EP2255894A1

EP2255894A1 - Delivery classification device, method for classifying delivery, program and computer readable recording medium

Info

Publication number: EP2255894A1
Application number: EP09717546A
Authority: EP
Inventors: Daisuke Nishiwaki
Original assignee: NEC Corp
Current assignee: NEC Corp
Priority date: 2008-03-05
Filing date: 2009-03-05
Publication date: 2010-12-01
Anticipated expiration: 2029-03-05
Also published as: JP5447366B2; JPWO2009110538A1; EP2255894A4; WO2009110538A1; EP2255894B1

Abstract

A delivery classification device, and delivery classifying method for reducing mis-sorting when deliveries stored in identifiable postbag are sorted, a program for carrying out the method and a recording medium for recording the program are provided.

The delivery classification device is provided with a registered characteristic data storage means for storing in advance a pair of first image data of one or more deliveries received in a specifiable receiver and address data of the deliveries, wherein the first image data are picked up by an image picking-up means, an image input means for inputting second image data of deliveries received in the receiver, wherein the second image data are picked up by an image picking-up means that is not always the same image picking-up means for the first image data, a retrieval image data storage means for storing the second image data inputted by an image input means, wherein the second image data correspond to all deliveries received in the receiver, and a checking means for retrieving the data stored at the registered characteristic data storage means with the second image data stored by the retrieval image data storage means as a retrieval key, looking for a degree of conformity between the first image data and the second image data, and extracting address data of the deliveries corresponding to the first image data with a high degree of conformity.

Description

TECHNICAL FIELD

The present invention relates to a delivery classification device, delivery classifying or sorting method for classifying or sorting deliveries such as postal matters stored in a storage body like as a postbag, a program for carrying out the method and a computer-readable recording medium for recording the program.

BACKGROUND ART

Conventionally, as this kind of delivery classification device, for example, a delivery classification device has been proposed, in which to the deliveries which are stored in an identifiable postbag, the target addresses of the deliveries are registered in associated with the numerical values calculated from images obtained by imaging the deliveries when the deliveries are collected. And then the deliveries are imaged to obtain another image once again, and the classification device calculates numerical values from the obtained images as a retrieval key for reading the target addresses to perform a classification or sort by the addresses when the deliveries are delivered. (Patent document 1, for example)
Here, a collection time means the time in which deliveries are collected to a collection place from senders, and to be sorted according to each target address of it. There is the delivery classification device at the collection place, and the target address of the delivery is registered in associated with the numerical value calculated from the obtained image by imaging the delivery. Target address data is generated automatically by an optical character recognition apparatus (OCR) at the time of imaging, or is inputted by a person using a keyboard to watch the image of a delivery. Each delivery is sorted to each convenient delivery place according to each target address of it, and is gathered into a postbag according to a prescribed number to each sorting. The postbag is conveyed to a delivery place with other postbags to the same delivery places. Although this postbag is generally managed by the number indicated on it, whether it can be recognized by a machine or not, will depend on the dispatching system of those skilled in the art.
Further, here, a delivery time means the time in which the deliveries are gathered into the postbag to arrive at the delivery place, and are sorted by each town, chome or number. There is also the same machine as the delivery classification device at the delivery place. The delivery classification device at the collection place or the delivery place has the same function, and the registered data can be accessed from the delivery classification devices at said two places. The delivery in the postbag will be imaged by the delivery classification device at the delivery place to take an image of it again, and a numerical data for retrieval will be calculated from the obtained image. The accord numerical data can be retrieved from the numerical data registered at the collection time for using the numerical data for retrieval as a key because the target address data has been stored in associated with the numerical data in advance to deliveries at the collection time. About a mail classifying method, additionally, for example, the patent documents 2, 3 propose it. Further, for example, the delivery is a mail, here, it means a thin individual. For example, a brochure package, a sealed letter and a postcard and the like.

