JPH05290166A - Segment recognition system - Google Patents

Abstract

PURPOSE:To extract more accurate segment data by recognizing the thickness of the segment. CONSTITUTION:A rectangle integration part 5 which extracts only black runs having length larger than a certain value and stores them in a black run memory 4 integrates black runs into a rectangle including all the black runs when the black runs are within a range having certain values determined individually in a main scanning direction and a subscanning direction and stores them in a rectangle memory 6. A segment extraction part 7 recognizes the segment by regarding rectangles as a new rectangle including all the rectangles when the rectangles are within a range having certain values determined respectively in the main scanning direction and subscanning direction and stores its in a segment memory 8. A segment thickness recognition part 9 outputs the thickness of the segment with the black runs and vertical length of the extracted segment rectangle.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a line segment recognition method, and more particularly to a line segment recognition method in document image processing, for example, line segment recognition processing in a binary image such as a document form.

[0002]

2. Description of the Related Art There is a strong demand for recognizing a line segment in a document image, inputting it as data, and reusing it. For example, when inputting a front manuscript output by a word processor, the characters inside the front
In addition to inputting with a CR (Optical character Reader), ruled lines forming a table may be input and reused at the same time. Therefore, in addition to reading the characters in the table, a method of recognizing a ruled line and inputting it as ruled line data is adopted. However, this method
Since the focus is on finding the coordinate values where ruled lines exist, it is not possible to find any information about the thickness of lines such as thick lines, medium lines, and thin lines. Therefore, for documents having different line thicknesses, these differences cannot be identified, and the demand for more faithful reproduction of the document cannot be met. Therefore, the present invention makes it possible to faithfully reproduce line segments in a document by obtaining the thickness of lines such as thick lines, middle lines, and thin lines, and incorporating these pieces of information.

[0003]

The present invention has been made in view of the above circumstances, and as a process of extracting a line segment in a document in a document image processing apparatus, the thickness of the line segment is recognized to obtain a more accurate line. The purpose of the present invention is to provide a line segment recognition method that enables extraction of minute data.

[0004]

In order to achieve the above object, the present invention provides (1)
In a line segment extraction method for extracting a line segment from a binary document image, a black run extracting means for extracting a black run having a predetermined value or more and the black run are separately defined in a main scanning direction or a sub scanning direction. If it is within the range of a certain value, the integrating means for integrating the black run into a rectangle that includes all of the black runs, and the rectangle are within a range of a certain value defined separately in the main scanning direction or the sub-scanning direction. If it is, the line segment extraction means for recognizing the line segment as a new rectangle including all the rectangles, and the line segment thickness for recognizing the thickness of the line segment by the length in the vertical direction of the line segment rectangle and the vertical direction. Or (2) in the line segment extraction method for extracting line segments from a binary document image, a black run extraction unit for extracting black runs of a certain value or more, and the black run extraction unit. Are divided into the main scanning direction and the sub-scanning direction. If it is within a predetermined range, the integration means for integrating the black run into a rectangle that includes all of the black runs and the rectangle within the predetermined range defined in the main scanning direction or the sub-scanning direction, respectively. If yes,
Line segment extraction means for recognizing a line segment as a new rectangle including all the rectangles, black pixel number counting means for counting the number of black pixels in the line segment rectangle, and the number of black pixels horizontally with the black run. A line segment thickness recognizing means for recognizing the thickness of the line segment by a value divided by the length in the direction, or (3) a line segment extraction method for extracting a line segment from a binary document image. In black run extraction means for extracting black runs equal to or greater than a certain value, and if the black runs are within a range defined by a separate value in the main scanning direction or the sub scanning direction, all the black runs are An integrating unit that integrates the included rectangle, a black pixel number counting unit that counts the number of black pixels when the black run is integrated into the rectangle, and the rectangle is within a fixed value determined separately in the main scanning direction or the sub-scanning direction. If it is within the range of, a new A line segment extracting means for recognizing a line segment as a shape, and a line segment thickness recognizing means for recognizing the thickness of the line segment by a value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction, or , (4) In a line segment extraction method for extracting line segments from a binary document image, a rectangle including a range in which black runs are connected is obtained, and the widths and heights of all the rectangles and their adjacent rectangles are obtained. Number counting means for counting the value of the distance, line segment extracting means for recognizing a line segment by integrating the respective rectangles when the respective cumulative values exceed a certain value, and the line segment rectangle A black pixel number counting means for counting the number of black pixels, and a line segment thickness recognizing means for recognizing the thickness of the line segment by a value obtained by dividing the black pixel number by a correction value of the black run and the length in the horizontal direction. Or (5) in the line segment extraction method for extracting line segments from a binary document image, A black pixel number counting means for counting the number of black pixels when a black run is integrated into a rectangle by obtaining a rectangle including a range in which runs are connected, and the widths and heights of all the rectangles and the distances between the adjacent rectangles. The number of black pixels is determined by the number of black pixels and the number of black pixels which is used for recognizing a line segment by integrating the rectangles when the cumulative value of each rectangle exceeds a certain value. And a line segment thickness recognizing means for recognizing the thickness of the line segment by a value divided by a correction value of the length in the horizontal direction. (6) In (1), the line segment rectangle A line type is determined by thresholding the black run and the length in the vertical direction, and (7) In (2), the value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction. To determine the line type by thresholding, and (8) in (3) above, A line type is determined by thresholding a value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction.
(9) In (4), the line type is determined by thresholding a value obtained by dividing the number of black pixels by a correction value of a black run and a horizontal length, and (10) above ( In 5), the line type is determined by thresholding a value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction. Hereinafter, description will be given based on examples of the present invention.

