JPS6394384A - System for deciding direction of character row - Google Patents

Abstract

PURPOSE:To automatically decide the direction of a character, to improve the performance of character recognition, and to decide the direction of a space and that of the character, by providing a means which blocks an image on the space according to its constituent, a means which extracts the feature of the block such as the position, the form, or the size, etc., and a means which decides the direction of the character based on the form of the space. CONSTITUTION:In an address read sorter which reads and sorts the address described on a mail, etc., directly by using an optical reader, the image of the whole of the space is inputted, and all of the described items are made into blocks, and feature information such as the position, the form, and the size of the block, etc., are extracted, and based on relation among those information, and the form of a sheet, the relation among the blocks is evaluated at step 13, and the direction of the character is decided at step 14. It is possible to decide the direction of the character of a domestic mail by the position and the size of a sender block 4. The sensor block 4 exists at an end part opposite to the blocks of the stamp, and the address number of the mail, and it is possible to decide that a receiver block 3 is formed suitable for vertical writing by a bit of information that a horizontal width approaches to the width of the mail.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides an OCR method for identifying addresses written on mail items and the like.
This invention relates to a character line direction determination method for an address reading and sorting machine that directly reads and sorts using an optical reader (optical reader).

[Conventional technology]

-m, some kind of printing was done on the paper in advance,
The paper is oriented according to that direction and the letters are written.

Furthermore, in printed matter, other information such as pages is written in addition to what needs to be read.

Conventionally, in OCR for reading lines of text, there is a condition that the orientation of the paper and the orientation of the characters match, so the orientation of the paper can be determined by detecting certain features on the paper. For example, in overseas address leagues, the orientation of characters was determined by detecting the position of stamps. In addition, Japanese postal code reading and sorting machines do not determine whether postal codes are written vertically because all postal codes are written horizontally.

[Problem that the invention seeks to solve]

Unlike those using the alphabet, character lines in Japanese and other languages are written both vertically and horizontally, so it is difficult to determine in which direction even the same line image is written unless there is a regulation. For example, in Japanese mail, there are three orientations for the same line image, as shown in FIGS. 3(a) to 3(C). Therefore, it is difficult to determine the orientation of the letters by simply detecting the position of the stamp, as is done with overseas address readers.

With current character recognition technology, it is difficult to recognize characters unless the orientation of the characters is known. Therefore, in OCR that reads character lines, it is necessary to determine the orientation of the characters, but as mentioned above, the orientation of the characters can be determined without reading the characters. It was difficult to determine the

SUMMARY OF THE INVENTION An object of the present invention is to provide a character line direction determination method that solves these problems and makes it possible to easily determine the orientation of characters.

[Means for solving problems]

The present invention provides a means for inputting an image on a paper surface and dividing the image into blocks according to its constituent elements, in a character line direction determination method for an optical reading device that optically reads printing or handwritten characters, and The present invention is characterized by comprising means for extracting characteristics such as position, shape, and size of each block, and means for determining character line orientation based on the characteristics of the blocks and the paper shape.

〔Example〕

Next, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a flowchart explaining one embodiment of the present invention;
Figures 2(a) and (b) show the mail items (
FIG. 2 is a plan view showing an image of a block unit of a sealed letter. The mail envelope shown in FIG. 2(a) is represented in units of blocks as shown in FIG. 2(b). That is, block 1 of the stamp and block 2 of the postal code frame are detected by the automatic postal code reading and sorting machine. Further, it can be seen that the receiver is the receiver's address, block 3 of the name, and block 4 of the sender's address and name, which is larger than the size and positional relationship of the blocks, is the receiver's address. but,
In this receiver, the direction of the letters of the recipient's address is shown in Figure 3 (a) - (
I don't know which one of C) is supported.

In this embodiment, in step 10, an image of the entire paper is input, and all written items from the entire paper are converted into blocks (step 11), and in step 12, characteristic information such as the position, shape, size, etc. of all these blocks is input. is extracted, and based on the relationship between these pieces of information and the paper shape, the relationship between blocks is evaluated in step 13, and the orientation of the characters is determined in step 14.

For example, in Japanese mail, stamps,
It is possible to detect postal code frames, sender's address and name, receiver's address and name, angle lines, surrounding frames, etc., and determine the direction of the characters based on information such as their positional relationships and shape relationships. Also, if the Angline and its position are detected, if the Angline is vertically on the left side with the stamp facing up, the orientation corresponds to Figure 3 (C), and if it is on the right edge, it will be in the 3rd position.
It can be seen that this corresponds to figure (b).

In this embodiment, by further checking the position and size of the sender block 4, the orientation of the characters can be determined.

This sender block 4 is located at the opposite end of the postal stamp and postal code blocks, and the recipient block 3 is written vertically (see Figure 3(a)
(equivalent to). A data survey of 6,000 emails found no exceptions.

〔Effect of the invention〕

As described above, according to the present invention, since the direction of a character can be determined from the relationship between each block, it is possible to reduce rejections and erroneous continuations due to characters being read backwards or sideways. Furthermore, the processing speed can be faster than the method of reading characters from three directions and finding the most likely one among them.

Furthermore, since the number of characters and words that appear in address recognition is small, using this information is expected to improve recognition performance, and if the orientation of the characters is known, it is possible to determine which direction the beginning of a character line is. It is.

As explained above, the present invention has the effect of improving character recognition performance by automatically determining the direction of characters, and also has the effect of improving the performance of character recognition by automatically determining the orientation of characters. The orientation and direction of characters can be easily determined.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the operating procedure of an embodiment of the present invention, FIGS. ) is a front view of a Japanese postal item showing the direction of each character written on the postal item. 1... Stamp block, 2... Postal code block, 3
... Recipient address block, 4... Sender address block, 11-14... Operation step. Ga1 mother foil 3 times

Claims (1)

[Claims] In a character line direction determination method for an optical reading device that optically reads printed or handwritten characters, means for inputting an image on a paper surface and dividing the image into blocks according to constituent elements;
The present invention is characterized by comprising means for extracting characteristics such as the position, shape, and size of each of the blocks, and means for determining the orientation of character lines based on the characteristics of the blocks and the paper shape. Character line direction determination method.