JP4005676B2 - Destination verification device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a destination verification device used when sorting deliveries.
[0002]
[Prior art]
The operation of sorting the delivery items is performed manually, and in recent years, automation has been promoted. In order to sort, it is necessary to read and recognize the destination, but there are many other things that cannot be read. ing.
[0003]
The manual addressing process is slower and more expensive than a machine because it is done by humans. Moreover, this work is not reduced and must be performed as long as the installed automatic sorting apparatus is in operation. Therefore, an apparatus that reduces this cost is proposed.
[0004]
In other words, the results input by the operator are registered in the extended database, so that know-how of people who cannot be imitated by machines can be obtained. Destination verification device that can be stored and expected to reduce processing that requires an operator Is desired.
[0005]
[Problems to be solved by the invention]
In this invention, the results input by the operator are registered in the extended database, so that know-how of people who cannot be imitated by machines can be obtained. Destination verification device that can be stored and expected to reduce processing that requires an operator The purpose is to provide.
[0006]
[Means for Solving the Problems]
A destination collation apparatus according to the present invention includes a destination reading unit that reads a destination of a delivery and outputs a read character string and an image, a destination dictionary that stores the destination, an output unit that displays whether the destination is read successfully, and an input Manual correction means for determining a destination matching the destination in the image by searching the destination dictionary for the destination that matches the destination in the image and specifying the character string entered by the operator from the destination dictionary; The extended destination dictionary that stores the record of the data that can be corrected by the correcting means, and whether or not the character string input from the destination reading means exists in the destination dictionary or the extended destination dictionary. Is output to the output means, and only when it does not exist, the manual correction means and all the data input from the destination reading means are searched. The highest score A collation unit for outputting the destination, and when the character string is specified by the manual correction unit as a result of reading the destination, the character string before correction by the manual correction unit and the destination of the character string after correction Corresponding addresses in the dictionary are recorded in the extended destination dictionary so that they can be used for the next collation by the collation means.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, a destination matching apparatus according to an embodiment of the present invention will be described with reference to the drawings.
[0009]
(First embodiment)
FIG. 1 shows a block of a destination collation apparatus for explaining a first embodiment of the present invention.
[0010]
FIG. 1 is a block diagram of a destination collation apparatus that promotes the automation of sorting work by registering and utilizing an incorrect destination and an accurate destination corresponding to this when assigning a read destination to a delivery item. With reference to FIG. 1, the flow of processing of the destination matching apparatus will be described.
[0011]
First, the outline of the apparatus will be described. In FIG. 1, first, using the destination reading device 1, an image of a delivery item and characters related to the destination described in the delivery item are acquired. The characters are labeled so that they can be identified as addresses or names. In order to check whether the address and name are correct, the collation device 2 searches the address and name using the destination dictionary 3 and the extended destination dictionary 5. If the searched address name can be specified by either the destination dictionary 3 or the extended destination dictionary 5 as the two dictionaries, the address name is correct, and the destination information is sent to the output device 6. If the result of the search cannot be specified, either the read result is incorrect or the written address is incorrect, so the read result and the image of the delivery destination (usually the image of the delivery surface) and the search The result (candidate remaining finally) is transferred to the manual correction device (VCS) 4. The operator infers an accurate destination from the transferred image, the final candidate, and the data in the destination dictionary 3, and transfers the destination to the output device 6. Further, the destination information inferred by the operator is stored in the extended destination dictionary 5 to be used for the next search. By using this apparatus, the frequency of using the manual reader gradually decreases, and the cost associated with correction can be reduced.
[0012]
Next, each device will be described in detail.
[0013]
The address reading device 1 is an apparatus that uses the mail address reading device described in Japanese Patent Application Laid-Open No. 5-217015 to decompose an optically read image into characters and configure it as a destination. The character string configured as the destination is transferred to the collation device 2 together with the image. The transferred image and its character string are shown in FIGS. The character string is output after being distinguished into a part that is the address of the destination and a part that is the name of the destination.
[0014]
The collation device 2 is a device that identifies a corresponding delivery destination from the destination dictionary 3 and the extended destination dictionary 5 based on the character string transferred from the destination reading device 1.
