JPS63249279A - Information recognizing device - Google Patents

Abstract

PURPOSE:To recognize a circle graph and semicircle graph by approximating an arc part of an arc pattern of inputted image data with straight line groups and calculating coordinates of the center and the radius of the arc based on these straight line group information and discriminating whether a noticed pattern is a circle or an arc. CONSTITUTION:At least an arc pattern in image data inputted by an input means is approximated with plural straight line groups by a straight line approximating means (circle graph recognizing part 4), and the radius of the arc is calculated based on these straight line groups, and it is discriminated based on calculated information by a discriminating means whether the noticed pattern is a circle or an arc. When it is recognized as a circle, the circle graph is developed in a VRAM 7 based on these parameters and is displayed on a CRT 9. When it is judged that the noticed pattern is not a circle at least, a semicircle is decided in accordance with the number of circle elements.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an information recognition device, and more particularly to an information recognition device that recognizes graphic data.

[Prior Art] Conventionally, there have been character recognition devices that recognize regular characters from character symbol patterns in image data input using a document reading device. The effect of this character recognition device is that it not only converts handwritten manuscripts into regular fonts, but also encodes the recognized character symbol patterns (J■S code in many cases). It also means compressing data into a small amount of data before storing it on a storage device.

By the way, it is assumed that only character and symbol patterns exist in the document or image data to be recognized by this type of apparatus. Therefore, for example, character symbol patterns and graphic patterns (e.g. pie charts, etc.)
If both are mixed, there is a drawback that the graphic pattern cannot be recognized.

Furthermore, you can input a graphic pattern, such as a pie chart, and perform image processing (for example, scaling processing, etc.) on the pie chart.
, or when data processing is performed, in the past, the image data itself is processed without graphic recognition, which has the disadvantage that the amount of data after such processing ends up being enormous.

[Problems to be Solved by the Invention] The present invention has been made in view of the above-mentioned prior art, and it is an object of the present invention to provide an information recognition device that recognizes a pie chart in input image data.

[Means for solving the problem] In order to solve this problem, the present invention has the following configuration.

In other words, an input means for manually inputting a manuscript image as image data;
a linear approximation means for approximating at least an arc portion of an arc pattern in human-generated image data to a group of straight lines; a calculation means for calculating center coordinates and a radius of the arc based on information on the group of straight lines approximated by the approximation means; [Operation] In the configuration of the present invention, at least the circular arc pattern in the image data manually entered by the input means is converted into a straight line. The approximation means approximates a plurality of straight lines, the calculation means calculates the radius of an arc using the approximated straight lines, and based on this calculated information, it is determined whether the figure of interest is a circle or an arc. This is determined by a determining means.

[Examples] Examples according to the present invention will be described in detail below with reference to the accompanying drawings.

[Description of Device Configuration (FIG. 1)] FIG. 1 is a block diagram of the information recognition device of this embodiment.

In the figure, 1 is a document, and 2 is an image reading device that optically reads the document. 3 is a buffer memory for storing read image data.

Reference numeral 4 denotes a pie chart recognition unit that recognizes a pie chart in the image data developed in the buffer memory 3, and the details of its processing will be described later. Reference numeral 5 denotes a character recognition unit that recognizes character symbol patterns in the image data developed in the buffer memory 3 based on information in the character font ROMB and information obtained by the pie chart recognition unit 4. The character pattern recognized by the character recognition unit 5 is read out from the character font ROM 6 and written into the VRAM 7, which is a display RAM. Also, this VRAM7
The various patterns stored in are displayed on the CRT 9 by the graphic controller GC8.

Furthermore, 10 is a CPU that controls the entire device, and includes a ROM 10a that stores a control program (flow chart in FIG. 2) for controlling the entire device, and a RAM 10b that is used as a work area for the CPU 10.

[Explanation of Processing Contents (FIGS. 2 to 9)] The basic processing contents of the present apparatus having the above configuration will be explained with reference to the flowchart shown in FIG. 2.

