JPH07120391B2 - Character cutting device - Google Patents

Description

The present invention relates to a character cutting device, and more particularly to a device for cutting out a character pattern optically read from a form or the like for each character (for example, an optical character reading device). In the above, the present invention relates to a character cutout device suitable for separating and cutting out a character and a character touched by a plurality of points.

[Conventional technology]

Conventionally, when a contact character is separated and cut out in an optical character reader or the like, first, it is determined whether the input character pattern is a contact character, and if it is determined that it is a contact character,
The touch character pattern is divided. Here, the judgment as to whether or not it is a touched character is made by comparing the X-direction width W of the input character pattern with the specified character entry frame width or pitch P. That is, if W ≧ KP (K: constant), it is determined that the character is touched. Generally, a value of K≈1.5 is used.

Also, the number of contact characters is calculated by N = W / P and rounded off to the right of the decimal point. In such a method, it is premised that the written characters are written at regular intervals, and there is a limit to the separation of contact characters having different written character widths or pitches with respect to a specified character pitch P. .

For example, in the clipping method disclosed in Japanese Patent Laid-Open No. 57-45676, a method is proposed in which the distance from the upper and lower frames to each character is calculated to detect a change point, and the contact character is cut based on this. However, there is a problem that when the characters overlap and come into contact with each other, the characters are cut at a point other than the original contact point.

Further, in the cutout method disclosed in Japanese Patent Laid-Open No. 57-5181, a method of cutting a touch character by an analog signal level difference using a multiplexed pattern is proposed. There was a problem that it was difficult to detect the cut point when the density was high.

[Problems to be solved by the invention]

After all, in any of the above-mentioned conventional techniques, when the characters of a form or the like are read by an optical character reading device or the like, the character entry frame must be large, and a space must be provided for each character. However, the size of the sheet becomes large and the efficiency of use of the sheet deteriorates.

In order to solve such a problem, the present applicant first divides the contour of the touched character into an upper contour and a lower contour of the touched character, and divides the upper contour and the lower contour. A character cutting method has been proposed in which a contact point is detected and a contact character is separated by utilizing a change in distance (see Japanese Patent Laid-Open No. 59-160274). In this clipping method, the distance between the upper contour and the lower contour of the character pattern is obtained for each point in the X direction, and a threshold value is determined from the minimum value of the contour difference thus obtained. A point that changes from below to a threshold value or more is a contact point.

The character cutout device according to the present invention is based on a system obtained by further improving this cutout system, and can deal with a case where characters are in contact with each other at a plurality of points. That is, the object of the present invention is to solve the above-mentioned problems in the prior art, to stably detect a contact point even when characters are in contact with each other at a plurality of points, and separate the characters into character units. It is to provide a character cutting device capable of cutting out.

[Means for solving problems]

The above object of the present invention is to detect whether or not a character pattern is a contact character by detecting the width (width in the X direction) of the character pattern in a character cutting device that cuts out a character pattern optically read from a form or the like for each character. A first detecting means for determining whether or not there is an inner contour, a second detecting means for detecting the presence or absence of the inner contour and the number of the inner contour, and a plurality of characters generated when the characters come into contact with each other at multiple points (a plurality of points). The third detecting means for detecting the Y-direction partial width of the character pattern from the inner contour of the character pattern and the outer contour of the character pattern, and the second pattern for the character pattern determined to be the contact character by the first detecting means. When the presence of the inner contour is detected by the detecting means, the change point of the Y-direction partial width of the character pattern detected by the third detecting means is set as the contact point based on the position and the number of the inner contour. Sentences cut out in single character units It is accomplished by character extraction apparatus characterized by having a cutout control unit.

[Action]

In the character cutting device according to the present invention, when the width of the character pattern in the X direction is equal to or larger than a certain value, it is determined that the character is a contact character, and it is checked whether an inner contour exists near the center of the character pattern. If there are n contours, it is determined that it is a character pattern that touches at (n + 1) points, and it is determined that these n inner contours have occurred due to the characters touching at multiple points. To do. Then, the outer contour of the character pattern determined to be in contact with the character at a plurality of points, the upper outer contour and the lower outer contour of the character pattern at the left end point and the right end point of this outer contour. Divide into

For the sake of simplification, when it is assumed that characters are in contact with each other at two points, an inner contour determined to be caused by contact between two characters is also similar to the outer contour. To divide into an upper inner contour and a lower inner contour. Next, the distance between the upper outer contour and the upper inner contour is calculated for each point in the X direction, and the upper contour difference is obtained. Similarly, the lower contour difference is obtained. Using the change point of the upper contour difference obtained by such a procedure, the upper contact point of the two contact points is detected, and the lower contact point is detected using the change point of the lower contour difference. To do. Then, the characters are separated for each character at the detected contact point.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an optical character reader according to an embodiment of the present invention. In the figure, 11 is a character pattern extraction unit that extracts one character pattern from a read pattern optically read from a form, and 12 is a contact character determination that determines whether or not the extracted character pattern is a contact character. Part, 13 is a contact character separation part that separates and cuts out character patterns that have been determined to be contact characters on a character-by-character basis, and 14 is a recognition that recognizes character patterns by matching unknown input character patterns with standard character patterns Reference numeral 15 denotes a dictionary portion in which standard character patterns are stored.

The outline of the operation of the optical character reading apparatus according to this embodiment configured as described above will be described below.

The characters read from the form or the like are sent to the character pattern extraction unit 11. The character pattern extraction unit 11 extracts one character pattern and sends it to the contact character determination unit 12. The contact character determination unit 12 determines whether the character pattern sent from the character pattern extraction unit 11 is a contact character by the conventional method described above, and the character pattern determined to be the contact character is a contact character. Send to the character separation unit 13. The character pattern that is determined not to be a contact character is sent to the recognition unit 14 and the dictionary unit 15
And is recognized as. The contact character separation unit 13 separates the character pattern sent from the contact character determination unit 12 into character units and sends it to the recognition unit 14.

Next, the operation of each unit will be described in detail.

First, the operation of the character pattern extraction unit 11 will be described.
An example of extracting the outline of a character pattern is shown in FIG. As shown in FIGS. 2A and 2B, the character pattern extraction unit 11 traces the contour of the read pattern 20 and extracts the outer contour 21 and the inner contour 31. The inner contour is extracted by the method described in JP-A-61-59568, page 3, upper left column, lines 4 to 7, for example. Next, as shown in FIG. 2C, the X and Y coordinates of each point on the extracted contour are
The outline coordinate table 100 is written as shown in FIG.

Next, the operations of the contact character determination unit 12 and the contact character separation unit 13 will be described. FIG. 3 is a flow chart showing the operation of the touch character determining unit 12, and FIG. 4 is a flow chart showing the operation of the touch character separating unit 13. Also, as an example here, the process until the characters touching at two points are separated is shown in FIGS. Below, from FIG.
Using the figure 10, the contact character determination unit 12 and the contact character separation unit
The operation of 13 will be described.

First, the operation of the contact character determination unit 12 will be described. The contact character determination unit 12 executes the processing of step 501 in FIG.
As shown in FIG. 5, from the difference between the X coordinate minimum value Xmin o22 and the X coordinate maximum value Xmax o23 of the outer contour 21 of the character pattern extracted by the character pattern extraction unit 11, the width in the X direction of the character pattern (hereinafter, Simply called "width"). And
As described above, when the width of this character pattern is equal to or larger than a certain value, the character pattern is determined as a contact character.

Next, in the process of step 502 in FIG. 3, the character pattern determined to be a contact character is a pattern in which characters are in contact with each other at one point, or characters are in contact with each other at multiple points. The pattern is checked by the number of inner contours near the center of the character pattern. That is, if there are n inner contours near the center of the character pattern, it is determined that they are in contact at point (n + 1).