Patent document 1: Japanese domestic publication of PCT international application Number 2003-510183
Patent document 2: Japanese Patent Application Laid-Open No. 2002-216074
Patent document 3: Japanese Patent Application Laid-Open No. 2000-262984

DISCLOSURE OF INVENTION

PROBLEMS TO BE SOLVED BY THE INVENTION

By this delivery classification device described in the above-mentioned patent document 1, the target address of the delivery is obtained by retrieving an accord numerical value among the numerical values calculated from the images obtained by imaging all deliveries registered in the postbag in advance at the collection time by using the numerical value calculated from the image obtained by imaging the delivery in the postbag at the delivery time as a retrieval key. This retrieval is sequentially-processed to all numerical data corresponding to the deliveries in the postbag. The target address registered in associated with the numerical data will be outputted at time immediately when the accord numerical data has been revealed.
However, in case of the method disclosed in the patent document 1, the target address registered in associated with the numerical data is outputted at the time immediately when the accord numerical data has been revealed, even if the delivery is same one, there is a low possibility that the numerical data calculated from the obtained image when it is collected and the numerical data calculated from the obtained image by re-imaging when it is delivered, are accord completely because of the influences of the individual specificity of the imaging device used for imaging the delivery at the collection time and the imaging device used for imaging the delivery at the delivery time, and physical damage like a broken of the delivery itself when it is conveyed, and a rotation of the delivery when it is imaged et, al. This is an ascription that the image obtained at the delivery time and the collection time from the same delivery can not be identical completely. Generally, the numerical data can be identical easily by using a numerical data calculated from the image more than by using the image itself.
In order to judge whether it is identical or not, a method of setting a margin to the difference in the numerical value is considered. However, for this margin, it will occur that the numerical data is possibly identical to the delivery corresponding to the target address of next different delivery again if the numerical data will be outputted at time immediately when the accord numerical data has been revealed like the method in the above-mentioned patent document 1. In fact, in the sequential retrieval, a check of many-to-one will be occurred. At that time, if next different delivery has a higher degree of conformity, it should be judged to be identical with the next one originally, so that a mis-check will be occurred. As a result, there is a problem that a mis-sorting of a delivery will be happened. Further, other patent documents 2, 3 do not describe about the technology which the target address and the image characteristic of the delivery are registered at the collection time, and the target address is extracted from the above-mentioned registered data by using the characteristic of image obtained by imaging the delivery at the delivery time as a retrieval key as described later by the present invention.
The present invention has been made in view of such real state, the object of the present invention is to reduce the mis-check caused by the check of the above-mentioned many-to-one, and to provide a high sorting performance of the delivery classification device, delivery classifying or classifying method , program and a computer-readable recording medium.

MEANS FOR SOLVING THE PROBLEMS

The delivery classification device according to the present invention is characterized by comprising
a registered characteristic data storage means for storing in advance the characteristic of first image data obtained by imaging, more than one deliveries which are stored in an identifiable storage body, by an image picking-up means , and storing the target address data of a delivery as a pair,
an image input means for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means for the first image data, to image the delivery stored in the storage body,
a retrieval image data storage means for storing the second image data inputted by an image input means, in which the image data of all deliveries stored in the said storage body can be stored,
a checking means for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity.
Further, the delivery classifying method according to the present invention is characterized by comprising
a registration step for storing in avdance first image data obtained by imaging more than one deliveries which are stored in an identifiable storage body by an image picking-up means , and the target address data of the delivery as a pair,
an image input step for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means for the first image data, to image a delivery stored in the storage body,
a storing step for storing the second image data inputted by an image input means, in which the image data of all deliveries stored in the said storage body can be stored,
an extracting step for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity, in which the target address of a delivery is extracted after target addresses have been corresponded to all of the deliveries stored in a storage body in this extracting step.
The program of the present invention can be executed by a computer, is characterized by comprising
a registration process for storing in advance first image data obtained by imaging more than one deliveries which are stored in an identifiable storage body by an image picking-up means , and the target address data of the delivery as a pair,
an input process for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means for the first image data, to image a delivery stored in the storage body,
a storing process for storing the second image data inputted by an image input means, in which the image data of all deliveries stored in the said storage body can be stored,
an extracting process for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity.
A computer-readable recording medium of the present invention is a computer-readable recording medium, in which the above-mentioned program is recorded.

