JPH07282191A - Table processing method - Google Patents

Abstract

PURPOSE:To make it possible to correctly recognize the frame and faithfully reproduce its image for a table whose corner is round. CONSTITUTION:A black run longer than a definite value on a binary image in a memory 2 is extracted by a black run extracting part 3, and ruled lines in a main and sub scanning directions are extracted by a ruled line rectangle extracting part 5 as a rectangle in which the connected black runs are integrated. A frame area extracting part 7 recognizes the frame of the table on the basis of the combination of the connected ruled lines and the ruled lines which are not connected but whose interval is below the definite value. The interval between the not connected ruled lines among the ruled lines to constitute the frame of the table is reproduced by a circular arc.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to processing a table in an image read by an image scanner or the like.

[0002]

2. Description of the Related Art When an image such as a document or a form is read and processed, an input image is often divided into a character area, a table area, a figure and other areas, and processing suitable for each area is performed. Also, when recognizing the characters written in the table on the form, the ruled line information is used as a clue to recognize the table frame,
Characters in each frame are recognized.

Conventionally, regarding the processing of a table, a black run having a length of a certain value or more is extracted from a binary image of a table area, a ruled line is recognized as an integrated rectangle of the black run, and the main scanning direction and the mutually connected main scanning direction are recognized. There is known a method of recognizing a frame of a table as an area surrounded by ruled lines in the sub-scanning direction (for example, Japanese Patent Laid-Open No. HEI 2-
264386, JP-A-3-172983, JP-A-3
-172984).

[0004]

The above-mentioned conventional method is premised on that all the ruled lines forming the table are made of straight lines and all the ruled lines are connected.

However, in the forms used in the business system, there are many cases in which the corners of the table are rounded for the purpose of relieving stiffness. In the case of such forms that include tables with rounded corners, black runs of sufficient length are not extracted at the corners of the table, so the ruled lines that make up the table should be separated at the corners of the table. To be extracted. In such cases,
According to the above-mentioned conventional method which is premised on that the ruled lines forming the table are connected, the frame area of the table cannot be reliably extracted.

Therefore, an object of the present invention is to provide a table processing method capable of accurately processing even a table having rounded corners, more specifically, a table frame even if the table has rounded corners. An object of the present invention is to provide a table processing method that enables accurate extraction of regions, accurate recognition of characters in a table, and more faithful reproduction of an image of the table.

[0007]

In order to achieve the above object, the invention of claim 1 extracts a black run having a length of a certain value or more from a binary image of an original and connects the extracted black runs. The process of extracting a rectangle that contains all the things as ruled lines is performed in the main scanning direction and the sub-scanning direction, and the connection relationship between the extracted ruled lines in the main scanning direction and the sub-scanning direction is checked. It is characterized in that ruled lines whose intervals are equal to or smaller than a certain value are regarded as being connected, and the frame of the table is recognized based on the connected ruled lines and the combination of the ruled lines which are considered to be connected.

According to a second aspect of the present invention, a black run having a length of a certain value or more is extracted from a binary image of a document, and a rectangle including all connected black runs is extracted as a ruled line. Perform in the main scanning direction and the sub scanning direction,
The connection relationship between the extracted ruled lines in the main-scanning direction and the sub-scanning direction is checked. At this time, the variation in the interval between the end points of the ruled lines that are not connected is checked, and if the variation is within a certain range, it is not connected. It is characterized in that the ruled lines are considered to be connected, and the frame of the table is recognized based on the connected ruled lines and the combination of the ruled lines considered to be connected.

According to a third aspect of the present invention, in the method of the first or second aspect of the present invention, when the image of the original is reproduced, if the ruled lines forming the frame of the table include unconnected ruled lines, It is characterized in that the corners of the frame of the table corresponding to the unconnected ruled lines are reproduced using arcs.

According to a fourth aspect of the present invention, in the method of the third aspect, as the arc used to reproduce the corner portion of the table frame, the main scanning direction ruled line and the sub-scanning ruled line corresponding to the corner portion are extended. It is characterized by using an arc having a radius equal to the distance between the intersection of the lines and one end of both ruled lines.

