JP2000298701A - Destination identification device - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a destination specifying device capable of preventing erroneous reading of unregistered data and similar customer names. An image pattern input unit (111) optically reads an image of a postal matter and stores it in an image buffer (121). The character position extraction unit 112 creates character pattern position information from the image in the image buffer 121 and stores the information in the character position buffer 122. The character recognition unit 113 performs character recognition on each character position candidate based on the image in the image buffer 121 and the position information in the character position buffer 122, and outputs the character candidate to the character candidate buffer 123. The destination candidate extraction unit 114 includes a destination information dictionary
The customer name of the destination information in 126 and the character candidate in the character candidate buffer 123 are collated to extract a destination candidate, and stored in the destination candidate buffer 124. The candidate verification unit 115 verifies whether the destination candidates stored in the destination candidate buffer 124 are to be adopted or rejected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a destination specifying device for specifying a delivery destination (a destination) of a delivery such as a mail in a mail processing device or the like.

[0002]

2. Description of the Related Art Recently, for example, postal information of postal matter is registered in advance, and handwritten or printed characters written on the destination surface of the postal matter are optically read to perform character recognition. A destination specifying device has been developed which specifies a destination of a postal matter by collating with destination information.

Conventionally, in such a destination specifying device, for example, a name character string such as a company name or an individual name in the destination information is obtained by the method described in Japanese Patent Application No. 9-256842. The registered name character string including each character of the character recognition result is extracted, and further, the reversal of the character position is checked, and the registered name character string having the highest similarity between the character recognition result and the registered name character string. The destination information is specified by the destination information.

[0004]

However, in the above-described conventional destination specifying device, the customer names having a high appearance frequency cannot be uniquely narrowed down, and the customer names having similar characters are erroneously identified as the collation results. There was a problem of outputting.

[0005] In addition, when a company name is registered, but not all employee names are registered, a mail addressed to an unregistered person is specified as another person having the same name. Was.

On the other hand, in the registered destination information, data may not be deleted in spite of the absence of a resident due to relocation or the like, data may not be registered, or a character may be erroneously input due to a data input error. .

Accordingly, an object of the present invention is to provide a destination specifying device capable of preventing unregistered data and similar customer names from being erroneously read.

Another object of the present invention is to provide a destination specifying device which can prevent a delivery addressed to an unregistered person from being specified by another person having the same name and the same name.

A further object of the present invention is to provide a destination specifying device capable of specifying a destination more reliably when code information (for example, bar code information) representing destination information is present in a delivery. I do.

[0010]

According to the present invention, there is provided an address specifying apparatus for reading an image on a destination surface of a delivery and inputting image data, and character recognition based on the image data input by the image input means. Character recognition means, a destination information storage means in which destination information including a customer name of a target destination and address information corresponding thereto is stored in advance, a recognition result of the character recognition means and the destination information storage Destination candidate extracting means for extracting a customer name and corresponding address information as destination candidates by comparing the customer name of the destination information stored in the means, and a destination candidate extracted by the destination candidate extracting means By comparing the address information specified by the customer name with the recognition result of the character recognition unit, the likelihood of the destination candidate extracted by the destination candidate extraction unit is determined. Constant and the adoption of the destination candidate, and a candidate verifying means for verifying the rejected.

Further, the destination specifying apparatus of the present invention comprises an image input means for reading an image on a destination face of a delivery and inputting image data, and a character recognition for performing character recognition from the image data input by the image input means. Means, destination information storage means for storing destination information including a customer name of a target destination and address information corresponding to the destination information in advance in association with layout information when the destination information is described,
By comparing the recognition result of the character recognition unit with the customer name of the destination information stored in the destination information storage unit, the customer name, the corresponding address information, and the layout corresponding to the destination candidate as the destination candidate Destination candidate extracting means for extracting information, and candidate verification for verifying the adoption or non-employment of the destination candidate extracted by the destination candidate extracting means according to the layout information corresponding to the destination candidate extracted by the destination candidate extracting means. Means.