First, in step S1, the image reading device 2 reads a document and develops the image data in the buffer memory 3. Next, in step S2, the input image data is thinned.

When this process is completed, a character pattern and a graphic pattern are discriminated in step S3. As a criterion for this determination, in this embodiment, when the length of the line segment (or vector) in the thinned image is short, it is determined that it is a character symbol.
On the other hand, if it is long, it is determined that it is a graphic pattern.

Now, if it is determined that it is a graphic pattern (pie chart), only that graphic pattern is extracted and transferred to a memory (not shown) in the pie chart recognition unit 4. The pie chart recognition unit 4 executes the processes from step 34 onwards.

In step S4, the thinned image is divided into point sequences, and linear approximation processing is applied to each of these point sequences to obtain linearized vector information. This situation is shown in Figures 3(a)-(
C), and the pattern shown in FIG. 3(C) is obtained by linear approximation.

[Explanation of Linear Approximation, Step 54] An example of the principle of this linear approximation process will be explained with reference to FIG.

Fixed point P on the circumference. Take a moving point P, which moves clockwise. And point P. When the tangent at is L, Po
L and P. The angle θ with Pl is the predetermined angle θ. P when it becomes. We decided to approximate P, (ie, PGF2) by a straight line. At the same time, during this process, PoPl and P
. The angle θ sandwiched by L is the predetermined angle θ. Before becoming P o
When the length of the line segment of P I exceeds the predetermined length,
We decided to linearly approximate P o P r at that point.

Hereinafter, similar processing is performed using the point P1 at which the linear approximation is made as the fixed point P0.

Note that a pie chart is composed of a circumferential portion that constitutes the pie chart, and a line segment C that connects a point on the circumference and the center of the circle. Fourth
In the figure, the rotation P1 passes through the end point P3 of this line segment 1 while moving to the point P2. Therefore, the moving destination of the moving point P is divided into two (branches). In this case, when the moving point P+ moves in the °C direction, it goes without saying that the angle θ0 described above will be satisfied when the moving point P moves to the point P4. Therefore, point P4
In order to prevent this from being selected, when the moving point PI moves and branches into several branches, the degree of change in the moving direction at these branch destinations (the degree of change is constant on the circumference) changes. If you do so, you will know that you should not proceed in that branching direction. However, the coordinate position of this branch point is stored in the RAM 10b. Now, after linearly approximating the circumference in this way, we will now proceed with linear approximation in one direction of the line segment from the coordinate position previously stored in RAM l0b, and when the linear approximation of the entire pie chart is completed, Figure 3 (
A pattern as shown in C) will be obtained. In the following explanation, the line segment approximated by a straight line on the circumference in Fig. 3 will be described as a circle element, and the line segment approximated by a straight line connecting the center position to a point on the circumference will be referred to as a straight line element. do.

[Description of calculation of center coordinates, step S5] Now, after the above-described linear approximation process is performed in step S4 of FIG. 2, the process moves to the next step S5, and the center coordinates of the pie chart are calculated.

The details of this calculation process will be explained below with reference to FIG.

In this embodiment, it is decided to calculate the coordinates Gl of the intersection of each circular element and the straight line element with the perpendicular bisector straight line of vector (1) when linear approximation is performed. Normally, the center position is calculated using the intersection coordinates of the perpendicular bisectors of only the vector of the circular element, but it cannot be determined at this point whether the vector of interest is a circular element or a straight line element. , the intersection coordinate G1 of the perpendicular bisector including the vector of the straight line element is calculated.

That is, assuming that the perpendicular bisector of each vector vI is 11, the intersection coordinates G1 of (jZ+, J22) are calculated, and below, (fLl, na) m (x l, x +), (
Il2, Il-3) ・・'(fi2.J2+)・
... Calculate the intersection coordinate Gl of the sum (i-1) of (IL+-+, u plate).