ADVANTAGEOUS EFFECT OF THE INVENTION

According to the present invention, mis-sorting of deliveries when they are delivered, can be reduced.

BEST MODE FOR CARRYING OUT THE INVENTION

The embodiment of the present invention will be described with the accompanying Figures as the followings.
The first embodiment of the present invention will be described with reference to the accompanying Figures. Fig. 1 is a block diagram conceptually view of a summary exemplary configuration of the delivery classification or sorting device according to this embodiment. Here, the delivery classification device is at a delivery place. By this delivery classification device, a target address of a delivery which is stored in a storage body or delivery receiver like a postbag sent from a collection place is detected, and the delivery is classified or sorted according to the address.
In Fig. 1, an image data (the second image data) of the delivery imaged by a camera at the delivery place is inputted to an image input unit 401. This image data is stored in the retrieval image data storage DB (database) 301. A numerical data created from the image data (the first image data) obtained by imaging the delivery using the camera at the collection place and the target address data are associated with each other, and are stored into the registered characteristic data storage DB 201 in advance. A check unit 101 creates the numerical data from the image data of the retrieval image data storage DB 301, and retrieves the registered characteristic data storage DB 201 using this numerical data as a retrieval key, and extracts the target address data of the deliveries. A controller 100 controls the whole and has a memory for storing CPU (Central Processing Unit) and programs et, al.
Fig. 2 is a flowchart showing an example of the delivery sorting process to be carried out at the delivery time by the dispatch classification device according to the embodiment of the present invention. This flowchart is a processing procedure of which the CPU executes a software and a program in the delivery classification device shown in Fig. 1.
In Step S1001, the number N of deliveries stored in an identical postbag sent from a collection place is inputted to the image input unit 401. At the collection time, the number N is stored in the registered characteristic data storage DB 201 in advance as shown in Fig. 3, for example. In Fig. 3, the target address data of each delivery, and number 1-N of the deliveries et, al. are registered to one postbag number. This image data is stored in the retrieval image data storage DB 301.
Next, in Step S1002, an image data of a first delivery among the deliveries of number N is inputted to the image input unit 401, and is stored in the retrieval image data storage DB 301. The input order of the image data does not need to be one by which the delivery is stored in the postbag, and it should be able to store the image data of number N in the retrieval image data storage DB 301 at one time.
Next, in Step S1003, the check unit 101 takes out an image data (one of a delivery) from the retrieval image data storage DB 301, and creates numerical data showing the characteristic of this image data. As shown in Fig. 4, the check unit 101 includes a characteristic extraction unit 103 for calculating numerical data, and an identification unit 102 for calculating the degree of conformity of the numerical data, and then extracting numerical data with the highest degree of conformity, and gets correspondence of the address data to all image data of N.
As a creating method of numerical data of the characteristic extraction unit 103, for example, a method described in "high-speed document image retrieval using a digital camera based on local arrangement of a characteristic point, the Institute of Electronics, Information and Communication Engineers of Japan, J89-D, No.9 and pp.2045-2054 (2006), Nakai and others" should be used. Further, the value of the pixel which a picture of image data is composed, can be used. In this case, above-mentioned Step S1003 can be omitted.
Next, in Step S1004, the identification unit 102 extracts a candidate of the numerical data from numerical data which has been registered in the registered characteristic data storage DB 201 according to the highest order of degree of conformity using the numerical data calculated in above-mentioned Step S1003 as a retrieval key. And then, the numerical data with highest degree of conformity should become to be the first candidate, others should be the second candidate, the third candidate according to the high order of the degree of conformity. The numerical data which has been registered in the registered characteristic data storage DB 201, are constituted as shown in Fig. 3.
