JP2001266069A - Reading method for optical character - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correctly read a character, even if the describing position of the character is deviated in optical character reading. SOLUTION: A segmentation control part 14 calculates segmentation start position and the number of segmentation characters for segmenting a character image, comprising its front and rear character positions based on the character start position and the effective number of characters to be read, which is registered in a slip form file 18 and controls a character-segmenting part 13. Then the part 13 segments the character image, having equal to or more than the prescribed effective number of characters, a character-recognizing part 15 recognizes the character and a character code is stored in a buffer memory 16. An output processing part 17 abandons invalid data such as 'null' in the character code stored in the buffer memory 16 and outputs only valid data for the portion of the valid number of characters as output data OUT.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reading method for reading a character entry area (hereinafter referred to as "field") in an optical character reading apparatus (hereinafter referred to as "OCR").

[0002]

2. Description of the Related Art Conventionally, when reading a character or a number described in a field of a form in an OCR, first, an image of the field is decomposed into pixels and read and temporarily stored in an image memory. Next, image data for a predetermined number of valid characters is cut out from a character start position predetermined as a position where a character or the like should be described. Further, character data is extracted from the extracted image data to perform character recognition, and a corresponding character code is output as a recognition result.

[0003]

However, the conventional OCR has the following problems. That is, for example, when numbers or the like are printed on a form using a printer or the like, the printing position of the first character may be shifted from the character start position in the field. In such a case, there has been a problem that valid characters having a predetermined number of digits cannot be read, and a digit drop occurs.

An object of the present invention is to solve the problems of the prior art and to provide an optical character reading method capable of reading characters accurately even if the printing position is shifted.

[0005]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a reading process in which an image of a field of a form to be read is decomposed into pixels and optically read; An optical character reading method for performing a cutout process of cutting out an image of the field based on the position and the number of valid characters, and a recognition process of recognizing the cutout image and outputting a character code of a recognition result; Then, based on the character start position and the number of valid characters, the image for the number of characters greater than the number of valid characters is cut out from the field, and in the recognition process, invalid data in the recognition result of the cut out image is discarded. Character codes for the number of valid characters are output.

According to the present invention, since the optical character reading method is configured as described above, the following operation is performed.
First, by a reading process, an image of a field of a form to be read is decomposed into pixels and optically read. Next, by a cutout process, a character image of a number of characters larger than the number of valid characters is cut out from the field based on a predetermined character start position and the number of valid characters. Further, character recognition is performed by a recognition process. At this time, invalid data is discarded in the recognition result of the image of the character extracted by the extraction processing, and only character codes corresponding to the number of valid characters are output.

[0007]

FIG. 1 is a block diagram of an OCR showing an embodiment of the present invention. The OCR includes an image scanner 11 that optically reads a form 1 to be read. A plurality of fields 2 are provided on the reading surface of the form 1.
Is provided. In each field 2, a predetermined character start position 3 for describing the first character is defined, and a predetermined number of valid characters are printed in a line at regular intervals starting from the character start position. I have.

The image scanner 11 is for decomposing the image of the field 2 of the form 1 into pixels and optically reading the image. An image memory 12 is connected to the output side of the image scanner 11. The image memory 12
The image of the field 2 is temporarily stored for recognition processing, and a character cutout unit 13 is connected to the image memory 12. The character extracting unit 13 extracts a character image from the image memory 12 based on an instruction from the extracting control unit 14 and outputs the character image to the character recognizing unit 15.

The cutout control unit 14 instructs the character cutout unit 13 to cut out a character image based on the character start position 3 of the field 2 in the form 1, the number of valid characters, the character size, and the like. is there. Character recognition unit 15
Extracts character data from the extracted character image and generates a corresponding character code by referring to a character dictionary or the like (not shown).

The output side of the character recognizing unit 15 is connected to a buffer memory 16 for temporarily storing the generated character codes for output processing of the recognition result. An output processing unit 17 is connected to the buffer memory 16. The output processing unit 17 discards invalid data in the recognition result and outputs only character codes for the number of valid characters as output data OUT.

Further, the OCR has a form format file 18 in which the form of the form 1 is registered in advance. Information such as the position of the field 2 in the form 1 to be read, the character start position 3, the number of valid characters, and the character size is registered in the form file 18. Then, from the form file 18, the position information of the field 2 is given to the image scanner 11, and information such as the character start position 3 and the number of valid characters is given to the cutout control unit 14. Further, information such as the number of valid characters is provided from the form format file 18 to the output processing unit 17.

