JPH01129358A - Arithmetic unit for table numerical value - Google Patents

Abstract

PURPOSE:To automate an input and to prevent an input miss by recognizing the frame line of a primitive image as the gather of a rectangular frame, identifying the position relation between the rectangular frames and operating a numerical value in the rectangular frame based on this relation. CONSTITUTION:A user executes the rough sketch of the desired frame line on the paper of a notebook or a sheet paper, etc. This original is loaded on an image scanner 5 and an image input instruction is given from a keyboard 3 or a mouse 4 to a control part 1. According to a program stored in a program memory 9, a CPU circuit 8 sends an image scan instruction to the image scanner 5. The image scanner 5 reads picture information on the original from the sensor of a CCD, etc., and stores the image in an image memory 1. A table frame line is recognized from the primitive image to be digital-inputted and further, a cell number is added to frame data with using the information of the position relation. A character image in respective cells is recognized and the desired operation between cells is executed. Thus, the easy data input and operation processing of a high speed can be executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a surface value calculation device, and particularly to a surface value calculation device that recognizes numerical values within a frame line of an original image and performs desired calculations between the numerical values. It is something.

[Prior Art] Conventionally, when manually inputting a digital image to an image processing device such as a computer, a desired image portion is captured using a pen.

It was necessary to specify and input the information using a tablet, etc. For this reason, all numerical calculations in tables were performed by computer and data was entered manually. Also,
Even when inputting using a character recognition device, the correspondence between each input value and the table frame had to be manually established.

[Problems to be Solved by the Invention] In the conventional example described above, the user uses a mouse and a pen to indicate a desired image portion, which makes the operation complicated and takes a long input time. In addition, input errors occurred depending on the person, making it impossible to input accurate images. Furthermore, there is a drawback that a large amount of work time is required due to input errors.

The present invention eliminates the drawbacks of the conventional example, and provides a surface value calculation device that recognizes the table frame of an original image and performs desired calculations on the numerical values within the frame.

[Means for Solving the Problem] As a means for solving this problem, the surface value calculation device of the present invention includes a frame line recognition means that recognizes the frame line of the original image as a set of rectangular frames. , a positional relationship identifying means for identifying the positional relationship between the rectangular frames, and a calculation means for calculating a numerical value within the rectangular frame based on the positional relationship between the rectangular frames.

[Operation] In this configuration, the positional relationship of the rectangular frames recognized by the frame line recognition unit is identified by the positional relationship identifying unit, and the numerical value within the rectangular frame is calculated based on the positional relationship of the rectangular frames.

- Below is a margin 1 [Example] Hereinafter, a frame line configuring device and a cover value calculating device according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

The following embodiments include: (1) A system that recognizes the frame lines forming a table from a hail image as a collection of rectangular frames and finishes the entire frame line into a table consisting of straight lines; (2) Furthermore, a method for cutting out an image within a rectangular frame, and (3) a method for recognizing a matrix of a rectangular frame and calculating numerical values within the frame will be explained in order.

FIG. 2 is an external view of a graphic system including a frame line configuring device according to an embodiment, in which 1 is a control unit, 2 is a CRT, 3 is a keyboard, and 4 is a mouse.

These constitute a so-called workstation. Further, 5 is an image scanner that inputs images by manipulating documents, and 6 is an image printer that hard copies images processed at the workstation onto paper. In the wooden frame line composing device, the handwritten original image is obtained by reading what is written on manuscript paper etc. with an image scanner 5 or by using a mouse 4.
Alternatively, a handwritten image may be formed on the CRTZ. The point is that the handwritten image may be stored in some form in the image memory section 10, which will be described later.

FIG. 1 is a block diagram showing the hardware configuration of the frame line configuring device of this embodiment. Workstation control unit 1
, a CPU circuit 8 that controls the entire system, a program memory section 9 consisting of a RAM, etc., which stores a program for recognizing frame lines, and an image memory section 1 that stores digital images.
Consists of 0.

FIG. 3 is a flowchart showing the concept of the flow of frame line recognition processing. Following this flowchart, the details of each step will be explained in order.

<Image input>...Step S31 First, the user roughly sketches a desired frame line on a paper such as a notebook or sheet paper.