According to a fifth aspect of the invention, in the method of the first or second aspect of the invention, as a ruled line forming a frame of the table, a mode in which only the connected ruled line is used and a connected ruled line not connected but connected are connected. It is characterized in that the mode using the ruled line regarded as is switched.

According to a sixth aspect of the present invention, in the method of the third or fourth aspect, character recognition is performed on the inside of the recognized table frame, and the recognized character, the recognized ruled line of the table and the corner portion of the table are recognized. It is characterized in that the data of the circular arc indicating is output to a device such as a DTP device and the original is reproduced.

[0013]

[Function] In the case of a table with rounded corners, ruled lines are not connected at the rounded corners. The spacing between the endpoints of such a ruled line is
Depends on the curvature of the table corners. According to the invention of claim 1, the ruled lines separated by the roundness of the corners of the table are connected by the threshold processing of the interval between the end points of the ruled lines that are not connected. Recognized,
It becomes possible to perform processing such as recognition of characters in the table.

When all or some of the corners of the table are rounded, it is common to make the curvature of those corners uniform.
Therefore, when a specific table is processed, the interval between the end points of the ruled lines corresponding to the rounded corners is usually a substantially constant value. Focusing on the small variation in the spacing between the end points of the ruled lines separated by the rounded corners of the table,
According to the invention of claim 2, without performing threshold processing of the interval,
By considering that the ruled lines separated by the rounded corners are connected, it is possible to recognize the frame of the table with rounded corners and perform processing such as recognition of characters in the table. Moreover, in the method of performing the threshold processing of the interval, there is a limitation in the range of the curvature of the corner of the table that can be dealt with depending on the used threshold value. It can handle various tables with rounded corners and various curvatures.

According to the invention of claim 1 and the invention of claim 2, the ruled lines separated by the rounded corners of the table can be recognized, which means that the table being processed has rounded corners. It is also possible to distinguish whether it is a table that has not been rounded or a table with rounded corners, so it is possible to distinguish the type of input form etc. and prevent processing errors due to input of non-target forms. Will also be possible.

According to the invention of claim 3, the roundness of the corners of the table included in the document such as a form can be reproduced, and according to the invention of claim 4, the curvature of the corners can be faithfully reproduced. The reproducibility of the image of a document such as a form including a rounded table can be significantly improved.

According to the fifth aspect of the present invention, when processing a form or the like that includes only a table with no rounded corners, the corners are switched by switching to a mode in which the frame of the table is recognized using only the connected ruled lines. It is possible to eliminate the processing related to the roundness, prevent the occurrence of errors due to the processing, and improve the processing efficiency. On the other hand, if the form, etc. to be processed contains or may contain a table with rounded corners, by switching to another mode, the table with rounded corners Can also be processed accurately.

According to the invention of claim 6, it is possible to faithfully reproduce the ruled lines of the table including the rounded corners of the table and the characters in the table even for a form or the like including the table with rounded corners. .

[0019]

Embodiments of the present invention will be described below with reference to the drawings.

<First Embodiment> FIG. 1 is a block diagram showing an example of the arrangement of a device according to the present embodiment. In this device configuration, the processing steps will be described step by step.

Step 1: By the binary image input unit 1,
Binary image data of a document (a document including a table or a form) is stored in the binary image memory 2. Binary image input unit 1
Is a scanner for reading a document original, a means for reading binary image data from a storage medium,
Alternatively, 2 from a facsimile terminal or the like via a communication line.
There are various means for receiving and restoring the compressed data of the value image data. Alternatively, the binary image data of the document may be stored in the binary image memory 2 after being reduced in size from the original document size.

Step 2: In the black run extraction unit 3, 2
By extracting black runs in the main scanning direction and the sub-scanning direction from the binary image data in the value image memory 2 and performing threshold processing on the lengths of these black runs, only black runs having a length equal to or longer than a predetermined threshold value are extracted. Is extracted as a valid black run. Then, valid black run data (for example, a set of coordinates of the start point and the end point, or a set of coordinates and the length of the start point) is stored in the black run memory 4.