FIG. 6 shows an example of plotting the intersection coordinates G obtained by the above-described processing. As explained earlier, since the intersection coordinates are also calculated for the perpendicular bisector of a straight line element,
Although variations occur, the original center coordinates of the circle (xMAX
It can be seen that the prowato points are concentrated in +MAX). Therefore, it can be seen that the coordinate position that appears most frequently should be set as the center coordinate of the circle.

Although this is a method for calculating the center coordinates of a circle, it is not limited to this method. For example, when vectorization is performed, both end positions of the vector are at least on the circumference, so an equation is found to calculate the coordinates that are the same distance from both end positions of at least three of the two vectors. I don't mind if you do it the same way. In this case as well, it is desirable to perform calculations not in one calculation but in multiple combinations and to obtain statistics.

[Recognition of circle and straight line elements, step S6] Now, once the center coordinates of the circle have been calculated as described above, it is now time to identify circular elements and straight line elements.

In this embodiment, as shown in FIG. 7, the angle α between the perpendicular bisector ρ of each vector and the line connecting the center of the vector and the center coordinates of the circle. is less than a predetermined value, the vector is recognized as a circular element. In reality, this angle is “0゛°
However, since the pixels developed in the memory are located at discrete locations, they may not necessarily become "0". Therefore, the angle α. is greater than a given value,
Vector 1 will be recognized as a line element. The above processing is executed for all vectors to distinguish between circular elements and straight line elements. At this time, it is assumed that at least the starting point and ending point positions of the vector recognized as a straight line element, or the angle information from the center coordinates of the straight line vector are stored in the RAM 10b.

Now, once each element has been recognized, it's time to calculate the radius of the circle.

As shown in FIG. 8(a), the calculation method is to calculate the distance 11 r r between the obtained center coordinate and the end point or starting point of the vector of each circular element. Next, the obtained rl
As shown in FIG. 8(b), a frequency distribution is calculated with distance on the horizontal axis and frequency on the vertical axis. In this case, it can be seen that there are two "mountains".In other words, the pie chart is made up of two circles.Therefore, these "mountains"
It is sufficient to use the maximum point of the part as the radius.

[Explanation of identification of circles and semicircles, step S7] Through the above processing, the center coordinates and radius of the pie chart, as well as the coordinate positions of the straight lines that partition each data within the pie chart, are obtained. There are two types of graphs: pie charts and semicircle charts.

Therefore, if only the above-described processing is performed, the result shown in FIG. 9(a).

It is not possible to identify pie charts and semicircle charts with the same radius and the same central coordinates as shown in (b).

Therefore, it is now necessary to distinguish between these two graphs.

When the linear approximation process as described above is performed, the circumferential portion is composed of a plurality of groups of straight lines (circular elements). Figure 9 (C) shows the relationship between the radius of the circle and the number of circle elements, which are the circle drawing parameters, and the circle with radius r is n (=f(r), f is a function). It means that it is composed of circular elements. In other words, when the radius is r and the number of circle elements is n, the pattern of interest can be recognized as a circle. For example, radius 12
mm, and if the number of circle elements is 8, the pattern of interest can be recognized as a circle.

Therefore, it can be seen that determining whether it is a pie chart or a semicircle chart can be done by counting the number of circle elements having the radius at the time the radius is determined. In the example above, the radius is 12mm
When the number of circle elements is 8, it is determined that it is a circle,
When there are around 4 pieces, which is half of that, it is judged to be a semicircle.

Therefore, in step S7, it is determined from the above-mentioned relationship whether the pattern of interest is a circle. With this judgment,
In the case of "Yes" (if it is recognized as a circle), the process moves to step S8, and this time a pie chart is developed on the V RAM M7 based on those parameters, and Stella;s 14
Then it is displayed on the CRT9.