FIG. 4 shows an example of a sheet with handwritten frame lines. This document is placed on the image scanner 5, and an image input command is given to the control section 1 from the keyboard 3 or mouse 4. Then, the CPU circuit 8 operates the image scanner 5 according to the program stored in the program memory section 9.
Send an image scan command to. The image scanner 5 reads image information on a document from a sensor such as a CCD, and stores the image in an image memory section 10.

<Preprocessing> Step S32 The digital image input from the image scanner 5 has a matrix structure shown in FIG. 5 on the image memory section 10. The image consists of a maximum of m images in the horizontal direction of the document (hereinafter referred to as the X direction) and a maximum of n images in the vertical direction of the document (hereinafter referred to as the y direction). ,
Corresponding to black, each pixel has a value of °'o" and "t°', respectively.

In other words, the coordinate value of the pixel is x=1.2°...1 in the horizontal direction.
m, vertically y=1.2. ..., when n is the position (
The pixel value p (x, y) of pixel (x, y) is expressed by the following formula% A table frame drawn by hand is expressed as a set of pixels where P (x, y) = 1. , P (x
, y)=1, noise removal and line thinning are performed as preprocessing for tracking the pixel column where ,y)=1. Noise can be removed by, for example, well-known smoothing. Further, for thinning, for example, a well-known technique such as degenerating black pixels in consideration of order, degree of connectivity, etc. can be used.

When this noise removal and line thinning are completed, the handwritten table frame becomes a set of continuous points with a width of 1. For example, in the frame line for handwritten input in FIG. 4, a line near the frame line portion 61 and a white portion 60 around it are shown in FIG. In FIG. 6, pixels with logical value "0" are white parts, and images with logical value "1" are black parts.

Vectorization>...Step S33 Each pixel P (1, 1) of the digital image thus obtained
, =-, P (x, y), ..., P (m.

By tracing the black pixels with respect to n), vectorization is performed, and the obtained coordinates are written into the coordinate type vector data table 91 shown in FIG.

Enter the x and X coordinates of the starting point in the starting point x coordinate field 93° starting point y.
In the coordinate field 94, enter the x and X coordinates of the end point in the end point X coordinate field 95. Fill in the end point y-coordinate column 96. Rusk image, end points, fold points.

The method of converting a feature point such as a branch point of an intersection into vector data having a starting point and an ending point is a well-known technique, so
I won't mention it here.

As shown in FIG. 7, the vector data V is the starting point coordinate Px
(Xi, ys) and end point coordinates P.

It consists of (xa, 3'a). Subscript S is 5tar
The first letter of t and the subscript e are the first letters of end.

The vector number 92 in the coordinate type vector table 91 is
A serial number given to each vector one by one, V, ,V
2. ..., is given as VIll.

Here it is the total number of vectors.

The angle θ97 of the coordinate type vector table 91 is an angle indicating the inclination of each vector, and can be determined from the coordinate values of the starting point and ending point of the vector using the following equation. That is, X, -X.

π (Xs ” Xi, '/e> ys) π - (Xs = X13'e < Vg) error
(x,=x□y@=ys) The inclination θ97 of each vector thus obtained is written into the coordinate type vector table 91.

In FIG. 9, the unit of θ is degree.

When the original image in FIG. 4 is manually created, it is vectorized as shown in FIG. The coordinate values are appropriately determined. The numerical values in the coordinate type vector table of FIG. 9 are the example of FIG. 8, and the total number m of vectors is 13 (1 to V13).

Note that for subsequent processing, vertical vectors are written in the coordinate-type vector table 91 with the point with the smaller X coordinate as the starting point, and vectors other than the vertical vectors are written in the coordinate type vector table 91 with the point with the smaller X coordinate as the starting point. The vertical vector is determined by an algorithm shown in FIG. 20, which will be described later.

Generation of relationship tables>... Step S34 generates relationship tables, based on the coordinate type vector data 91 of the frame line image obtained in step 333, three data tables are created that are tables of relationships between vectors, points, and coordinates. , that is, the 10th
The point vector table 101 in Figure (a) and Figure 10 (
This process generates the point data table 105 in b) and the vector list table 110 in FIG. 10(c).