Step 3: In the ruled line rectangle extraction section 5,
With reference to the black run data in the black run memory 4, the black run connection is checked in each of the main scanning direction and the sub-scanning direction, and all connected black runs are integrated and included in a rectangular shape (main (Rectangle rectangles in the scanning direction and the sub-scanning direction) are extracted. Then, the data of the ruled line rectangle (for example, the coordinates of two diagonal vertices of the rectangle) is stored in the ruled line rectangle memory 6.

Step 4: In the frame area extracting section 7, the ruled line rectangle data in the ruled line rectangle memory 6 is referred to, and four ruled line rectangles having a connection relationship are arranged vertically (upper and lower left and right in the main scanning direction and left and right). Two sub-scanning ruled line rectangles) are found, a frame area having these ruled line rectangles as four sides is extracted, and the data is output.

When extracting the frame area, even if the noticed main scanning direction ruled line rectangle and the sub-scanning ruled line rectangle are not completely connected, if the interval between both ruled line rectangles is less than or equal to a predetermined threshold value, they are connected. It is regarded as doing. An example of a method of determining this interval will be described with reference to FIG. In FIG. 2, the main scanning direction ruled line (rectangle) BC and the sub scanning direction ruled line (rectangle) DE are not connected, but the intervals AB and AD between the intersection points A of the extension lines of both ruled lines and the respective ruled line end points are both threshold values. In the following cases, it is considered that both ruled lines (rectangles) are connected, and the frame area is extracted.

For example, in the case of a table with rounded corners, a ruled line relationship as shown in FIG. 2 appears, but by the processing described above,
The frame area of the table can be extracted without any trouble.

In the present embodiment, whether all the frame areas extracted in step 4 have four completely connected ruled line rectangles (the table is composed of only straight ruled lines), If not (the table with rounded corners is included), identify the type of input document (form),
The result is also output. By this identification processing, it becomes possible to prevent an error due to the inclusion of a non-target form.
However, this identification process may be performed as necessary, and is omitted when there is no need.

<Embodiment 2> FIG. 3 is a block diagram showing an example of the apparatus configuration according to the present embodiment. In this device configuration, the same reference numerals as those in FIG. 1 indicate corresponding parts. In FIG. 3, the frame area memory 8 and the document reproducing section 9 are parts not included in FIG. Hereinafter, the steps of the process will be described step by step.

Steps 1 to 4 are the same as steps 1 to 4 of the first embodiment. However, the frame area data extracted by the frame area extracting unit 7 in step 4 is stored in the frame area memory 8.

Step 5: The document reproducing unit 9 refers to the contents of the frame area memory 8 and the ruled line rectangle memory 6 and instructs a document preparation device such as a DTP device to express the connected ruled line rectangles as line segments. If there is a ruled line rectangle that does not have a connection relationship among the ruled line rectangles that form the upper, lower, left, and right sides of the frame region, all four corners of the frame region are rounded using arcs to represent the DTP device. Etc. By doing so, the reproducibility of the image of a form or the like including a table with rounded corners is improved.

When a more faithful reproduction of the table image is expected, it is determined whether or not the four corners of the table area are rounded, and the ruled lines forming the four sides of the frame area are determined according to the result. The method for reproducing the rectangle is instructed to the DTP device or the like. By doing this, it is possible to faithfully reproduce a table with only some corners rounded. This method will be described with reference to FIG.

First, the upper right corner of the frame area. That is, among the ruled-line rectangles that form the four sides of the frame area, FIG.
As shown on the left side of (A), when the ruled line rectangle 20 on the upper side and the ruled line rectangle 21 on the right side are not connected, the ends of these ruled line rectangles 20 and 21 are moved to the right side of FIG. 4 (A). As shown, it is reproduced by connecting it using the first quadrant of the arc. When the ruled line rectangles 20 and 21 are connected, they are reproduced as they are as line segments, but the same applies to other corners.