[0015]
The collation process by this collation apparatus 2 is demonstrated using the flowchart of FIG. 3, FIG.
[0016]
That is, a character string and image data are acquired from the destination reading device 1 in step 1. In step 2, this image data is temporarily stored in a buffer (not shown). In step 3, a search is performed with the address name of the destination. After determining whether the search is successful or unsuccessful in step 4, if the search is successful, the fact that the search was successful in step 5 is output and the process ends. If it fails in step 4, the failure in step 6 and the data remaining as candidates until the end, the input character string and the image in the buffer are output to the manual correction device 4 and the process ends.
[0017]
The search process is shown in steps 7 to 12. In step 7, candidate data is extracted from the address dictionary 3. The extraction method is to extract data including a certain amount of characters in the character string to be searched. In this case, the address field of the destination dictionary 3, the address equivalent part of the input data, the name field, and a part of another article (Oguchi name, Oguchi name 2) are compared with the name equivalent part of the input data to check the number of characters hit.
[0018]
In step 8, the extracted data group is weighted. The weight is given according to the number of hit characters. In step 9, these candidates are narrowed down and data with the largest weight is extracted. If there is one data extracted in step 10, the search is successful and the process proceeds to step 11. If there is no data extracted in step 10, it cannot be specified, and the process proceeds to step 12.
[0019]
In a successful step 11, the identified result and the identified result are output and the search is terminated. In step 12, it is checked whether or not there is a data record that matches all the input character strings in the extended destination dictionary 5 that has accumulated experience so far. The result (the destination obtained by subtracting the destination dictionary 3 from the record number on the destination dictionary 3 included in the subtracted record) is output in step 11. If it does not exist, it means that the search has failed, the fact that it failed in step 14 and the result narrowed down in step 9 are output and the process ends.
[0020]
Whether the character string inputted from the destination reading device 1 in the collation device 2 exists in the destination dictionary 3 or the extended destination dictionary 5 is matched with the inputted character string and the character string to be searched. There is only one candidate in the two dictionaries of the destination dictionary 3 and the extended destination dictionary 5 to be searched. However, in the case where the IDs of candidate destination dictionaries 3 subtracted from the extended address dictionary 5 are common, it is not counted that there are a plurality of candidates.
[0021]
The destination dictionary 3 stores a pair of an address of the target district and a list of residents living there. The format of data to be stored is shown in FIG. The record number, address, name, and other items included in the ID are included. From the name of the prefecture to the issue, the name includes the head of the household, the family name, and the other item includes a large flag, large name, building name, and room number. included.
[0022]
The manual correction device 4 is a device in which the operator manually determines the destination when the collation device 2 cannot identify the destination. This apparatus has a display unit for displaying an input image and a character string, and an input unit (mainly a keyboard) for inputting a destination. Note that FIG. 10 described later corresponds to an example of the display unit.
[0023]
Manual correction processing by the manual correction device 4 will be described with reference to the flowchart of FIG.
[0024]
That is, in step 21, the collation device 2 inputs a character string related to the destination, the image of the delivery, and the result of the narrowed destination. In step 22, an image of the delivery is displayed, and at the same time, narrowed-down destination candidates are displayed. The operator inputs in step 23 whether any of the displayed candidates is correct or not at all. When data in the candidate is selected (step 24), the character string related to the input destination and the record number of the selected candidate are sent to the extended destination dictionary 5 and recorded (step 25). After this, the candidate selected in step 29 and success are output to the output device 6 and the process is terminated.
[0025]
If it is input in step 23 that there is no correct answer in the candidate (step 24), the process proceeds to step 26, and the operator inputs a character string that seems to be correct. In step 27, this character string is searched from the destination dictionary 3, and if it exists in step 28 (successful search), the character string input in step 21 in step 25 and the record number of the destination dictionary 3 selected by the operator are stored. It is sent to the extended destination dictionary 5 and recorded. If the search is not successful in step 28, the process proceeds to step 29 and the fact that the search is unsuccessful is transferred to the output device 6 and the process ends.
[0026]
Note that the search method is equivalent to the method drawn in step 1 of FIG. 4 of the flowchart showing the operation of the collation device 2.
