JP3242241B2 - Picture magnification / angle measuring device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern magnification / angle measuring apparatus, and more particularly, to a first pattern formed on a printed document and a layout of the first pattern in a printing process. A second pattern drawn as a line drawing on the layout designation paper,
And a device for measuring a relative magnification and a rotation angle of the camera.

[0002]

2. Description of the Related Art In general, printed matter is constituted by allocating elements such as pictures, figures, and characters at predetermined positions, in predetermined sizes, and in predetermined directions (rotation angles). 2. Description of the Related Art In recent years, electronic plate making technology using a computer has been advanced, and a pattern on a printed document (for example, a photograph) is captured as image data by a scanner device, and plate collection processing is performed by the computer. In order to smoothly and accurately perform the plate collection processing using the computer, it is preferable to take in the image data in consideration of the actual layout state. For example, when a picture on a photograph is taken in by a scanner device, it is preferable to set the taking magnification and the rotation angle as image data to match the size and the rotation angle on a full-size printed matter. For this reason, conventionally, an assignment designation sheet showing the state where each pattern is assigned to the actual size is prepared, and the pattern on the photograph is projected on the assignment designation sheet using a projector, and the projected image is drawn on the assignment designation sheet. The projection magnification and the direction of the photograph were adjusted so as to match the pattern indicating the assigned layout mode, and the appropriate magnification and direction (rotation angle) were measured. For example,
JP-A-60-112039 discloses a measuring device for measuring such magnification and angle.

[0003]

The measurement of the magnification and the angle using the above-described projector is performed manually by a skilled worker. That is, the worker sets the photograph on the projector, manually adjusts the projection magnification and the angle of the photograph while projecting the image of the picture on the layout designated paper,
After visually confirming the matching of the pictures, the work of measuring the projection magnification and the angle of the photograph at that time is performed.
In addition, this operation is performed in a dark place to make the projected image easier to see. Since the work in such a dark place places a heavy burden on the worker, the work time is increased and a work error is likely to occur.

Accordingly, an object of the present invention is to provide a magnification / angle measuring device capable of automatically measuring the magnification and angle of a picture.

[0005]

[Means for Solving the Problems]

(1) The first invention of the present application is directed to an apparatus for measuring a relative rotation angle of two patterns having similar characteristics, wherein a pattern to be measured is converted from a binary image represented by black pixels and white pixels. Image capturing means for capturing as line image data, and for the line image data captured by the image capturing means, a center of gravity calculating means for calculating the barycentric position of the black pixel, and radially at a predetermined angle Δθ interval around the barycentric position. Contour pixel determining means for defining a plurality of half-lines and determining a black pixel at a position farthest from the center of gravity on each half-line as a contour pixel; and a contour distance defined as a distance between the center of gravity and each contour pixel. , And the vertical axis defines the angular position of the half line, and the vertical axis defines the outline distance obtained for the half line. Create a graph showing the relationship between, first obtained based on the first picture to be measured
Is shifted by a predetermined amount of shift in the horizontal axis direction with respect to the second graph obtained based on the second pattern to be measured, and the vertical axis value of the first graph on the same horizontal coordinate is The variance calculation means for calculating the ratio of the second graph to the vertical axis value for a plurality of different abscissa positions and calculating the variance for the obtained ratios, and setting a plurality of different shift amounts,
The variance for each shift amount is obtained by the variance calculation means, and the shift amount that minimizes the variance is determined by the second amount of the first pattern.
And an angle determining means for determining a rotation angle with respect to the pattern.

(2) The second invention of the present application is directed to an apparatus for measuring a relative magnification of two pictures having similar characteristics, wherein a picture to be measured is a binary image represented by black pixels and white pixels. Image capturing means for capturing as line drawing data composed of an image, barycentric calculating means for calculating the barycentric position of the black pixel for the line drawing data captured by the image capturing means, and a predetermined angle Δθ interval around the barycentric position A plurality of radial rays are defined, and a contour pixel determining means for determining a black pixel located farthest from the center of gravity on each half line as a contour pixel is defined as a distance between the center of gravity position and each contour pixel. By defining the contour distance calculating means for calculating the contour distance, and defining the angle position of the half line on the horizontal axis and the contour distance obtained for the half line on the vertical axis, the angle position of the half line A graph showing the relationship with the contour distance is created, and a first graph obtained based on a first pattern to be measured is converted into a second graph obtained based on a second pattern to be measured. Is shifted in the horizontal axis direction by a predetermined shift amount, and the ratio between the vertical axis value of the first graph and the vertical axis value of the second graph on the same horizontal coordinate is obtained for a plurality of different horizontal coordinate positions, A variance calculation means for calculating the variance for the plurality of ratios obtained, and a plurality of different shift amounts are set, and the variance for each of the shift amounts is obtained by the variance calculation means. Magnification determining means for determining an average value of the ratios as a magnification of the first pattern with respect to the second pattern.

(3) The third invention of the present application is the measurement apparatus according to the first or second invention, wherein the variance calculating means obtains the first variance by the first calculation, and then calculates the first variance. A second variance is obtained by performing a second calculation using only the value of the ratio within a range of a predetermined variation determined based on the rotation angle, and the rotation angle determining unit or the magnification determining unit includes:
The determination using the second variance is performed.