Further, the destination specifying apparatus of the present invention reads an image of the destination surface of a delivery having destination information and code information representing the destination information represented by a code on the destination surface, and inputs image data. Image input means, code information reading means for reading the code information from the image data input by the image input means, character recognition means for performing character recognition from the image data input by the image input means,
A destination information storage unit that stores destination information including a customer name of a target destination and address information corresponding to the customer name in advance, and a destination information storage unit that responds based on the code information read by the code information reading unit. A destination determining unit that determines whether to adopt or reject the extracted destination information by comparing the extracted destination information with the recognition result of the character recognition unit.

Further, the destination specifying apparatus of the present invention reads an image of the destination surface of a delivery having destination information and code information representing the destination information in codes, and inputs image data. Image input means; code information reading means for reading the code information from the image data input by the image input means; character recognition means for performing character recognition from the image data input by the image input means; A destination information storage unit in which destination information including a destination customer name and corresponding address information is stored in advance, and corresponding destination information from the destination information storage unit based on the code information read by the code information reading unit. By extracting the extracted address information and comparing the extracted address information with the recognition result of the character recognizing means, it is possible to adopt or reject the extracted address information. Address determining means for determining the address information specified by the code information read by the code information reading means and recognition by the character recognizing means when the extracted destination information is determined to be rejected by the destination determining means. By searching the candidate storage means for storing the result and the candidate storage means, the address information specified by the code information and the recognition result of the character recognition means that are equal to each other are larger than a preset threshold. And destination information updating means for updating the destination information in the destination information storage means, if any.

According to the present invention, a customer name similar to the character recognition result is identified by collating the character recognition result of the destination information described on the delivery with the customer name of the destination information registered in advance. Unregistered data is extracted by comparing the address information specified by the extracted customer name with the customer name of the registered destination information and verifying whether the extracted customer name is adopted. And misreading of similar customer names can be prevented, and more reliable destination identification can be performed.

Further, according to the present invention, when registering the destination information, whether the destination information is a company name, a store, an individual name, a building, an apartment, etc. According to the attribute of the customer, layout information indicating whether additional information such as a company name, a department name, a building name, etc. is added before and after the customer name is added in advance, and the destination information described on the delivery is determined by the layout information. By checking whether there is a character candidate between and after the position of the extracted customer name and the position of the extracted customer name, it is possible to verify the validity of the data,
In the case where the extracted destination information is a personal name, but a company name is written on the read delivery, the candidate is rejected, so that misreading can be reduced. In particular, when the company name is registered but not all the employee names are registered, it is possible to avoid a problem that a delivery to an unregistered person is specified by another person having the same name and the same name.

Further, according to the present invention, when code information (for example, bar code information) indicating destination information is described in a delivery, the destination information specified by the code information and the description on the delivery are described. By collating with the character recognition result of the specified destination information, more reliable identification can be performed.

If the matching result between the destination candidate based on the character recognition and the destination specification based on the customer name and the destination candidate based on the code information are different, the destination information based on the code information and the character recognition candidate are stored in the same manner. When a large number of results are accumulated, the registered contents of the destination information are updated based on the accumulated data, so that it is possible to correctly specify the subsequent deliveries without code information.

[0018]

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

First, a first embodiment will be described.

FIG. 1 schematically shows a configuration of a destination specifying apparatus according to the first embodiment. The destination specifying device according to the first embodiment registers address information of a mail in advance, performs optical recognition of handwritten or printed characters written on the address surface of the mail, and performs character recognition. The mail address is specified by comparing it with the address information registered in advance. This will be described below as an example.

First, the outline of the apparatus will be described. In FIG. 1, an image pattern input unit 111 such as a scanner as an image input unit optically reads handwritten characters or printed characters written on a destination surface of a postal matter as a delivery, and sets a predetermined threshold value to two. By performing the binarization, a binary image is stored in the image buffer 121 as image data.

The character position cutout unit 112 is provided in the image buffer 1
By extracting the character delimiter positions of the binary image stored in 21, position information of the character pattern is created and stored in the character position buffer 122.