[0027]
The extended destination dictionary 5 stores a character string obtained as a result of reading by the collation device 2 and a corresponding record number of the destination dictionary. The format of data to be stored is shown in FIG. The output device 6 is a device for confirming the collated result. Usually, it is a CRT or a printer. Further, the delivery items can be quickly classified by linking with the output device 6 and the delivery sorting device that actually sorts the delivery items.
[0028]
This is the end of the description of each device. Next, the operation of the present apparatus will be described using the specific example of FIG. When the delivery shown in FIG. 8 is input to the destination reader 1, the destination reader 1 outputs the image corresponding to FIG. 8 and the character string shown in FIG. In this figure, these characters are displayed so that the address and name can be seen. When this character string is input to the collation device 2, the collation device 2 uses the destinations stored in the destination dictionary 3 and the extended destination dictionary 5 as the address “7, Yanagimachi, Saiwai-ku, Kawasaki-shi” and as the name “Taro Shiba” Search for "sama".
[0029]
The search method is based on matching of all character strings, but suffixes such as “Sama” and “Tono” are not included in the match. In addition, a score is given based on how much the input character string is hit, and a destination group (final candidate) having the best score is held. In this case, if the destination dictionary 3 includes the address "70 Yanagimachi, Saiwai-ku, Kawasaki-shi, Kanagawa" and the name "Taro Shiba", this score has 10 characters out of 15 characters. Therefore, it becomes 67 points. Similarly, if the address includes “17, Yanagimachi, Saiwai-ku, Kawasaki-shi” and the name includes “○ Taro Shiba”, the same score is displayed (see FIG. 10).
[0030]
After searching the target address name in the destination dictionary 3, if the corresponding destination cannot be narrowed down, the extended destination dictionary 5 is searched. In this example, it is assumed that the corresponding destination is not found in the extended destination dictionary 5 under the all character matching condition. At this time, the automatic collation has failed, and the image of the deliverable as shown in FIG. 8, the character string shown in FIG. 9, and the candidate group (final candidate) having the highest score are input to the manual correction device 4.
[0031]
The manual correction device 4 has a display device as shown in FIG. 10, and displays the input image, character string, and final candidate. After display, the operator selects the correct candidate when it exists. In this case, select number 1 that seems to be correct. At this time, the character string of the read result and the record number on the selected destination dictionary 3 (not displayed on this display unit but set to 134525) are stored in the extended destination dictionary 5. An example of the stored record is shown in FIG. In addition, the success and its destination are output to the output device 6 and the process ends.
[0032]
If the correct answer is not in the final candidate, select failure. At this time, a new address search is performed by the operator. The search screen at this time is shown in FIG. As shown in the figure, when the operator inputs 70 Yanagimachi, Saiwai-ku, Kawasaki City, and Taro Shiba in the respective fields, the search is performed from the destination dictionary 3 under the all-character matching search condition. In this case, the prefecture name is not included, but since the score becomes 100 by interpolation, this is displayed. If there is no corresponding candidate as a result of the execution, the failure is selected, the failure of the output device 6 is displayed and the process ends.
[0033]
If successful, the record number “134525” on the destination dictionary 3 and the character string of the read result are stored in the extended destination dictionary 5 as shown in FIG. The destination is output to the output device 6 and the process ends.
[0034]
When succeeded, from the next time, there will be data registered in the record number “010253” in the extended dictionary 5 when a delivery with the address: 7 Yanagimachi, Saiwai-ku, Kawasaki City, name: ○ Taro Shiba Therefore, this can be automatically collated and output to the output device 6 without bothering the operator.
[0035]
Therefore, according to the first embodiment described above, the operator's know-how that cannot be imitated by the machine can be stored by registering the result input by the operator in the extended database, as compared with the conventional manual correction device. it can. By accumulating and utilizing this know-how, work that requires an operator is reduced. This increases the number of parts to be processed by the machine and enables high-speed processing. In addition, personnel costs can be reduced.
[0036]
(Second Embodiment)
A second embodiment of the present invention will be described below with reference to FIG.
