JPS616778A - Recognizing method of character - Google Patents

Abstract

PURPOSE:To recognize simply a character with the small capacity of a memory by setting up plural testing points on the edge parts of a character pattern and a reference character pattern and comparing the changed values of two-dimensional plane coordinate positions between these testing points. CONSTITUTION:In the two-dimensional plane of a character to be recognized, twelve testing areas 21-32 e.g. are set up and the two-dimensional coordinate positions of the edge parts are detected from vertical and horizontal directions. Plural testing points of the edge parts of the character to be recognized and that of the reference character pattern are formed and the changed values of the two-dimensional plane coordinate positions between the testing points are compared with each other to recognize the character to be recognized.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a character recognition method for recognizing printed characters, marks, etc.

[Prior art]

Conventionally, as a method for recognizing printed characters, marks, etc., first, a reference character pattern, which serves as a reference for characters to be recognized, is stored in a first image memory 1 as shown in FIG.
Thereafter, the characters to be recognized are read by an imaging device or the like and stored in the second image memory 2 as shown in FIG.
, z ato and columns 1 to L, 1° are sequentially designated by the address signals, the reference character pattern information and the character pattern information to be recognized are read out, the matching relationship between the pattern information of the same address is determined, and the matching relationship is further determined. A widely used method is to determine whether a character to be recognized corresponds to a reference character by determining whether the number of addresses is equal to or greater than a predetermined number.

However, in this conventional method, the storage addresses of the reference character pattern information and the character pattern information to be recognized must almost match, and there is no flexibility in how to allocate the addresses of the image memory 1.2. This has the disadvantage that it cannot be used efficiently and as a result, a large amount of image memory is required.

(Summary of the Invention) The present invention has been made in order to solve the above-mentioned drawbacks, and its purpose is to efficiently utilize image memory and to recognize characters to be recognized with a small memory capacity. The purpose is to provide a recognition method.

To this end, the present invention sets a plurality of inspection points at the edge portions of the character pattern to be recognized and the reference character pattern,
The character to be recognized is recognized by comparing the increase/decrease values of the two-dimensional plane coordinate positions between the inspection points.

[Embodiments of the invention]

Hereinafter, the present invention will be described in detail based on examples.

2 and 3 are diagrams showing an example of setting an inspection area used for character recognition according to the present invention, and FIG. 3 is an enlarged view of the inspection area 21 in FIG. 2.

In these figures, a total of 12 inspection areas 21 to 32 are set on the two-dimensional plane of the character "3" to be recognized, and the two-dimensional coordinate positions of the edge portions are detected from the top, bottom, and left and right directions.

Specifically, to describe the inspection area 21 in the row direction as a representative, there are a plurality of inspection points 50 to 53 as shown in FIG.
is set, and the two-dimensional coordinate position 5 of these inspection points 50 to 53 is set.
0 (19,35), 51 (8,28), 52 (7,
21), 53 (23°14) are detected.

Specifically, this coordinate position is detected by addresses in the row and column directions of the image memory.

As a result, for the inspection area 21, 50 (19, 35
), 51 (8,28), 52 (7°2'l), 5
3 (23, 14) coordinate position data groups are obtained.

Next, based on the coordinate position data group obtained in this manner, the increase/decrease value of the coordinate positions between the inspection points is determined. For example, consecutive inspection points 50 to 5 in the inspection area 21 in the row direction
For 3, the increase/decrease value of the coordinate position in the row direction is 1+, Iz+1
3 can be found as follows.

1, -19-8=11 ■2-8~7=1 1s=1 23=16 In the same way, inspection areas 22-29 are inspected in the row direction.
Also, for the inspection areas 24 to 32 where inspection is performed in the column direction, increases and decreases in the coordinate positions in the row and column directions between successive inspection points are determined.

In this case, if the increase/decrease value is less than a certain value, for example "2" or less, the edge portion is considered to be changing smoothly and is discarded.

Therefore, for the inspection area 21, the increase/decrease value in the row direction is I+=11. I3-16 will finally be obtained.