If it is determined in step S7 that it is not a circle, the process moves to step s9, and it is determined whether the number of circle elements is n/2. If “Yes”, the pattern of interest is determined to be a semicircle, and step S
Expand the semicircle on VRAMT with step S
14, it will be displayed on the CRT9. Furthermore, step s
If the determination in step 9 is "NO", the pattern of interest is not a semicircle (a fan shape other than a semicircle), so error processing is performed in step Sit.

On the other hand, for the image data determined to be characters in step S3, the character recognition unit 5 performs character symbol pattern cutting processing (step 512). Next, in step S13, normal character recognition processing is executed, and the normal pattern of the recognized character is read out from the character font ROM 6.
VRAMT corresponding to the position of the character in the input image
Expand to the top position.

As described above, according to this embodiment, it is possible to recognize pie charts in the same way as character symbols. Particularly with regard to pie charts, it becomes possible to distinguish circles and semicircles, and it becomes possible to identify and recognize pie charts and semicircle graphs.

In addition, when displaying recognition results, character symbols are drawn and displayed based on codes, and pie charts and semicircle graphs are drawn and displayed based on parameters expressing them.
The amount of data can be minimized. In particular, when the recognition results are stored in an external storage device, since they are not stored as image data, it is possible to reduce the amount of information to an extremely small amount.

In particular, for pie charts, the pie chart is displayed using the coordinate position or direction of the line segment representing the radius and the boundary of each data in the pie chart, so the display is clearer than when displaying the image data itself manually. becomes possible.

Furthermore, when displaying a pie chart, its display position, size, etc. can be easily changed.

Note that the relationship between the radius and the number of circular elements in Figure 9 (C) changes depending on the standard length of the straight line connecting the fixed point P0 and the moving point P when performing linear approximation. Therefore, the present invention is not limited to this.

[Effects of the Invention] As explained above, according to the present invention, it becomes possible to identify and recognize circles and arcs, and it becomes possible to recognize pie charts and semicircle graphs.

[Brief explanation of the drawing]

Figure 1 is a block configuration diagram of the information recognition device in this embodiment, Figure 2 is a flowchart showing the overall processing procedure in this device, and Figures 3 (a) to (C) are straight lines in the pie chart recognition unit. Figure 4 is a diagram to explain the progress of the approximation process, Figure 4 is a diagram to explain the process of linear approximation, Figure 5 is a diagram to explain the calculation of the center coordinates of a circle, and Figure 6 is a diagram to explain the calculated center coordinates of the circle. Figure 7 is a diagram showing the distribution, Figure 7 is a diagram explaining the process of recognizing circular elements and straight line elements that make up a pie chart, Figure 8 (a) is a diagram explaining the calculation of the radius of a circle, Figure 8 (b) ) is a diagram showing the frequency of calculated circle radii, Figures 9 (a) and (b) are diagrams showing the difference in the number of circle elements between circles and semicircles, and Figure 9 (C) is a diagram showing circles and circle elements. It is a figure showing the relationship with the number of. In the figure, 1... original, 2... image reading device, 3...
・Buffer memory, 4... Pie chart recognition unit, 5...
Character recognition unit, 6...Character font ROM, 7...u
RAM, 8...GC, 9...CRT, 10"-CP
U, 10a...ROM, 10b...RAM. (721 threads), - River Figure 1 Figure 3

Claims (4)

[Claims] (1) An input means for inputting a document image as image data, a straight line approximation means for approximating at least the arc portion of the circular arc pattern in the input image data to a straight line group, and a straight line group information approximated by the approximation means is based on An information recognition device comprising: calculation means for calculating the center coordinates and radius of a circular arc; and determination means for determining whether a figure of interest is a circle or an arc based on the calculated information. (2) The information recognition device according to claim 1, further comprising character recognition means for recognizing character symbols in image data. (3) The information recognition device according to claim 1, wherein the arc pattern is a circle or an arc. (4) The information recognition device according to claim 1, wherein the discriminating means discriminates based on the number of straight line groups.