Here, it is determined that the two points are in contact with each other. It should be noted that this determination that the inner contour exists near the center of the character pattern is made by, for example, the minimum X coordinate value Xmin i of the inner contour 31.
32 and the maximum X coordinate value Xmax i33 are obtained, and the minimum X coordinate value Xmin o22 and the maximum X coordinate value Xmax of the outer contour 21 of the character pattern are obtained.
It is assumed that the relationship with o23 satisfies the following expression, that is, the inner contour is applied to the center of the character pattern in the X direction.

Xmin i ≤ (Xmin o + Xmax o) / ≤ Xmax i ...... (1) If three or more characters are in contact, Xmin i ≤ (Xmin o + α) / ≤ Xmax i ...... (2) You can Here, α can be a one-character entry frame width based on predesignated format information. It is also possible to add a condition in the Y direction to the determination of the position of the inner contour.

Next, the operation of the contact character separation unit 13 will be described. The contact character separation unit 13 executes the processing of step 601 in FIG.
The outer contour 21 of the character pattern determined to be in contact at two points by the contact character determination unit 12 is shown in FIG.
As shown in (B), the upper outer contour 41 and the lower outer contour 42 are divided according to the X coordinate minimum value Xmin o22 and the X coordinate maximum value Xmax o23.

Next, as with the outer contour 21, the inner contour 31 generated by the contact between the characters at a plurality of points, that is, the inner contour 31 satisfying the above-described formula (1) or (2), also has the upper inner contour. 51 and the lower inner contour 52.

Next, by the processing of step 602 in FIG. 4, as shown in FIG. 7 (A), the Y of the upper outer contour 41 and the upper inner contour 51 is compared.
The directional distance (Y-direction partial width) is calculated from the X coordinate minimum value Xmin i32 to the X coordinate maximum value Xmax i33 of the upper inner contour. The upper contour difference 61 thus obtained is shown in FIG. 7 (C). Similarly, the lower-side contour difference 62 is obtained (see FIGS. 7B and 7D).

Next, as shown in FIG. 8 (A), the minimum value of the upper contour difference 61 is detected by the processing of step 603 in FIG.
The upper threshold T _U 71 is determined by the following formula.

T _U = minimum value of upper contour difference 61 + β (3) Here, β is a constant value.

Next, by the processing of step 603 in FIG. 4, points 73 to 74 at which the above-described upper contour difference 61 becomes equal to or less than the above-mentioned upper threshold T _U 71.
(See FIG. 8A) is detected. And this point 73 ~
Among the 74, the point 74 closest to the center 72 of the character pattern is detected as the upper contact point (see FIG. 9A). Less than,
Similarly, for the lower contact point, the lower contact point 84 is detected using the lower outer contour 42 and the lower inner contour 52 (see FIGS. 8 (A) and 9 (A)). .

Finally, at the upper contact point 74 and the lower contact point 84, the character pattern is separated into two characters. Figure 10 (A), (B)
The separated one-character patterns 91 and 92 shown in FIG.

According to the above-described embodiment, even when characters are in contact with each other at two points, it is possible to stably separate and cut out in units of one character.

In the above embodiment, the separating operation when the characters are in contact with each other at two points has been described. However, the separating operation when the characters are in contact with each other at only one point is as described above. It goes without saying that the method described in JP-A-59-160274 can also be used.

Next, another embodiment of the present invention will be described. The present embodiment relates to a separating operation when characters are in contact with each other at three or more points.

In FIG. 11, “0” and “3” indicate that the number of inner contours is n = 2,
That is, an example of a pattern in which (n + 1) = 3 points of contact is shown, and FIGS. 12 and 13 are diagrams for explaining such separation of multi-point contact characters.

