DE10064529C1 - Mail address post code reading method has address scanned in parts when full address cannot be clearly read - Google Patents

Abstract

The invention relates to a method and a device for reading the postcodes of addresses on mail, said postcode (300) pinpointing the address to a part of a large road or several small roads. After an unsuccessful first read attempt (310), in which addresses could not be read clearly in the OCR unit (130), only the parts of the postcode (300) that roughly identify the addresses are first video-encoded (320) by operators employed for this purpose. The respective encoded result is then fed to the OCR unit (130) for a second read attempt (340), which takes into account the video-encoded result. If the read result of the OCR unit (130) is again unclear, the remaining parts of the postcodes (300) that identify the addresses in more detail are video-encoded (330) by operators employed for this purpose and in order to validate the result are respectively fed to the OCR unit (130) for a third read attempt (350), which takes into account the video-encoded result of the complete postcode (300).

Description

The invention relates to a method and an apparatus ge according to the preambles of the independent claims.

Automatic address reading (OCR) systems are on well known in the field of letter processing and e.g. B. in the DE 195 31 392 C1 described. With modern OCR letter sorting plants are processing rates of 10 letters per second, d. H. 36,000 letters per hour and more reachable. The he however, reliability of identification varies greatly with the Font and overall quality of the surface of the Letters attached address information. In case of success The letter in question can be sufficiently recognized with a ma bar-readable barcode. This barcode he permits further mechanical processing up to one be dear desired sort order. In particular enables the use of barcodes sorting the letters up to the sorting level of the post office, for letters accordingly sorted in the order of their distribution by the deliverer become.

Since the recognition rates of the automatic reading systems are very high riieren, it is necessary to use different forms of manual intervention.

The entered address is encoded and by means of a Directory into a machine-readable sort code changed, which is applied to the shipment.

In this case, video images of the programs are used in video coding systems Operators demonstrated for coding. This can be done online and done offline. With online video coding systems (OVS) the video image is shown to the surgeon while the physical  Consignment is held in delay lines. In these Delay routes are usually the shipment in Bewe held for a period of time sufficient for the OVS operator the necessary sorting information for the enters the image in question. The usual delay lines allow a delay between 10 and 30 seconds. ever the longer the delay line, the greater the costs as well as the maintenance and physical size requirements the plant.

When using OVS, the available one is sufficient Often time only for a careful entry of the postal code (PK) off.

In offline coding systems, such as in US Pat. No. 4,992,649 will be sent to shipments with unrecognized addresses additional information, tracking identification (TID). The programs that are not recognized become external saved while the pictures of these broadcasts Operatoreu are presented for coding, but no temporal ones There are restrictions. Then the shipments Demonstrated TID readers. The TID is entered with the Address information linked. Based on this, a usual barcode sorting information on the shipment be brought so that the shipment in question as usual OCR-read programs can be processed. Even though the offline video encoding method is an effective way to Represents coding of all address components, become additional capacities for further processing with non- mailed addresses and a corresponding complex logistics demands.

In order to further shorten the reading times, WO 97/49503 A1 suggested that the image of everyone within a certain not clearly recognized by video coding complete address information about recognized address parts of the  OCR unit for further automatic evaluation Address interpretation is fed.

This method is used for addresses where only the Destination also as a readable postal code (e.g. post code) is applied to the surface of the shipment.

But there are also countries in which other parts of the address are shown in the postal code. Can this post code then in the OCR unit cannot be read automatically, this is done so far a video coding of the complete post kodes. The input is made with the left and with the right hand.

The invention specified in claims 1 and 10 lies based on the task of the video coding process for reading a Postcode not clearly recognized by an OCR unit ner shipment address, which the address at least to the street eats marks to shorten.

This is achieved by putting the postal code in the address roughly descriptive part and a description of the address the remaining part is divided. Every part is made by for this part intended operators video-coded. First is attempted after an unsuccessful attempt to read the address the OCR unit at the corresponding video coding stations Address roughly descriptive left part of the videoko postal code dated and the OCR unit for a second reading attempt leads. Only if this attempt to read is negative, the remainder of the right part of the postal code is used for this seen surgeons at corresponding video coding stations vi deocoded and the OCR unit receives in addition to the respective the video and the first part of the postcode remaining part of the postal code, so that the corresponding Ab picture of the respective shipment with the transmitted, complete video-encoded postcode in the OCR unit a third Times to validate the postal code.  

