JPH02148170A - Table structuralizing system - Google Patents

Abstract

PURPOSE:To easily execute a search of a line and a row, a search of a frame designated by a mouse, etc., and a table calculation, etc., by constituting the system so that a table of a non-fixed form type can also be managed in the same way as the case of a table of a fixed form type. CONSTITUTION:A character/graphic data processing system consists of a scanner 10 for reading character/graphic data containing a table, an image memory 11, a CPU core block 2, an OCR (opto-character reader) 13, and a document editing device 14. In this state, by plural virtual lines partitioned by a virtual line, the same frame number is given to that which originally belongs to the same frame, and existence information of the frame for showing whether the frame is a frame which exists actually, or a frame which is generated by only the virtual line and does not exist actually, and connecting information for showing a connecting state of each frame are allowed to have at every frame, and a table of a non-fixed form type is structuralized as table data of a fixed form type having a simple frame array. In such a way, in the same way as the case of the table of the fixed form type, a search in the line and the row directions in the table, a table calculation, etc., can be executed easily.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides ruled line data (starting points of ruled lines,
The present invention relates to a table structuring method that creates table structured data that defines the table structure based on a set of data (including end points and line types).

In the field of image processing, when the object to be recognized is a table, the table is extracted from the binary image data obtained by reading the table with a scanner, and the ruled line data (starting point of the ruled line,
(end point and line type), and then create table structured data that defines the table structure from the ruled line data, thereby making it possible to freely handle text and graphic data including tables. become. In the present invention, the table structuring method creates table structured data that defines the table structure from the ruled line data (starting point, end point, and line type of the ruled line) that constitutes the table, assuming that it has already been obtained. It is related to.

[Conventional technology]

FIG. 5 is a block diagram showing an example of the configuration of a character and graphic data processing system that handles character and graphic data including tables. In the figure, 10 is a scanner that reads text and graphic data including tables, 11 is an image memory, 2 is a CPU core block (includes the CPU, ROM, RAM, peripheral Ilo, etc., and processes and transfers data), 13 14 is an OCR (opto character reader), and 14 is a document editing device.

The scanner 10 reads text and graphic data including tables 2
It is converted into a value and taken into the image memory 11 as image data.

The CPU core block 12 retrieves image data from the image memory 11, cuts out a table, creates table structured data, recognizes characters and figures using the OCR 13, and sends the results to the document editing device 14. send.

[Problem to be solved by the invention]

It has been announced that the tables that can be handled conventionally in such character and graphic data processing systems are fixed-form tables.

FIG. 6 is an explanatory diagram showing a fixed form table. The table shown in the figure has frames partitioned in the shape of so-called base grids, and the number of frames is the same no matter which column is taken in the vertical direction, and the number of frames is the same in any row direction in the horizontal direction. There are five tables, and although the area of each frame may vary, it is a regular table with an orderly arrangement of frames.

On the other hand, FIG. 7 is an explanatory diagram showing a non-standard type table. -As you can see, even in the same horizontal direction, there are 5 horizontal frames in the bottom row, but in the top row,
There are only 4 frames.

In addition, there are four frames in the column for the first person, but three frames do not exist in the column for the first person.

FIG. 8 is also an explanatory diagram showing a non-standard type table.

It has been announced that the tables that can be handled conventionally in text and graphic data processing systems are fixed-form tables as shown in Figure 6, but an example of handling non-standard tables as shown in Figure 7 is now available. has not been announced (a non-standard table can be drawn on the display surface, but it cannot be treated as a table with meaning in the vertical and horizontal positional relationship, so searches in the row and column directions in the table, spreadsheet calculations, etc. I couldn't do it).

An object of the present invention is to provide a table structuring method that enables a character/graphic data processing system to handle a non-standard table as shown in FIG. 7 in the same way as a regular table. Our goal is to provide the following.

[Means to solve the problem]

In order to achieve the above object, the present invention assumes a virtual line in an amorphous table structure in which the number of frames differs between rows and columns, so that each row and column have different numbers of frames. Assuming a fixed table structure with the same number of frames, we assign the same frame number to multiple virtual frames separated by virtual lines that originally belong to the same frame, and also By including the step of registering frame presence/absence information indicating whether the frame exists in the frame or a frame that does not actually exist created only by virtual lines, and connection information indicating how the frames are connected to each other, the fixed form This makes it possible to create table-structured data using the same method as when creating a table structure.

FIG. 1 is an explanatory diagram in which what was originally an amorphous table as shown in FIG. 7 has been transformed into a regular table by imagining virtual lines as shown by broken lines. Similarly, FIG. 2 is an explanatory diagram in which what was originally a non-standard table as shown in FIG. 8 has been transformed into a regular table by imagining virtual lines as shown by broken lines.