FIG. 2 is a flowchart showing the operation of FIG. 1, and FIG. 3 is an explanatory diagram of the processing of the cutout control unit 14 in FIG. Hereinafter, with reference to FIGS. 2 and 3,
The operation of FIG. 1 will be described. When the OCR of FIG. 1 is activated and the form 1 is set on the image scanner 11, the image reading process in step S1 of FIG. 2 is started.

In step S1, the operation of the image scanner 11 is started, and an image of the field 2 of the form 1 is read based on the field position information in the form file 18. The read image of the field 2 is stored in the image memory 12.

In step S2, the operation of the cutout control unit 14 is started, and based on the information of the character start position 3 and the number of valid characters of the field 2 in the form file 18,
For example, the cutout start position and the number of cutout characters of the character image are calculated as in the following equation. Extraction start position = Read start position-1 Number of extracted characters = Number of valid characters + 2

That is, in this step S2, as shown in FIG. 3, the extraction of the character image is started from a character position one character left of a predetermined character start position. The position has been set. Furthermore,
The number of extracted characters is set so that a character image that is two characters larger than the predetermined number of valid characters is extracted from field 2. The cutout control unit 14 controls the character cutout unit 13 based on the cutout start position and the number of cutout characters thus calculated.

In step S 3, the operation of the character extracting section 13 is started, and a character image in the image memory 12 is extracted based on an instruction from the extracting control section 14 and output to the character recognizing section 15. In step S4, the operation of the character recognizing unit 15 is started, the characteristic data of the character image provided from the character extracting unit 13 is extracted, and a corresponding character code is generated by referring to a character dictionary or the like.

In step S5, the character code of the recognition result generated by the character recognition unit 15 is stored in the buffer memory 1.
6 is stored. In step S6, the output processing unit 1
7 is started, and the invalid data is discarded. That is, of the character codes stored in the buffer memory 16, the character code corresponding to "blank" in which nothing is described is discarded. Further, in step S7, only valid data corresponding to the number of remaining valid characters is output as output data OUT.

As described above, the OCR according to the present embodiment cuts out a character image larger than the number of valid characters from the image memory 12 based on the character start position and the number of valid characters registered in the form file 18 in advance. It has a cutting control unit 14 for controlling. Furthermore, the character recognition unit 1
An output processing unit 17 that discards invalid data from among the many character codes recognized in step 5 and outputs only character codes of a predetermined number of valid characters as output data OUT. Thereby, even if the position of the character described in the field 2 of the form 1 is deviated from a predetermined character start position, the character is accurately read without generating a digit loss and the character code having a correct number of digits is output. There is an advantage that can be.

The present invention is not limited to the above embodiment, and various modifications are possible. For example, there are the following modifications (a) and (b). (A) The OCR shown in FIG. 1 includes a character extracting unit 13, a clipping control unit 14, and a character recognizing unit 1 for clarifying the description of each process.
5 and an individual processing unit such as the output control unit 17, but may be performed by program control using a microcomputer or the like.

(B) In the description of FIG. 3, a character image is cut out by an extra two characters before and after a predetermined character position, but more character images are cut out. You may do it.

[0021]

As described above in detail, according to the present invention, based on a predetermined character start position and the number of valid characters, an image is extracted from a field by extracting an image of a number of characters larger than the number of valid characters. I have. As a result, even if the position of the character described in the field of the form is deviated from the predetermined character start position, the character can be read without causing digit dropout or the like. Further, a recognition process of discarding invalid data in the recognition result of the cut-out image and outputting character codes corresponding to the number of valid characters is performed. Thereby, even if the position of the character described in the field of the form deviates from the predetermined character start position, it is possible to read the valid number of valid characters.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an OCR showing an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of FIG.

FIG. 3 is an explanatory diagram of a process of a cutout control unit 14 in FIG. 1;

[Explanation of symbols]

 1 Form 2 Field 3 Character Start Position 11 Image Scanner 12 Image Memory 13 Character Extraction Unit 14 Extraction Control Unit 15 Character Recognition Unit 16 Buffer Memory 17 Output Processing Unit 18 Form Format File

Claims (1)

[Claims] 1. A reading process in which an image of a character entry area of a form to be read is decomposed into pixels and optically read, and an image of the character entry area is determined based on a predetermined character start position and the number of valid characters. In an optical character reading method that performs a cutout process of cutting out and a recognition process of recognizing a cutout image and outputting a character code of a recognition result, the cutout process includes the steps of: An image corresponding to the number of characters greater than the number of valid characters is cut out from the character entry area. In the recognition process, invalid data in the recognition result of the cut-out image is discarded, and a character code corresponding to the number of valid characters is output. An optical character reading method, comprising: