JPH01161575A - Money order processing system - Google Patents

Abstract

PURPOSE:To make the investing amount to a whole money order processing system lower and to prevent wrong inputs to the system by transferring read money transfer requesting data to the processor of a money order center and completing transaction data by causing an operator to make a supplementary input. CONSTITUTION:A money transfer requesting slip 1 is read by an OCR 2 and an unrecognizable character is converted into a specific code. An OCR adaptor 3 stores the received data after compression and, at the same time, calls a district center. The call is connected to the MPX/TRC 8 of the processor 6 of the money order center through a public network 4 and EPBX 5. At the processor 6, a CPU 9 decodes the received data and stores the decoded results in a file section 11. At a data extending section 12 the data are recognized by a Kanji (Chinese character) character recognizing section 13 after the data are reproduced. Regarding the unrecognizable character, an intelligent process is performed in such a way that the CPU 9 retrieves the file 11 and deduces a relevant character from the context. The operator of the district center supplementarily inputs an unrecognizable character, etc., through a keyboard and makes a slip completed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an exchange processing method that automatically reads and processes transfer request form data used in transfer processing of exchange transactions in financial institutions.

[Conventional technology]

A major part of the operations of financial institutions such as banks is processing related to deposits, loans, and foreign exchange transactions. Among these, foreign exchange transactions involve financial institutions intervening between the client and the payee or payer, and the processing of receivables and exchange transactions between the two parties. As a debt settlement intermediary, it plays an important and essential fund settlement function in the economy and society.

On the other hand, with the development of the economic trading society, the volume of foreign exchange transactions has been rapidly increasing, and the daily volume of domestic exchange transactions has reached 1.5 million transactions, especially on the last day of the month, when 3 million transactions are handled. The issue has also come up.

By the way, it is well known in this industry that there are two types of foreign exchange: teleexchange and document exchange, but teleexchanges using the gold-silver system account for the majority of transactions.

Teleexchange includes remittances, transfers, salary transfers, and payment collections, and most of these are handled at the counters of financial institutions, but over 80% of the total amount of exchanges handled is through remittances. The increase rate is 10% compared to the previous year.
%, there is an urgent need to streamline transfer operations at financial institutions' branches.

FIG. 8 is an explanatory diagram showing an example of the mechanism of transfer operations currently carried out by banks under such circumstances, and this will be explained below.

(1) First, the sender (transfer requester) fills out the transfer request form by hand with the necessary information such as the recipient and the amount to be transferred, and then submits the transfer request form along with the transfer funds and fees to the sending branch (sales branch). Make a transfer request.

(2) The sending store that has received the request confirms the contents of the transfer request form and issues a full transfer receipt to the transferee in exchange for the transfer funds and fees.

(3) The sending store sends a money transfer notification to the destination store.

This process is actually inputted at the destination store terminal, district center or destination bank center, sent to the gold and silver center, and sent from the gold and silver center via the destination bank center to the destination store (the other party's business branch). sent to.

(4) The destination store receives the exchange notification from the destination store.

(5) The destination store deposits the transfer funds into the recipient's deposit account based on the exchange notice received from the destination store.

The above is the transfer processing procedure for foreign exchange transactions, but in this series of steps, data entry processing (1) and (3) in particular requires human labor, and there is an urgent need to streamline this data entry processing. It becomes.

In addition, the data in (1) and (3) above include numbers,
Although kana is used, there are cases where recipients differ depending on homophones in the name, so a processing system using kanji has been introduced. Data input has become complicated, resulting in a decline in processing capacity.

Therefore, conventional measures to address these issues include developing multi-functional automatic transaction machines (multi-functional ATMs) that add a transfer processing function to automated equipment, or centralizing administrative processing instead of processing transfer request forms at branch offices. We have a FAX10CR system that uses an optical character reader (OCR) to collectively process transfer request forms sent by facsimile (FAX) to district centers, etc., and received at the district centers, etc., and handwritten kanji OCRs are installed at branches for data processing. Handwritten kanji OCR
We have begun to introduce systems, but each has the following problems.