The character recognition unit 113 includes an image buffer 121
By performing character recognition for each character position candidate based on the binary image in the above and the position information of a plurality of character patterns in the character position buffer 122, for example, the top ten character candidates are output to the character candidate buffer 123. .

The destination candidate extraction unit 114 extracts a destination candidate by comparing the customer name of the destination information in the destination information dictionary 126 as the destination information storage means with the character candidates in the character candidate buffer 123. Destination candidate buffer 1
24. Here, as a method for extracting a destination candidate, for example, a method described in Japanese Patent Application No. 9-256842 is used.

The candidate verifying unit 115 stores the destination candidate buffer 1
In order to verify whether the destination candidate stored in No. 24 is to be adopted or not to be adopted, the character recognition result of the written address including the town name and the block information is compared with the address information of the destination candidate. Verification of the adoption / non-adoption of destination candidates is performed based on differences in structure due to codes, the number of characters, delimiters, and the like.

Next, the overall processing flow will be described with reference to the flowchart shown in FIG. First, step 201
Then, the image pattern input unit 111 takes in the image of the written character on the mail. Next, in step 202, the character position extracting unit 112 extracts a character position from the input image. Next, in step 203, the character recognition unit 113
Perform character recognition.

Next, in step 204, the destination candidate extracting unit 114 compares the character recognition result with the destination information in the destination information dictionary 126 to extract a destination information candidate. next,
In step 205, the candidate verifying unit 115 compares the extracted residence code, which is the address information of the destination information candidate, with the character recognition result of the destination address. Next, step 20
6, if the number of candidates is reduced to one, the destination information dictionary 12
6 is output as a recognition result.

Next, the collation processing of the address information in the candidate verification section 115 will be described in detail with reference to the flowchart shown in FIG. First, in step 301, among the address information of the destination candidates stored in the destination candidate buffer 124, the block information including the street, number, and number information is set in the memory G1, and the number of the destination address described in the mail is set. The information is set in the memory G2.

The block information of the destination address in the memory G2 exists on the right side in the case of vertical writing and on the upper side in the case of horizontal writing, based on the character recognition results collated to select the destination candidates stored in the destination candidate buffer 124. The category is a character candidate for numbers, and a plurality of numbers separated by delimiter information such as hyphens are extracted as the same hierarchy.

Next, in step 302, the memories G1 and G
The number of layers of information in 2 is compared. For example, when the block information is “123-4”, the number of layers is “2”, and “1-2-3”.
In the case of “−101”, the number of layers is “4”. If the two layers have the same number, it is checked in step 303 whether the number of characters in each layer and the degree of coincidence of the characters satisfy a preset condition.

For example, "123-4" and "128-
4 ”and so on,
If the difference is one character, the condition is satisfied. In this case, the process proceeds to step 305 to determine the adoption of the destination candidate. Also, for example, if the characters differ by two characters, such as “123-4” and “323-1,” the condition is not satisfied.

In step 304, it is determined whether or not the number of characters ignoring the hierarchy and all character codes match. If the number of characters matches, the decision is made in step 305 to adopt a destination candidate. Step 306
To determine the rejection of the destination candidate. For example, if the registered address in the destination information dictionary 126 is “123-4” and the character recognition result is “1234”, it is adopted because there is a possibility that the delimiter “−” has not been detected.

As described above, according to the first embodiment,
By matching the character recognition result of the customer name of the mail address information with the address information registered in the address information dictionary 126, a customer name similar to the character recognition result is extracted. By comparing with the address information of the destination and determining whether to use the extracted customer name, it is possible to prevent data not registered in the destination information dictionary 126 or misreading of similar customer names.

Next, a second embodiment will be described.

Since the configuration of the destination specifying apparatus according to the second embodiment and the overall processing flow are the same as those of the first embodiment, description thereof will be omitted, and only different parts will be described. I do.