Here, as the method of calculating degree of conformity, the numerical data of one delivery will be as one vector, for example, the similarity and the interval scale should be used. The similarity corresponds to a cosine (Cos θ )of the vectors, and the similarity will be 1 if the similarity is accord completely. Accordingly, when the similarity is used, it should judge that one near 1 is the one with highest degree of conformity. Although Euclidean distance is general in case of the interval scale, a square root should be calculated to a square sum of a difference between the elements to said vectors. Accordingly, when the interval scale is used, it should judge that one near 0 (zero) is the one with highest degree of conformity. On the other hand, as a numerical data, when image data is used, a normalized correlation technique in addition to the above-mentioned calculating method of degree of conformity, "D.I.Barnca and H.F.Silverman, A Class of Algorithms for Fast Digital Image Registration, IEEE Trans. of Computer, Vol.c-21,No.2,pp.179-186 and 1972." or the like can also be used.
Next, in Step S1005, the identification unit 102 adopts said first candidate as a check result. The input image data and ID of the first candidacy and the value of the degree of conformity are stored in the identification unit 102 as a set.
Next, in Step S1006, the identification unit 102 confirms whether ID has been used (it has corresponded) to other image data or not. However, it is obvious that's not necessarily to need a check for the first one among image data of N. If ID is not used to other image data (step S1006/NO), the identification unit 102 will return to above-mentioned Step S1002, and repeats the step from above-mentioned S1003 to above-mentioned step S1006 to the next image data. Further, if ID has been used (step S1006/YES), the identification unit 102 will come on to Step S1007.
In this case, a plurality of image data will compete for one ID. In order to resolve this competition, the identification unit 102 compares the degree of conformity of the image data which this ID has been assigned first, and the one of the present image data, changes the assigning of this ID to the one with a higher degree of conformity in Step S1007. At that time, when the ID is going to be assigned to the present image data, not to the image data assigned first, the ID to the image data assigned first is changed to next candidate, in this case, is changed to the second candidate.
Next, in Step S1008, the identification unit 102 stores the ID that has been changed to assign again to two of image data of the competition target each other, and degree of conformity to the identification unit 102 as a set.
Next, in Step S1009, the check unit 101 confirms whether the process to all image data has been ended or not. If the process has not ended (step S1009/NO), the check unit 101 will return to above-mentioned Step S1002, and repeats the subsequent steps of above-mentioned Step S1003 to the next image data. If the process has ended (step S1009/YES), the set data of the image data and ID and degree of conformity stored in the identification unit 102 will be outputted.
According to the above mentioned handling process, when the delivery is delivered, the delivery can be sorted based on the address data if the target address data corresponding to the ID assigned to image data is taken out because the input order of the image data is the same as the order which the delivery is supplied to the delivery classification device of this embodiment.
Next, the second embodiment of the present invention is described with reference to the accompanying Figures. Fig. 5 is a flowchart showing an example of the delivery classification device according to this embodiment for carrying out a delivery sorting process when deliveries are delivered. This flowchart shows a processing procedure of software and a program executed by CPU of the delivery classification device shown in Fig. 1,
In Step S2001, the number N of deliveries stored in the identical postbag is inputted to the image input unit 401. At the collection time, the number N is stored in the registered characteristic data storage DB201 as shown in Fig. 3, for example.
Next, in Step S2002, an image data of the first delivery among number N is inputted to the image input unit 401, and is stored in the retrieval image data storage DB 301. The input order of the image data does not need to be the order of which the delivery is stored in the postbag. The retrieval image data storage DB 301 should be able to store image data of N at one time.
Next, in Step S2003, the check unit 101 takes out an image data from the retrieval image data storage DB 301, and creates numerical data from this image data.
Next, in Step S2004, the check unit 101 calculates the degree of conformity of the input image data and all of the numerical data registered in the registered characteristic data storage DB 201, and stores the input image data and the degree of conformity as a set in the check unit 101.
Next, in Step S2005, the check unit 101 confirms the processed number of the image data. If the number has not reached to N (step S2005/NO), the check unit 101 will return to above-mentioned Step S2002, and repeats the steps from above-mentioned Step S2002 to above-mentioned Step S2004. If the number has reached to N (step SS005/YES), the check unit 101 will go to Step S2006.
Next, in Step S2006, the check unit 101 obtains a combination which the total of the degree of conformity calculated from the image data of N becomes to be biggest.