As described above, the present invention assumes a fixed table structure with the same number of frames between rows and between columns, assuming virtual lines.

[Operation] The same frame number is given to multiple virtual frames separated by virtual lines that originally belong to the same frame, and along with that,
Each frame has frame presence/absence information indicating whether the frame actually exists or a frame that does not actually exist created only by virtual lines, and connection information indicating how the frames are connected to each other. In this way, a non-standard table can be structured as a standard table data with a simple frame arrangement, and you can easily search in the rows and columns of the table, perform spreadsheet calculations, etc., just as you would with a standard table. etc. will be possible.

〔Example〕

The present invention will be explained in detail below.

FIG. 3 is an explanatory diagram showing an example of a table structure as a target for which table structured data is to be created.

Assuming that the reference point in the same table is point A in the upper left corner, reference point A is at a position slightly shifted from the origin (0,0). It is recognized that the front is of an amorphous type, with frame numbers 1 to 16. (This is why we chose such a table to create table-structured data.)

Further, FIG. 4 shows a format for entering the created table structured data.

The formant shown in Figure 4(a) is a format for entering basic data, and its name is HNO(*).
Enter numbers 1 to 8 in place of (*). That is, HNO(1)
is the table number (there may be more than one table displayed on one display surface, so in that case,
number), HNO (2) is the slope of the table, HNO (3)
is the offset in the reference point AOX axis direction, HNO (4) is the offset in the reference point AOY axis direction, HNO (5) is the width of the entire back in the X axis direction, HNO (6) is the width of the entire back in the Y axis direction, HNO ( 7) is the number of frames in the X-axis direction, H
NO(8) is data representing the number of frames in the Y-axis direction, and is entered in each column 1 to 8 constituting the format.

The formant shown in FIG. 4(b) is a format for writing in the X-direction ruled line data, and its name is DOY (*,
*) In the place of (*, *), the first * is a number that identifies the left and right columns, and the next * is a number in the top and bottom directions. The line type of the ruled line is entered in the left column, and the width in the Y-axis direction is entered in the right column.

The line type of the ruled line is represented by a 5-bit configuration: a bit is set in the 0th digit to represent a solid line, a bit is set in the 1st digit to represent a broken line, a bit is set in the 2nd digit to represent a dotted line, and a bit is set in the 2nd digit to represent a dotted line.
A bit is set in the digit to represent a one-dot chain line, and a bit is set in the fourth digit to represent a two-dot chain line.

The formant shown in FIG. 4(c) is a format for entering the Y-direction ruled line data, and its name is DOX (*,
*), and the line type and width in the X-axis direction of the ruled line are entered in the same way.

The format shown in Figure 4(d) is a format for entering frame number data, and its name is DOB (2, *,
*), and the format shown in Figure 4(e) is
This is a format in which connected data for each frame is entered, and its name is DOB (1, *, *), where (*, *) are filled with numbers corresponding to each array.

Now that the format for entering table structured data has been explained, the procedure for actually creating table structured data will be explained next.

■ With the scanner lO shown in FIG. 5, a table figure as shown in FIG. 3 is read and taken into the memory 11 as image data,
The starting point, end point, and line type (ruled line data) of the ruled lines constituting the table are determined from the image data (this table cutting method is not a problem in the present invention, and is performed using an existing method).

■Calculate the slope of each line segment from the obtained ruled line data, ±4
After defining those within 5 degrees as horizontal lines and the others as vertical lines, it is checked whether the angle of each line segment is within a threshold value. For example, if the threshold value is ±5°, the horizontal line is within 0″±50±50, within 90″±5″,
Other line segments are determined to not be ruled lines composing a table and are removed.

Then, statistics on the slope of each line segment are taken, and the most frequent slope value is registered as data HNO(2) in column 2 of the format shown in FIG. 4(a).

■ Based on the slope obtained in above ■, correct the slope of the coordinates of the starting point and end point of the line segment using the following formula, and then intersect A between the vertical line at the minimum X coordinate and the horizontal line at the minimum Y coordinate. Find the coordinates to the intersection point as the X-axis offset and Y-axis offset in columns 3 and 4 of the format shown in Figure 4(a),
They are registered as data HNO(3) and HNO(4), respectively.

X' = Xsin θ+Y cos θY' = Xco
s θ−Ysin θ where X and Y are coordinates before correction, x' and y' are coordinates after correction, and θ is the slope.

■ Find the spacing between each line segment (ruled line) and DOX the vertical line segment.
(2,*”) is registered in the right column of the format shown in FIG. 4(C), and the horizontal line segment is registered as DOY (2°*) in the right column of the format shown in FIG. 4(b).

However, like line segment BC and line segment DB in Figure 3, Y''
Horizontal line segments with the same coordinate values or vertical line segments with the same X° coordinate values are grouped together as one line segment.