FIG. 4 shows an example of the configuration of the destination information dictionary 126. As the destination information, a structure flag is set in advance in addition to the house code and the customer name as the address information. The structure flag indicates whether there is a possibility that additional information such as a building name, a company name, or a department name may be written between the town name and the block information and the customer name, and after the customer name (layout). Information). Candidate verification unit 115
Then, the final candidate is verified by the determination according to the value of the structure flag.

FIG. 5 shows an example of what the value of the structure flag indicates. For example, if the value of the structure flag is "1",
For example, personal names that do not have an attached string before and after the customer name,
When the value of the structure flag is “2”, it is a company name in which an attached character string such as an employee is added after the customer name. When the value of the structure flag is “3”, if the attached information such as the name of the store is in front of the customer name and there is no character string after it, the value of the structure flag is “4” and the customer name is the building name. It indicates that there is a possibility that the building name may be written before the customer name and the affiliation or personal name in the company may be written after the customer name.

According to the values of these structure flags, the position of the character candidate string of the town name and the district information, the position of the character candidate string corresponding to the customer name stored in the destination candidate buffer 124, and the character candidate information before and after the position The candidate that satisfies the condition is determined to be adopted, and the candidate that does not satisfy the condition is determined to be not adopted.

FIG. 6 shows the flow of processing in the candidate verification unit 115. First, in step 601, a structure flag of the extracted destination candidate is extracted. Next, step 6
In step 02, it is checked whether or not a line exists before the customer name by using the extracted structure flag. If there is a line, in step 603, a character string corresponding to the extracted destination candidate and the destination address part are checked. It is determined whether or not there is a description of a character string in between.

The destination address portion is a character candidate string of numeric characters whose category exists on the right side in vertical writing and on the upper side in horizontal writing. If the structure flag indicates that there is a line before the customer name, but the address information part immediately before the customer name is rejected in step 607.

Next, in step 604, if a line exists after the customer name due to the structure flag, step 6
At step 05, it is determined whether or not a character string exists after the character string corresponding to the extracted destination candidate. If no character string exists, it is rejected at step 607. If the condition of the structure flag is satisfied, it is adopted as a destination candidate in step 606.

As described above, according to the second embodiment,
Depending on the attributes of the customer name registered as the destination information, such as company name, store, personal name, building, apartment, etc., the company name is added before and after the customer name. Information on whether additional information such as name, department name, building name, etc. is added is set in advance, and character candidates are placed between and after the location of the street name and street information described on the mail and the location of the extracted customer name. By checking whether or not the data exists, the validity of the data can be verified. If the extracted destination information is a personal name but the read mail has a company name written on it, candidates are Rejection can reduce misreading. In particular, it is possible to avoid a problem that a mail addressed to an unregistered person is specified as another person with the same last name when, for example, the company name is registered but not all the employee names are registered.

Next, a third embodiment will be described.

FIG. 7 schematically shows a configuration of a destination specifying apparatus according to the third embodiment. The destination specifying device according to the third embodiment registers the destination information of postal matter in advance and, for postal matter with bar code information, optically converts handwritten characters and printed characters together with the described bar code information. Character recognition is performed by reading the address information, the destination information specified from the bar code information reading result is collated with the character recognition result to specify the destination, and the destination information is determined by a character recognition result different from the destination information registered in advance. This will be described below as an example.

First, the outline of the apparatus will be described. In FIG. 7, an image pattern input unit 711 such as a scanner as an image input unit optically reads handwritten characters, printed characters, and the like written on a destination surface of a mail as a delivery, and uses a preset threshold value. By performing binarization,
A binary image is stored in the image buffer 721 as image data.

The character position extracting section 712 is provided in the image buffer 7.
By extracting the character delimiter position of the binary image stored in 21, position information of the character pattern is created and stored in the character position buffer 722.

The character recognition unit 713 includes an image buffer 721
By performing character recognition for each character position candidate based on the binary image in the above and the position information of a plurality of character patterns in the character position buffer 722, for example, the top ten character candidates are output to the character candidate buffer 723 .