Next, in Step S2007, the check unit 101 confirms whether ID described in the first embodiment is overlapped to assign to plural image data or not. If the ID is overlapped to assign (step S2007/YES), this competition is resolved in Step S2008. Specifically, it should give the ID which is being competed to the image data with a high degree of conformity as described in Step S1007 of the first above-mentioned embodiment. If the ID is not overlapped to assign (step S2007/NO), it goes to Step S2009.
In Step S2009, the check unit 101 outputs the set data of the input image data and degree of conformity stored in the check unit 101 to the input image of N.
By the above mentioned handling process, when the delivery is delivered, the delivery can be sorted based on the address data if the target address data corresponding to the ID assigned to image data is taken out because the input order of the image data is the same as the order which the delivery is supplied to the delivery classification device of this embodiment.
Further, the delivery classification device for realizing the practice of this invention of the above-mentioned first embodiment and second embodiment, can also realize the practice of this invention by which has a configuration as shown in Fig. 6. Fig. 6 shows a delivery classification device which has a check result correction unit 501 newly, and it can be realized by a display apparatus and an input apparatus of a computer. The process step of the check result correction unit 501 corresponds to Step S1007 of Fig. 2 or Step S2008 of Fig. 5, and an operator resolves the above-mentioned competition. The competing image data is shown on a display apparatus of a computer, specifically, shown on a display and simultaneously, a table of Fig. 3 stored in the registered characteristic data storage DB 201 is shown on it together. An operator can see it to input right corresponding relationship by an input apparatus like as a mouse, a keyboard specifically.
An exemplary configuration of the delivery classification device according to this embodiment is shown in Fig. 7. The delivery classification device includes a supply unit 601 for supplying a delivery taken out from a postbag to the scanner unit 602,
a scanner unit 602 having a camera for obtaining an image data from the delivery supplied from the supply unit 601,
an image check unit 603 for extracting a corresponding target address data by using the image data taken out by the scanner unit 602 by the means described in first and second above-mentioned embodiment,
a sorting unit 604 for sorting deliveries according to each address using the target address data extracted by the image check unit 603, and
a carrier 605 for transferring deliveries received from the scanner unit 602 to the sorting unit 604.
Deliveries including 1 of the postal matters (delivery) (1)... N of the postal matters (1), are stored in the postbag (1).
The delivery classification device obtains the target address of these deliveries, sorts them according to the address, and stores them to sorting sets 1, 2... M, and transfers them. If an electronic tag (RF tag) and a barcode are given to a postbag, the number of the postbag should be read automatically by the supply unit 601. If a numeral and symbol printed in a postbag are used, that should be inputted manually. For this, for example, as shown in Fig. 8, the supply unit 601 should include a postbag number input unit 611 and delivery supply unit 621, The delivery supply unit 621 takes out the delivery from the postbag, and supplies it to the scanner unit 602, and this can be realized by a supply unit of the existing delivery classification device. The postbag number input unit 611 of deliveries should use a usual electronic tag (RF tag) reader or a barcode reader. Further, when the numeral and symbol printed in the postbag are inputted, an optical character recognition apparatus can be used. Furthermore, when those are inputted manually, an operator should input by a keyboard, and also, it is possible that a sound of the numeral and symbol printed in the postbag read by an operator, can be inputted according to sound recognition.
All of the deliveries stored in the postbag are sorted after all of the target address data has been obtained by the delivery classifying method of this embodiment. Accordingly, a storage means for holding data of all deliveries stored in the postbag in the computer is needed, and simultaneously, a holding means for holding deliveries until address data is obtained to all deliveries stored in the postbag, should be also needed.
In order to secure these means, for example, in the carrier 605, a delay unit should be installed for securing time until the target address data has been obtained to all deliveries stored in the postbag. This delay unit should adopt to make the belt to be longer until it reaches to the sorting unit 604 if the carrier 605 is composed by a transfer belt. Or, the moving length of the packet should be done lengthily if each of the delivery is stored into a packet at the carrier 605 to transfer, and also, a conveying path of a packet may be composed like a loop.