■ Subtract 1 from each of the horizontal lines and vertical lines found here, and use the remainder as the number of frames in the X direction and the number of frames in the Y direction, and write them in columns 7 and 8 of the formant shown in Figure 4 (a). They are registered as data HNO(7) and HNO(8), respectively.

■ Find the intersection points of all line segments from the corrected coordinates found in step (■) above, calculate the frames, and assign frame numbers starting from 1 in order. This is converted into the format shown in Figure 4(d) by DOB (2, *
, *'). Here, frame 9 seen in Figure 3
For a modified frame such as , a virtual line is assumed between the CDs to divide it into a plurality of frames, and each frame is given the same number. Furthermore,
Between the plurality of frames, connection information indicating that they are connected to form one frame is registered as a connection flag in the format shown in FIG. 4(e).

Here, the concatenation information has a 4-bit configuration, with a bit set in the O digit to indicate a right direction concatenation frame, a bit set in the 1st digit to indicate a left direction concatenation frame, and a bit set in the 2nd digit to indicate a downward concatenation frame. If there is a frame, set a bit in the third digit to indicate that there is an upward connecting frame.

In frame 9 in Fig. 3, a connection flag indicating that there is a downward connection frame is set for the upper frame partitioned by the imaginary line connecting CDs, and an upward connection frame is set for the lower frame. A concatenation flag is set to indicate the existence of the concatenation.

To summarize the above, the following data has been created as table structured data for the table shown in FIG.

Number of frames in X direction HNO (7) Number of frames in Y direction HNO (8) 01゜ Frame number　　　　04゜ DOB (2**) 08° 12° 14° =4 =5 02, 02. 0 05, 06. 0 09, 10. 1 09, 13. 1 15, 16. 1 Consolidation information DOB (1**) IH.

0° 4H.

8H.

0° 2H.

0° 0° IH.

IH.

H H Here, IH corresponds to hexadecimal 1 (0001 in binary) and means "with right direction connecting frame", and 2H corresponds to hexadecimal 1 (0001 in binary), and 2H corresponds to hexadecimal ), which corresponds to 2 (0010 in binary notation) and means "with left direction connecting frame". 4H corresponds to 0100 in binary system and means "with downward connecting frame", and 8H corresponds to 1 in binary system.
It corresponds to 000 and means "with upper connecting frame".

■Next, the image data within each frame is determined using the intersection coordinates and table tilt data and transferred to the 0CR13, and the frame number is added to the recognition result and transferred to the document editing device 14 along with the table structured data. .

When the table for which you want to create table-structured data is a table like the one shown in Figure 2, that is, when it contains a non-existent frame surrounded only by imaginary lines, the frame is a frame that actually exists. The presence/absence information of the frame is also registered, which indicates whether the frame is created only by virtual cotton or does not actually exist.

〔Effect of the invention〕

As explained above, according to the present invention, non-standard tables can be managed in the same way as regular tables.
It has the advantage of making it easier to search rows and columns, search frames specified with a mouse, etc., and perform spreadsheet calculations.

[Brief explanation of the drawing]

Figures 1 and 2 are explanatory diagrams in which what was originally a non-standard table has been transformed into a regular table by assuming virtual lines as shown by broken lines according to the present invention, and Figure 3 is a table structured table. An explanatory diagram showing an example of a table structure as a target for creating data. Figure 4 is an explanatory diagram showing a formant for filling in the created table structured data. Figure 5 is an illustration of characters and figures including tables. A block diagram showing an example of the configuration of a character and graphic data processing system that handles data, Figure 6 is an explanatory diagram showing a fixed form table, Figure 7
8 are explanatory diagrams each showing a non-standard type table. Description of symbols 10...Scanner for reading character and graphic data including tables, 11...Image memory, 12...CPU core block, 13...OCR (objector), 14...Document editing device.

Claims (1)

[Claims] 1) A table structure in which ruled line data (including the start point, end point, and line type of the ruled line) constituting a table figure is given, and then table structured data that defines the table structure is created. In the system, the table figure does not have a fixed table structure in which each row and column in the table have the same number of frames, but has a different number of frames in each row and column. When adopting a non-standard table structure, by assuming an imaginary line in the table, a fixed table structure with the same number of frames between rows and columns, and by imaginary lines The same frame number is given to multiple virtual frames that are partitioned and originally belong to the same frame, and at the same time, it is also necessary to indicate whether the frame actually exists or actually exists but was created only by virtual lines. By including the step of registering frame presence/absence information indicating whether the frame is a blank frame, and connection information indicating the degree of connection between frames, table structured data is created using the same method as when using a fixed form table structure. A table structuring method that is characterized by making it possible to