DE19911116C1 - Method of sorting sequences of transmissions with two-part addresses - Google Patents

Abstract

The method involves automatically reading the destination address acquired at the transmission point for each message and video encoding using the associated image of the message in the event of an ambivalent automatic reading result, sorting at the transmission point and further distribution point before the destination en route according to the read first part of the message, transporting the sorted message to the destination and sorting the messages in sequence at the destination according to the read second part of the destination address and distributing to the addressees. The second parts of the addresses are read automatically using only the possible addressees and the directories contg. their mail address for each destination for address interpretation. If the second part cannot be automatically read unambiguously, at least the second part is video encoded at the destination.

Description

The invention relates to a method and an arrangement for Order sorting of consignments with two-part destination addresses sen from a first, at least indicating the destination Part and a second, the addressee and its dispatch part at the destination. The destination is the last sorting point on the distribution route.

According to the state of the art, the place of dispatch after the task be an image of the recipient's address at the post office and the consignment surface of the Broadcasts recorded. Then the automatic evaluation takes place process based on the image. Do the following steps carried out: determination of the destination address block, segmentation in lines, words and / or single characters and address interpreter tion based on a national dictionary. The Le The process is either carried out entirely at the place of dispatch, if there is enough storage space, even online in one Sorting system, or the shipments are taken after the picture me and reading the zip code with an ID code verse hen and shipped.

The evaluation of the complete recipient address, with a necessary videoco If the automatic Le Sevorganges also carried out in a video coding center can be. Then the address information with the the associated ID marking is transferred to the recipient location, and the consignment in question is sorted accordingly and distributed (EP 0 424 728 B1).

A similar method is also known, in which mail items to be forwarded are preferably video-coded centrally and in which destination addresses are read a second time and in the case of destination addresses which are not clearly read, these are video-coded a second time (DE 196 441 163 A1). The disadvantage here is that the address space for the address reader, under which a single address must be selected from the dictionary, is very large (e.g., nationwide). Wei terhin is disadvantageous that with a centralized video coding the coding forces have no knowledge of the special conditions and the delivery logistics at the destination. Since the video encoders work according to rules that dispel all cases of doubt, the following facts can often be noted:

• - For a large part of the broadcasts are from the video encoder Requires information, d. H. reached a coding depth, who do not need to know the circumstances at the destination would have been agile.
• - The Co The rules often do not rule out, so that a manual follow-up tation becomes necessary.
• - A complex extraction coding is required creation is a considerable effort.

The invention specified in claims 1 and 5 is the Task, a method and an arrangement for the passage to create sequential consignments that reduce Enable video coding effort.

The invention is based on the idea of the address address and the video encoders to be read for evaluation the second part of the destination address by restriction to greatly reduce the destination and thus the relationship to use that under otherwise identical conditions (e.g. consignment quality, evaluation algorithms) the rejection rate depends the information content, which is determined by the logarith distinguishable states.

This relationship exists because at each decision level between two states, i.e. H. possible address variants, a be agreed rejection rate occurs, and the overall rejection rate solution rate with, for example, approximately the same rejection guess themselves from multiplication at each decision level the rejection rates of the decision levels and there  with proportional to the logarithm of the number of differ conditions.

Advantageous embodiments of the invention are the Unteran to take sayings.

So it is possible, according to claim 2, the first part of the goal address at the place of dispatch and the second part of the destination address in automatically evaluate a second reading process at the destination.

Another advantageous embodiment according to claim 3 including an image of the shipment surface at the point of dispatch the destination address to record the shipment with an identifier chen, e.g. B. an identification number in the form of a bar codes, and the image outside of each Read destination.

This reading process can take place in two stages:

• 1. Read the first part of the destination address (zip code) on respective place of dispatch;
• 2. Read the second part of the destination address, e.g. B. also at the sending location or in an evaluation center, to the Ab 1st stage images and reading results are transmitted and transferring the reading result to the destination Sort by.

Furthermore, according to claim 4, the first part of the destination address also using video encoding, both at the sending location or can also be carried out in the evaluation center, be read if the automatic reading in the OCR reader does not gave a clear result.

The invention is described in an exemplary embodiment below explained in more detail with reference to the drawing.

It shows

Fig. 1 is a functional block diagram of the distribution and sorting process with reading of the first part of the destination address at the sending and reading of the second part of the destination address at the destination

Fig. 2 is a functional block diagram of the distribution and sorting process with reading the first part of the destination address at the sending location and reading the second part of the destination address in an evaluation center

First of all, the consignments of a batch are separated and aligned at the point of dispatch. Then the image of the surface of the destination address of each shipment 1 is recorded using a known camera device. The Ziela dresse has a first part characterizing the destination, ie the last distribution station, in the form of z. B. the postcode zip code and a second, the addressee and his shipping address at the destination containing part such. B. Name, street, house number, apartment number, post office box.

This image is then evaluated in a known OCR reader with regard to the postcode / reading the postal code number 2 .

If the automatic reading process, consisting of the known steps: layout analysis (localization of the target address, segmentation), recognition of the individual characters, address interpretation, was successful and led to a clear result, this item is sorted according to the postcode 4 . If the result of the automatic reading device 2 is not clear, the postal code is video-coded 3 and the reading result from the video coding is then used to sort 4 by destination. The entire process is shown in online mode, which means that the item is read and separated into a buffer after it has been separated. After the reading result is available, the consignment is sorted out from the buffer according to the zip code. This ensures the assignment of the shipment and the application of an identification number (ID number) is not necessary. If online operation is not possible in every case, e.g. B. because the reading time is greater than the maximum buffering time, it follows an off-line operation, in which the programs have an ID no. received and thus with a later processing and only then available reading result with the ID no. an assignment can be made.

The distribution of the shipments to the destination with the reading processes can of course take place in several stages. After Sortie ren 4 , the consignments are transported to the destination 5 and there is the automatic reading again before the sequence sorting 8 , but in this case with regard to the second part of the destination address, the identification of the addressee at the destination, e.g. B. Name, street, house number, apartment number or post office box number. The address interpretation during the reading process is carried out with a restricted dictionary that only contains the possible addressees of the destination. The result of this is that the rejection rate drops compared to a country-wide dictionary (as has been customary hitherto), since the distinguishable address states have been considerably reduced and thus also the error or rejection rate. With a realistic reduction from 30% to 10%, only 10% of the programs read have to be video-coded. This video coding 7 also takes place at the destination.

It can be assumed that the deliverer or their Supervisors use the breakpoints for an optimal gear sequence voices. You don't set too many points, but neither do you too few, because they know best what division opti times for delivery. Assuming that this In formation about the optimal gear sequence the reading system (OCR- Address reader and video coding system) as sequence lists are known, and that the video coding forces these Lieferbe zirke know, so the video coding with a maximum throughput and error-free. Another advantage consists of shipments that were previously automatic Processing failed (due to rejections  finally manual processing), again in the automati attributed stream of mail and automatically after the Sort order of gears.

In a further embodiment variant ( FIG. 2), the second part of the destination address is read automatically in an evaluation center.

After separation at the place of dispatch, an ID number is assigned to each shipment 10 and the shipment is identified 11 using this ID number, for example using a barcode printer. There is also an image 12 of the transmission surface with the destination address and then, as already described, an automatic reading 13 of the first part of the destination address, the postcode zip code in a known OCR reader. If the result is ambiguous, the postal code number is video-coded 15 , only this time in an evaluation center. The shipments are then sorted 14 by destination (postcode) and transported to the destination 18 . After the image acquisition 12 , the images in the evaluation center are assigned the ID no. assigned saved 16 . Furthermore, under this ID no. the zip code determined is also saved. At the destination, the ID no. read every broadcast 19 . The read ID no. is transferred to the evaluation center and there it is registered under this ID no. stored image and the associated zip code are retrieved and automatically read with an OCR reader with regard to the second part of the destination address, the addressee identification (name, street, house number, apartment number, post office box) 17 . A restricted dictionary that is only valid for the respective zip code is used, so that the reading rate is higher than that of a global dictionary.

If the result is unambiguous, the corresponding data is sent to a sorter at the destination, where the ID no. identified shipment is sorted according to the course sequence 22 . After the sorting run on the sorter has ended, the mail items are distributed 22 to the addressees. If, on the other hand, the automatic reading of the second part of the target address was unsuccessful, the image is sent to a video coding station at the target location and video-coded there 20 . The result is then also transmitted to the sorter so that the consignment can be sorted with the associated ID number.

Claims (5)

1.Procedure for sorting consignments of consignments with two-part destination addresses, which consist of a first part indicating at least the destination and a second part containing the addressee and his shipping address at the destination, the destination identifying the last sorting point on the distribution route with the Steps:

• 1. automatic reading of the destination address of each shipment recorded at least at the sending location and video coding on the basis of the corresponding image of the shipment if the result of the automatic reading is not clear,
• 2. Sort at the place of dispatch and other distribution points before the destination on the way there according to the first part of the destination address read,
• 3. Transport the sorted shipments to the destination,
• 4. sorting the consignments at the destination according to the second part of the destination address read and distributing them to the addressees,

characterized in that the second parts of the destination addresses are read automatically using the dictionaries containing only the possible addresses and their mailing addresses of the respective destination for address interpretation and, if the second part of the destination address is not clearly automatically read, at least this second part of the destination address is video-coded at the destination. 2. The method according to claim 1, characterized records that at least the first part of the destination addresses at the place of dispatch and at least the second part of the destination dresse is read automatically at the destination. 3. The method according to claim 1, characterized records that each consignment has an identifier at the place of dispatch Chen receives and assigns the identifiers of the shipments images of the destination addresses outside the respective target lortes are read automatically.   4. The method according to any one of claims 1 to 3, because characterized in that at not a clear result of the automatic reading of the first part a video coding of the first part of the destination address Destination address is outside of the destination. 5. Arrangement for performing the method according to claim 1, with

• 1. means for recording the images of the mail item surfaces containing the destination addresses,
• 2. means for automatic reading of the destination address of each shipment recorded at least at the sender and for video coding on the basis of the associated image of the shipment if the result of the automatic reading is not clear,
• 3. Means for sorting at the sending location and further distribution locations before the destination on the way there according to the first part of the destination address read,
• 4. means for sorting the consignments at the destination according to the second part of the destination address read,

characterized in that the second parts of the destination addresses are readable automatically using the dictionaries containing only the possible addresses and their shipping addresses of the respective destination for address interpretation, and if the second part of the destination address is not clearly automatically read, at least this second part of the destination address can be video-coded at the destination.