Next, the increase/decrease value of the coordinate position between the inspection points in the row direction or column direction obtained in this way is distinguished into an increment value and a decrement value, and the increment values and decrement values are added together. Ru.

As a result, for inspection area 21, IP21=11, I
A total value IP21 of the increment values and a total value IM21 of the decrement values of M21--16 are obtained.

Similarly, for the remaining inspection areas 22 to 32, the total value IP22 of the increment value and decrement value of the coordinate position between the inspection points is calculated.
．． IP23. - IP32°IM'22. IM23. −
-IM'32 is determined.

As a result, in a total of 12 inspection areas 21 to 32, 24 number strings IP21. IM21°IP22. IM22. −
IP32. IM32 is obtained.

FIG. 4 is a flowchart showing the procedure for obtaining such a number string IP21 to IM32, in which lPx in one inspection area X.

When obtaining IMx (x=21 to 32), the values IPx and IMx in the previous inspection area are cleared, and then the increase/decrease value In of the coordinate positions between adjacent inspection points among the n inspection points is Desired. Next, this increase/decrease (l!!I n
It is determined whether is a positive value or a negative value, and if it is a positive value, a predetermined value "
2" is accumulated as the total value IPx. Conversely, if the value is negative, it is accumulated as the total value IMx on the condition that it is smaller than the predetermined value "-2".

Such processing is performed until inspection point n is reached. This results in a total value IPx of increases and decreases in coordinate positions between inspection points in one inspection area.

IMx is obtained.

By performing such processing for all inspection areas, 24 number strings IP21 to IM32 representing the characteristics of the edge portion of the recognition target character "3" are obtained as described above.

Therefore, by pre-registering the reference character pattern as such a number string in the image memory, and comparing it with the number string representing the characteristics of the edge portion of this character pattern when the character to be recognized is read, it is extremely easy to read the character pattern. It is possible to recognize which character the recognition target character corresponds to. Moreover, since the characteristics of the character pattern are represented by a string of numbers, storage areas within the image memory can be freely allocated. As a result, the image memory can be used efficiently, and the capacity of the image memory can be reduced.

In addition, in the example, only character recognition was explained, but
Objects whose characteristics are expressed by edges, such as marks and figures, can be recognized in the same way.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, a plurality of inspection points are set at the edge portions of the character pattern to be recognized and the reference character pattern, and the increase/decrease value of the two-dimensional plane coordinate position between the inspection points is calculated. Since the comparison is made, there is an effect that characters to be recognized can be recognized extremely easily with a small memory capacity.

[Brief explanation of the drawing]

FIG. 1 is a character pattern diagram for explaining a conventional character recognition method, FIGS. 2 and 3 are diagrams showing an example of setting an inspection area and inspection points used in character recognition according to the present invention, and FIG.
3 is a flowchart showing a procedure for obtaining a number string representing the characteristics of an edge portion in one inspection area. 1.2... Image memory, 21-32... Inspection area,
50-53...Inspection points. Agent: Masuo Oiwa (2 others) Figure 4 Procedural amendment (voluntary) 1. Indication of case: Patent Application No. 128044/1982 2.
Invention name character recognition method 3, person making the amendment Representative Hitoshi Katayama Part 5, column for detailed explanation of the invention to be amended. 6. Contents of amendment +11 Specification, page 2, line 12 [address line G1゜a
-aJ is corrected to "address in rows a to a". that's all

Claims (2)

[Claims] (1) By successively comparing the character pattern to be recognized stored in the image memory with multiple reference character patterns registered in advance corresponding to multiple reference characters, the character pattern to be recognized is determined to be a plurality of reference characters. In a character recognition method that recognizes which of the following applies, multiple inspection points are set at the edge portions of the character pattern to be recognized and the reference character pattern, and the increase/decrease value of the two-dimensional plane coordinate position between these inspection points is calculated. A character recognition method characterized by recognizing characters to be recognized by comparing the characters. (2) The increase/decrease value is divided into an increment value and a decrement value, and the comparison of the increase/decrease value regarding the standard character pattern and the character pattern to be recognized is performed based on the total value of only the increment value and the total value of only the decrement value. A character recognition method according to claim 1, characterized in that the character recognition method is carried out.