[0037]
Figure 13 teaches the correct address to the sender who posted the delivery at the wrong destination, thereby reducing the number of erroneous deliveries from the root and promoting the automation of the sorting work. It is a block diagram of the destination collation apparatus aiming at doing.
[0038]
With reference to FIG. 13, the processing flow of the destination matching apparatus will be described.
[0039]
First, an outline of the destination verification device will be described.
[0040]
In FIG. 13, first, an extended destination reader 21 is used to acquire an image of a delivery item and characters related to the destination attached to the delivery item. Unlike the address reading device 1 of the first embodiment, not only the address name of the address but also the address name of the sender is read. In order to check whether the address and name are correct, the collation device 2 searches for the address and name of the destination using the destination dictionary 3.
[0041]
If the searched address name exists in the destination dictionary 3, the destination is correct and the destination information is sent to the output device 6. If the result of the search cannot be specified, either the read result is incorrect or the written address is incorrect, so the read result and the image of the delivery destination (usually the image of the delivery surface) and the search Is transferred to the extended manual correction device 22. The operator infers an accurate destination from the transferred image, the data in the destination dictionary 3 and the final result, and transfers the destination to the output device 6.
[0042]
Further, the operator transfers the read sender's address name and recipient's address name to the reply label creating device 23, and creates a label indicating that the sender has a wrong address. Creating and posting a delivery with this label contributes to reducing fundamental address mistakes, gradually reducing the frequency of using a manual reader, and reducing the cost of correction.
[0043]
Next, each device will be described in detail.
[0044]
The extended destination reading device 21 is an extension of the destination reading device 1 and transfers an image and a character string to the collation device 2, but the data to be output as the character string is different, and the address name of the sender is output in addition to the destination. To do. Examples of images and character strings output from the extended destination reader 21 are shown in FIGS. The character strings are output with meanings assigned to the destination address, destination name, sender address, and sender name, respectively.
[0045]
The extended manual correction device 22 is a device obtained by expanding the manual correction device 4 in which an operator manually specifies a destination when the destination cannot be specified by the collation device 2. The portion where the extended manual correction device 22 is different from the manual correction device 4 is a portion that does not create data for the extended destination dictionary 5 but instead outputs the source and destination toward the reply label creation device 23. The extended manual correction device 22 has a display unit for displaying an input image and a character string, and an input unit such as a keyboard for inputting a destination. An example of the display unit is shown in FIG. The extended manual correction process of the extended manual correction device 22 will be described with reference to the flowcharts of FIGS.
[0046]
That is, in step 31, the image of the delivery item and all the character strings are input from the collation device 2 to separate what each data is. In step 32, the image, the reading result, and the final candidate are displayed, and the operator's input in step 33 is prompted. The operator confirms whether or not the corresponding data exists in the final candidate, and if it exists, selects the corresponding number.
[0047]
In step 34, it is confirmed whether or not the operator's input is “failure”. If anything other than “failure” is selected, the process proceeds to step 35, and flag 1 for determining success or failure is set to OK. The candidate selected at the same time is set as the correct destination. Thereafter, after confirming the state of flag 1 in step 41, if it is OK, the process proceeds to step 42 to check whether the sender's address name is present in the input from the collation device 2.
[0048]
When it exists, it progresses to step 43 and outputs a sender address name and an exact destination toward the reply label production apparatus 23. Further, the process proceeds to step 44 and the success is output to the output device 6 and the process is terminated. If the sender's address name does not exist in step 42, there is no destination even if it is input to the reply label creating apparatus 23, and therefore, the process proceeds to step 44. Only the fact that the search is successful is output to the output apparatus 6 and the process ends. .
[0049]
If the operator determines in step 33 that the corresponding address does not exist as a final candidate and “failure” is input, the process proceeds to step 36 in step 34 where the operator inputs an accurate name and searches (step 37). The search method is the same as that of the manual correction device 4. If the search is successful, it is determined in step 38 that the search is successful, the process proceeds to step 40, flag 1 is set to OK, and the search character string is set as the correct destination. Thereafter, the process proceeds to step 41 and the operations from step 42 to step 45 as described above are performed.
[0050]
If it is determined in step 38 that the search is unsuccessful, flag 39 is set to NG in step 39, then the process proceeds to step 41, and the state of flag 1 is checked. A message to the effect is output to the output device 6 and the process is terminated.
[0051]
The reply label creating device 23 is a device that creates a label for notifying the sender that the destination is wrong, and is a label printing device that outputs two sets of a destination label and a text label as a set. The reply label creating device 23 is divided into a data shaping unit, a printing unit, and a clock. The data shaping unit is a device that processes the input data so that the printing unit can print, and the printing unit is equivalent to a normal printer. . Therefore, the description of the operation of the printing unit and the clock will be omitted, and the data shaping unit will be described using the flowchart of FIG. 18 and the printing examples of FIGS.
[0052]
That is, when the sender's address name and correct address are acquired from the extended manual correction device 22 in step 51, an image of the address label as shown in FIG. In FIG. 19, a solid line portion 31 indicates a printing range, a dotted line portion 32 indicates an address field, and a dotted line portion 33 indicates a name field. The sender address and the sender name are adjusted and imaged in this frame, respectively.
[0053]
In step 53, a text label image as shown in FIG. 20 is created. A solid line portion 41 indicates a printing range, and a fixed form document other than the name of the destination address is entered. The dotted line part 42 shows an address field, and the dotted line part 43 shows a name field. Each field contains the address and name of the exact destination. The dotted line 44 is filled with a date obtained from a clock inside the apparatus as a date / time field.
[0054]
After each image is created, each image is input to a printing unit inside the apparatus and output as a label in step 54. Labels are affixed to postcards and delivered, so that attention is paid to how to fill in deliverables from the next time and correct addresses are filled in. As a result, manual labor can be reduced.
[0055]
Since the collation device 2, the destination dictionary 3, and the output device 6 are the same as those in the first embodiment, the description thereof is omitted.
[0056]
This is the end of the description of each device. Next, the operation of the present apparatus will be described using the specific examples of FIGS. When the delivery shown in FIGS. 14A and 14B is input to the extended address reader 21, the extended address reader 21 matches the image and character string shown in FIGS. Output to. In FIGS. 14A and 14B, these characters are displayed so that the address and name can be understood. When this character string is input to the collation device 2, the collation device 2 searches the address stored in the destination dictionary 3 for the address in the name of Yanagimachi, Saiwai-ku, Kawasaki-shi as the address, and for Mr. Taro Shiba as the name. The search method is the same as the method described in the first embodiment. In this example, it is assumed that the automatic collation has failed, and the image and character string shown in FIGS. 14A and 14B are input to the extended manual correction device 21.
[0057]
The extended manual correction device 21 has a display device as shown in FIG. 15, and displays an input image and a character string. At the same time, the final candidate sent from the verification device 2 is displayed. It is assumed that after the display, an input by the operation evening is performed, and in this case, No. 2 is selected. At this time, the extended manual correction device 22 sends the information of the successful identification to the output device 6 at the same time, and the address name (exact destination) of the second candidate selected and the extended destination reader 21 in FIG. a) The sender's address name shown in (b) is transferred to the reply label creating device 23.
[0058]
The reply label creating device 23 prints the name and address of the input sender address and the correct destination on the label as shown in FIGS. 19 and 20, and ends. When “failure” is selected by the operator, a search by an operator input is performed as in the case where the operator selects “failure” in the manual correction device 4, and processing for specifying a destination is performed.
[0059]
The search method is the same as that of the manual correction device 4 and is omitted. By attaching the printed label to a postcard or the like and delivering it, attention can be drawn from the next time, and as a result, the frequency of bothering the operator is reduced.
[0060]
Therefore, according to the above-described second embodiment, it is possible to expect a reduction in processing that requires an operator by alerting the sender of a delivery that has been difficult to read, from the next writing. In addition, by correcting the entry of the sender's address and name, it is possible to reduce the number of deliveries of lost children whose destinations are unknown and whose return origin is unknown, and it is possible to reduce the costs associated with storing these lost children's deliveries.
[0061]
【The invention's effect】
As described above in detail, according to the present invention, the results input by the operator are registered in the extended database, so that know-how of a person who cannot be imitated by a machine can be obtained. Destination verification device that can be stored and expected to reduce processing that requires an operator Can provide.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of a destination collation apparatus for explaining a first embodiment of the present invention;
FIG. 2 is a diagram showing an image transferred by the destination reading device shown in FIG. 1 and its character string.