Among the ruled line rectangles forming the four sides of the frame area, as shown on the left side of FIG.
When 0 and the ruled line rectangle 22 on the left side are not connected, the ends of these ruled line rectangles 20 and 22 are connected using the second quadrant of the arc as shown on the right side of FIG. 4B. Reproduce in shape. Among the ruled line rectangles that form the four sides of the frame area, as shown on the left side of FIG.
2 and the ruled line rectangle 23 on the lower side are not connected, the ends of these ruled line rectangles 22 and 23 are connected using the third quadrant of the arc as shown on the right side of FIG. 4 (C). Reproduce in shape. Similarly, when the right side ruled line rectangle 21 and the lower side ruled line rectangle 23 are not connected as shown on the left side of FIG.
As shown on the right side of (D), it is reproduced in the form of connection using the fourth quadrant of the arc.

An example of a method of determining the radius and center of the arc connecting the ruled lines will be described with reference to FIG. FIG. 5 shows an example in which the ruled line rectangle BC on the upper side of the table area and the ruled line rectangle DE on the left side are connected by an arc BD. In the method described here, the intersection point of the extension line of the ruled line rectangle BC and the ruled line rectangle DE is A, and the distance A from the end point B or D of the ruled line rectangle BC or DE to the intersection point A.
Let B or AD be the radius of the arc BD. Then, a straight line parallel to the ruled line rectangle BC passing through the point D and a ruled line rectangle D passing through the point B
The intersection F with a straight line parallel to E is the center of the arc BD.

According to such a method, since the corner roundness can be reproduced with an arc having a curvature corresponding to the degree of roundness of the corners of the surface area, the curvature of the corner roundness differs from that of the method using the arc having a constant curvature. The image can be reproduced more faithfully in response to various tables.

<Embodiment 3> The apparatus configuration according to the present embodiment may be the same as the apparatus configuration of the first embodiment, and therefore the processing contents of the present embodiment will be described with reference to FIG.

Step 1: By the binary image input unit 1,
Binary image data of a document (a document including a table or an image of a form) is stored in the binary image memory 2.

Step 2: In the black run extraction unit 3, 2
By extracting black runs in the main scanning direction and the sub-scanning direction from the binary image data in the value image memory 2 and performing threshold processing on the length of these black runs, only black runs having a length equal to or longer than a predetermined threshold value are extracted. Is extracted as a valid black run. Then, the valid black run data is stored in the black run memory 4.

Step 3: In the ruled line rectangle extraction section 5,
With reference to the black run data in the black run memory 4, the black run connection is checked in each of the main scanning direction and the sub-scanning direction, and all connected black runs are integrated and included in a rectangular shape (main (Rectangle rectangles in the scanning direction and the sub-scanning direction) are extracted. Then, the ruled line rectangle data is stored in the ruled line rectangle memory 6.

Step 4: In the frame area extracting section 7, the ruled line rectangle data in the ruled line rectangle memory 6 is referred to, and four ruled line rectangles having a connection relationship are arranged vertically (upper and lower left and right). Two sub-scanning ruled line rectangles) are found, a frame area having these ruled line rectangles as four sides is extracted, and the data is output.

At the time of extracting the frame area, the processing for obtaining the interval between the main scanning direction ruled line rectangle and the sub-scanning ruled line rectangle which are not connected is performed for the four corners (three corners or two corners) of the table ( Interval thresholding as in the first embodiment is not performed). Then, if the obtained interval is a substantially constant value (a slight variation in the interval value is allowed), it is considered that the ruled line rectangles are connected and the frame area is extracted.

In the method of performing the threshold processing of the interval of the ruled line rectangle, one fixed threshold value can be applied only to the table in which the curvature of the roundness of the corner is within a certain range, so that it corresponds to various tables. It is necessary to prepare some threshold values in advance and then select and use an appropriate threshold value. On the other hand, in the present embodiment, since the threshold processing of the interval is not performed, it becomes possible to deal with various tables having rounded corners without performing a special operation.