[Problem that the invention seeks to solve]

First, in the case of the multi-function ΔTM, data input processing is performed using a display with a touch panel, but since it takes a long time to operate, it is not suitable for large-volume processing, and since transfer processing is only one function of the device, it is difficult to perform counter operations. Although it will reduce the burden, we will use a multi-functional ATV to further streamline the transfer process.
The problem is that adding more ATVs is not worth it considering that ATVs are expensive.

In addition, with the FA Since there is no function to detect data errors in data transmission control, if noise generated from the communication network is output as dirt etc. on the transfer request form received by fax, the character recognition rate when data is processed using OCR. There is a problem that the amount decreases to 60 to 70%, making it impractical.

As a way to improve the character recognition rate, it is possible to combine OCR with a knowledge processor and read pre-registered kanji combinations such as bank names, branch names, etc. from a word dictionary file, but in this case. But FA
If the numbers or katakana on the received transfer request slip are smudged, it may cause misreading, and especially misreading the transfer amount will cause serious consequences, so there is a problem of lack of reliability.

On the other hand, in the case of handwritten kanji OCR, the handwritten kanji recognition unit is expensive, so the equipment investment is high and it is not worth it. Since supplementary inputs such as, etc., are required, if the handling volume is large, the workload of the operator at the branch increases, and there is a problem that it becomes difficult to cope with the situation, especially on peak days.

The present invention was made in order to solve these problems, and its purpose is to realize a money exchange processing system that is excellent in transfer processing ability, investment effectiveness, reliability, and operability.

[Means for Solving the Problems] In order to achieve the above-mentioned object, the invention of the present application is installed in a branch of a financial institution, and optically records data on transfer request slips in which data for exchange transactions are written by hand. A branch terminal device that has an image reader section that reads data automatically, a character recognition section that recognizes characters from the data read by the image reader section, and a data transfer section that transfers the character code and image data recognized by the character recognition section. and a data transmitting/receiving unit installed at a district center that handles administrative processing for multiple branches, connected to the branch terminal, and a character recognition unit that recognizes characters from data transferred from the branch terminal. The system includes a money exchange center processing device having a bank transfer center, and a workstation placed under the control of the money transfer center processing device and having an input section and a display section for data supplementation.

Another invention of the present application includes a determination unit that determines whether or not a management code is printed on the transfer request form based on data read by an image reader unit; A printing unit for printing information is added to the branch terminal device.

[Effect]

The former invention having the above-mentioned configuration reads the data entered in the first area of the transfer request form using the branch terminal device, performs character recognition, and reads the character recognition result and the second area of the transfer request form. The read image data of the third area is transferred to the exchange center processing device.

The exchange center processing device performs character recognition based on the read image data of the second area, and also sends the character recognition results of the first and second areas and the image data of the third area to the workstation. furthermore, the workstation displays the image data of the third area and the character recognition results on a display section;
The operator completes the input transaction data of the transfer request form by using the input unit to input the data entered in the third area and the unrecognizable characters that could not be recognized by the branch terminal device.

Further, the data of the completed transfer request form is transferred to the gold and silver center via the exchange center processing device, and the gold and silver center notifies the branch of the counterparty's financial institution.

In the latter invention, the branch terminal device first determines whether or not a management code is printed at a predetermined position on the transfer request form, and if the management code is already printed, error processing is performed.

If the management code has not been printed yet, the data entered in the first area of the transfer request form is read and the characters are recognized, and the control code is printed at the position of the transfer request form by the printing section, and then Similar to the former invention, the input transaction data is completed and finally notified to the branch of the counterparty's financial institution.

As explained above, according to the present invention, handwritten kanji OCR
The data of the transfer request form read by the branch terminal device is transferred to the money transfer center processing device, and supplementary input is performed by the operator at the workstation under the control of this money transfer center processing device, and the data of the transfer request form is processed without using Since the transaction data is completed, the terminal equipment at the branch is low-cost, and the investment amount for the entire system can be kept low. erroneous input can be prevented and reliability is improved.

In addition, since the transaction data for the transfer request form is completed on the workstation as described above, the only operation required by the operator when inputting the transfer request form on the terminal device at the branch is to set the transfer request form to the OCR. In addition to being highly efficient, it can significantly reduce the work burden on branch operators, making it possible to greatly contribute to overall rationalization and to be able to adequately handle foreign exchange transactions on peak days.