Point-type vector table 101 differs from coordinate-type vector table 91 in that point-type vector table 101 is different from coordinate-type vector table 91.
It is expressed as 3,104. In addition, the point data table 105 shows the point X of each point number.

The y coordinate values 107 and 108 are shown. The vector list table 110 stores all vectors connected to each point in a list structure.

In the example of FIG. 8 shown in FIGS. 10(a) to (C), there are nine point data P1 to P9, a vector list, to L21
There will be 26 pieces of 1.

First, according to the flowchart shown in FIG.
1 and the algorithm for determining the point data table 105 will be explained. In step 5110, the counter 1 is initialized to the total number n=o of 09 points, and in step 5111, the coordinates of the start point and end point of the vector with vector number Vl in the coordinate type vector table 91 are obtained. In step 5112, it is checked whether the starting point has been registered in the point data table 105. In the case of vector number v1, the starting point (xs, ys
) has not yet been registered in the point data table 105, the process advances to step 5113, where it is registered as a point with point number P1, and the total number n of points is incremented by 1. and step 5
115, this point number P1 is stored in the point type vector table 10.
1 to the starting point column 103.

Next, steps 8116 to 5 for the end point (xe, ya)
Similar processing is performed in step 119. The end point (Xs-ya) is registered as a point with number P2, and the total number of points n is increased by one. In step 5120, i is incremented by 1, and in step 5121, i is
>m, that is, whether all vectors have been processed. If not, return to step 5111 and repeat steps 5111 to 5121. In this way, the above processing is performed for each vector Vl (i=1 to 13).

If a point is already registered in the point data table 105, such as the starting point of vector V3, this point number P1 is read out in step 5114 or step 5118 and written in the starting point column 103 or ending point column 104 of the point type vector table 101. Just that is fine. Figure 10(a),(b)
) are written in the point type vector table 101 and the point data table 105 for the frame line example in FIG.

Next, an algorithm for obtaining the vector list table 110 from the point type vector table 101 and the point data table 105 will be explained according to the flowchart of FIG.

For all point data PJ (j=1 to n), n=9 in this example, and the following processing is performed. First, step 5130
The counter is initialized to j=1 in step 5131, and the point P is set in step 5131.
A vector having J as the starting point or ending point is found from the point type vector table 101. If there is a vector that satisfies the above conditions, in step 5132 the vector list number L
1 is registered in the vector list number column 111 of the vector list table 110, and the vector number is registered in the vector number column 112. In step 5133, it is checked whether there is the next connection vector, and if there is, in step 5134, it is entered in the NEXT column 113 of the vector list table 110.
Enter the next vector list number. If there is no other connection vector, φ is entered in step 5135.

In step 5136, j is incremented by 1, and in step 5137, it is checked whether j>n, that is, all points have been processed.
If j>n is not satisfied, the process returns to step 5131 and steps 8131 to 5137 are repeated. For example, in the case of FIG. 8, the vector connected to point P1 is V +
, Vl.

■It becomes 4.

Frame line extraction>...Step 335 Coordinate type vector table 91° in Fig. 9 Fig. 10 (a)
~(C) Point type vector table 101, point data table 105. An algorithm for extracting frame line data from the vector list table 110 will be explained.

Since a frame line is a set of single or plural rectangular frame data, a vector set forming each rectangular component is determined in the form of a rectangular frame vector table 130 in FIG. 13. The rectangular frame vector table 130 includes upper frame vectors 132 . left vector 133
．． Right vector 134. It consists of a lower vector 135.

Furthermore, in the case of a frame including a diagonal line, the lower right diagonal vector 136. An upper right diagonal vector 137 is added.

Using the flowcharts in FIGS. 16(a) and (ti),
An algorithm for obtaining the rectangular frame vector table 130 will be explained.

In step 5160, the counter is initialized to i=1. In step 5161, it is checked from the coordinate type vector table 91 whether the vector V1 (i=t to m) is a horizontal vector. The algorithm for horizontal vector determination is described in the 19th
This is shown in the flowchart in Figure. In step 5190, the angle θ
It is checked whether the absolute value 1θ1 of the height is smaller than a predetermined threshold value θ, and if it is smaller, the vector V1 is determined to be a horizontal vector in step S19,1. In addition, θ
, may be set to about 5° to 10°.