[0008]

The present invention is an improvement over the basic invention disclosed in Japanese Patent Application No. 5-251125. In this basic invention, contour pixels forming the contour line of each picture are extracted, and a contour distance that is a distance between the position of the center of gravity of the picture and each contour pixel is obtained. Then, based on the contour distances in the respective angular directions (a plurality of directions radially outward from the position of the center of gravity) obtained for the two pictures, the magnification and the rotation angle for these two pictures are obtained.

The present invention discloses a method of obtaining the information on the contour distance in each angle direction and then accurately obtaining the magnification and the rotation angle of two pictures. Now, it is assumed that the contour distance in each angular direction has been determined for the first picture and the second picture of a similar shape obtained by multiplying the first picture by X. At this time, if the rotation angles (directions) of both pictures are the same, the contour distance of the latter should be X times the contour distance of the former at each angular position. In other words, the ratio of both contour distances should always be constant X at any angular position. However, if there is a deviation in the rotation angle (direction) of the two patterns, the ratio of the two contour distances will not be constant unless the pattern is a perfect circle, and will have different values for each angular position.

The present invention focuses on this point and obtains the rotation angle of both pictures. That is, a graph is created by defining the angle position on the horizontal axis and the contour distance on the vertical axis, and the first graph obtained based on the first pattern to be measured is the measurement target. The vertical axis value of the first graph and the vertical axis value of the second graph on the same abscissa are shifted by a predetermined amount in the horizontal axis direction with respect to the second graph obtained based on the second pattern. With respect to a plurality of different abscissa positions. If the directions of the two pictures coincide with each other by shifting by a predetermined shift amount, the ratio of the vertical axis values obtained at that time should always be a constant value. Therefore, if the variance of the ratio value is obtained and the amount of deviation that minimizes the variance is obtained, this amount of deviation becomes the rotation angle between the two pictures. Also, the value of the ratio at this time (which is theoretically constant but is actually averaged because there is variation) is the magnification between the two patterns.

[0011]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. FIG. 1 is a diagram showing a magnification / size of a picture according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a basic configuration of the angle measuring device. The measuring device is configured to measure a relative position between a first pattern drawn on a print original A and a second pattern drawn as a line drawing on a layout designation sheet B to indicate a layout state of the first pattern. It has a function of measuring a typical magnification / angle. The image input device 10 has a function of scanning the print document A and inputting the first pattern drawn here as image data indicating the density value of each pixel. Similarly, the line image input device 20 has a function of scanning the layout designation paper B and inputting the second pattern drawn here as line image data indicating the density value of each pixel. As described later, in this embodiment, a color positive photograph is used as the print document A, and the image input device 10 performs a process of inputting a first pattern from the photograph. On the other hand, the second pattern drawn on the layout designation sheet B is a line drawing indicating the layout mode of the first pattern, and the line image input device 20 performs a process of inputting the line drawing. Of course, in practice, the image input device 10 and the line image input device 20 may be shared by the same scanner device.

The computer 100 includes an image input device 10
And the line image data input from the line image input device 20, and performs various arithmetic processing described later on these data to obtain a first pattern drawn on the print original A and the layout designation sheet B. Has a function of calculating a relative magnification and a rotation angle with respect to the second pattern drawn in the above. Although the computer 100 is actually a general-purpose computer with dedicated software incorporated therein, in FIG. 1, for convenience of description, the computer 100 is represented as an apparatus including five functional blocks. The edge detection unit 110 performs an edge detection process for detecting an edge portion where a density value changes rapidly with respect to image data input from the image input device 10. Print manuscript A used in this embodiment
Is a photographic original. By performing edge detection processing on image data obtained from the photographic original, an edge portion in which the density value changes rapidly, that is, line image data obtained by extracting only the boundary line of the picture is obtained. can get. As such an edge detection process, for example, various methods such as a mask process using a 3 × 3 Laplacian operator are known, and a detailed description thereof will be omitted.

The binarization processing section 120 includes an edge detection section 11
The binarization processing is performed on the line image data (the pattern of the print original A) subjected to the edge detection processing by 0 and the line image data (the pattern of the layout designation paper B) input by the line image input device 20. Has functions. For example, the line image data is set to 0 for each pixel arranged in the vertical and horizontal directions.
When a density value of ~ 255 is defined, a pixel having a density value equal to or higher than a predetermined threshold value (for example, density value 128) is a white pixel, and a pixel having a density value lower than the threshold value is black. Binarization processing is performed by defining pixels. Here, the pixels forming the line portion of the line drawing are called black pixels, and the pixels forming the background portion are called white pixels.

The contour line extraction unit 130 includes a binarization processing unit 12
The line image data binarized with 0 has a function of extracting an outline, and the magnification / angle calculation unit 140 has a function of calculating the magnification and angle of two patterns based on the extracted outline. Have. The processing in each of these units is the center of the technical idea of the present invention, and will be described in detail below with reference to specific examples.