A different exemplary configuration of the delivery classification device according to this embodiment is shown in Fig. 9. A temporary collection unit 606 for temporarily collecting deliveries of which the image data has been extracted by the scanner unit 602, is installed newly to the configuration of mentioned above Fig. 7. By this configuration, the time until the target address data has been obtained to all deliveries stored in the postbag, can be secured by collecting a delivery temporarily by installing the temporary collection unit 606. About the sorting, it should secure the space for piling up the deliveries one by one because the scanned order by the scanner unit 602 should not be changed. Specifically, for example, deliveries which have been scanned by the scanner unit 602 will be guided to the above-mentioned space in the carrier 605 in turn. By opening a bottom, deliveries are supplied to the sorting unit 604 without changing order when the deliveries are supplied to the sorting unit 604. The above-mentioned space may be reversed for supplying. In this case, the corresponding with the target address data in the sorting unit 604 should be reversed because the order is reversed.
A more different exemplary configuration of the delivery classification device according to this embodiment is shown in Fig. 10.
A scanner array unit 607 instead of the scanner unit 602, is installed newly to the configuration of mentioned above Fig. 7. The scanner array unit 607 is composed by installing a plurality of scanner units 602 in line, and can obtain image data to a plurality of deliveries at one time. It can be realized by conveying deliveries from the supply unit 601 to each scanner unit in the carrier 605. In this configuration, the time until the target address data has been obtained to all deliveries stored in the postbag, can be secured by securing a plurality of paths to the sorting unit 604 of the deliveries in parallel.
The first effect by the embodiment of the present invention is to reduce mis-sorting of the delivery when it is delivered. The reason is that the sorting of deliveries in the postbag is not performed to each of the delivery sequentially, the retrieval is performed after the correlating of address data to all deliveries in the postbag has been completed by extracting the target address data of the delivery corresponding to the numerical data with high degree of conformity.
The second effect is the improvement of the sorting rate when deliveries are delivered. The sorting rate is that the obtained check number of each delivery and target address data is divided by the total number of the deliveries processed when the deliveries are delivered. It is a right sort if each delivery can obtain the address data of the delivery itself, otherwise it is a mis-sort.
The following works out.
The sort rate = right sort rate + mis-sort rate. The Reason of the improvement of the sorting rate is that, for example, there is one hundred deliveries in a postbag, if the corresponding address data to the ninety-nine deliveries of it has been determinate, it is possible to judge the address data of the remaining one delivery to the remaining one delivery which the correspondence can not be obtained.
Further, a computer program for performing the process of the flowchart shown on Fig. 2 and Fig. 5 and a recording medium in which this program is recorded, compose a program according to the present invention and a computer reading and recording medium. As a recording medium, a storage device such as a readable and writable memory or a hard disk apparatus can be used. The objects of the present invention can be achieved by supplying the above-mentioned program, reading it to the check unit 101 including CPU of the machine, and carry out it. Further, the object of the present invention can be also achieved by supplying this program to an apparatus and a system different from the machine of the present invention.
Although the postbag as the storage body, postal matter as the delivery has been described, the present invention can be also applied to other deliveries, storage body.
This application is based on and claims the benefit of priority from Japanese Patent Application No. 2008-055057, filed on March 5th, 2008 , the disclosure of which is incorporated herein in its entirely by reference.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a block diagram showing a summary exemplary configuration of the delivery classification device according to the embodiment.
Fig. 2 is a flowchart showing an example of operation of the delivery classification device according to the embodiment.
Fig. 3 is a configuration diagram showing the structure of the data stored in the registered data storage DB according to the embodiment.
Fig. 4 is a block diagram showing a summary exemplary configuration of the check unit according to the embodiment.
Fig. 5 is a flowchart showing an example of operation of the delivery classification device according to the embodiment.
Fig. 6 is a block diagram showing a summary exemplary configuration of the delivery classification device according to the embodiment.
Fig. 7 is a block diagram showing an exemplary configuration of the delivery classification device according to the embodiment.
Fig. 8 is a block diagram showing a summary exemplary configuration of supply unit according to the embodiment.
Fig. 9 is a block diagram showing a different exemplary configuration of the delivery classification device according to the embodiment.
Fig. 10 is a block diagram showing a more different exemplary configuration of the delivery classification device according to the embodiment.

