JP3243000B2 - Mail processing apparatus and mail processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a mail processing apparatus such as a mail address automatic reading / sorting machine and a mail processing method .

[0002]

2. Description of the Related Art In general, a mail address automatic reading and sorting machine recognizes mail addresses by optical means and automatically sorts the mail into sorting boxes according to the recognition result. In recognition of an address by such a mail address automatic reading / sorting machine, it is necessary to recognize a town name and a house number.
Conventionally, recognition of these town names and house numbers has been performed by cutting out portions where these are described from postal matter and calculating recognition candidates for the cut out portions by, for example, a composite similarity method or the like.

In such a recognition method based on the composite similarity method or the like, a plurality of recognition candidates may be calculated particularly for a house number, which is a combination of a maximum of four digits. Needs to uniquely determine a recognition candidate.

However, conventionally, when a plurality of recognition candidates are calculated, a method of rejecting the data at the expense of the segmentation rate or a method of appropriately selecting misreading with due consideration have been adopted.
There has been a problem that the processing efficiency and reliability are reduced.

[0005]

As described above, in the mail address automatic reading / sorting machine, when recognizing a town name and a house number, a plurality of recognition candidates are particularly recognized for a house number which is a combination of a maximum of four digits. May be calculated,
Conventionally, the recognition candidates are uniquely determined by a method of rejecting the data at the expense of the segmentation rate or a method of appropriately selecting misreading with the intention of being prepared, resulting in a problem that processing efficiency and reliability are reduced.

Accordingly, an object of the present invention is to provide a mail processing apparatus which can uniquely determine a recognition candidate without reducing processing efficiency and reliability.

[0007]

Means for Solving the Problems The present invention for solving the above problems, in the above mail processing apparatus recognizes the house number of the mail, in conjunction with house number and frequency and misreading rate of house numbers a database for storing the evaluation value obtained by weighting the frequency in misreading rate for each street, said mailpiece
Perform character recognition for
Means for selecting a range of house numbers in the street name;
A number character recognition candidate in the house number range
Means for selecting numeric character recognition candidates; selecting said numeric character recognition candidates
By detecting means, the more digits character recognition candidate is elected
Based on the occurrence frequency corresponding to each numeric character recognition candidate, obtain evaluation values of each numeric character recognition candidate from the database, and an evaluation unit that uniquely determines a numeric character recognition candidate based on these evaluation values. Is provided. Also book
The invention relates to a mail processing method for recognizing a mail house number.
Character recognition for the mail,
The frequency of occurrence and misreading rate of the house number
Stores the evaluation value obtained by weighting the raw frequency with the misreading rate for each town name
House number for a given town name from the database
Select a range and enter a number in the selected house number range.
Select character recognition candidates and select multiple numeric character recognition candidates
To refer to the database based on
, Based on the occurrence frequency corresponding to each numeric character recognition candidate.
The evaluation value of each digit character recognition candidate
And obtain numeric character recognition candidates based on these evaluation values.
It is uniquely determined.

[0008]

According to the present invention, a database is prepared which stores in advance the occurrence frequency, the misreading rate, the occurrence frequency weighted by the misreading rate, and an evaluation value weighted by the misreading rate for each town name in association with each house number of each town name . Perform character recognition and read the database
Select the house number range for the specified street name
In the house number range, select numeric character recognition candidates.
Here, when a plurality of numeric character recognition candidates are selected, the evaluation value of each numeric character recognition candidate based on the occurrence frequency corresponding to each numeric character recognition candidate is obtained from a database, and the numeric character recognition candidates are obtained based on these evaluation values. Recognition candidates are uniquely determined. Therefore, in comparison with the conventional technique in which the recognition candidate is uniquely determined by a method of rejecting at the expense of the segmentation rate or a method of appropriately selecting misreading with due consideration, the recognition candidate is uniquely identified without deteriorating the processing efficiency and reliability. Can be determined. In addition, by updating the frequency of occurrence of the database each time the house number of the mail is recognized, the evaluation performance improves as the mail is used.