FIG. 1 is a block diagram for explaining an embodiment (embodiment 1) of a line segment recognition system according to the present invention.
Is a binary image input unit, 2 is a binary image memory, 3 is a black run extraction unit, 4 is a black run memory, 5 is a rectangular integration unit, 6 is a rectangular memory, 7 is a line segment extraction unit, and 8 is a line segment memory. , 9 are line segment thickness recognition units. A document or form is read using a binary image input device such as a scanner and stored in the binary image memory 2. From the read binary image, the black run extraction unit 3 extracts only black runs having a length of a certain value or more,
Stored in the black run memory 4 (thick solid line in FIG. 6). In the rectangle unifying unit 5, if the black runs are within a predetermined range defined in the main scanning direction or the sub-scanning direction with respect to the black runs extracted above, a rectangle including all the black runs. , And store it in the rectangular memory 6 (dotted rectangle in FIG. 6). In the line segment extraction unit 7, if the rectangles are within a predetermined value defined separately in the main scanning direction or the sub scanning direction with respect to the rectangles extracted above, a new rectangle including all the rectangles The line segment is recognized as and is stored in the line segment memory 8 (solid line rectangle in FIG. 6). The line segment thickness recognizing unit 9 recognizes the thickness of the line segment based on the length of the extracted line segment rectangle in the vertical direction with respect to the black run, and outputs the value as the line segment thickness.

FIG. 2 is a diagram showing another embodiment (embodiment 2) of the line segment recognition method according to the present invention. In FIG. 2, 10 is a black pixel number counting unit, 11 is a black pixel number memory, and the like. Portions having the same functions as those in FIG. 1 are designated by the same reference numerals. The process from reading a document using a binary image input device such as a scanner to extracting the line segment and storing it in the line segment memory is the same as in FIG. The number of black pixels is counted from the binary image memory 2 with reference to the extracted coordinate values on the line segment memory 8 and stored in the black pixel number memory 11 (the number of black pixels in the line segment rectangle in FIG. 7). Is counted). The line segment thickness recognizing unit 9 refers to the line segment memory 8 and the black pixel number memory 11 to calculate the line segment thickness. The line segment thickness is calculated as follows. The line segment thickness Thick is obtained from the black pixel number Pixel obtained above, the black run of the line segment rectangle, and the horizontal length Width. Thick = Pixel / Width Line thickness Thick value is output as the thickness of the line. When this line segment thickness Thick is used, for a document having a skew as shown in FIG. 7A, a line segment thickness close to the actual line segment thickness as shown in FIG. 7B is obtained. ..

FIG. 3 is a diagram showing yet another embodiment (third embodiment) of the line segment recognition method according to the present invention. From the read binary image, the black run extraction unit 3 extracts only black runs having a length of a certain value or more,
Stored in the black run memory 4. The black pixel number counting unit 10 accumulates the number of black pixels in the black pixel number memory 11 for black runs that can be integrated into a rectangle. The rectangular integration unit 5 integrates the black runs and stores them in the rectangular memory 6. In the line segment thickness recognition unit 9, the number of black pixels obtained above
The line thickness Thick is calculated from the Pixel, the black run of the line segment rectangle, and the horizontal length Width. Thick = Pixel / Width Line thickness Thick value is output as the thickness of the line.