<Embodiment 4> In the present embodiment, the first mode in which only a table consisting of only straight lines (a table without rounded corners) is processed, and a table with rounded corners is also processed. The second mode is different from the above-described embodiments in that the second mode can be switched. Therefore, the device configuration and the processing contents in the second mode are the same as those in the above-described embodiments. When the first mode is selected, unlike the above embodiments, the frame area extracting unit 7 extracts the frame area using only the connected ruled line rectangles as in the conventional case.

The advantage of enabling such mode switching is that by selecting the first mode when it is known that the target table corners are not rounded, the table corners can be rounded. It is possible to eliminate the related processing and improve the processing efficiency, and to avoid erroneous extraction of the frame area that may occur due to an error in the processing. <Embodiment 5> FIG. 6 shows an example of a device configuration according to the present embodiment. In this device configuration, the same reference numerals as those in FIG. 1 indicate corresponding parts. 6, the frame area memory 8 and the character recognition unit 10 are parts not included in FIG.

The details of the processing will be described below. The processing up to the extraction of the frame area by the frame area extraction unit 7 is the same as in the first embodiment. However, the extracted frame area data is stored in the frame area memory 8.

In the character recognition unit 10, the frame area memory 8
Also, referring to the contents of the binary image memory 2, character recognition is performed within the frame area of the input form image, and the character code obtained as a recognition result is output. This character recognition may be performed by the same processing method as general character recognition except that character cutting is performed only within the frame area.

According to the present embodiment, good character recognition is possible even for a table with rounded corners.

<Embodiment 6> FIG. 7 shows an example of the apparatus configuration according to the present embodiment. In this device configuration, the same reference numerals as those in FIG. 6 indicate corresponding parts. In FIG. 7, the recognition result memory 11 and the document reproducing unit 12 are parts not included in FIG.

Regarding the processing contents, the processing up to character recognition by the character recognition unit 10 is the same as that in the fifth embodiment. However, the character code obtained as a result of character recognition is stored in the recognition result memory 11.

In the document reproducing section 12, the ruled line rectangle memory 6, the frame area memory 8 and the recognition result memory 11 are referred to, and the ruled line data, the arc data and the character data are output to a document preparation device such as a DTP device, and the above-mentioned operation is carried out. As in the case of the document reproduction unit 9 (FIG. 3) of Example 2, the document creation device is instructed to directly express the ruled line rectangles having the connection relationship as a line segment, and forms the frame of the table although not having the connection relationship. The document creating apparatus is instructed to connect the expressed ruled lines with arcs to be expressed, and further, the document creating apparatus is instructed to display the character of the recognition result in the corresponding frame.

According to this embodiment, it is possible to faithfully reproduce a form including a table with rounded corners, including characters.

In the present embodiment, the same mode as in the fourth embodiment may be switchable, and the method of frame area extraction processing does not perform the interval threshold processing as in the third embodiment. It may be a method.

[0053]

As described above in detail, according to the present invention, the following effects can be obtained. According to the invention of claim 1 and the invention of claim 2, it becomes possible to accurately recognize a frame and perform processing such as recognition of characters in a table even with rounded corners. In particular, according to the second aspect of the invention, since the threshold value processing of the interval between the end points of the ruled line is not performed, it is possible to process various tables having rounded corners with various curvatures.

According to the invention of claim 1 and the invention of claim 2, whether the table being processed is a table with rounded corners,
Since it is possible to identify whether or not the table has rounded corners, it is possible to discriminate the type of the input form or the like and prevent the processing error due to the input of the non-target form or the like.

According to the invention of claim 3, the roundness of the corners of the table included in the document such as a form can be reproduced, and according to the invention of claim 4, the curvature of the corners can be faithfully reproduced. The reproducibility of the image of a document such as a form including a rounded table can be significantly improved.