[0009]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 is a front view of a mail address automatic reading / sorting machine according to an embodiment of the present invention.

In FIG. 1, reference numeral 1 denotes an operation panel for designating a reading mode, and 2 denotes a supply unit for supplying mail 3.

In the supply unit 2, the mails 3 taken out one by one from the supply table 4 by the take-out unit 5 are sent out to the reading unit 7 through the transport path 6.

The reading unit 7 recognizes the address written on the mail 3 and the recognized mail 3 is collected by the stacking unit 8.
Will be sent to

In the stacking unit 8, the mail 3 passing through the conveyance path 9
Are sorted into sorting boxes 10, 10,... According to the address based on the recognition result in the reading unit 7.

FIG. 3 is a block diagram showing the configuration of the reading section 7. As shown in FIG.

As shown in FIG. 1, the entire image of the mail 3 picked up by the photoelectric conversion unit 11 is converted into an electric signal and transmitted to the image processing unit 12.

The image processing unit 12 detects an address area, performs character detection cutout, and transmits a character detection cutout candidate to the character recognition unit 13.

The character recognizing unit 13 performs recognition in accordance with the transmitted character detection cut-out candidates, performs evaluation integration, finally determines a recognition result uniquely, and transmits the result to the MCPU 15 via the LCPU 14.

Then, the MCPU 15 controls the driving unit 16 of each unit of the apparatus to sort the mail 3 into a specific sorting box 10.

FIG. 4 is a block diagram showing the configuration of the character recognition unit 13.

The address recognition unit 17 shown in FIG. 2 calculates an address character recognition candidate by the composite similarity method according to the character detection cut-out candidates transmitted from the image processing unit 12, and further refers to the address information database 18. Then, address character recognition candidates are narrowed down to certain reliable recognition candidates. The result of recognition by the address recognition unit 17 is transmitted to the number recognition unit 19.

The number recognizing unit 19 calculates a numeric character recognizing candidate from the numeric character portion following the recognition candidate of the address recognizing unit 17 by using the combined similarity method and the structural feature matching method in accordance with the character detection cutout candidate. Then, referring to the house number information database 20, the number character recognition candidates are narrowed down to the number character recognition candidates of the existing numbers.

The number recognizing unit 19 refers to the house number frequency database 21 to uniquely determine the final answer.

FIG. 1 is a block diagram showing the configuration of the numeral recognition unit 19.

The recognition target 22 shown in FIG. Here, since there are a plurality of detection cutout candidates, a plurality of recognition results are calculated.

Further, the numeric character recognizing section 23 reads out only the actually existing house number as the recognition result 24 by referring to the house number information database 20.

Since the numeric character recognizing section 23 cannot uniquely determine the recognition result, the recognition result is transmitted to the recognition result evaluating section 25.

The recognition result evaluating section 25 refers to the house number frequency database 21 to uniquely determine a final answer. The house number frequency database 21 stores the frequency of occurrence and misreading rate of each house number of each town name, and the frequency of occurrence.
The evaluation value of the evaluation function f (x) weighted by the misread rate is
It is stored as residence number frequency information 29 . Note that the house number frequency database 21 includes a recognition result evaluation unit 25.
Automatically updated according to the recognition result.

The recognition result evaluator 25 evaluates the evaluation function f (x)
With reference to the house number frequency database 21, the evaluation value is acquired.

Then, for example, if the evaluation calculation result for the recognition result 24 is f (c) <f (b) <f (a) and f (a) ≧ TH (TH is a threshold value), A in the recognition result 24 becomes a recognition result 26 that is uniquely determined.

FIG. 5 is a diagram showing the structure of the house number frequency database 21.

As shown in the figure, town name information 27 is entered in the house number frequency database 21, and a predetermined town name is selected based on the recognition result by the address recognition unit 17. The house number range information 28 is linked to each of the town names in the town name information 27, and the numeral recognition unit 1
9, a predetermined house number range is selected based on the recognition result.

