JPH1097588A - Ruled-line recognizing method, table processing method, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately extract only dotted lines even when characters in similar shapes are arranged or when there are a plurality of dotted lines with narrow line spacing. SOLUTION: A document 2 is read as a binary value image, which is stored in a memory 2. A rectangle extraction part 3 generates rectangles containing all connecting black pixels by referring to the memory 2 and stores only a rectangle of size corresponding to a dotted line in a rectangle memory 4. A black pixel rate calculation part 5 counts the black pixels in the rectangle and calculates the number of black pixels to the rectangle area as a black pixel occupation rate. A dotted-line element decision part 6 decides whether or not a dotted line is constituted on the basis of the black pixel occupation rate of the rectangle and a dotted ruled line extraction part 8 integrates only adequate rectangles among rectangles decided to be dotted-line elements as a dotted line on the basis of the intervals between the dotted-line elements to extract the dotted ruled line.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of recognizing a character and a ruled line of a character image including a ruled line such as a table or a form, a table processing method, and a recording medium.

[0002]

2. Description of the Related Art In general, when a document is processed by a character recognition apparatus, a document image is often classified into a character area, a table area, a figure, and other areas, and processing according to each area is often performed. Among them, as a method of recognizing ruled lines constituting a table, in addition to a process of recognizing ruled lines composed of solid lines, a method of recognizing dotted lines is a method of recognizing dotted lines by integrating rectangles constituting black pixel connected components. Is known (for example, see Japanese Patent Application Laid-Open No. 7-230525).

[0003]

However, according to the above-described method, when a rectangle or a part of a character other than a dotted line and a rectangle having a similar shape is printed at regular intervals, the dotted line is erroneously printed. When there are a plurality of narrow dotted lines between rows, there are cases where the dotted lines are erroneously recognized in a direction orthogonal to the dotted lines.

[0004] The object of the present invention is to provide a method in which even if characters having similar shapes are lined up or a plurality of dotted lines with a narrow line spacing exist.
An object of the present invention is to provide a ruled line recognition method, a table processing method, and a recording medium that enable accurate extraction of only dotted lines.

[0005]

In order to achieve the above object, according to the first aspect of the present invention, a rectangle including all connected black pixels is extracted from a binary image, and the rectangle is integrated to form a dotted ruled line. , Wherein the validity as a dotted ruled line is determined based on the ratio of black pixels in the extracted rectangle.

According to a second aspect of the present invention, there is provided a rule recognition method for extracting a rectangle including all connected black pixels from a binary image, and integrating the rectangle to extract a dotted rule. Label the rectangle used to extract the dotted ruled line in the scanning direction or sub-scanning direction, and do not use the labeled rectangle when extracting the dotted ruled line in the sub-scanning direction or the main scanning direction. It is characterized by.

According to the third aspect of the present invention, a black run having a length equal to or longer than a predetermined threshold is extracted from the binary image, and when the extracted black runs are within a distance within a predetermined threshold, the black run is detected. The method is characterized in that a process of integrating them and extracting them as a solid ruled line and a process of extracting a dotted ruled line by the method of claim 1 or 2 are performed.

According to a fourth aspect of the present invention, the processing according to the third aspect is performed in the main scanning direction and the sub-scanning direction, and a ruled line in the main scanning direction and a ruled line in the sub-scanning direction are extracted. Are combined to recognize the frame.

According to a fifth aspect of the present invention, a character in the frame is extracted from the frame region extracted by the processing of the fourth aspect, and the character is recognized.

According to a sixth aspect of the present invention, the coordinate values and the ruled line types of the solid ruled line and the dotted lined rule extracted by the processing of the third aspect are output, and a document corresponding to the ruled line type is reproduced. And

According to a seventh aspect of the present invention, the character code recognized by the processing of the fifth aspect is output, and at the same time, a ruled line is output by the processing of the sixth aspect to reproduce the original. And

According to an eighth aspect of the present invention, there is provided a program for causing a computer to extract a rectangle including all connected black pixels from a binary image, and to integrate the rectangle to extract a dotted ruled line. A recording medium in which a program for realizing a function of determining validity as a dotted ruled line based on a ratio of black pixels in the extracted rectangle is recorded.