FIG. 3 is a flowchart for explaining collation processing by the collation apparatus shown in FIG. 1;
4 is a flowchart for explaining collation processing by the collation apparatus shown in FIG. 1; FIG.
FIG. 5 is a diagram for explaining the format of data stored in the destination dictionary shown in FIG. 1;
FIG. 6 is a flowchart for explaining manual correction processing by the manual correction device shown in FIG. 1;
7 is a diagram for explaining the format of data stored in the extended destination dictionary shown in FIG. 1. FIG.
FIG. 8 is a diagram showing a specific example of a delivery used in the destination matching apparatus shown in FIG. 1;
FIG. 9 is a diagram showing a character string transferred by the destination reading device shown in FIG. 1;
FIG. 10 is a diagram showing a display example on the display device of the manual correction device shown in FIG. 1;
FIG. 11 is a diagram showing a display example on the display device of the manual correction device shown in FIG. 1;
12 is a diagram for explaining an example of data stored in the extended destination dictionary shown in FIG. 1. FIG.
FIG. 13 is a block diagram showing a schematic configuration of a destination collation apparatus for explaining a second embodiment of the present invention.
FIG. 14 is a diagram illustrating an image output by the extended destination reading device illustrated in FIG. 13 and its character string.
FIG. 15 is a diagram showing a display example on the display unit of the extended manual correction device shown in FIG. 13;
FIG. 16 is a flowchart for explaining an extended manual correction process of the extended manual correction device shown in FIG. 13;
FIG. 17 is a flowchart for explaining extended manual correction processing of the extended manual correction device shown in FIG. 13;
FIG. 18 is a flowchart for explaining reply label creation processing of the reply label creation device shown in FIG. 13;
FIG. 19 is a diagram illustrating a print example of a destination label in the reply label creating apparatus illustrated in FIG. 13;
FIG. 20 is a diagram illustrating a print example of a text label in the reply label creating apparatus illustrated in FIG. 13;
[Explanation of symbols]
1 ... Destination reading device (destination reading means)
2 ... Collation device (collation means)
3 ... Destination dictionary
4. Manual correction device (manual correction means)
5 ... Extended destination dictionary
6 ... Output device (output means)

Claims (6)

From the destination reading means for reading the destination of the delivery and outputting the read character string and image, the destination dictionary storing the destination, the output means for displaying the success or failure of the destination reading, and the input image and character string, Manual correction means for determining a destination that matches the destination in the image by searching and specifying the input character string and the character string input by the operator from the destination dictionary, and data corrected by the manual correction means Check whether the extended destination dictionary for storing the record and the character string input from the destination reading means exist in the destination dictionary or the extended destination dictionary, and output the result to the output means, Further and a collating means for outputting all the data and the search result most score for higher was the destination that is input from the destination reading means in said manual correction means only when no, destined When the character string is specified by the manual correction means as a result of reading the character string, the extended destination is associated with the character string before being corrected by the manual correction means and the address of the corrected character string in the destination dictionary. A destination collation apparatus characterized in that it is used for collation from the next time by the collation means by being recorded in a dictionary. 2. The destination collation apparatus according to claim 1, wherein all characters of the character string to be retrieved coincide in the retrieval by the collating means.   2. The destination collation apparatus according to claim 1, wherein the record of data in the extended destination dictionary is a destination record number searched from the destination dictionary and a character string of a read result.   A score indicating whether or not the character string inputted from the destination reading means in the collating means exists in the destination dictionary or the extended destination dictionary and the degree of coincidence between the inputted character string and the character string to be searched 2. The destination matching apparatus according to claim 1, wherein there is only one candidate having a score equal to or higher than a predetermined score in the two dictionaries of the destination dictionary to be searched and the extended destination dictionary. .   5. The destination matching apparatus according to claim 4, wherein the score equal to or higher than the predetermined value is 80% or more. 2. The destination matching apparatus according to claim 1, wherein the manual correction unit specifies one of a plurality of character strings.

Images