If Vl is a horizontal vector, step 3161a
Enter the vector number V1 in the upper vector column 132,
Next, in step 5162, it is checked whether there is a vertical vector connecting the starting point of 1 to the starting point. FIG. 17 shows a flowchart representing the algorithm for searching the vertical vector on the starting point side. In step 5175, vectors connected to the starting point are found from the vector list table 110. In step 5176, it is checked whether the vector is a vertical vector, and in step 5177, it is checked whether the point of interest is the starting point of the vertical vector. If the point of interest is a vertical vector starting from the point of interest, step 517
8. Let this vector be the perpendicular vector to be found. Check whether all vectors that share the point of interest have been checked in step 5179, and if not, step 5175
Return to and repeat steps 8175 to 5179.

Here, vertical vector determination will be explained according to the flowchart of FIG. 20. In step 5200, it is checked whether the absolute value 1θ11 of the angle θ1 of vector 1 is larger than a predetermined threshold value θ, and if it is, step 520
1 makes this vector VI a vertical vector. Furthermore, θ,
4 may be approximately 80° to 85°.

If there is a vertical vector on the starting point side, step 516
3, enter the vector number in the left vector column 133 of the rectangular frame vector table.

Next, in step 5164, the end point side is examined. 1st
FIG. 8 shows a flowchart representing the algorithm for searching the vertical vector on the end point side.

In step 5180, vectors connected to the end point are determined from the vector list table.

In step 5181, it is checked whether the vector is a vertical vector, and in step 3182, it is checked whether the point of interest is the starting point of the vertical vector. If the point of interest is a vertical vector starting from the point of interest, step 5183 is checked. Let this vector be the perpendicular vector to be found. Step 5184
Check whether all vectors that share the point of interest have been checked, and if not, return to step 3180 and repeat steps 5180 to 3184.

If there is a vertical vector on the end point side, step 516
5, the right vector column 134 of the rectangular frame vector table
Enter the vector number in .

Once the left vector and right vector have been determined, in step 5166, a vector whose end points are the end point of the left vector and the end point of the right vector is stored in the point vector table 10.
If the vector exists, the vector number is entered in the lower vector column 135 of the rectangular frame vector table 130 in step 5167.

Next, let's explore diagonal lines. In step 8168, the point-type vector table 105 is searched for a vector whose end points are the start point of the left vector and the end point of the right vector, and if it exists, the vector number is entered in the lower right diagonal vector field 136 in step 5169. Similarly, in step 5170, the point type vector table is searched for a vector whose end points are the end point of the left vector and the start point of the right vector, and if it exists, the vector number is entered in the upper right diagonal vector field 137 in step 5171. Fill out.

As described above, in step 5172, the rectangular frame number W1131 is registered as a rectangular frame for a frame in which at least the upper, lower, left, and right vectors exist.

Next, in step 5173, add 1 to i, and in step 5174
Check whether i>m, that is, all vectors have been processed, and if not, return to step 5161 and proceed to step 51.
Repeat steps 61-5174.

By repeating the above process, the rectangular frame vector table 13
0 is obtained. Note that if the vector ■ is not a horizontal vector in step 5161, or if the vector
If there are no left, right, or lower vectors in steps 5164 and 5166, the process advances to step 5172 and checks the next vector.

Next, since it is difficult to reconstruct the frame lines using the rectangular frame vector table 130 as is, each rectangular frame is reexpressed using coordinate data, size data, and type.

That is, as shown in FIG. 14, the X coordinate 14 of the upper left point
2. X coordinate 143. Frame width w 144° Frame height h 145,
A frame table 140 expressed by five parameters of fffl class kind 146 is obtained. The type of frame represents one of the four types shown in Fig. 15, each of which is K. , to,
, are distinguished by the value of 2°K3.

Value K. The value represents a frame 150 without a diagonal line, the value represents a frame 151 with only a left diagonal line, the value 2 represents a frame 152 with only an upper right diagonal line, and the value 3 represents a frame 153 with only a lower right diagonal line and an upper right diagonal line. shall be taken as a thing.