In addition, most of the supplementary input by the operator at the workstation is unrecognized characters and the names (addresses) of the transfer requestor and recipient.
It is possible to prevent incorrect input due to large numbers of kana characters being corrected, which improves data reliability, and the ability to continuously process transfer request forms on branch terminals improves processing capacity. can be increased.

Furthermore, in the present invention, a management code can be printed on the transfer request form, and the management code can be checked on the branch terminal, so that double transfers can be prevented, and
When the contents displayed on the display section are unclear, the operator of the workstation can reliably request the branch office to resend the data of the transfer request form whose contents are unclear or inquire about the contents using the management code.

〔Example〕

Examples will be described below with reference to the drawings.

Figure 1 is a connection diagram of the system used in the first embodiment of the present invention, Figure 2 is a format diagram showing an example of the transfer request form in Figure 1, and Figure 3 shows the functions of the main devices in Figure 1. The block diagram shown in FIG. 4 is an explanatory diagram showing an example of the format of data communicated by the OCR adapter of FIG. 3.

In the same, 1 is a transfer request form, 2i is a small optical character reader (OCR) for desktop use, and 3 is an OCR adapter.The 0CR2 and OCR adapter 3 are used as a branch terminal device to be used at the counter of a branch such as a bank. Multiple machines are installed, and each operator (teller) who is a counter staff operates the machine.

The transfer request form 1 has the format shown in FIG. 2,
The O1△ and × marks in the diagram are the client's entry areas. Here, the O-marked area is the number/kana character area (first area) to write the date, transfer amount 2, account number 1 name in furigana, and the Δ-marked area is the kanji character to write the transfer destination bank name and branch name. The area (second area) and the area marked with an X are image areas (third area) where the names (addresses) of the recipient and client are written in kanji.

Current handwritten character recognition technology can achieve a recognition rate of over 99% for numbers and kana characters, and 0CR2 as an automatic character input device is at a practical level, but compared to that, kanji characters are difficult to recognize in terms of font and size. Since there are considerable individual differences in the recognition rate, the recognition rate is low, which poses significant practical limitations.

On the other hand, the combinations of bank names and branch names corresponding to the Δ marks are determined, so if all bank names and branch names are registered in advance in the file section of the exchange center processing device, which will be described later, it will be possible to Even if there is an unrecognizable character among the characters, it is possible to compensate for the unrecognizable character based on the relationship between the characters.

For example, if it is a "Hi?, Tani" branch, ? The character is "ratio"
It is easily inferred that this is the case.

Therefore, by adding the above-mentioned knowledge processing to handwritten kanji recognition, the bank names and branch names marked with Δ can be made to a level that is sufficiently practical.

In the past, kanji processing technology has been rapidly advanced in order to prevent receivers from misrecognizing homonyms caused by data communication of kanji written on a transfer request form l in kana characters. In this case, it is necessary to convert the name (address) of the recipient and the client in the X-marked area into kanji code data, but the technology for handwritten character recognition in the X-marked area is different from the processing in the Δ-marked area. As already mentioned, the actual situation is that the system has not reached the level of practical use, and the operator has no choice but to make supplementary inputs.

On the other hand, 0CR2 is the transfer request form l as shown in Figure 3.
an image reader unit 2a for optically reading the data; a character recognition unit 2b for converting alpha numeric kana characters (ANK characters) into character codes; Hopper part 2
c. It is equipped with a stacker section 2d for accumulating the read transfer request form 1, an operation section 2e1, an interface section 2f, and an OCR control section 2g for controlling these, and when the transfer request form 1 is inserted, Each area is read, the area marked with ○ is converted into character code data based on a preset document format, and the area marked with Δ and the area marked with X are each transferred to the OCR adapter 3 as image data. It has become.

However, in order to make it easier to perform OCR processing on the transfer request form 1, the formats of the areas marked with ◯, the areas marked with Δ, and the areas marked with × are changed.

Further, the OCR adapter 3 has an interface section 3a connected to the interface section 2f of the OCR2.

A file section 3b, a data compression section 3c, an operation section 3d, and a communication section 3e as a data transfer section connected to a public network to be described later.
, and an OCR adapter control unit 3f that controls these, and when data is received from the 0CR2, the image data therein is compressed at a specific compression rate by the data compression unit 3c, and then all the data is transferred to the file unit 3b. Store it in
At the same time, it instructs the OCR control unit 2g to read the next transfer request form 1.

Here, the file capacity of the file section 3b of the OCR adapter 3 can store a maximum of 8 copies of the total data of about 100 KB including the image data of the transfer request form 1, and this is the buffer capacity that enables continuous reading of the transfer request form 1. It is fulfilling its role.

Further, after storing the data of the first transfer request form l in the file section 3b, the OCR adapter 3 starts the process of receiving the data of the next transfer request form l, but at the same time communicates with the higher-level device, that is, the exchange center processing device, which will be described later. Data communication is performed via the line in a specific format shown in FIG. 41.

As shown in the figure, this format 1- consists of a control command and the first to fourth blocks, where the first block is an ANK area for account number, date, amount, etc., and this ANK area is recognized as a control code. Contains character code data, and the second block is A when there are unrecognized characters.
The NK area contains control codes and unrecognizable character image data.

The third block is a kanji area for the bank name and branch name, and this kanji area includes control codes and image data.
Furthermore, the fourth block is an image area for the name (address).
This image area includes control codes and image data.

In the case of a device configuration such as ``-'', the character recognition of 0CR3 is only handwritten ANK characters, and the handwritten ANK version used in the recognition unit 2b is extremely inexpensive compared to the kanji board. In the case of a kanji board, the recognition unit is expensive, and a supplementary input device is required, and the adapter is also considerably more expensive than the OCR adapter 3 shown in this embodiment. This is contrary to the fact that the number of terminal devices installed in a branch office is large, so they need to be priced low in terms of investment efficiency. In other words, the device configuration of this embodiment has good investment efficiency and is advantageous as a terminal device installed in a business office.

In addition, in the device configuration of this embodiment, since the transfer request form l is directly read using 0CR2, it is possible to
In addition to obtaining an extremely high character recognition rate compared to the R system, in the conventional F A There will be no misunderstanding.

Furthermore, in this embodiment, the buffering function of the OCR adapter 3 makes it possible to read the transfer request form 1 continuously, resulting in high processing capacity and, in the case of data input, the operator at the counter of the branch can easily read the transfer request form 1 at the OCR2. The operation is extremely simple since all you need to do is set the transfer request form 1.

In addition, all the data of the transfer request form 1 is divided into multiple blocks and communicated by the data transmission control of the OCR adapter 3, but data conversion due to noise etc. is detected by an error detection function (not shown) and the block data is retransmitted. in,
It also becomes possible to perform data communication reliably.

Now, let's return to the explanation of FIG. 1 again.

In FIG. 1, 4 is the communication section 3 of each OCR adapter 3.
This public network 4 is compatible with FAX network, telephone network, digital data exchange network (DDXIM), or service integrated digital & M (ISDNWI), and is a high-speed digital network with a maximum speed of 64 KI3/S. Can communicate.

5 is the above public 11M? Electronic exchange (EPB) connected to I4
X), 6 is a money transfer center processing device, and 7 is a work station (WS), which are installed at a bank's district center or business center (receiving bank center).

Nowadays, medium-sized and larger banks have established regional centers to process the administrative work generated at their branches by region, with the aim of streamlining administrative processing through centralized processing. This district center normally handles office work for 10 to several dozen sales offices, and is located between the sales offices and the office center.

The exchange center processing device 6 includes a multi-channel communication section (MPX/TRC) 8 as a data transmission/reception section, a central processing unit (CPU) 9, a memory 10, a file section 11, a data decompression section 12, a kanji character recognition section 13, and an IMB. /S high-speed workstation interface (WS/TF) 14
, a communication section 15, and an operation section 16, and is connected to the EPBX 5 through the MPX/TRC 8, and connected to a host such as an office center through the communication section 15.

Furthermore, the WS7 is a data supplementary input device consisting of a keyboard as an input section, a display as a display section, a data decompression section, etc., and is connected to the WS/IF 14 in plurality and is under the control of the currency exchange processing device 6. each WS7 edits specific format data,
When compressed image data is received, the data is decompressed by a data decompression unit at a specific decompression rate, reproduced as a character image, and displayed on a display. It has a function of completing the transaction data by the operator inputting necessary supplementary data by operating the keyboard, and transmitting the completed transaction data to the exchange center processing device 6.