On the other hand, when bar code information as code information is described in a postal matter from an image captured by the image pattern input unit 711, the bar code reading unit 714 as code information reading means The code information is read, and the read result is stored in the destination code candidate buffer 7.
24. Here, it is assumed that the barcode information is described by coding an address code, for example.

If the read bar code information (destination code candidate) is stored in the destination code candidate buffer 724, the destination determination unit 715 searches the destination information dictionary 726 based on the destination code candidate to respond. The destination information is obtained, the candidate is verified together with the character candidates in the character candidate buffer 723, and the destination information is output.

If there is no destination code candidate in the destination code candidate buffer 724 because bar code information is not described, for example, the character candidate in the character candidate buffer 723 is compared with the destination information in the destination information dictionary 726 to determine the destination. Determine and output information. In this case, the destination information is determined by, for example, the same method as described in the first embodiment.

When the customer name and the character candidate in the destination information are compared, if they do not match at all, the candidate storage unit 727 stores the destination code candidate in the destination code buffer 724 together with the destination code candidate in the destination code buffer 724. Save the character candidates (character recognition results). Also, when the destination code candidate is not registered in the destination information dictionary 726,
The destination storage unit 72 stores the destination code candidate together with the character recognition result.
7

The destination information dictionary updating unit 716 searches the candidate storage unit 727 to determine if the same destination code and the same character recognition result data are present in excess of a preset threshold value. Update the data of. If the destination code has already been registered and the customer name is different, rewrite the customer name,
If the destination code has not been registered, the destination code is newly registered together with the customer name.

FIG. 8 shows a flow of a process of storing data in the candidate storage unit 727 in the destination determination unit 715 and the destination information dictionary update unit 716. First, step 801
Then, it is checked whether there is a bar code information read result. If there is a bar code information read result, in step 802, the destination information dictionary 726 is searched to correspond to the bar code information read result. Retrieves destination information and stores it in memory A
Set to 1.

Next, in step 803, destination information is extracted from the character recognition result and set in the memory A2. Here, it is assumed that the destination information includes address information and a customer name. Next, in step 804, the memories A1 and A2
In step 805, the destination information in the memory A1 is stored in the candidate storage unit 72 in step 805.
Save to 7.

Next, in step 806, the destination information dictionary updating unit 716 checks whether or not the same data has been stored in the candidate storage unit 727 more than a preset threshold value. Updates the data in the destination information dictionary 726 in step 807.

As described above, according to the third embodiment,
When the barcode information is described in the mail, collation between the address information specified by the barcode information and the character recognition result enables more reliable specification. In addition, storing unregistered data and error data,
When a large number of data are stored, the reliability of the destination information database is increased by updating the destination information dictionary based on the stored data. The identification can be performed.

[0057]

As described above in detail, according to the present invention, it is possible to provide a destination specifying apparatus capable of preventing unregistered data and a similar customer name from being erroneously read.

Further, according to the present invention, it is possible to provide a destination specifying device which can prevent a delivery addressed to an unregistered person from being specified by another person having the same name and the same name.

Further, according to the present invention, it is possible to provide a destination specifying device capable of specifying a destination more reliably when code information (for example, bar code information) representing destination information is present in the delivery.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a configuration of a destination specifying device according to first and second embodiments of the present invention.

FIG. 2 is a flowchart for explaining the flow of overall processing according to the first embodiment;

FIG. 3 is a flowchart illustrating a flow of processing of a candidate verification unit according to the first embodiment.

FIG. 4 is a diagram showing a configuration example of a destination information dictionary according to a second embodiment of the present invention.

FIG. 5 is a diagram showing an example of the contents of a structure flag according to the second embodiment.

FIG. 6 is a flowchart illustrating a flow of a process performed by a candidate verification unit according to the second embodiment.

FIG. 7 is a block diagram schematically showing a configuration of a destination specifying device according to a third embodiment of the present invention.

FIG. 8 is a flowchart illustrating a flow of processing of a destination determining unit and a destination information dictionary updating unit according to the third embodiment.