FIG. 4 is a diagram showing still another embodiment (embodiment 4) of the line segment recognition method according to the present invention, in which 12 is a rectangular number counting unit, 13 is a histogram memory, and FIG. Portions having the same function are designated by the same reference numerals. A document or form is read using a binary image input device such as a scanner and stored in the binary image memory 2. The black run extraction unit 3 extracts a black run from the read binary image and stores it in the black run memory 4. In the rectangle integration unit 5, if there are black runs connected to the black runs extracted above, they are integrated into a rectangle including all the black runs and stored in the rectangle memory 6. The rectangle number counting unit 12 counts the values of the widths and heights of all the rectangles and the distances between the rectangles adjacent to each other, calculates the cumulative value of each, and stores them in the histogram memory 13. In the line segment extraction unit 7, when the histogram obtained as described above has gained a certain value, the rectangles are integrated to extract line segments, which are stored in the line segment memory 8. The number of black pixels is counted from the binary image memory 2 with reference to the extracted coordinate values on the line segment memory 8 and stored in the black pixel number memory 11. The line segment thickness recognizing unit 9 refers to the line segment memory 8 and the black pixel number memory 11 to calculate the line segment thickness. The line segment thickness is calculated as follows. The line segment thickness Thick is obtained from the black pixel number Pixel obtained above, the black run of the line segment rectangle, and the horizontal length Width. Thick = Pixel / (Width-CONST) (CO
(NST is a constant) This correction is for broken line and broken line segments. Line thickness Thick value is output as the line thickness.

FIG. 5 is a diagram showing still another embodiment (embodiment 5) of the line segment recognition method according to the present invention. The black run extraction unit 3 extracts a black run from the read binary image and stores it in the black run memory 4. The black pixel number counting unit 10 accumulates the number of black pixels in the black pixel number memory 11 for the black runs that can be integrated into a rectangle. The rectangular integration unit 5 integrates the black runs and stores them in the rectangular memory 6. In the line segment thickness recognition unit 9, the number of black pixels obtained above
The line thickness Thick is calculated from the Pixel, the black run of the line segment rectangle, and the horizontal length Width. Thick = Pixel / (Width-CONST) (CO
(NST is a constant) This correction is for broken line and broken line segments. Line thickness Thick value is output as the line thickness.

Next, a sixth embodiment will be described. The line segment memory 8 is used to extract line segments from the reading of the document using the binary image input device such as the scanner in the first embodiment.
It is the same until it is stored in. Line thickness recognition unit 9
In, the length of the extracted line segment rectangle in the vertical direction with respect to the black run is set to a thick line, a middle line,
It is classified into three types of thin lines and this line type is output.

Next, a seventh embodiment will be described. Example 2
The process from reading a document using a binary image input device such as a scanner to the process of extracting line segments and storing them in the line segment memory 8 is the same. The value of the line thickness Thick is classified into three types of line types, for example, a thick line, a medium line, and a thin line, by using a two-step threshold value, and this line type is output.

Next, an eighth embodiment will be described. The line segment memory 8 is used to extract the line segment from the reading of the document using the binary image input device such as the scanner in the third embodiment.
It is the same until it is stored in. The value of the line thickness Thick is classified into three types of line types, for example, a thick line, a medium line, and a thin line, by using a two-step threshold value, and this line type is output.

Next, a ninth embodiment will be described. The line segment extraction is performed from the reading of the document using the binary image input device such as the scanner in the fourth embodiment, and the line segment memory 8
It is the same until it is stored in. The value of the line thickness Thick is classified into three types of line types, for example, a thick line, a medium line, and a thin line, by using a two-step threshold value, and this line type is output.

Next, a tenth embodiment will be described. The process from reading a document using a binary image input device such as a scanner in the fifth embodiment to extracting the line segment and storing it in the line segment memory 8 is the same. Line thickness Thick
The value of is, for example, a two-step threshold value
It is classified into three types of thin lines and this line type is output.

[0015]

As is apparent from the above description, the present invention has the following effects. (1) The effect of the first embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment in the document. It becomes possible to extract minute data. As is clear from the above description, the present invention has the following effects. (2) The effect of the second embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment from the document. Therefore, it becomes possible to obtain a more accurate value. (3) The effect of the third embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment in the document. Therefore, a more accurate value can be obtained. Further, since the number of black pixels is counted when the runs are integrated, higher speed processing becomes possible. (4) The effect of the fourth embodiment is that it is possible to obtain the line segment thickness even for a broken line segment such as a dotted line or a broken line, whereby the type of line segment is allowed. The frequency will increase. (5) The effect of the fifth embodiment is that it is possible to obtain the thickness of a line segment even for a broken line segment such as a dotted line or a broken line, which allows the type of line segment to be permitted. The frequency will increase. Further, since the number of black pixels is counted when the runs are integrated, higher speed processing becomes possible. (6) The effect of the sixth embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment in the document. It becomes possible to express minute data. (7) The effect of the seventh embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment in the document. It becomes possible to express minute data. In particular,
It is possible to express more accurate line segment data for a document with skew. (8) The effect of the eighth embodiment is that the thickness of the line segment can be output as an absolute value on the actual image in the process of extracting the line segment in the document. It becomes possible to express minute data. In particular,
It is possible to express more accurate line segment data for a document with skew. Further, since the number of black pixels is counted when the runs are integrated, higher speed processing becomes possible. (9) The effect of the ninth embodiment is that the line segment thickness can be obtained even for a broken line segment such as a dotted line or a broken line, which allows the type of line segment to be permitted. The frequency will increase. (10) The effect of the tenth embodiment is that it is possible to obtain the thickness of a line segment even for a broken line segment such as a dotted line or a broken line. The frequency will increase. Further, since the number of black pixels is counted when the runs are integrated, higher speed processing is possible.