The coordinates x, y, width W, and height h are determined by the following formula.

x = (VIIP,)x, y= (Vl]Pi)
yw −(VIIPs) x−(VLIP−) Xh
== (VLEFTa)y - (VLEFT,)y However, (VUP,)X is the X coordinate of the starting point of the upper vector (VL
IPjy is the X coordinate of the starting point of the upper vector (VtlPe)
x is the X coordinate of the end point of the upper vector (VLEFTa) y
is the X coordinate of the end point of the left vector (VLEFT), and 1a is the X coordinate of the start point of the left vector.

The frame type kind may be determined based on the presence of the lower right diagonal vector and the upper right diagonal vector as shown in the flowchart shown in FIG. Step 5210 in FIG. 21 checks whether there is a lower right diagonal vector, and steps 5211 and 5212 check whether there is an upper right diagonal vector. From that combination, step 5213 sets the type to 3, step 52
In step 14, the type is set to 1, in step 5215, the type is set to 2, and in step 5216, the type is set to 0. In step 5217, it is checked whether all rectangular frames have been processed, and if not, the process returns to step 5210 and step 8
Repeat steps 210-5217.

Frame line configuration> If output to a printer or floppy disk is performed based on the frame table 140 obtained in step 33, the frame lines manually drawn by hand or the like can be reconstructed as straight lines.

In the embodiment described above, the frame line data was obtained by focusing on the horizontal vector, but the frame line data can be similarly obtained using an algorithm that focuses on the vertical vector by replacing the horizontal vector and the vertical vector in the algorithm.

As explained above, by selecting a rectangular vector group from the vector set obtained by vectorizing the frame line pixels in the input original image and using this as frame line data, it is possible to create a complex structure. It is now possible to recognize frame lines.

Next, FIGS. 23 and 24 show an example of cutting out an image within a frame line by applying the above-described frame line extraction method.

This will be explained using FIG. 25.

In FIG. 23, an auxiliary storage section 11 for exchanging data with the CPU circuit 8 is further added to FIG. 1. In the flowchart of FIG. 24, steps 531 to 535 are the same as in FIG. 3, and step 536a is added next to step S35 in FIG. 3 instead of step S36.

Image extraction> Step 36a Step 536
In step a, the cut out image is stored in the auxiliary storage section 11.

Image cutting based on the frame table 140 obtained by frame line extraction will be explained with reference to the flowchart of FIG. 25.

First, in step 5250, the upper left coordinate data X1, 3'+ and Wl of the gold frame of the frame table 140 are obtained.

Find the coordinates of the frame line from the height value. Step 52
At 51, a rectangular image data section surrounded by this frame line is read out to the CPU circuit 8 from among the digital image data stored in the image memory section 10.

Here, the image corresponding to the frame line including the diagonal line as described above may not be read out.

The rectangular image data section cut out in step 5252 is displayed on the CRT display 2 for confirmation by the inputter.
In step 5253, it is checked whether to cut out or not. When cutting out, in step 5254, this rectangular image data section is stored in the auxiliary storage section 11 (for example, a floppy disk).

(optical memory, etc.) and store it. In step 5255, it is checked whether all rectangular frames have been processed. If not, the process returns to step 5250 and steps 3250 to 5255 are repeated.

If necessary, image data can be displayed on the CRT display 2 or output to the printer section 6 by the CPU circuit 8 by giving instructions from the keyboard and mouse 3.4.

By cutting out the main image, the input original image is processed to generate a table frame line made of straight lines, and by extracting a rectangular area with this table frame line as a boundary, the image within the frame line can be processed without human intervention. can be cut out and input to the image processing section.

Next, a characteristic example in which the above-mentioned cutting out of an image within a rectangular frame is applied, and an example of calculation of numerical values within the frame will be described with reference to FIG. The tree table numerical calculation device has the configuration shown in FIG. 23, and is provided with a cell table 290 and a character code table 300, which will be described later, in the auxiliary storage section 11.

FIG. 26 is a flowchart showing the calculation processing procedure for numerical values within the frame in this example. Here, steps S31 to S35 are the same as in FIG. 3. Step 536b is added instead of step 336, and thereafter steps S37 to S39
It continues.

Cell numbering>...Step 536b Step S3
1 to S35, when the extraction process using the rectangular frame is finished,
Next, in step 536b, a cell number is added to each piece of frame data.

FIG. 27 shows the input image of this example. Figure 27 shows a table frame of 5 width x 3 height, and as shown in the figure, each cell has cell number C1,
I, C1,2,...CI,5, C2,+ +・
..., C3,5. Here, C1,l is a frame line including diagonal lines as described above. In addition, C, , X
represents the row number, and y represents the column number.

The algorithm for the cell numbering process in step 536b of FIG. 26 is shown in the flowchart of FIG. First, in step 8280, each rectangular frame data is sorted and rearranged in ascending order with respect to the X coordinate of the upper left point of the rectangular frame. The sorted rectangular frame data is y1≦y2≦...≦
The order is yllll. m is the total number of frame lines. Next, in step 3281, the control variable is set to =1. j=1. i=1
Initialize.

At step 5282, yl of each rectangular frame data.

y2. ..., calculate yl, 1-yI for y,
Compare with ε. ε is a preset threshold,
The number of pixels may be set to a value smaller than the width and height of the frame, which corresponds to 3 mm to 5 mm. For example, for a digital image with 8 pixels per mm, 5×8=40.

If y+++-3'+ is greater than the threshold,
It is assumed that the i+1th frame is the frame of the next row, and the process proceeds to step 3283, where the jth to ith frames are respectively set as Ck, l, Ck, 2++Ck, +1-J, etc.). The thus obtained frames in the same row are then sorted in ascending order with respect to the X coordinate in step 5284. Again, in the sorted order, Ck, I, Ck,
2,..., Ck, +1-J. 1. shall be.

In step 5285, the control variables j and i are incremented by 1, and the process returns to step 3282 to process the next row.

If y+++-yi is less than or equal to the threshold value in step 5282, i is incremented by 1 in step 5286, and step 828
At step 7, it is checked whether all rectangular frames have been processed, and if not, the process returns to step 3282 and the next rectangular frame is checked. Once all the rectangular frames have been processed, there is no frame to compare next, so step 3288.5289 is followed by step 3283.
It performs the same processing as 284 and returns.

<Arithmetic numerical value input>...A cell table 290 resulting from cell numbering in step S37 is shown in FIG. Frame number 2
91 cell table 29 showing correspondence with cell number 292
0 is obtained.

From this table, the frame number i corresponding to the cell numbers x and y can be obtained, and furthermore, the frame number i corresponding to the cell numbers
The coordinate values of the character image area within the cell can be calculated from 0, and the character image area can be cut out from the original image. Since the method of extracting and recognizing each character from a character image containing a plurality of character strings is a well-known technique, it will not be described here. The numerical character data obtained after character recognition is stored in cell 301 shown in Figure 30.
and a character code 302 are stored in a character code table 300 showing the correspondence between the character code 302 and the character code 302.

The user manually inputs calculation formulas between cells using input devices such as keys and a mouse. For example, a + b from the keyboard
+ c + d manually.

Calculation processing>...Step S38 In the case of the example shown in FIG. 27, from the character code table 300, 10+20+30+40" and
20” is calculated.

<Computation result output>...Step S39 result no" 10
0" and 5o" are displayed on the CRT display 2 or recorded on the printer section 6. Note that the output of the calculation result may be a blank space in the same row of the input table, or may be output by adding a frame.

According to this embodiment, table frame lines are recognized from a digitally input original image, cell numbers are added to the frame data using the positional relationship information, character images in each cell are recognized,
By performing desired inter-cell calculations, it becomes possible to perform simple and high-speed data input and calculation processing.

[Effects of the Invention] According to the present invention, it is possible to provide a front value calculation device that recognizes the front frame of an original image and performs desired calculations on the numerical values within the frame.