[Explanation of symbols]

111, 711... Image pattern input unit (image input unit), 112, 712... Character position extraction unit (character recognition unit), 113, 713... Character recognition unit (character recognition unit), 114. Extraction unit (destination candidate extraction unit), 115... Candidate verification unit (candidate verification unit), 12
6,726... Destination information dictionary (destination information storage means), 7
14... Bar code reading section (code information reading means), 7
15 destination determining unit (destination determining unit) 716 destination information dictionary updating unit (destination information updating unit) 727 candidate storage unit (candidate storage unit)

Claims (7)

[Claims] 1. An image input means for reading an image on a destination surface of a delivery and inputting image data; a character recognizing means for performing character recognition based on the image data input by the image input means; A destination information storage unit that stores destination information including a customer name and address information corresponding to the customer name in advance; a recognition result of the character recognition unit; and a customer name of the destination information stored in the destination information storage unit. By extracting the customer name and the address information corresponding to the customer name as the destination candidate, and the address information specified by the customer name as the destination candidate extracted by the destination candidate extracting means. By comparing the recognition result of the character recognition unit with the recognition result, the likelihood of the destination candidate extracted by the destination candidate extraction unit is determined. The destination specifying device, wherein the candidate verifying means certifying, by comprising a. 2. An image input means for reading an image on a destination surface of a delivery and inputting image data; a character recognizing means for performing character recognition based on the image data input by the image input means; Destination information storage means which stores in advance destination information including a customer name and address information corresponding to the customer information in association with layout information when the destination information is described; and a recognition result of the character recognition means. A destination candidate extracting unit for extracting a customer name as a destination candidate, address information corresponding to the customer name, and layout information corresponding to the destination candidate by comparing the customer name of the destination information stored in the destination information storage unit. Adopting the destination candidate extracted by the destination candidate extracting means according to the layout information corresponding to the destination candidate extracted by the destination candidate extracting means; A candidate verifying means for verifying the adopted destination specifying device, characterized by comprising a. 3. The layout information is information indicating whether or not another character string is described between a customer name to be described and address information to be described. The destination identification device as described. 4. The destination specifying apparatus according to claim 2, wherein the layout information is information indicating whether another character string is described after the customer name to be described. 5. An image input means for reading an image on the destination surface of a delivery having destination information and code information representing the destination information represented by a code on the destination surface and inputting image data; Code information reading means for reading the code information from the image data input by the means, character recognition means for performing character recognition from the image data input by the image input means, and a customer name of a target address and correspondence thereto A destination information storage unit that stores destination information including address information obtained in advance, and extracts corresponding destination information from the destination information storage unit based on the code information read by the code information reading unit,
By comparing the extracted address information with the recognition result of the character recognizing means, the adoption of the extracted address information,
A destination specifying device, comprising: destination determining means for determining rejection. 6. An image input means for reading an image on the destination surface of a delivery having destination information and code information representing the destination information represented by a code on the destination surface, and inputting image data; Code information reading means for reading the code information from the image data input by the means, character recognition means for performing character recognition from the image data input by the image input means, and a customer name of a target address and correspondence thereto A destination information storage unit that stores destination information including address information obtained in advance, and extracts corresponding destination information from the destination information storage unit based on the code information read by the code information reading unit,
By comparing the extracted address information with the recognition result of the character recognizing means, the adoption of the extracted address information,
Destination determining means for determining rejection; address information specified by the code information read by the code information reading means and the character recognizing means when the destination information extracted by the destination determining means is determined to be rejected; A candidate storage unit for storing the recognition result of the character recognition unit. By searching the candidate storage unit, the address information specified by the code information and the recognition result of the character recognition unit that are equivalent to each other are determined by a predetermined threshold value. And a destination information updating unit that updates destination information in the destination information storage unit when there are many destination specifying devices. 7. The destination information updating means, when address information specified by code information in the candidate storage means is stored in the destination information storage means, recognizes a customer name by the character recognition means. If the address information specified by the code information in the candidate storage unit is not stored in the destination information storage unit, the address information and the customer name specified by the code information are newly registered. 7. The destination specifying device according to claim 6, wherein: