JPH0715702B2 - Character pattern cutting device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a character pattern cutout device that separates and cuts out characters that have entered each other.

[Conventional technology]

FIG. 2 is a block diagram of a conventional device of this type disclosed in, for example, Japanese Patent Publication No. 57-51144. In the figure, 1 is a sheet, 2 is a scanning means for optically scanning and photoelectrically converting a character string written on the sheet 1, and 3 is a pattern of the photoelectrically converted character string (hereinafter referred to as "character string pattern"). ) Is stored in the character string pattern storage means 4, and the character string pattern is horizontally scanned from above to detect the coordinates of the left end and the right end of an area where white pixels are continuous (hereinafter referred to as "white area"). Then, the white area detecting means 5 stores the continuity of the vertically adjacent white areas based on the information of the white areas detected by the white area detecting means 4, and detects all the scans from the uppermost stage to the lowermost stage. Continuity check means for selecting and storing columns of white areas that are continuous with each other; 6 is a pattern of character strings stored in the character string pattern storage means based on information of columns of white areas selected by the continuity checking means. Character pattern for each character from A character separation means for clipping the output.

FIGS. 3A to 3C are views for explaining the operation sequence of the conventional device shown in FIG. In the figure, 7 is an example of a character string pattern “23”, 8 is a white pixel, 9 is a black pixel, 10 is an example of a white area, 11 is an example of a column of a selected white area, and 12 and 13 are examples of an area to be cut out. Is.

Next, the operation of the conventional device of this type shown in FIG. 2 will be described with reference to FIGS.

First, the character string on the sheet 1 is photoelectrically converted by the scanning means 2 and stored in the character string pattern storage means 3. Next, the character string pattern “23” 7 in the character string pattern storage means 3 is passed to the white area detection means 4.

The white area detecting means 4 scans the character string pattern "23" 7 in the left-right direction to detect the coordinates of the left edge and the right edge of the white area where white pixels are continuous. That is, the character string pattern "2
In the first scan for "3" and 7 ", the white areas whose left end and right end coordinates are x1 and x3, x6 and x8, respectively, are detected. Similarly, the white area in the third scanning, ..., The white area in the tenth scanning are sequentially detected, and the character string pattern “23” 7 to each of the scans as shown in the example of FIG. 3B. The white area 10 in is obtained and sent to the continuity checking means 5.

Next, the continuity checking means 5 selects white areas which are continuous with each other between all scans from the uppermost scan to the lowermost scan based on the information of the white areas detected by the white area detecting means 4. Then, it is transferred to the character separating means 6.

That is, the continuity checking means 5 is configured to determine the continuity of the white area depending on whether or not there is an overlapping area in the section indicated by the coordinates of the left end and the right end of the two white areas. Then, in the continuity checking means 5, first, based on the information of the white area 10 shown in FIG.
The continuity of the white area in the second and second scans, the second
Continuity of white areas in the 3rd and 3rd scans, ...,
Is sequentially checked to detect a continuous white area between adjacent scans, and then the white area in the first scan to the white area in the tenth scan are continuous with each other (FIG. 3C). The column 11 in the white area shown in the example of is selected.

Finally, in the character separation means 6, the patterns corresponding to the areas 12 and 13 separated by the white area 11 shown in the example of FIG. By reading from the pattern storage means 3, the character "2" and the character "3" are separated and cut out and output to an external device.

[Problems to be solved by the invention]

The conventional character extraction device examines the continuity in the vertical direction of the white area extracted from the character string pattern as described above, and the areas separated by the white area that is continuous between all the scans from the top row to the bottom row. Since it is configured to cut out as individual character areas, a character composed of a plurality of areas (hereinafter referred to as “black connected components”) in which black pixels are connected in eight connections (hereinafter referred to as “separated character”). There is a problem that when a character is cut out from a character string including "," each connected component is separated and cut out.

For example, FIGS. 4 (A) and 4 (B) are diagrams showing an example of a processing result of a conventional device for a character string containing a separation character “shi”, and 14 in the figure is an example of a character string pattern “nashita”, Reference numeral 15 is an example of a row of white areas which are continuous between all the scans selected from the character string pattern "Nashita" by the continuity check means 5, 16
.About.20 are examples of areas cut out from the character string pattern "Nashita" 14.

In the conventional device, the white string 15 continuous in the vertical direction shown in FIG. 4 (B) is detected from the character string pattern “Nashita” 14 shown in FIG. 4 (A), and five white regions 15 separated by this white region 15 are detected. The areas "na" 16, "ヽ" 17, "ヽ" 18, "no" 19, and "ta" 20 are respectively cut out as character areas.

In this way, with conventional devices, individual characters can be cut out correctly from character strings such as numbers without separation characters, but characters can be correctly cut out from general Japanese character strings consisting of kana, kanji, etc. that contain separation characters. There was a problem that it did not exist.

The present invention has been made to solve the above problems, and is an apparatus capable of correctly cutting out individual characters even from a character string in which separated characters such as "shi" and "word" are included and overlap each other. Aim to get.

[Means for solving problems]

Therefore, in the character slicing device according to the present invention, the white area obtained along the first scanning line is the separator area for the first scan, and the second to nth separator areas are the nth separator area. When a white area is adjacent to a plurality of (n-1) th separator areas, a separator area detecting means for determining the nth white area and the (n-1) th separator area as common areas is provided. It is characterized in that a character pattern is cut out based on.

[Action]

In the separator area detecting means according to the present invention, the white area obtained along the first scanning line is the separator area for the first scanning, and the second to nth separator areas are the nth white areas. When the area is adjacent to the plurality of (n-1) th separator areas, the nth white area and the (n-1) th separator area are determined as common areas. Then, the character pattern is cut out based on this separator area.

〔Example〕

 The details will be described below with reference to the drawings.

FIG. 1 is a diagram showing an embodiment of the present invention. 1 in the figure
6 is the same as the conventional device shown in FIG.
Reference numeral 21 in the drawing denotes a separator area detecting means for detecting and storing the separator area based on the information of the white area detected by the white area detecting means 4.

FIG. 5 shows the character string pattern “Nashita” shown in FIG. 4 (A).
It is a figure which shows the example which detected the white area | region from 14, and 22 in the figure is an example of the detected white area.

FIG. 6 shows the character string pattern “Nashita” shown in FIG. 4 (A).
It is a figure which shows the example which detected the separator area | region from 14, 23 is an example of the detected separator area | region.

FIG. 7 is an explanatory diagram for explaining an algorithm for detecting the separator area by the separator area detecting means. In the figure, 24 to 26 are examples of separator areas in the (n-1) th scan, 27 and 28 are examples of white areas obtained in the nth scan, and 29 to 31 are examples of separator areas in the nth scan. Is.

Next, the character string pattern "Nashita" shown in the example of FIG.
The operation up to the separator area detecting means of the embodiment shown in FIG. 1 will be described with reference to FIGS. 5 to 7 by taking the case of cutting out characters from 14 as an example.

First, the character string written on the sheet 1 is photoelectrically converted by the scanning means 2, and the resultant character string pattern "Nashita" 14
Is passed to the character string pattern storage means 3.

Next, the character string pattern "Nashita" 14 is passed to the white area detecting means 4. Further, the white area detecting means 4 scans the character string pattern "Nashita" in the left-right direction to detect the white area 22 shown in the example of FIG. 5, and the coordinates of the left end and the right end of the white area 22 are set to the separator. Transfer to the detection means 21.

The separator area detecting means 21 sequentially determines the separator area in each scan based on the information of the white area 22 detected by the white area detecting means 4. That is, first, the area indicated by the coordinates of the left edge and the right edge of the white area obtained by the first scan is stored as the separator area of the first scan. Next, the continuity between all the separator areas of the first scan obtained as described above and each white area obtained in the second scan is checked, and the continuity with the plurality of separator areas in the first scan is determined. When a white area of the second scan is detected, the white area is divided corresponding to the plurality of separator areas, and each area is determined as a separator area in the second scan. When the second white area which is not continuous with the plurality of separator areas in the first scan is detected, this white area is determined as one separator area. Similarly, the separator area in the third scan is sequentially determined to the separator area in the tenth scan, and the separator area 23 shown in the example of FIG. 6 is detected.

Next, the method of extracting the separator area will be described in more detail using an example. For example, FIG. 7 is a diagram showing an example of detecting the separator area in the nth scan from the information of the separator area in the (n-1) th scan and the white area obtained in the nth scan. . In FIG. 7, the white area 27 obtained in the nth scan is continuous with only one separator area 24 in the (n-1) th scan. Therefore, the white area 27 is determined as the separator area 29 for the nth scan, and the coordinates of its left end and right end are stored. The white area 28 in the nth scan is continuous with the two separator areas 25 and 26 in the (n-1) th scan. This is a phenomenon that occurs when a black connected component that is a part of a separated character exists in the area existing between the separator area 25 and the separator area 26. Therefore, the separator area detecting means 21 detects the white area 28 of the nth scan and the nth scan.
The white area 28 is divided by obtaining a common area between the two separator areas 25 and 26 of the -1st scan, and this is determined as the nth separator area 30 and the separator area 31.
Then, the coordinates of the left end and the right end of each separator area are stored.