Here, more specifically, an example using a print original A as shown in FIG. 2A and an assignment designation sheet B as shown in FIG. 2B will be described below. I will. The printed document A is a color positive photograph in this embodiment, and a pattern a is shown. On the other hand, the layout designation paper B is the same size paper as the actual printed matter, and has a pattern b for indicating the actual layout state of the pattern a. As shown in the figure, usually, the relative magnification of the picture a and the picture b are different, and the directions (relative rotation angles) are also different. In the example of FIG. 2, the pattern b is an enlarged version of the pattern a. If the lower side A1 of the print document A and the lower side B1 of the layout designation paper B are defined as reference sides, respectively, the lower side B2 of the pattern b is defined as the reference side B1.
(That is, the rotation angle is 0), but the lower side A2 of the pattern a is inclined by the rotation angle α with respect to the reference side A1. An object of the present invention is to automatically determine a relative magnification and a relative angle (angle α) between a picture a and a picture b.
Conventionally, a print original A is set on a projector, a picture a is projected on an allocation designation sheet B, and a projection magnification and a direction of the print original A are adjusted so that a projected image of the picture a and a picture b coincide. Was performed manually, and the magnification and the angle α were measured.
However, as described above, such a work is heavy and an error is likely to occur.

The use of the measuring apparatus shown in FIG. 1 makes it possible to measure the magnification and the angle by the following simple operations. First, the operator inputs a picture a on the print document A by the image input device 10 (scanner device). At this time, the orientation of the print document A may be based on, for example, the reference side A1 (in other words, if inputting as image data in the XY coordinate system, the input is performed with the reference side A1 as the X-axis direction). Should be done). Subsequently, the operator inputs the picture b on the layout designation paper B by the line image input device 20 (scanner device). At this time, the orientation of the layout designation paper B may be, for example, based on the reference side B1 (in other words, if input as image data in the XY coordinate system, the reference side B1 is set in the X-axis direction). Just enter it). The work performed by the worker is now completed.
After that, the computer 100 automatically determines the magnification and the angle. Hereinafter, processing in the computer 100 will be described in detail.

The image data captured by the computer 100 is subjected to a binarization process in a binarization processing section 120. However, first, the edge detection unit 110 performs an edge detection process on the image data of the picture a on the print document A. This is because, as shown in FIG. 2, the design b on the layout designation paper B is a line drawing, whereas the design a on the print original A is a so-called photographic image. By performing the edge detection processing, a line image in which only the boundary line portion is extracted from the photograph-like image can be obtained. When the binarization process is performed by the binarization processing unit 120, each of the patterns a and b becomes a line drawing composed of two types of pixels, a black pixel and a white pixel. FIG. 3 is an enlarged view of a part of the line drawing thus obtained. In the computer, each pixel is represented as “0” or “1”. In this specification, as described above, a pixel constituting a line portion of a line drawing is a black pixel, and a background portion is a pixel. The constituent pixels are called white pixels. In the example shown in FIG. 3, hatched squares are black pixels, and white squares are white pixels.

Contour line extraction processing is performed on the line drawing data that has been subjected to the binarization processing in the contour line extraction unit 130. Here, in order to facilitate understanding of this contour line extraction processing, as shown in FIG. 4, the contour line extraction unit 130 is further divided into four functional blocks (a centroid calculation unit 131, a contour pixel determination unit 132, and a contour distance calculation unit). 133, contour correction unit 134)
Think separately.

The center-of-gravity calculating section 131 has a function of calculating the position of the center of gravity of the black pixel for the given line drawing data. For example, consider a case where each pixel is arranged in j rows and i columns on an XY coordinate system, and the density value f (i, j) of each pixel is given as line drawing data. The density value f (i,
j) is either “0” or “1”. here,
Moment M (p, q) of the image becomes what, M (p, q) = Σ i, j f (i, j) · i If p · ^{j q} and definition, the position coordinates of the center of gravity _(X G, Y _G ) is given by X _G = M (1,0) / M (0,0) Y _G = M (0,1) / M (0,0) The center-of-gravity calculation unit 131 has a function of performing such a calculation to determine the position of the center of gravity G. For example, for the line drawing data of the picture a, the barycenter G of the black pixel is obtained at a position as shown in FIG.

The contour pixel determining section 132 has a function of determining a contour pixel using the center of gravity G thus obtained. Now, as shown in FIG. 6, the center of gravity G obtained for pattern a as the end point, to define a semi-straight line L ₀ in the direction parallel to the reference side A1. The direction of the half-line L ₀ as a reference 0 ° angle, below, and to measure the angle counterclockwise, to defining an angle of up to 360 °. Here, with the center of gravity G as an end point, a half line L _θ arranged at an arbitrary angular position θ is defined, and a black pixel at the position farthest from the center of gravity G on this half line L _θ is determined as a contour pixel. It is. In the example of FIG. 6, on the half-line L _theta is two black pixels Q1, Q2 exists, black pixel Q2 of these farther from the center of gravity G is the contour pixel. In the case where the thickness of the line is equal to a plurality of pixels, a plurality of black pixels are included in the position indicated by the point Q2 in FIG. in this case,
Among them, the black pixel farthest from the center of gravity G is the contour pixel. For example, in the example shown in FIG. 7, there are about 3 pixels line thickness, on a semi-straight line L _theta is three black pixels adjacent P1, P2, P3 is present, the furthest in this The black pixel P3 at the position becomes a contour pixel.

In FIG. 6, the angle position θ is set to a predetermined angle Δθ.
In each case, a semi-straight line having the center of gravity G as an end point is obtained. For example, if Δθ = 1 °, 360 half-lines can be defined radially at a predetermined angle of 1 ° about the center of gravity G, such as half-lines L ₁ , L ₂ ,..., L ₃₆₀ . Then, an outline pixel is determined for each of the half lines. If a contour is constituted by a set of 360 contour pixels determined in this way, for example, a contour as shown in FIG. 8 will be obtained. In a strict sense, a complete outline cannot be obtained unless the angle Δθ is set to infinity. In the case of a roughness of about Δθ = 1 °, rather than an outline, “ Although it should be called a "set," in the present specification, such a rough one will be referred to as a contour line for convenience. FIG. 8 shows a half line L _θ at an arbitrary angle position θ and a contour pixel P _θ on the half line L _θ . Compared to the original line drawing shown in FIG. 6, it can be seen that there is no line segment passing through the point Q1. As described above, the contour pixels extracted by the contour pixel determining unit 132 indicate only the contour outside the picture. In the case where the line has a thickness of a plurality of pixels as in the example shown in FIG. 7, the obtained contour is a line having a thickness of substantially one pixel as shown in FIG.

As shown in FIG. 8, the contour distance calculation unit 133 has a function of calculating a distance r _θ between the contour pixel P _θ and the center of gravity G. As described above, if 360 half-lines are defined radially at a predetermined angle of 1 °, a contour distance such as r ₁ , r ₂ ,..., R ₃₆₀ is obtained for each half-line.

The contour correcting section 134 has a function of correcting the position of a contour pixel by performing a smoothing process on the contour line obtained as described above. This contour correction function is effective in the following cases. For example, FIG.
As shown in the image data obtained from the original pattern,
Consider a case where the noise pixel P4 is mixed. Various factors can be considered as a cause of the generation of the noise pixel P4, such as mixing of noise at the time of input by the scanner device and adhesion of dust to the surface of the print document A. Now, the contour pixel determining unit 132, as described above, among the pixels existing on the half line L _theta defined at each angular position, to determine the farthest pixel as a contour pixel from the center of gravity G. Therefore, FIG.
In the case of the pattern shown in (1), the noise pixel P4 may be selected as the contour pixel. In such a case, the contour obtained is as shown in FIG. 11, and the contour includes a defect. The contour correcting unit 134 performs a process of correcting such a contour defect. That is, by smoothing the outline, the position of the outline pixel P4 in FIG. 11 can be corrected to the position of the pixel P5. Since various methods such as a method using a median filter are known as the contour correction processing based on such smoothing, a detailed description thereof is omitted here.

The outline extracting unit 130 has been divided into four functional blocks, and the individual functions of each functional block have been described. Next, the operation of the outline extracting unit 130 as a whole will be described with reference to the flowchart of FIG. It will be described based on the following. First, in step S1, the center of gravity of the given line drawing data is calculated. That is, the position of the center of gravity G is obtained by the center of gravity calculation unit 131. Subsequent steps S2 to S7 are processes performed by the contour pixel determination unit 132 and the contour distance calculation unit 133. That is, in step S2, the angle parameter θ is set to the initial value θ = 0 °, and in the next step S3, θ is increased by the predetermined angle Δθ.
Then, the half line _Lθ is defined in step S4, and the contour pixel _Pθ is determined in step S5. Further, in step S6, the contour distance _rθ is calculated. The above procedure is repeated in step S7 until θ ≧ 360 °. Therefore, for example, if Δθ = 1 ° is set, the parameter θ becomes 1 °, 2 °,.
The same procedure will be repeated 360 times while being updated to °. By this repetition procedure, the contour distances r ₁ ,
r ₂ ,..., r ₃₆₀ are obtained.

In the following step S8, a contour correction process is performed. That is, when there is a defect in the outline, the outline correction unit 134 performs correction based on the smoothing of the outline. Then, in step S9, a new center of gravity is calculated. The center of gravity calculation in step S1 described above is the center of gravity of the original line drawing (for example, the line drawing as shown in FIG. 5) before determining the contour pixels, but the center of gravity calculation in step S9 is The center of gravity. Then, in step S10, the old center of gravity (the center of gravity obtained in step S1) and the new center of gravity (step S9)
It is determined whether or not the deviation from the center of gravity obtained in (1) is within a preset allowable range. If the deviation is within the allowable range, the operation by the contour line extraction unit 130 ends as it is. In this case, the contour distances r ₁ , r ₂ ,..., R ₃₆₀ obtained based on the old center of gravity are used as they are in the next calculation processing (processing in the magnification / angle calculation unit 140). If the deviation exceeds the allowable range, the procedure from step S2 is performed again. However,
The target of this redo is the image of the contour line that has been subjected to the contour correction processing in step S8. Pixel P shown in FIG.
If a large number of noise components such as No. 4 are included and the correction process is performed in step S8, the position of the center of gravity may change significantly. In such a case, step S10
Then, the process returns to step S2, and the process of _obtaining the contour distance rθ again based on the correct center of gravity is performed.

Now, in the apparatus shown in FIG. 1, the contour line extracting section 130 performs the contour distances r ₁ , r _{2 ,.}
Is obtained as described above. In addition, the process of obtaining the contour distance is performed for a line drawing based on the first pattern a on the print document A shown in FIG. 2A, and also on the layout designation paper B shown in FIG. This is also performed for a line drawing based on the second pattern b. Here, the contour distance obtained by the former process is _represented by ra _θ (ra ₁ ,
_{ra 2,} ..., ra ₃₆₀₎ and calls, rb contour distance obtained by the processing for the latter _{θ (rb 1, rb 2,} ...,
rb ₃₆₀ ). FIG. 13 is a graph obtained by defining the angular position θ on the horizontal axis and the contour distance ra _θ obtained for the angular position on the vertical axis. Figure 14 is a graph obtained in the same manner for the contour distance rb _theta. In other words, FIG. 13 shows FIG.
14 is a graph obtained based on a pattern a shown in FIG.
Is a graph obtained based on the picture b shown in FIG.