According to the ninth aspect of the present invention, there is provided a program for causing a computer to extract a rectangle including all connected black pixels from a binary image, and to integrate the rectangle to extract a dotted ruled line. In the recording medium on which recording is performed, a label is attached to a rectangle used for extracting a dotted ruled line in the main scanning direction or the sub-scanning direction of the image,
When a dotted ruled line in the sub-scanning direction or the main scanning direction is extracted, a program for realizing a function that does not use the labeled rectangle is recorded.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. <Embodiment 1> FIG. 1 shows the structure of Embodiment 1 of the present invention.
FIG. 2 is a processing flowchart of the first embodiment of the present invention. Using a binary image input unit 1 such as a scanner, an original such as a document or a form is read as a binary image and stored in a binary image memory 2 (step 101). Rectangle extraction unit 3
Scans the binary image memory 2, generates a rectangle that includes all the connected black pixels, and generates rectangular data of an appropriate size as an element forming a dotted line from the rectangle (the coordinates of the start point, end point, etc.). ) Is stored in the rectangular memory 4 (step 102).

The black pixel ratio calculating section 5 counts the number of black pixels in the rectangle read from the rectangular memory 4 and calculates the rectangular area (Ar
ea), the number of black pixels (Pixel1) is calculated as a black pixel occupation ratio (Black Ratio) (step 103). BlackRatio = Pixel / Ar
ea The dotted line element determination unit 6 determines whether or not the rectangle forms a dotted line based on the black pixel occupancy of the rectangle, and stores the rectangle determined to be a dotted line element in the dotted line element memory 7 (step 104). That is, in the case of a rectangle forming a dotted line as shown in FIG. 3B (rA to rD), it is considered that the image inside the rectangle has many black pixels. On the other hand, in the case where characters of the same shape (left parenthesis in this example) are arranged side by side as shown in FIG. Absent. Therefore, by utilizing this property, it is possible to accurately determine whether or not the rectangle forms the dotted line. Further, in determining whether or not the element constitutes the dotted line, whether or not the black pixel occupation ratio is, for example, 0.9 or more is used as a criterion for determination. It should be noted that, in addition to the above, an element that can be a dotted line element may be extracted based on the width of the rectangle.

The dotted ruled line extracting unit 8 reads out dotted line elements from the dotted line element memory 7 and integrates valid dotted lines based on the intervals between the dotted line elements to extract dotted line rules (step 105). In this process, as shown in FIGS. 15, 16, and 17 of the above-mentioned publication, by performing the integration process in multiple stages, the ruled line can be more accurately extracted.

Embodiment 2 FIG. 4 shows the configuration of Embodiment 2 of the present invention, and FIG. 5 is a processing flowchart of Embodiment 2 of the present invention. The binary image input unit 21, the binary image memory 22, the rectangle extraction unit 23, and the rectangle memory 24 are the same as those in the first embodiment, and thus description thereof will be omitted. The processing up to the rectangle extraction processing in step 202 is the same as that of the first embodiment.

First, the dotted ruled line extracting unit 25 refers to the rectangular memory 24, integrates rectangles in the main scanning direction, extracts dotted ruled lines, and stores them in the dotted ruled line memory 26 (step 203). The labeling unit 27 performs labeling on the rectangle forming the ruled line in the main scanning direction (step 204). FIG. 6 is a diagram illustrating labeling for a rectangle. h1 is a solid line in the main scanning direction, h2 to h4
Are the extracted dotted ruled lines in the main scanning direction.

As described above, the dotted ruled lines h2 to h4 in the main scanning direction are extracted, and the labeling unit 27 extracts the rectangles (rectangular rectangles enclosed by oblique lines in the figure) constituting these. Label.

Next, the dotted ruled line extraction unit 28 refers to the rectangular memory 24, integrates rectangles in the sub-scanning direction, and extracts dotted ruled lines (step 205). that time,
The rectangles that have been previously labeled by the labeling unit 27 are excluded from integration, and integration is performed based on the intervals between the other rectangles (vertical lattices in FIG. 6). By the processing of the second embodiment, as shown in FIG. 6, the generation of a pseudo ruled line such as R1 is suppressed, and only a target dotted ruled line in the sub-scanning direction such as v1 is extracted.

<Embodiment 3> FIG. 7 shows the configuration of Embodiment 3 of the present invention, and FIG. 8 is a processing flowchart of Embodiment 3 of the present invention.

An original such as a document or a form is read as a binary image by a binary image input unit 31 such as a scanner.
It is stored in the value image memory 32 (step 301).
The black run extraction unit 33 extracts a black run equal to or larger than a predetermined threshold value from the binary image memory 32 and stores the data (the coordinate values of the start point and the end point) in the black run memory 34 ( Step 302).

The solid line rule recognition unit 35 includes a black run memory 34
It is checked whether or not the black runs extracted within the threshold value are within a predetermined threshold value. All the black runs within the threshold value are integrated and extracted as a solid ruled line.
(Step 303).

Hereinafter, steps 304 to 307 will be described.
Then, a dotted ruled line is extracted in the same manner as in the first or second embodiment, and in step 308, the extracted solid ruled line and the dotted ruled line are output together.

<Embodiment 4> FIG. 9 shows the configuration of Embodiment 4 of the present invention. The solid ruled line recognizing unit 41 includes a binary image input unit, a binary image memory, a black run extracting unit, a black run memory, and a solid line ruled line recognizing unit according to the third embodiment. It is composed of a rectangle extraction unit, a rectangle memory, a dotted line extraction unit (main scanning direction), a dotted line memory, a label assigning unit, and a dotted line extraction unit (sub scanning direction). FIG. 10 is a processing flowchart according to the fourth embodiment of the present invention.

The processing from step 401 to step 407 (dotted ruled line extraction) is the same as in the third embodiment. These processes are performed in both the main scanning direction and the sub-scanning direction.
The frame recognizing unit 43 refers to the solid ruled line and the dotted ruled line in the main scanning direction and the sub-scanning direction, extracts a frame region surrounded by four sides, and stores the frame region in the frame region memory 44 (step 408).

<Embodiment 5> FIG. 11 shows Embodiment 5 of the present invention.
This is a configuration in which a character recognition unit 55 is added to the configuration of the fourth embodiment. FIG. 12 shows Embodiment 5 of the present invention.
It is a processing flowchart of.

The processing up to the frame recognition processing in step 508 is the same as in the fourth embodiment. The character recognition unit 55 includes
The character recognition area is determined by referring to the binary image memory corresponding to 4 and the frame area, character recognition is performed on this area, and the area is stored in the character recognition result memory 56 (step 509).

<Embodiment 6> FIG. 13 shows Embodiment 6 of the present invention.
In this example, the frame recognizing unit of the fourth embodiment is replaced with a document reproducing unit 63. The description of the same parts as in the fourth embodiment is omitted. FIG. 14 is a processing flowchart of the sixth embodiment of the present invention.

The processing up to step 607 for obtaining the solid ruled line and the dotted ruled line is the same as that of the third embodiment. This processing is performed in both the main scanning direction and the sub-scanning direction, and solid lines / dotted lines in both the main / sub-scanning directions are extracted.

The document reproducing section 63 reproduces the document while distinguishing the ruled line data from solid lines and dotted lines, and outputs the ruled line data to a DTP device or the like, for example. As a reproducing method, the extracted ruled line data (range of ruled lines) is vectorized, and information such as the thickness of the line segment is added.

<Embodiment 7> FIG. 15 shows Embodiment 7 of the present invention.
The block 71 has the configuration shown in FIG. 11, and is connected to the block 71 by a document reproducing section 72.
The result of character recognition and the ruled line data are input to the document reproducing section 72. FIG. 16 is a processing flowchart according to the seventh embodiment of the present invention.

The description of the same parts as in the fifth embodiment is omitted.
The processing up to the processing of step 709 for obtaining the solid ruled line, the dotted ruled line and the character recognition result is the same. The manuscript reproducing unit 72
The extracted ruled lines and character codes are used for reproducing the document. The reproduction of the ruled line is the same as that of the sixth embodiment.
The position is calculated from the absolute coordinates existing on the value image memory. Both characters and ruled lines on the manuscript are D
This is effective when inputting to a TP or the like.

Embodiment 8 FIG. 17 shows Embodiment 8 of the present invention.
Is shown. This embodiment is an embodiment realized by software, and prepares a general-purpose processing device including a CPU 81, a memory 82, a hard disk 83, an input device 84, a CD-ROM drive 85 and the like. CD-
A recording medium 86 such as a ROM stores a program for realizing the processing functions and processing procedures of the ruled line recognition method and the table processing method of the present invention. Document images such as documents and forms are stored in the hard disk 83 or the like, for example. The CPU 81 reads a program for realizing the above-described processing functions and procedures from the recording medium 86, executes the program one by one, and recognizes and outputs ruled lines and the like.

[0035]

As described above, according to the first and eighth aspects of the present invention, a document in which characters or a part of characters are printed side by side can be erroneously extracted as a dotted line. , The dotted line can be recognized.

According to the second and ninth aspects of the present invention, even for a document having a plurality of dotted lines in a table with a narrow line or the like, only the original dotted lines are accurately recognized without erroneous extraction in the orthogonal direction. be able to.

According to the third aspect of the present invention, the dotted line can be accurately recognized simultaneously with the conventional solid line recognition.

According to the fourth aspect of the present invention, in the conventional table processing method for performing frame recognition, processing can be performed on a table including a dotted line, and the number of originals to be processed increases.

