In addition, these figures are 41W to the extent that this invention can be understood.
This is only an illustration in terms of PI, and the algorithm described in the example and the means for realizing this algorithm are not limited to the illustrated example, and the same components in these figures are the same. It is shown with a code.

Embodiment 2 FIG. 2 is a block diagram showing a hardware configuration of an example of a detection device suitable for use in the image pattern edge position detection method of the present invention.

In FIG. 2, reference numeral 31 denotes an imaging device composed of, for example, a CCD or the like for converting optical information of an image pattern into an image signal (density). Reference numeral 33 denotes an A/D converter, which converts an analog image signal into a digital signal, for example, 8 bits per pixel. 35 indicates an image memory, A/
The image signal converted by the D converter 33 is stored. Reference numeral 37 denotes a control circuit which reads image signals corresponding to a plurality of pixels of interest from the one-image memory 35, for example, the j-th line, and outputs them to the differentiation circuit 39. The differentiating circuit 39 uses the above-mentioned equation (3) d(X,Y)-1g(X+1.Y)-g(
X-1, Y) I Therefore, the difference value d(Xi, Y'') of the pixel x1 in the j-th row operation direction (X direction) is sequentially calculated and this value is output to the edge position detection circuit 41. . This edge position detection circuit 41 calculates the difference value between a threshold value smaller than a preset maximum difference value near the edge position and the difference values of a plurality of pixels, and for pixels for which this difference value is a positive value. The center of gravity position H of the j-th row is calculated from the following equation (5).

H” = [Green Xi (d(Xt, Yj) −Tt)]/
[Σ (d (Xi, Yj) -Ttll...'
(5) By setting this centroid position Hj as the edge position, the edge position can be detected with a resolution of a fraction of one pixel.

Note that when calculating the difference value, the difference ([ may be calculated by subtracting this threshold value from the difference value indicating a value larger than a predetermined threshold value).

The operation of the edge position detection circuit 41 will be explained in detail below with reference to FIG.

First, through the control circuit 51, the 75-register No. 53 is sent to the counter 55, the start pixel position xS for which edge position detection is to be performed, and the end γ pixel for which edge position detection is to be completed to the second register 57. Each position Xe is set. In addition, a clear signal is supplied from the control circuit 51 to the third register 58 and the fourth register 61 to clear the first register 58 and 61.
Perform initialization.

In the differential circuit 39 [see Figure 2], the above-mentioned (3)
The difference value d (Xi, Y'') calculated according to the formula is input to the subtracter 63. This subtracter 63 calculates the difference ([d
(xl, Y") stored in the first register
j (subtract aTL) difference value (d (Xt, Yi)
TL). Furthermore, the sign 65 of this difference value is output to the control circuit 51. If this sign 65 is positive, the pixel position at this time and the difference value of that pixel are the subject of gravity center calculation, so the first adder 67 is used to store the difference value at this time and the third register 53. The sum of the calculated difference values is calculated, and the sum of the calculated difference Σ(d (Xi, Y'') -Tu is sent to the third register 59 according to the instruction (clock signal) of the control circuit 51.
Store in.

- On the other hand, in the multiplier B9, the output value of the subtracter 63 and
Multiplication is performed by the pixel position x1 output from the counter 55, and Xi (d (Xi, Yj) - Tt) null (i'
iwo? G le.

Next, using the second adder 71, calculate the sum of the multiplier value of this multiplier 69 and the value stored in the fourth register 61. ,Y
j) Store -Ttl in the fourth register 61.

Note that the code 65 output from the subtracter 63 to the control circuit 51
If is zero or negative, the above calculation is not performed because the center of gravity is not subject to calculation.

Next, by incrementing the counter 55 by one, the difference value of the pixel advanced by one in the scanning direction is transferred from the differentiating circuit 39 to the subtracter 63.
At the same time, the pixel position advanced by one is input to the multiplier 69.
Output to. Furthermore, the sign 65 of the subtraction result of the subtractor 83
Accordingly, perform the operations described above.

The series of operations described above are performed cyclically until the pixel position Xi output from the counter 55 reaches the final r pixel position Xe. The comparator 73 uses the pixel position HXi output from the counter 55 and the value rI! stored in the second register. I
lj elementary position Xe is compared at any time, and when XH=Xe, a coincidence signal S1 is output to the control circuit 51.

Σi d (Xi
.

Yi) -Tel calculation result and ΣXi (d (Xi, Yj) -Ttl stored in the fourth register 81
(7) Total ff results! = is output to the divider 75. The divider 75 calculates the center of gravity position H according to equation (5).
'' is determined, and this center of gravity position Hj is determined as the edge position of the pattern.

FIG. 4 is a block diagram showing seven stages for realizing the image pattern edge position detection method of the present invention using a microprocessor (hereinafter sometimes referred to as a microcomputer). Further, FIG. 5 is an operation flowchart for explaining the functions of the microcomputer.

The imaging device 'f!' shown in FIG. ! 31, A/D converter 33
The components of the image memory 35 and the image memory 35 are the same as those described in the first embodiment, so the explanation thereof will be omitted.

81 indicates a microcomputer, inside which is a ROM (REME).
MORY) 85. This ROM 83 stores a program for executing the calculations of equations (3) and (5) described above. In addition, the RAM 85 stores the calculation result (
This is a memory that stores the calculation results of equations 3) and (5).

Next, the edge position detection procedure by the microcomputer 81 will be explained with reference to FIG.

It is assumed that the image signal (S degrees) for one image-row, for example, the j-th row, is stored in the array M of the image memory. Let I be K.

Also, as shown in equation (5), the molecule Σ(d (xl, Yj) −
A register that stores the calculation result of Tt) is defined as R1.

(5) The denominator Xt(d(Xt, Y+) −T
t)+7) Define the register that stores the calculation result as R2.

Note that the RAM 85 may be used for the R1 and R2 registers.

First, the contents A and B of the R1 and R2 registers are initialized (step 91).Next, using the parameter indicating the pixel position in the scanning direction of the j-th row of the image, first, I=
2, and the second pixel in the j-th row in the scanning direction is set as the pixel of interest (step 93).

Next, the difference (f
Calculate iD (step 95) Next, a threshold value T that is preset smaller than the maximum difference value near the edge position is determined.
Compare I and the difference (gi D) (step 87)
.

Here, in the case of D>Tt, this pixel position and the difference value are values that interfere with the calculation of the center of gravity, so the threshold value T1 is subtracted from the difference value to obtain the difference value D'', and this difference value D
A=A+IXD' corresponding to the numerator of formula (5) using '
and B=B+D' corresponding to the denominator (step 99). Next, the next pixel in the scanning direction is processed (step 101).

If D≦TE, the calculation for determining the center of gravity described in L is not performed, and the next pixel in the scanning direction is processed (step 101).

Next, in step 103, it is determined whether all processing of the image signal in the j-th row has been completed. If the processing has not been completed (INK), the process returns to step 85 and the processing is performed sequentially.

In addition, if the distance is ≧, the center of gravity position H is calculated according to equation (5) (step 105).

As described above, the method for detecting the position of the image putter 7 of the present invention can also be implemented using a microcomputer.

Incidentally, in each of the embodiments described above, processing of an image signal for one line of a gray scale image obtained by reading at the f reading stage was explained.
By sequentially processing the image signals of the remaining rows constituting the image according to the procedure described above, it is possible to detect the edge position of the pattern of the entire image.

(Effects of the Invention) As is clear from the above explanation, according to the method for detecting the edge position of an image pattern according to the present invention, the edge position detection method of the image pattern uses the difference value of the density indicating A1 which is larger than a predetermined threshold value. The position of the center of gravity is determined, and the difference value used to calculate the center of gravity for each edge position is replaced with the difference value obtained by subtracting a predetermined threshold 1 from this difference value. .

Therefore, regardless of whether or not the pixel 1 located at the end of the target area in the pixel position direction of the target area for calculation of the centroid 1, for example, the above-mentioned pixel The difference value (d
(Xp)-Tt), the influence on the result of center of gravity calculation is reduced.

Therefore, even if the threshold 4rl T L is changed, C
It is possible to provide a pattern edge position detection method that can stably determine the edge position of a light image pattern.

[Brief explanation of drawings]

1 (A,) and (B) are diagrams for explaining the edge position detection method of an image pattern according to the present invention, and FIG. 2 is a diagram for explaining the first embodiment, which is used in the present invention. FIG. 2 is a configuration diagram showing an example of a suitable edge position detection device. FIG. 3 is a block diagram showing an edge position detection circuit built into the device shown in FIG. 2. FIG. 4 is a diagram illustrating a second embodiment of the present invention, and is a configuration diagram showing means for realizing the present invention using a microcomputer. FIG. 5 is an operation flowchart for explaining the second embodiment of the present invention, and FIG. 6 (A) and CB) and FIG. 7 are diagrams for explaining the conventional method and the present invention. FIGS. 8(A) and 8(CB) are explanatory diagrams of a conventional edge position detection method. 21.23... Difference A1 (value obtained by subtracting the threshold value from the difference value) 31... Reading means (imaging device) 33... A/D converter, 35... Image memory 37.
...control section, 39...differentiation circuit 41...
Edge position detection circuit 51...control circuit, 53...first register 55...counter, 57...first register 59...third register, 61...fourth register 63...subtractor , 65... Sign of the subtraction result 87.
...first adder, 69...multiplier 71...
Second/Ill calculator, 73... Comparator 75...
Divider, 81... microprocessor 83
...ROM, 85...RAM. Patent applicant Oki Denki Gyokugyo Co., Ltd. 21.23: Difference pixel i fL! Pixel level! Diagrams suitable for explaining this invention: Fig. 1: A block diagram of the second embodiment of the invention; Fig. 2: a block diagram of the second practical example of the invention; The theory of Kousa 11 August Figure 5 Conventional 11L hi To explain this invention, the polishing wheel is turned around Figure 6 Pixel level! Conventional disorder ``Drawing 7 of this invention will make it clearer!Explanatory diagram of the way the conventional edge is placed and hardened.Figure S''

Claims (1)

[Claims] (1) When detecting the edge position of the pattern of the grayscale image input by the imaging device, the density difference values of the multiple pixels of interest are determined, and a predetermined threshold value is determined from the multiple density difference values near the edge. A method for detecting an edge position of a pattern, comprising: calculating a difference value by subtracting the difference value, determining a center of gravity position from a positive difference value among these difference values, and setting the position of the center of gravity as an edge position of the pattern.