Here, when the characteristics of both graphs are compared, it can be seen that almost the same pattern appears. This is because the position and size of the pattern b to which the pattern a should be allocated in the first place,
This is because the pattern a and the pattern b have similar characteristics because they are drawn as a line drawing indicating an angle.
Assuming that the operator has ideally drawn the pattern b, the two patterns have similar shapes and differ only in size (magnification) and direction (angle). Therefore, how the difference between the magnification and the angle appears on both graphs will be examined.

First, it can be seen that the magnification appears as a ratio of the contour distances. Assuming that both patterns have an ideal similar shape, the position of the center of gravity G is also similar. Therefore, a difference based on the similarity ratio between the two patterns should appear in the contour distance defined as the distance between the center of gravity G and the contour line. For example, if the pattern b is 1.5 times the pattern a, the corresponding contour distance may be considered to be 1.5 times. Therefore, what is obtained by multiplying the pattern shown in FIG. 13 in the vertical direction by 1.5 times should be the pattern shown in FIG. After all, if there are two of the picture orientation, the ratio of the contour distance at each angular position in each graph, if calculating the rb _theta / ra _theta, the both the pattern of the relative magnification this finding.

On the other hand, it can be understood that the angle appears as a phase difference between the two graphs. In FIG. 2, the pattern b
Paying attention to, it can be understood that a rotation obtained by adding an angle −α to this corresponds to the pattern a. This appears as a phase difference between the two graphs. That is, pattern b in the graph (hereinafter, referred to as a graph rb _theta) shown in FIG. 14 paying attention to, that shifting the graph rb _theta by only the phase displacement angle -α is a graph of a picture a shown in FIG. 13 (Hereinafter, referred to as a graph ra _θ ). Therefore, the graph rb _θ
Is shifted in the horizontal axis direction by a predetermined shift amount, and if it is checked whether or not the features of both graphs overlap on the same abscissa, the shift amount when the features overlap is the phase difference between the two graphs.
That is, it is the relative rotation angle between the two pictures to be obtained.

This will be described with reference to specific examples shown in FIGS. Now, the graph ra _θ and the graph rb _θ
Are superimposed on the same coordinate system. At this time, various changes are made to the phase difference between the two graphs. In other words, one of the graphs is shifted in the horizontal axis direction by a predetermined shift amount. In this case, the degree of overlap between the two graphs changes according to the shift amount. For example, the overlap may be as shown in FIG. 15 or the overlap may be as shown in FIG. As can be seen from FIG. 15, the characteristics of both graphs overlap in the horizontal axis direction. In this manner, the amount of shift for bringing the characteristics of the graphs into an overlapping state is the phase difference between the two graphs, that is, the relative rotation angle between the two pictures to be obtained. In order to obtain this specific shift amount, the shift amount is set to 1 °, 2 °, 3 °,.
Then, the degree of overlap between the characteristics of the two graphs may be checked.

The gist of the present invention is that the variance value of the contour distance ratio is used to determine the degree of overlap between the features of the two graphs. This method will be described with reference to FIGS. Now, as shown in FIG. 15, graph ra _theta
Wherein the graph rb _theta and is, consider the overlapping state somewhere south respect horizontal axis. In this case, when we seek the ratio of the contour distance in a certain angular position (rb θ / ra _θ), at any angular position, the value of this ratio will be appreciated that substantially constant. In the example shown in FIG.
It is shown that the value of the ratio at each of the 0 °, 180 °, and 270 ° angular positions is 1.5, but the value of the ratio is also 1.5 at any other angular position. become. Although a certain degree of variation actually occurs, the value of the ratio at each angular position should be substantially constant. On the other hand, as shown in FIG. 16, let us consider a state in which the characteristics of the graphs ra _θ and rb _θ are shifted in the horizontal axis direction. Also in this case, the ratio (rb _θ) of the contour distances at a certain specific angular position
/ Ra _θ ), it can be seen that the value of this ratio varies depending on the angular position. In the specific example shown in FIG. 16, the ratio values at the respective angular positions of 90 °, 180 °, and 270 ° are 1.0, 2.6, and 2.1, respectively. Here, when the variance of the contour distance ratio at the plurality of angular positions is obtained, the variance becomes very small when the characteristics of both graphs overlap perfectly as shown in FIG. 15, but are shown in FIG. Thus, it can be seen that the variance is large when the characteristics of both graphs are shifted. Therefore, if a certain amount of deviation is set and both graphs are overlapped, and the variance of the contour distance ratio at each angular position in that state is obtained, it is possible to determine whether or not the characteristics of both graphs are overlapped by the set amount of deviation. it can. The present invention is based on such a principle, the amount of deviation that can overlap the characteristics of both graphs,
That is, the relative rotation angle between the two patterns is obtained. Note that the relative magnification of the two patterns is obtained as the contour distance ratio in a state where the features of both graphs overlap. Specifically, the contour distance ratio of 1. in the state as shown in FIG.
5 (actually, an average value is taken because there is variation) is the desired magnification.