According to the fifth aspect of the present invention, in a conventional table processing method for recognizing characters written in a frame after performing frame recognition, processing on a table including a dotted line becomes possible. The number of target document types is significantly increased.

According to the sixth aspect of the present invention, based on a document printed on paper, the format information consisting of ruled lines
When an input is made to P or the like, a dotted ruled line that could not be recognized conventionally can be recognized, so that the original can be reproduced more faithfully.

According to the seventh aspect of the present invention, when a document on which ruled lines such as a table and characters are written is input to a DTP or the like based on a document printed on paper, a dotted line which cannot be recognized conventionally can be used. It is possible to input a document that has been hindered by character recognition due to the presence of ruled lines or dotted lines, so that various documents can be reproduced more faithfully.

[Brief description of the drawings]

FIG. 1 shows a configuration of a first exemplary embodiment of the present invention.

FIG. 2 shows a processing flowchart of Embodiment 1 of the present invention.

FIG. 3 is a diagram illustrating a black pixel occupancy.

FIG. 4 shows a configuration of a second exemplary embodiment of the present invention.

FIG. 5 shows a processing flowchart of Embodiment 2 of the present invention.

FIG. 6 is a diagram illustrating labeling for a rectangle.

FIG. 7 shows a configuration of a third embodiment of the present invention.

FIG. 8 shows a processing flowchart according to a third embodiment of the present invention.

FIG. 9 shows a configuration of a fourth embodiment of the present invention.

FIG. 10 shows a processing flowchart according to a fourth embodiment of the present invention.

FIG. 11 shows a configuration of a fifth embodiment of the present invention.

FIG. 12 shows a processing flowchart of Embodiment 5 of the present invention.

FIG. 13 shows a configuration of Embodiment 6 of the present invention.

FIG. 14 shows a processing flowchart of Embodiment 6 of the present invention.

FIG. 15 shows a configuration of Example 7 of the present invention.

FIG. 16 is a flowchart illustrating a process according to a seventh embodiment of the present invention.

FIG. 17 shows a configuration of Example 8 of the present invention.

[Explanation of symbols]

 Reference Signs List 1 binary image input unit 2 binary image memory 3 rectangle extraction unit 4 rectangle memory 5 black pixel ratio calculation unit 6 dotted line element determination unit 7 dotted line element memory 8 dotted line ruled line extraction unit

Claims (9)

[Claims] 1. A rule recognition method for extracting a rectangle including all connected black pixels from a binary image, and extracting a dotted rule by integrating the rectangles. A ruled line recognition method characterized by determining the validity of a dotted ruled line based on a ratio. 2. A rule recognition method for extracting a rectangle including all connected black pixels from a binary image, and extracting a dotted rule by integrating the rectangle, wherein a main scanning direction or a sub-scanning direction of the image is provided. Labeling a rectangle used for extracting the dotted ruled line, and extracting the dotted ruled line in the sub-scanning direction or the main scanning direction without using the labeled rectangle. Weaving method. 3. A black run having a length equal to or greater than a predetermined threshold is extracted from the binary image. When the extracted black runs are within a distance within a predetermined threshold, the black runs are integrated to form a solid line. 3. A ruled line recognizing method characterized by performing a process of extracting a ruled line and a process of extracting a dotted ruled line by the method according to claim 1 or 2. 4. The processing according to claim 3, which is performed in the main scanning direction and the sub-scanning direction, extracts ruled lines in the main-scanning direction and ruled lines in the sub-scanning direction, and recognizes a frame by combining the extracted ruled lines. A table processing method. 5. A table processing method comprising: extracting characters in a frame from a frame region extracted by the processing according to claim 4; and recognizing the characters. 6. A table processing method, comprising outputting coordinate values and ruled line types of a solid line ruled line and a dotted line ruled line extracted by the processing according to claim 3, and reproducing a document corresponding to the ruled line type. 7. A table processing method comprising: outputting a character code recognized by the processing according to claim 5; and simultaneously outputting a ruled line by processing according to claim 6, to reproduce a document. 8. A recording medium for recording a program for causing a computer to realize a function of extracting a rectangle including all connected black pixels from a binary image, and extracting a dotted ruled line by integrating the rectangle. A recording medium for recording a program for realizing a function of determining validity as a dotted rule line based on the ratio of black pixels in the extracted rectangle. 9. A recording medium storing a program for realizing a function of extracting a rectangle including all connected black pixels from a binary image, and extracting a dotted rule by integrating the rectangle. A label is attached to the rectangle used to extract the dotted line in the main scanning direction or the sub-scanning direction of the image, and the label is attached when extracting the dotted line in the sub-scanning direction or the main scanning direction. A recording medium on which a program for realizing a function not using a rectangle is recorded.