The time to video encode each part of the postal code is here at less than if the postal code total of one ope rateur would have been video coded since each surgeon only had one Must enter part of the characters. Another acceleration The effect is due to the fact that after the video coding the left part of the post roughly describing the address codes in the OCR unit with this result a second Le is carried out. If this is successful, the rest of the postal code no longer has to be video encoded the.

Advantageous embodiments of the invention are in the depend represented claims.

It is advantageous if the address that roughly describes the address Part of the postal code, the destination and the address her descriptive part of the postal code the street, partial street on large streets or several small streets.

In a further advantageous embodiment, the video are dated postal part codes based on identification numbers to be assigned to respective programs or their images.

For this it is advantageous to put the identification numbers on the Bring up shipments and read them together with the addresses sen, it being particularly advantageous to identify machine-readable numbers, especially as bar codes gene.

It is also advantageous to take pictures of Sen application surfaces under identification numbers.

To increase the coding speed, it is advantageous to use a preview coding method in which several Ab images of several consecutive programs at the same time can be displayed on the screen.

The invention is then based on the drawings in egg NEM embodiment explained in more detail.

Show

Fig. 1 is a schematic representation of an apparatus for reading Postkodes,

Fig. 2 shows the procedural flow of a flow image.

Fig. 1 shows a schematic representation of a mail distribution system with which the inventive method is carried out who can. The OCR letter sorter 100 consists of a feed device 110 which successively withdraws mail items from a magazine and transports them to a high-resolution video scanner 120 at about 10 mail items per second. The mail items are then transported in a delay line 121 . The programs usually have address information on their surface. In an OCR unit 130, the address information of the images of the programs that were obtained from the video scanner 120 is evaluated. In the case of a complete evaluation, a bar code printer 150 is activated and the consignment is provided with a corresponding bar code for the subsequent sorting into sorting compartments 160 . The OCR unit 130 consists of one or more microprocessors 131 with an associated memory 132 for storing images of the programs. Furthermore, the OCR unit contains an address database 134 with postal codes, city names and street names and possibly further address-related information. When evaluating the images containing address information, a feature-controlled reduction of the entry obtained from the address database 134 takes place, so that a type of partial dictionary is generated. Credibility is assigned in individual entries, so that a number of data from correctly recognized addresses are generated during the evaluation. The device also contains an image controller 170 and a number of video coding stations 200 which are connected directly to the image controller 170 or are connected by a local area network (LAN) 171 . If the OCR evaluation of an image was not, ie completely successful, this image is transferred from the OCR unit 130 to the image controller 170 , which on the one hand controls the TID barcode printer 151 and on the other hand sends the corresponding image to one of the video coding locations 200 . The TID barcode printer 151 applies an identification code TID to the corresponding mailing, which makes it possible at a later time to link the evaluated address information with the physical mailing. In this case, the images are preferably evaluated off-line, although in principle an online evaluation by video coding is also possible with a sufficiently long delay time. In the latter case, the TID can also be applied to the programs at a later point in time, ie if the video coding has not led to a complete evaluation within a specific predetermined time.

To carry out the further method shown in more detail, the image controller 170 is designed such that address information which has not been fully evaluated by the video coding is fed to a further automatic address evaluation using the results of the video coding in the OCR unit 130 .

The recorded and stored image of the respective consignment with the address, containing a postal code 300 , which identifies the destination and the street, the part of a street on large streets or several small streets, becomes a first automatic read attempt by the OCR unit 130 fed. In this example, this postal code 300 is a 6-digit code in canonical form with the following character string ANA NAN, where A is a letter (alphabetical character) and N is a number (numerical character).

The specific implementation of the postal code is of course not on limited to this example.

If the address can be clearly read in the OCR unit 130 in the first read attempt 310 , the recognition task is solved. If not, the image with the post code 300 for video coding the left ANA part, which characterizes the destination, is transmitted to a video coding station 200 , where an operator for the left post code parts video-codes 320 the left ANA part. For this purpose, this left post code be specially marked on the screen of the video coding station 200 , for. B. by a colored background o as shown in Fig. 2 by a border. The coding result is then transmitted to the OCR unit 130 for carrying out a second read attempt 340 with the coded ANA part, in which the search range in the address database is restricted to the NAN part due to the video-coded ANA part. If the post code and address can be read clearly, the recognition process is finished. If this is not the case, the image with the post code for video coding the right-hand NAN part, which characterizes the streets, is transmitted to a video coding station 200 , at which a responsible operator video-coded 330 the right-hand NAN part. This too In this case, the part of the postal code to be coded is advantageously marked on the screen. The result of the video encoding is then transmitted to the OCR unit 130 for a third read attempt 350 . Since this reading attempt 350 is now carried out with the entire entered postal code, the task of the OCR unit 130 in this case is to compare the postal code with the content of the address database and possibly with the other read address parts and to validate whether the postal code 300 has been entered correctly.

The division of the postal code 300 into two parts, which are separately video coded, and the link with the OCR reading process result in the following advantages:

• 1. The shorter sequence of characters to be encoded for an Opera expensive (video coding station) has a much lower error quote and does the preview encoding where the address pictures of several consecutive programs be displayed simultaneously, easier.
• 2. The necessary training time to enter the characters in the Preview mode is greatly reduced.
• 3. By the OCR retry after coding the ANA portion of the postal code can improve reading process can be achieved without the coding rhythm of the surgeon.

Claims (10)

1. A method for reading postal codes of addresses of consignments, in which the images of the addresses have the mailing surfaces recorded, stored and supplied to an OCR unit ( 130 ) for recognizing the addresses and generating corresponding address codes, in which in In the event of ambiguous reading results by the OCR unit ( 130 ), the images in question are video-coded at at least one video coding location ( 200 ) and in which the postal code ( 300 ) identifies the address up to part of a larger street or a number of small streets, where

• - After in the OCR unit ( 130 ) addresses in a first reading attempt ( 310 ) could not be read unambiguously, first the corresponding images of the transmission surfaces with these addresses are transmitted to video coding places ( 200 ), where only the addresses Roughly descriptive parts of the postal codes ( 300 ) are video-coded ( 320 ) by the provided surgeons and then the respective coding result is fed to the OCR unit ( 130 ) for a second reading attempt ( 340 ) including the video de-coding result,
• - And with a again ambiguous reading result of the OCR unit ( 130 ) the images of the transmission surfaces with the appropriate addresses are transferred to further video coding locations ( 200 ), where the addresses describing the addresses of the remainder of the postal code ( 300 ) are coded by the intended operators ( 330 ) and the OCR unit ( 130 ) are each fed to a third reading attempt ( 350 ) including the video coding result of the complete postal code ( 300 ) for validating the result.

2. The method according to claim 1, wherein the part of the postal code ( 300 ) roughly describing the address identifies the destination. 3. The method according to claim 1, wherein the part of the postal code ( 300 ) describing the address identifies the street, part street on large streets or several small streets. 4. The method according to claim 1, wherein the video-coded post partial codes are assigned to the respective programs or images using identification numbers and are combined to complete postal codes ( 300 ). 5. The method of claim 4, wherein the identification number applied to the shipments and together with the Addresses can be read. 6. The method of claim 4, wherein the recorded Figil that of the shipment surfaces under identification numbers can be saved. 7. The method of claim 5, wherein the identification number on the shipments as machine-readable codes be brought. 8. The method of claim 7, wherein the identification number The bar codes can be applied to the items. 9. The method according to claim 1, wherein in the video coding a preview coding method is used. 10. Device for reading postal codes of addresses of consignments with an automatic address reading system, the

• - A device for obtaining images of broadcasts ( 120 ),
• - an OCR unit ( 130 ) for the automatic evaluation of the images of the surfaces of the mailings containing the address information,
• - Video coding stations ( 200 ) for video coding the images containing the address information of the broadcast surfaces,
• - An image controller ( 170 ) for controlling the data flow between the OCR unit ( 130 ) and the Videoko dierplatz, wherein
• - After in the OCR unit ( 130 ) addresses in a first reading attempt ( 310 ) could not be read unambiguously, first the corresponding images of the transmission surfaces with these addresses are transmitted to video coding places ( 200 ), where only the addresses Roughly descriptive parts of the postal codes ( 300 ) are video-coded by the provided surgeons ( 320 ) and then the respective coding result of the OCR unit ( 130 ) is fed to a second reading attempt ( 340 ) including the video de-coding result
• - And with a again ambiguous reading result of the OCR unit ( 130 ) the images of the transmission surfaces with the appropriate addresses are transferred to further video coding locations ( 200 ), where the addresses describing the addresses of the remainder of the postal code ( 300 ) are coded by the intended operators ( 330 ) and the OCR unit ( 130 ) are each fed to a third reading attempt ( 350 ) including the video coding result of the complete postal code ( 300 ) for validating the result.