The magnification / angle calculator 140 calculates the magnification and the angle based on such a principle. This magnification /
The angle calculation unit 140 can be functionally divided into a variance calculation unit 141 and a magnification / angle determination unit 142. The variance calculator 141 calculates the graph ra obtained for the first pattern.
_theta and, a graph rb _theta obtained for the second pattern, shifted by a predetermined shift amount, a plurality of different ratios respectively between the vertical axis value and the vertical axis values of the graph rb _theta graphs ra _theta on the same abscissa Has a function of calculating the abscissa position of, and calculating the variance of the obtained ratios. Further, the magnification / angle determination unit 142 sets a plurality of different shift amounts for the variance calculation unit, receives the variance for each of the shift amounts as a calculation result, and determines the shift amount that minimizes the variance by two picture patterns. And the function of determining the average value of the ratio obtained for the amount of deviation as the magnification of the two pictures.

Next, the calculation in the magnification / angle calculator 140 will be described more specifically. First, the magnification / angle determining unit 142 sets a predetermined shift amount j, and gives the shift amount j to the variance calculation unit 141 to calculate variance.
The variance calculation unit 141 obtains a contour distance ratio r (i, j) expressed by an expression of r (i, j) = rb ((i + j) mod M) / ra (i) for the given shift amount j. . Here, ra (i) is the value of the i-th contour distance ra _θ , and rb ((i + j) modM) is ((i + j) mo
dM) th is a value of the contour distance rb _theta. M is
Is the number of contour distance values, Δθ as in the previous embodiment.
= 1 °, and when 360 contour distance values are obtained, M = 360. i is an angular position corresponding to the abscissa value of the graph, and 1 ≦ i ≦ M, that is, i
= 1, 2, 3,..., 360. Also, j is
As described above, the shift amount set by the magnification / angle determination unit 142 is a parameter indicating "how much one graph is shifted with respect to the other graph".

For example, if the shift amount j is set to 1, the variance calculation unit 141 calculates i = 1 to 3 in the above equation.
R (1,1) = rb (2) / ra (1) r (2,1) = rb (3) / ra (2) r (3,1) = rb (4) / ra (3)……… r (359,1) = rb (360) / ra (359) r (360,1) = rb (1) / ra (360) Performing 360 calculations: 360 The value of the ratio r
(1,1) to r (360,1) are obtained.

Subsequently, the variance calculation unit 141 calculates the average r _av (j) and the variance σ (j) ² for the obtained M (360 in the above example) ratio values by r _av (j) = (1 / M) Σ _{i =} 1 to M r (i, j) σ (j) ² = (1 / M) Σ _{i =} 1 to M (r (i, j) −r _av (j) ) ² consisting calculated by the formula, reports the calculation result to the magnification / angle determination unit 142.

The magnification / angle determination unit 142 changes the set value of j from 1 to 360, and receives a report of the calculation result from the variance calculation unit 141 for each value of j. Then, the smallest one of the obtained 360 variances σ (1) ^{2 to} σ (360) ² is searched. Then, the shift amount j that gives the minimum variance
= J is output as the relative rotation angle of the two patterns,
The average r _av (J) of the ratio for this shift amount j = J is
Output as the relative magnification of the two pictures.

In the above-described embodiment, the "statistical variance" is used to determine the shift amount j = J. However, in the present invention, it is not always necessary to use the "statistical variance". Instead, any value may be used as long as the value indicates the degree of variation of a plurality of values. For example, _{値 i = 1 to M} (r (i, j) −r _av (j)) ² , a statistical standard deviation (square root of variance), and an average deviation (difference between each value and the average) (A value obtained by averaging the sum of the absolute values of As described above, the term “variance” in this specification is not limited to “statistical variance”, but broadly includes a value indicating the degree of variation of a plurality of values.

Incidentally, in actual measurement, there is a possibility that extremely large errors are included in the data of the contour distance obtained by the contour line extracting unit 130. For example, in FIG.
Two graphs ra _theta, but a laminated state rb _theta is shown, in the data constituting the graph rb _theta, since it contains error data E, the value of the graph rb _theta in angular position E _theta However, it has become an extremely outlier. If any data having such a large error exists, the variance calculated by the variance calculator 141 is greatly affected. That is, as shown in FIG.
If the characteristics of both graphs overlap in the horizontal axis direction, the contour distance ratio should be substantially constant at any angular position. However, the angular position E _θ
Since only the contour distance ratio at is different from the substantially constant contour distance ratio at other angular positions, an unexpectedly large variance value is obtained. This is a problem in determining the correct rotation angle.