First, the operation of the contact character determination unit 12 will be described. Since there are n inner contours (230, 240) satisfying the above-described formula (1) or formula (2), the contact character determination unit 12 determines that the character pattern is a character pattern in which characters contact each other at (n + 1) points. To do. Next, C ₁ (230), C ₂ (240), ..., Cn in order from the top with respect to the position in the Y direction on the above-mentioned inner contour (230, 240).
And sign. At this time, the outer contour of the character pattern 210
Is C ₀ .

Next, the operation of the contact character separation unit 13 will be described. The touch character separation unit 13 determines the outer contour C ₀ (210) of the touch character pattern.
N inner contours C ₁ (230), C ₂ (24
0), ..., Cn are set to the upper contour C _0U (21
1), C _1U (231), C _2U (241), ... Cn _U and lower contour C _0L (2
12), C _1L (232), C _2L (242), ..., Cn _L , as shown in FIG. 12 (see FIG. 12), the contour difference is calculated, the threshold value is set, and the contact point is set. To detect.

At this time, as shown in FIGS. 13A, 13B, and 13C, the combination of contours for obtaining the contour difference is C _0U (211) −C _1U
(231), C _1L (231) -C _2U (241), ..., Cn _L -C _0L (212)
As a result, (n + 1) contact points can be detected. Here, the X-direction range for obtaining the contour difference is a range in which both contours exist in the combination of the contours described above.

According to the present embodiment, even when characters are in contact with each other at three or more points, it is possible to separate and cut out in units of one character.

In addition, the above embodiment shows an example of the present invention,
It goes without saying that the present invention should not be limited to this.

For example, in the above-described embodiment, the contact point can be detected even when the contour difference does not change, so that the Y coordinate at the minimum point and the maximum point of the X coordinate of the inner contour is set to infinity. It is also possible to force a change in the contour difference. This is because the minimum point and the maximum point of the X-coordinate of the inner contour are often contact points, and even if the contour difference does not change by such processing, the contact point is determined. It becomes detectable.

〔The invention's effect〕

As described above in detail, according to the present invention, even when characters are in contact with each other at a plurality of points, the contact points can be stably detected, and the characters can be separated into individual characters and cut out. This has the remarkable effect of realizing a simple character cutting device.

[Brief description of drawings]

FIG. 1 is a block diagram of an optical character reader according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of extracting the outline of a character pattern, and FIG. 3 is a flow chart for explaining the operation of a contact character determination unit. FIG. 4 is a flow chart for explaining the operation of the contact character separation unit, FIGS. 5 to 10 are diagrams for explaining the process until the two-point contact character is separated, and FIGS. 11 to 13 are many. It is explanatory drawing in the case of a point (3 points) contact character. 11: character pattern extraction unit, 12: contact character determination unit, 13: contact character separation unit, 14: recognition unit, 15: dictionary unit.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Toshihiro Hananoi 2880 Kokufu, Odawara City, Kanagawa Prefecture Odawara Plant, Hitachi, Ltd. (56) References JP 57-45676 (JP, A) JP 57 -5181 (JP, A) JP-A-59-160274 (JP, A) JP-A-61-59568 (JP, A)

Claims (1)

[Claims] 1. A character slicing device for slicing out, for each character, a character pattern optically read from a form or the like, and detecting the width (width in the X direction) of the character pattern to determine whether or not it is a contact character. First detecting means for detecting the presence or absence of inner contours and second detecting means for detecting the number of inner contours, and a plurality of inner contours generated when characters are in contact with each other at multiple points (a plurality of points) And a contour of the character pattern, the third detecting means for detecting the Y-direction partial width of the character pattern, and the second detecting means for the character pattern determined to be the contact character by the first detecting means. When the presence of the inner contour is detected by, the change point of the Y-direction partial width of the character pattern detected by the third detecting means is used as the contact point based on the position and the number of the inner contour, and the character unit is used. Character cut-out system Character extraction apparatus characterized by comprising means.