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram of the frame line configuration device of the embodiment, FIG. 2 is an external perspective view when the frame line configuration device of the embodiment is used as a graphic processing system, and FIG. 3 is a block diagram of the frame line configuration device of the embodiment. A flowchart showing an outline of the processing procedure; FIG. 4 is a diagram showing an example of a frame line image; FIG. 5 is a diagram showing a space in the image memory; FIG. 6 is a diagram showing image data stored in the image memory; Figure 7 is a diagram explaining the names of each part of the vector, Figure 8 is a diagram showing a vectorized frame line image, Figure 9 is a diagram showing a coordinate type vector table that stores coordinate type vector data, and Figure 10. (a) is a diagram showing a point type vector table that stores point type vectors, Figure 10 (b) is a diagram showing a point data table that stores point coordinates, and Figure 10 (C) is a diagram showing a point data table that stores point coordinates. Figure 11 is a flowchart showing an algorithm for creating a point-type vector table and point data table from a coordinate-type vector table, and Figure 12 is an algorithm for creating a vector list table. FIG. 13 is a diagram showing a rectangular frame vector table, FIG. 14 is a diagram showing a frame table, FIG. 15 is a diagram showing types of frames, and FIGS. 16 (a) and (b) are rectangular frames. A flowchart showing an algorithm for creating a vector table, Fig. 17 a flowchart showing a subroutine for tracking vertical vectors on the starting point side, Fig. 18 a flowchart showing a subroutine for tracking vertical vectors on an endpoint side, and Fig. 19 a flowchart showing a subroutine for tracking vertical vectors on the endpoint side. FIG. 20 is a flowchart showing the algorithm for determining vertical vectors. FIG. 21 is a flowchart showing the algorithm for determining the type of frame. FIG. 22 is an example of a frame line image with narrow frame intervals. 23 is a block diagram of a device for cutting out images within the frame and calculating numerical values in the table in the embodiment; FIG. 24 is a flowchart showing an outline of the processing procedure for cutting out the image within the frame in the embodiment; Fig. 25 is a flowchart showing the algorithm for cutting out the image within the frame. Fig. 26 is a flowchart showing an outline of the processing procedure for calculating numerical values in a table in the embodiment. Fig. 27 shows an example of a table in which calculated numerical values are entered. 28 is a flowchart showing an algorithm for creating a cell table, FIG. 29 is a diagram showing a cell table, and FIG. 30 is a diagram showing a character code table. In the figure, 1...Control unit, 2...CRT, 3...Keyboard, 4...Mouse, 5...Image scanner, 6...Image printer, 8...CPU circuit, 9
. . . program memory section, 10 . . . image memory section, 11 . . . auxiliary storage section. Figure 1 @ Figure 3y Figure 7 Figure 8 Figure 4 Figure 6 Figure 5 Figure 9 (C) Figure 10 Figure 17 Figure 18 Figure 19 Figure 20 Figure 22 Figure 23 Figure 24 Figure 26 Figure 27 Figure 29 Figure 30

Claims (5)

[Claims] (1) a frame line recognition means that recognizes the frame line of the original image as a set of rectangular frames; a positional relationship identification means that identifies the positional relationship between the rectangular frames; and based on the positional relationship between the rectangular frames, 1. A tabular numerical calculation device comprising: calculation means for calculating numerical values within a rectangular frame. (2) The frame line recognition means includes a vectorization means for converting the original image into a set of vector line segments whose starting points and end points are feature points such as end points, contact points, intersections, and branching points, and each Based on the connection relationship between vector lines,
2. The table numerical calculation device according to claim 1, further comprising rectangle extraction means for extracting a group of vectors forming a rectangle. (3) The rectangle extraction means includes a horizontal vector search means for searching for a horizontal vector from a vector line segment, and a vertical vector tracking means for tracing a vertical vector connected to the horizontal vector, and extracts a group of vectors forming a rectangle. A table numerical calculation device according to claim 2, characterized in that: (4) The rectangle extraction means includes a vertical vector search means for searching for a vertical vector from a vector line segment, and a horizontal vector tracking means for tracing a horizontal vector connected to the vertical vector, and extracts a group of vectors forming a rectangle. A table numerical calculation device according to claim 2, characterized in that: (5) The table numerical value set forth in claim 1, wherein the positional relationship identification means includes row identification means for identifying rows of the rectangular frame, and column identification means for identifying columns of the rectangular frame. Computing device.