In order to solve such a problem, when calculating the variance in the variance calculation unit 141, the following two-stage process may be performed. That is, first, the first variance is obtained by the first calculation, and then the second calculation is performed by using only the value of the ratio within the range of the square root (standard deviation) of the variance to obtain the second variance. This second variance is used as the value of the variance reported to the magnification / angle determination unit 142. For example, when 360 values r (1, j) to r (360, j) are obtained for a predetermined set value j, r _av (j) = (1/360) · Σi _{= 1 to 360} r (i, j) _σ by ^{(j) 2 = (1/360)} · Σ i = 1~360 (r (i, j) -r av (j)) 2 made wherein first, the first dispersion σ (j) ² is calculated. Subsequently, the above-described 360 values of the ratio values r (1, j) to
For each of r (360, j), the first variance σ
It is determined whether it is within the range of (j). And r
_The ratio value outside the range of _av (j) ± σ (j) is excluded. As a result, the value of the ratio at the angular position _Eθ shown in FIG. 17 is excluded. Thus, the average and the variance are calculated again using only the ratio values within the range of r _av (j) ± σ (j), and the second variance is reported to the magnification / angle determination unit 142. . In this way, if the variance is obtained in a two-step process, even if there is data having a large error,
It is possible to obtain the correct dispersion.

In the above embodiment, the standard deviation is used to exclude the error data. However, a value of an integral multiple of the standard deviation or a value of the average deviation may be used instead. In short, any value may be used as long as it can define a predetermined range of variation.

In this way, the magnification and the angle obtained by the magnification / angle calculation unit 140 are output from the computer 100. Eventually, the worker inputs the image input device 10
Just by performing an operation of inputting an image from the line image input device 20, the necessary magnification and angle are automatically obtained.

Although the present invention has been described based on the illustrated embodiment, the present invention is not limited to this embodiment, but can be implemented in various other modes. For example, in the above-described embodiment, M is set to 360 by setting Δθ = 1 °. However, the set value of Δθ is arbitrary and does not necessarily have to be a constant value. For example, Δθ = 1 ° and Δθ = 2 ° may be alternately repeated.

In the above embodiment, the computer 1
All processes in 00 are automatically performed, and there is no need for the operator to perform any procedure. However, the operator may perform a check if necessary. For example, when the relative magnification and angle between the pattern a and the pattern b are determined in the computer 100, the operator can check the magnification and the angle before immediately outputting the magnification and the angle. Specifically, the image a is enlarged by the obtained magnification, rotated by the obtained angle and displayed on the display, and the image b is also displayed on the same screen, so that the operator can confirm that they match. I just need. Also,
In the above-described embodiment, the contour correction processing is also performed automatically, but it is also possible to leave the matters such as whether or not the correction is required to the judgment of the operator.

In the above embodiment, this apparatus has been described as a measuring apparatus for measuring a magnification / angle. When used as a single measuring device, the computer 1
A printing device such as a printer may be connected to the subsequent stage of 00, and the obtained magnification and angle may be printed out. In contrast,
When this device is used by connecting it to another device, it is only necessary to connect necessary interfaces to the subsequent stage of the computer 100. For example, if a device for creating plate making information using the obtained magnification and angle is added at the latter stage of the computer 100, this device can be incorporated into a part of the plate making system and used. Specifically, this device can be incorporated into a part of the drum scanner device. That is, the image input device 10 and the line image input device 20 in this device are used as a pre-scan device, and the magnification and the angle when the print document A is input by the drum scanner device are obtained. Computer 100
In the plate making information creating apparatus connected to the subsequent stage, the range of scanner disassembly and the coordinates of feature points used for alignment at plate collection can be added to the obtained magnification and angle.

Alternatively, the image input device 10 may be replaced by a drum scanner device having a sufficient resolution (a resolution sufficient for practical use even after performing scaling processing such as enlargement or reduction or rotation processing). In this case, the present apparatus can be used as a part of a fully automatic plate collecting apparatus. In this case, the image input device 10 functions not as a pre-scanning device but as a formal scanning device, and the scanner input operation for the print document A only needs to be performed once. In this case, it is preferable to make fine adjustments to the values of the magnification and the angle obtained by the calculation so that the appearance of the final finish is adjusted. So, computer 1
It is advisable to add a function for performing this fine adjustment to the plate making information creating apparatus connected to the stage subsequent to 00.

[0046]

As described above, the magnification of the picture according to the present invention /
According to the angle measuring device, the contour distance from the center of gravity of the picture to the contour pixel is obtained, a graph in which this contour distance is plotted at each angular position is created, and the variance of the contour distance ratio in the graph for the two pictures is calculated. Thus, since the relative magnification and angle of both pictures are determined, the magnification and angle of the pictures can be automatically measured.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a basic configuration of a picture magnification / angle measuring apparatus according to one embodiment of the present invention.

FIG. 2 is a diagram showing a specific example of a print original A and a layout designation sheet B shown in FIG.

FIG. 3 is a partially enlarged view of a picture a shown in FIG. 2;

FIG. 4 is a block diagram in which the function of a contour line extraction unit 130 in the apparatus shown in FIG. 1 is exploded into four functional blocks.

FIG. 5 is a diagram showing a center of gravity G obtained by a center of gravity calculation unit 131 shown in FIG.

FIG. 6 is a diagram illustrating a method of determining a contour pixel by a contour pixel determining unit 132 illustrated in FIG. 4;

FIG. 7 is a partially enlarged view of a picture a shown in FIG. 6;

FIG. 8 is a diagram showing an outline obtained for the picture shown in FIG. 5;

9 is a partially enlarged view of the contour line shown in FIG.

FIG. 10 is a diagram showing an example of a picture in which a noise pixel P4 exists.

FIG. 11 is a view showing a contour obtained based on the picture shown in FIG. 10;

FIG. 12 shows a contour line extraction unit 130 in the apparatus shown in FIG.
6 is a flowchart showing the operation procedure of FIG.