By using the common area as the separator area only when the plural characters are adjacent to each other in this way, it has the effect of connecting the separated parts only to the separated characters, and may connect the characters that are trapped and overlap each other. No trouble occurs.
For example, since the white area of x15, x16 in the fourth scan of FIG. 5 is adjacent to the white area of the previous scan only one, it becomes the separator area as shown in FIG. It can be cut out correctly without being combined with "ta". In addition, even if the character string shown in Fig. 3 is included, although it is not a separation character, x7 in the 9th scan
Since the white area of ~ x12 is adjacent to the white area of the previous scan only one, it becomes the separator area as it is and "2"
It can be cut out correctly without being connected with "3".

As described above, the information of the separator area in each scan obtained by the separator area detecting means 21 of the embodiment shown in FIG.
Then, the data is transferred to the continuity checking means 5. Then, the continuity checking means 5 checks the continuity of the separator areas between the upper and lower scans as in the conventional apparatus, and selects the separator areas which are continuous between all the scans from the uppermost stage to the lowermost stage. For example, FIG. 8 is a diagram showing an example of the columns of the separator areas selected by the continuity checking means 5 from the separator areas 23 shown in the example of FIG. 6, and 32 and 33 in the figure show the selected separator areas. Row of, 34
Numerals to 36 are areas separated by the separator areas 32 and 33.

Finally, the character separating means 6 selects the pattern "na" corresponding to the areas 34 to 36 separated by the columns 32 and 33 of the separator areas of FIG. 8 selected by the continuity checking means 5.
The characters "shi" and "ta" are read from the character string pattern storage means 3 and output to an external device.

In the above embodiment, the case where the character is cut out from the character string of katakana has been described, but the present invention is not limited to this, and may be used when cutting out the character from a character string in which English characters, hiragana, kanji, and the like are mixed.

Further, in the above embodiment, the case of cutting out a character from a horizontally written character string has been described, but the present invention is not limited to this and may be used when cutting out a character from a vertically written character string.

〔The invention's effect〕

As described above, in the character slicing device according to the present invention, the white area obtained along the first scanning line is used as the first scanning separator area, and the second to nth separator areas are the second scanning area. When the n-th white area is adjacent to a plurality of the (n-1) th separator areas, a separator area detecting unit that determines the n-th white area and the (n-1) th separator area as a common area is provided. Since the character pattern is cut out based on this separator area, there is an effect that a device that can correctly cut out individual characters from a character string that contains a separated character such as "shi" and that overlaps and overlaps is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a conventional device, and FIG.
4 (A), (B), (C) and FIGS. 4 (A), (B) are diagrams for explaining an example of the operation of the conventional device, FIG. 5, and FIG.
8 and 9 are diagrams for explaining an example of the operation of the apparatus of the present invention, and FIG. 7 is a diagram for explaining an algorithm for detecting a separator area. In the figure, 1 is a sheet, 2 is a scanning means, 3 is a character string pattern storing means, 4 is a white area detecting means, 5 is a continuity checking means,
6 is a character separating means, and 21 is a separator area detecting means. Note that the same or corresponding parts in the drawings are denoted by the same reference numerals.

Claims (1)

[Claims] 1. A character pattern slicing device that optically scans a character string written on a sheet of paper or the like and divides a pixel into a white region and a black region along a scanning line to cut out a character pattern for each character. The white area obtained along the first scan line is used as the separator area for the first scan, and for the second to nth (n is an integer) scan lines, the white area is determined by the nth scan line. For each of the obtained white areas, the continuity with the (n-1) th separator area is examined, and when one nth white area is adjacent to a plurality of (n-1) th separator areas, the nth A region in which the one white region and the adjacent n-1th separator region are common is determined as a separator region, and the nth one white region is the nth region.
In the case where it is adjacent to only one -1st separator area, a separator area detection means for determining the nth one white area as a separator area is provided, and a character pattern is cut out based on this separator area. A device for cutting out character patterns.