DESCRIPTION OF THE CODES

100: Controller
101: Check unit
102: Identification unit
103: Characteristic extraction unit
201: Registered characteristic data storage DB (database)
301: Retrieval image data storage DB
401: Image input unit
501: Check result correction unit
601: Supply unit
602: Scanner unit
603: Image check unit
604: Sorting unit
605: Carrier
606: Temporary collection unit
607: Scanner array unit
611: Postbag number input unit
621: Delivery supply unit

Claims

A delivery classification device is characterized by comprising:
registered characteristic data storage means for storing in advance first image data obtained by imaging more than one deliveries which are stored in an identifiable storage body by an image picking-up means , and the target address data of the delivery as a pair,

image input means for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means, to image a delivery stored in the storage body,

retrieval image data storage means for storing the second image data inputted by an image input means, in which the image data of all deliveries stored in the said storage body can be stored, and

checking means for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity.
The delivery classification device according to claim 1 is characterized in that the checking means extracts the target address data of the delivery after the target addresses have been corresponded to all of the deliveries stored in said storage body.
The delivery classification device according to claim 1 is characterized in that an electronic tag is used for the identification of said storage body.
The delivery classification device according to claim 1 is characterized in that a barcode is used for the identification of said storage object.
The delivery classification device according to any one of claims 1 to 4 is characterized in that the checking means comprises:
characteristic extraction means for creating a numerical data showing the characteristic of said second image data,

identification means for performing a check using said numerical data created by said characteristic extraction means.
The delivery classification device according to any one of claims 1 to 5 is characterized by further comprising:
check result correction means for correcting a check result of a delivery stored in said storage body.
The delivery classification device according to any one of claims 1 to 6 is characterized by further comprising:
temporary collection means for temporarily collecting a delivery which the image input has finished by said image input means.
The delivery classification device according to any one of claims 1 to 7 is characterized by installing a plurality of said image picking-up means in line, and imaging a plurality of deliveries at one time.
A method of sorting delivery is characterized by comprising:
registration step for storing in advance first image data obtained by imaging more than one deliveries which are stored in an identifiable storage body by an image picking-up means, and the target address data of the delivery as a pair,

image input step for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means for the first image data, to image a delivery stored in the storage body,

storing step for storing the second image data inputted by an image input means, in which the image data of all delivery stored in the said storage body can be stored, and

extracting step for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity, in which the target address of a delivery is extracted after target addresses have been corresponded to all of the deliveries stored in a storage body in this extracting step.
The method according to claims 10 is characterized in that said extracting step comprising:
characteristic extraction step for creating a numerical data showing the characteristic of said second image data, and

identification step for performing a check using said numerical data created by said characteristic extraction means.
The method according to claim 10 is characterized by further comprising:
correction step for correcting a check result of a delivery stored in said storage body.
A program can be executed by a computer, is characterized by comprising:
registration process for storing in advance first image data obtained by imaging more than one deliveries which are stored in an identifiable storage body by an image picking-up means , and the target address data of the delivery as a pair,

input process for inputting second image data obtained by an image picking-up means that is not always the same image picking-up means for the first image data, to image a delivery stored in the storage body,

storing process for storing the second image data inputted by an image input means, in which the image data of all delivery stored in the said storage body can be stored, and

extracting process for taking out a degree of conformity of the first image data and the second image data by retrieving the registered characteristic data storage means using the second image data stored by the retrieval image data storage means as a retrieval key, and extracting the target address data of a delivery corresponding to the first image data with a high degree of conformity.
The program executed by a computer according to claim 12 is characterized in that said extracting process comprising:
characteristic extraction process for creating a numerical data showing the characteristic of said second image data,

identification process for performing a check using said numerical data created by said characteristic extraction means.
The program according to claim 13 can be executed by a computer, is characterized by comprising:
correction process for correcting a check result of a delivery stored in said storage body.
A computer-readable recording medium in which a program of any one of claims 12 to 14, is recorded.