[Brief description of drawings]

FIG. 1 is a configuration diagram for explaining an embodiment of a line segment recognition method according to the present invention.

FIG. 2 is a diagram showing another embodiment of a line segment recognition method according to the present invention.

FIG. 3 is a diagram showing still another embodiment of the line segment recognition method according to the present invention.

FIG. 4 is a diagram showing still another embodiment of the line segment recognition method according to the present invention.

FIG. 5 is a diagram showing still another embodiment of the line segment recognition method according to the present invention.

6 is an explanatory diagram of line segment thickness in FIG. 1. FIG.

7 is an explanatory diagram of line segment thickness in FIG. 2. FIG.

[Explanation of symbols]

1 ... Binary image input section, 2 ... Binary image memory, 3 ... Black run extraction section, 4 ... Black run memory, 5 ... Rectangular integration section, 6 ...
Rectangular memory, 7 ... Line segment extraction unit, 8 ... Line segment memory, 9 ... Line segment thickness recognition unit.

Claims (5)

[Claims] 1. A line segment extraction method for extracting a line segment from a binary document image, and a black run extracting unit for extracting a black run having a predetermined value or more, and the black run having a main scanning direction or a sub-scanning direction. In the range within a predetermined value separately set, the integration means for integrating into a rectangle including all the black runs, and the rectangle within a predetermined value separately set in the main scanning direction or the sub-scanning direction , The line segment extraction means for recognizing the line segment as a new rectangle including all the rectangles, and the thickness of the line segment by the black run of the line segment rectangle and the length in the vertical direction. A line segment recognition method comprising a line segment thickness recognizing means for recognizing. 2. A line segment extraction method for extracting a line segment from a binary document image, and a black run extracting unit for extracting a black run having a predetermined value or more, and the black run having a main scanning direction or a sub-scanning direction. In the range within a predetermined value separately set, the integration means for integrating into a rectangle including all the black runs, and the rectangle within a predetermined value separately set in the main scanning direction or the sub-scanning direction Line segment extraction means for recognizing a line segment as a new rectangle including all the rectangles, and black pixel number counting means for counting the number of black pixels in the line segment rectangle. A line segment recognition method, comprising line segment thickness recognition means for recognizing the thickness of a line segment by a value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction. 3. A line segment extraction method for extracting line segments from a binary document image, and a black run extraction unit for extracting a black run having a predetermined value or more, and the black run having a main scanning direction or a sub-scanning direction. In the range within a fixed value determined separately for each, the integrating means for integrating the black run into a rectangle including all the black runs, and the number of black pixels for counting the number of black pixels when integrating the black run into the rectangle Counting means and line segment extracting means for recognizing a line segment as a new rectangle including all of the rectangles when the rectangle is within a range defined within the predetermined value in the main scanning direction or the sub-scanning direction And a line segment thickness recognizing means for recognizing the thickness of the line segment by a value obtained by dividing the number of black pixels by the black run and the length in the horizontal direction. 4. A line segment extraction method for extracting a line segment from a binary document image, wherein a rectangle including a range in which black runs are connected is obtained, and the widths and heights of all the rectangles and their adjacent rectangles are obtained. A rectangle number counting means for counting the value of the distance to the line segment, a line segment extraction means for recognizing a line segment by integrating the rectangles when the respective cumulative values exceed a certain value, and the line segment rectangle A black pixel number counting means for counting the number of black pixels in the line, and a line segment thickness recognizing means for recognizing the thickness of the line segment by a value obtained by dividing the black pixel number by a correction value of the black run and the length in the horizontal direction. A line segment recognition method comprising: and. 5. A line segment extraction method for extracting a line segment from a binary document image, wherein a rectangle including a range in which black runs are connected is obtained, and the number of black pixels is integrated when the black runs are integrated into the rectangle. Black pixel number counting means for counting, rectangular number counting means for counting the width and height of all rectangles and the distance values between adjacent rectangles, and when the cumulative value of each exceeds a certain value A line segment extracting means for recognizing a line segment by integrating each rectangle, and a line segment thickness for recognizing the thickness of the line segment by a value obtained by dividing the number of black pixels by a correction value of the black run and the length in the horizontal direction. A line segment recognition method characterized by comprising a height recognition means.