JPS6355683A - Address reader - Google Patents

Abstract

PURPOSE:To collate a place name whose storage includes plural columns by defining connection information in the unit of word, and storing it as a plural rows key word dictionary for a place name probable to be written over plural rows. CONSTITUTION:A character reader part 1 outputs read character strings 11 which are divided/integrated by a word extraction part 2 in order to produce words 21. The words 21 are collated with the keywords stored in the plural rows keyword dictionary storage part 4, and if a prescribed keyword is detected, a connection information 41 is outputted. By using this connection information 41, the word-connection range 31 is determined in a word connection range control part 3, and outputted to an address collation part 5. In the address collation part 5, the read character string and the place name 61 that is stored in an address dictionary storage part 6 are collated with each other, and the result of collation 51 is outputted, so that the address is discriminated in an address discrimination part 7 and outputted as the address 71.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an address reading device for reading addresses written on paper sheets such as mail.

[Conventional technology]

Conventionally, in this type of address reading device, when a read character string is compared with an address dictionary, only character strings in the same line are compared. For example, as a recognition/reading device for the European address written on the mail etc.,
No. 90663 was proposed by the present inventors.

[Problem that the invention seeks to solve]

The above-mentioned address reading device stores the district name over multiple lines as in the Angline below, in order to match only character strings in the same line when comparing the read character string with the address dictionary. In this case, the disadvantage is that it is not possible to match place names.

rBLK169 A and -Mo Ki Rainbow Tndus
trial Park5 Sigapore 0913J On the other hand, if the matching range of character strings is expanded unconditionally to multiple lines, there is a drawback that the combinations of character strings become enormous and the processing time increases.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an address reading device which eliminates these drawbacks and is capable of collating place names stored in a plurality of columns and shortening the collation processing time.

[Means for solving problems]

The address reading device of the present invention has a structure including a character reading means for reading a character string related to address notation written on a sheet of paper, a word extraction means for editing the character string output of the character reading means into words, and the address a keyword dictionary storage means that stores connection conditions for keywords spanning multiple lines in accordance with the notation;
a word string connection detection means for detecting the keyword connection condition from the output of the word and extraction means and extracting a destination word string divided into a plurality of lines according to the connection information; and storing an address character string corresponding to the address display. address dictionary storage means;
The present invention is characterized by comprising address matching means for matching the address character string stored in the address dictionary storage means and the output of the word string connection detection means.

〔Example〕

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

FIG. 1 is a block diagram of one embodiment of the present invention.

In this embodiment, the character reading section 1. Word extraction part 2. Word connection range control unit 3. Multi-line keyword dictionary storage section 4, address matching section 5. Address dictionary storage unit 6. It is composed of an address discrimination section 7. The character reading section 1 outputs a read character string 11, and the word extracting section 2 divides and integrates the read character string 11 into words 21. These words 21 are compared with keywords stored in the multi-line keyword dictionary storage section 4, and if a predetermined keyword is detected, connection information 41 is output.

The word connection range 31 is determined by the word connection range control section 3 based on this connection information 41, and is output to the address matching section 5. In the address matching section 5, the read character string is matched against the district classification 61 registered in the address dictionary storage section 6, the verification result 51 is outputted, and the address is determined in the address classification section 7, and the address is registered as an address 71. The address is output as address 71.

As this connection information 41, for example,
OJ stores connection information "X" such that it is connected with one word before and four words after. This kind of word connection information is stored in the keyword dictionary storage section 4, and the content and the word connection range are compared in the address matching section 5 with the contents of the address dictionary storage section 6. It is possible to reduce the number of address verifications.

〔Effect of the invention〕

As explained above, the present invention defines connection information for each word for a district name that may be written across multiple lines, and stores it as a multi-line keyword dictionary. Addresses that span multiple lines can also be matched, and since the connection range is limited by this keyword, this has the effect of efficiently matching addresses.

[Brief explanation of drawings]

FIG. 1 is a block diagram of one embodiment of the present invention. DESCRIPTION OF SYMBOLS 1...Character reading unit, 2...Word extraction unit, 3...Word connection range control unit, 4...Multi-line keyword dictionary connection unit, 5...Address matching unit, 6...Address Dictionary storage section, 7
... Address discrimination section, 11 ... Read character string, 21 ...
Word, 31... Connection range, 41... Connection information, 51
...Verification result, 61... District name, 71... Address output.

Claims (1)

[Claims] A character reading means for reading a character string related to an address notation written on a sheet of paper, a word extracting means for editing the character string output of this character reading means into words, and a keyword extracting means spanning multiple lines corresponding to the address notation. a keyword dictionary storage means for storing connection conditions; a word string connection detection means for detecting the keyword connection condition from the output of the word extraction means and extracting a destination word string divided into a plurality of lines according to the connection information; and the address. It is characterized by comprising an address dictionary means for storing an address string corresponding to the display, and an address matching means for comparing the address string stored in the address dictionary storage means with the output of the word string connection detecting means. address reader.