According to the fifth aspect of the present invention, when processing a form or the like that includes only a table with no rounded corners, by switching to a mode in which the frame of the table is recognized by using only the connected ruled lines, Eliminates the processing related to rounded corners,
It is possible to prevent the occurrence of errors due to the processing and improve the processing efficiency. On the other hand, when the form etc. to be processed contains or may contain a table with rounded corners. By switching to another mode, you can perform accurate processing even for tables with rounded corners.

According to the invention of claim 6, it is possible to faithfully reproduce the ruled lines of the table including the rounded corners of the table and the characters in the table even for a form or the like including the table with rounded corners. .

[Brief description of drawings]

FIG. 1 is a block diagram illustrating an example of a device configuration according to first and third embodiments.

FIG. 2 is an explanatory diagram of an interval between end points of a ruled line.

FIG. 3 is a block diagram illustrating an example of a device configuration according to a second embodiment.

FIG. 4 is an explanatory diagram of reproduction by an arc in the upper right corner of the table (A). (B) It is explanatory drawing of the reproduction | regeneration by the circular arc of the upper left corner of a table. (C) It is explanatory drawing of the reproduction | regeneration by the circular arc of the lower left corner of a table. (D) It is explanatory drawing of the reproduction | regeneration by the circular arc of the lower right corner of a table.

FIG. 5 is an explanatory diagram of a method of determining a radius and a center of an arc representing a corner of a table.

FIG. 6 is a block diagram illustrating an example of a device configuration according to a fifth embodiment.

FIG. 7 is a block diagram illustrating an example of a device configuration according to a sixth embodiment.

[Explanation of symbols]

 1 Binary image input unit 2 Binary image memory 3 Black run extraction unit 4 Black run memory 5 Ruled line rectangle extraction unit 6 Ruled line rectangle memory 7 Frame area extraction unit 8 Frame area memory 9 Original reproduction unit 10 Character recognition unit 11 Recognition result memory 12 Document reproduction section

Claims (6)

[Claims] 1. A process of extracting a black run having a length of a certain value or more from a binary image of a document and extracting a rectangle including all connected black runs as ruled lines in the main scanning direction and the sub direction. Conducted in the scanning direction, and examined the connection relationship between the extracted ruled lines in the main-scanning direction and the sub-scanning direction.At this time, it is considered that the ruled lines that are not connected but the interval between the end points is a certain value or less are connected, A table processing method characterized in that a frame of a table is recognized based on a combination of connected ruled lines and ruled lines that are considered to be connected. 2. A process of extracting a black run having a length of a certain value or more from a binary image of a document and extracting a rectangle including all connected black runs as ruled lines in the main scanning direction and the sub direction. Performed in the scanning direction to check the connection relationship between the extracted ruled lines in the main scanning direction and the sub-scanning direction. At this time, check the variation in the spacing between the end points of the unconnected ruled lines, and if the variation is within a certain range. Is a method of processing a table, characterized in that it regards unconnected ruled lines as connected, and recognizes the frame of the table based on the combination of the connected ruled lines and the ruled lines that are considered connected. 3. When reproducing an image of an original, when the ruled lines forming the table frame include unconnected ruled lines, the corners of the table frame corresponding to the unconnected ruled lines are changed. The reproduction is performed by using an arc.
Or the table processing method described in 2. 4. A distance between an intersection of the main scanning direction ruled line and the extension line of the sub-scanning ruled line corresponding to the corner and an end point of both ruled lines as an arc used to reproduce the corner of the frame of the table. 4. The table processing method according to claim 3, wherein an arc having a radius equal to is used. 5. A mode in which only a connected ruled line is used as a ruled line forming a frame of a table and a mode in which a connected ruled line and a ruled line which is not connected but is considered to be connected are used are switched. The table processing method according to claim 1. 6. A document is recognized by performing character recognition inside the recognized table frame, and outputting the data of the recognized characters, the recognized ruled lines of the table and the arcs representing the corners of the table to a device such as a DTP device. The table processing method according to claim 3 or 4, wherein