House number frequency information 29 is linked to each house number range in the house number range information 28, and the occurrence frequency and the evaluation value of the evaluation function f (x) are acquired from the house number frequency information 29. . When the corresponding house number frequency information 29 is referred to at the same time, the number of occurrences is set to 1
Voted. Further, the total number of street names is automatically updated without being referred to. The error rate is information given in advance, and if the expected value of the evaluation value cannot be obtained, the value may be reconsidered.

As described above, in the present embodiment, the house number frequency information 2 stored in the house number frequency database 21 is used.
9, the evaluation value of each numeric character recognition candidate is obtained , and the numeric character recognition candidate is uniquely determined based on these evaluation values. Therefore, the recognition candidate is uniquely determined without lowering the processing efficiency and reliability. can do.

[0035]

As described above, according to the present invention, the number character recognition candidate is uniquely determined from the evaluation value in which the occurrence frequency is weighted by the misreading rate, so that the processing efficiency and reliability are reduced. And a recognition candidate can be determined uniquely.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a number recognition unit according to an embodiment of the present invention.

FIG. 2 is a front view of a mail address automatic reading / sorting machine according to an embodiment of the present invention.

FIG. 3 is a block diagram illustrating a configuration of a reading unit according to an embodiment of the present invention.

FIG. 4 is a block diagram illustrating a configuration of a character recognition unit according to one embodiment of the present invention.

FIG. 5 is a diagram showing a configuration of a house number frequency database according to one embodiment of the present invention.

[Explanation of symbols]

20: house number information database, 21: house number frequency database, 22: recognition target, 23: numeric character recognition unit, 24: recognition result, 25: recognition result evaluation unit, 26: evaluation result.

Continuation of the front page (56) References JP-A-3-80389 (JP, A) JP-A-3-291777 (JP, A) JP-A-57-209575 (JP, A) JP-A-4-111186 (JP, A) JP-A-3-198180 (JP, A) JP-A-3-17663 (JP, A) JP-A-2-277187 (JP, A) JP-A-1-282693 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06K 9/00-9/72 B07C 3/14

Claims (4)

(57) [Claims] 1. A mail processing apparatus for recognizing a house number of a postal matter, wherein a frequency of occurrence of a house number, an erroneous reading rate, and an evaluation value obtained by weighting the occurrence frequency with an erroneous reading rate are associated with the house number for each town name.
And character recognition for the postal matter, the database
Means for selecting a house number range for a given town name
And in the selected house number range,
The numerical character recognition candidate selecting means to be selected, and each numerical character based on the occurrence frequency corresponding to each numerical character recognition candidate based on the plurality of numerical character recognition candidates being selected by the numerical character recognition candidate selecting means. A mail processing apparatus comprising: evaluation means for acquiring evaluation values of recognition candidates from the database and uniquely determining numeric character recognition candidates based on the evaluation values. 2. The mail processing apparatus according to claim 1, further comprising means for updating the frequency of occurrence of the database each time the house number of the mail is recognized. 3. A postal matter processing method for recognizing a house number of a postal matter, wherein character recognition is performed on the postal matter and the house number is determined in advance.
The occurrence frequency of the house number, the misread rate, and the evaluation value obtained by weighting the occurrence frequency with the misread rate are stored for each town name.
Select a house number range for a given town name from the database
And select a number character recognition candidate in the selected house number range.
Selected, and based on the plurality of numerical character recognition candidates are selected, by referring to the pre-SL database, the evaluation value of each digit character recognition candidates based on the occurrence frequency corresponding to each number character recognition candidates A mail processing method characterized by acquiring from a database and uniquely determining a numeric character recognition candidate based on these evaluation values. 4. The mail processing method according to claim 3, wherein the frequency of occurrence of the database is updated every time the house number of the mail is recognized.