FIG. 13 is a graph showing the relationship between the contour distance / angle position obtained for the picture a shown in FIG. 2 (a).

FIG. 14 is a graph showing the relationship between the contour distance / angle position obtained for the picture b shown in FIG. 2 (b).

And graphs ra _theta shown in FIG. 15 FIG. 13 is a diagram showing a state where the horizontal axis direction position is overlapped with the shift amount to match the characteristic portion of the graph rb _theta shown in FIG. 14, a graph.

And graphs ra _theta shown in FIG. 16 FIG. 13 is a diagram showing a state where the horizontal axis direction position is overlapped with the shift amount does not match the characteristic portion of the graph and chart rb _theta shown in FIG. 14, a.

FIG. 17 is a diagram showing a case where data E having a large error is included in the graph shown in FIG. 15;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Image input device 20 ... Line image input device 100 ... Computer 110 ... Edge detection part 120 ... Binarization processing part 130 ... Contour line extraction part 131 ... Center of gravity calculation part 132 ... Contour pixel determination part 133 ... Contour distance calculation part 134 ... Contour correction unit 140 ... Magnification / angle calculation unit 141 ... Dispersion calculation unit 142 ... Magnification / angle determination unit A ... Printed manuscript (photograph) B ... Designated layout sheet E ... Data including error G ... Center of gravity _Lθ ... Semi-straight line P1~P3, Q1, Q2 ... black pixels P4 ... noise pixel P5 ... white pixel Pshita ... contour pixels a, b ... picture R.theta ... contour distance ra _θ, rb θ ... contour distance graph

Continuation of front page (56) References JP-A-7-83633 (JP, A) JP-A-6-265325 (JP, A) JP-A-62-186383 (JP, A) JP-A-3-100777 (JP, A) JP-A-2-158888 (JP, A) JP-A-2-179401 (JP, A) JP-A-62-263578 (JP, A) JP-A-2-91778 (JP, A) Arai, Aizu , Kudo, Takagi, Estimation of affine transformation parameters for line images, Proc. Of the 1993 IEICE Spring Conference, Volume 7, Japan, The Institute of Electronics, Information and Communication Engineers (58). . ^7, DB name) G01B 11/00 - 11/30 G03F 1/00 G06T 1/00 - 7/60

Claims (3)

(57) [Claims] 1. A device for measuring a relative rotation angle of two pictures having similar characteristics, wherein the picture to be measured is a line drawing data consisting of a binary image in which black and white pixels are expressed. Image capturing means for capturing, as for the line drawing data captured by the image capturing means, a center of gravity calculating means for calculating the position of the center of gravity of the black pixel, and a plurality of radially spaced at a predetermined angle Δθ around the center of gravity. Contour pixel determining means for defining a half-line, and determining a black pixel at the position farthest from the barycentric position on each half-line as a contour pixel; a contour distance defined as a distance between the barycentric position and each contour pixel By defining an angle position of a half line on the horizontal axis and a contour distance obtained for the half line on the vertical axis,
A graph showing the relationship between the angular position of the half line and the contour distance is created, and the first graph obtained based on the first pattern to be measured is obtained based on the second pattern to be measured. A predetermined amount of shift in the horizontal axis direction with respect to the obtained second graph, and different ratios between the vertical axis value of the first graph and the vertical axis value of the second graph on the same abscissa. And a variance calculating means for calculating the variance for the plurality of ratios obtained, and setting a plurality of different shift amounts, obtaining the variance for each of the shift amounts by the variance calculating means, Angle determining means for determining the amount of deviation that reduces the rotation angle of the first pattern with respect to the second pattern. 2. An apparatus for measuring a relative magnification of two patterns having similar characteristics, wherein the pattern to be measured is line drawing data consisting of a binary image expressed by black pixels and white pixels. Image capturing means for capturing, centroid calculating means for calculating a barycentric position of a black pixel with respect to the line drawing data captured by the image capturing means, and a plurality of halves radially spaced at a predetermined angle Δθ around the barycentric position. A straight line is defined, and a contour pixel determining means for determining, as a contour pixel, a black pixel located at a position farthest from the barycenter position on each half line, a contour distance defined as a distance between the barycenter position and each contour pixel. By defining the contour distance calculating means to calculate, and the horizontal axis the angle position of the half line, and the vertical axis the contour distance obtained for the half line, respectively,
A graph showing the relationship between the angular position of the half line and the contour distance is created, and the first graph obtained based on the first pattern to be measured is obtained based on the second pattern to be measured. A predetermined amount of shift in the horizontal axis direction with respect to the obtained second graph, and different ratios between the vertical axis value of the first graph and the vertical axis value of the second graph on the same abscissa. And a variance calculating means for calculating the variance for the plurality of ratios obtained, and setting a plurality of different shift amounts, obtaining the variance for each of the shift amounts by the variance calculating means, The average value of the ratios obtained for the shift amount in which
And a magnification determining means for determining a magnification of the pattern with respect to the second pattern. 3. The measuring apparatus according to claim 1, wherein the variance calculating means obtains a first variance by a first calculation, and then calculates the first variance within a predetermined variation range determined based on the variance. A second calculation is performed by using only a value of a certain ratio to obtain a second variance, and the rotation angle determining means or the magnification determining means makes a determination using the second variance. To measure the magnification / angle of the picture.