The number of WS7 installed is determined by the number of cases handled at the district center or administrative center, and the number of WS7 installed is determined by the number of cases handled at the district center or administrative
The number of channels in C8 is determined by the number of branches, communication speed, number of transactions, etc.

Next, the exchange processing method implemented by using such a system will be explained.

First, at the sales branch (destination branch), the operator at the counter receives the transfer request form 1 brought by the employee.
After confirming the details, issue a full transfer receipt to the chain in exchange for the transfer funds and fees.

Then, when the transfer request form 1 is accumulated for a predetermined time or a certain amount, the operator sets it in the hopper section 2C of the 0CR2 and processes it.

That is, the image reader section 2a of the OCR1 is
Each area of the transfer request form L is read according to the instructions of the I-Gori section 2g, and the read data in the area marked O shown in Fig. 2 is converted into character code data by the character recognition section 2a and recognized based on the preset form format. Unrecognized characters are converted into image data using a specific code, and the character code data and image data of unrecognized characters are transferred to the OCR adapter 3, and the read data in the area marked with Δ and the area marked with X are each converted into image data. OCR
Transfer to adapter 3.

After the reading is completed, the transfer request form 1 is accumulated in the stacker section 2d.

When the OCR adapter 3 receives data from the 0CR2, the data compression unit 3C compresses the image data at a specific compression rate, stores all the data in the file unit 3b, and at the same time sends the next transfer request form to the OCR control unit 2g. 1 and call the main number of the district center or office center.

This call signal is sent to the EPBX5 via the public M44, and the representative number is processed by this EP[3X5 and connected to an empty channel of the MPX/TRC8 of the exchange center processing device 6.

Thereafter, the data of the transfer request form l stored in the file section 3b is read out by the OCR adapter control section 3f, and is communicated from the OCR adapter 3 to the exchange center processing device 6 as specific format data shown in FIG. Ru.

The exchange center processing device 6 converts the specific format data of the transfer request form 1 received from the OCR adapter 3 into a CP.
tJ 9 deciphered the 1st block with 1 transfer request form~
The fourth block is divided into blocks and stored in the file section 11 via the memory 10 as original data for the transfer request form.

Thereafter, similarly, the format data sent from the OCR adapter 3 is sequentially stored as the original transfer request form data.

Here, the file section 11 is a large-capacity file of several 10 to several 100 megabytes, and can store several 100 to 1000 copies of the original data file of the transfer request form 1, as well as 7 aisles of word dictionaries for kanji recognition. , a character dictionary file, and an exchange data logging file for host-to-host communications.

When the original data of the transfer request form is stored in the file section 11, a call is made from the exchange center processing device 6 to the WS7 via the WS/IF 14, which allows the operator to
If you perform the specified operation in step 7, that operation will be used as a kick and CP
U9 starts processing the transfer request form original data in the file section 11 in the order in which they are stored.

That is, first, the data of the third block of the first transfer request form l is extracted, and the data decompression section 12 performs specific data decompression to reproduce the image data of the bank name and branch name in the 0CR2, and then the kanji characters are reproduced. It is transferred to the recognition unit 13. □ This Kanji character recognition unit 13 recognizes Kanji characters such as bank names and branch names, and sends them back to the CPU 9 as Kanji code data. If there is a character that cannot be recognized at this time, the character is The specified code is returned.

Based on this, the CPU 9 searches the word dictionary file in the file section 11, and performs knowledge processing to determine the applicable character from the relationship before and after the unrecognizable character.

For example, as mentioned above, if it is recognized as the "Hi?Tani" branch, then if it is recognized as the "Hi?Tani" branch, then the "R"? It can be seen that the "Tani" branch indicates the "Hibiya" branch.

In this way, in this embodiment, the image data obtained by directly reading the transfer request form 1 using 0CR2 is transferred to the OCR adapter 3.
This is compressed by the exchange center processor 6 and decompressed by the exchange center processing device 6 to accurately reproduce the original image data.
Even if the Kanji character recognition unit 13 is located in a remote location, it is possible to obtain the same Kanji recognition rate as that of the device built into the 0CR2.

Now, following the above processing, if there is data of the first block and second block, these are extracted together with the data of the fourth block, and the data of the first block, the data of the second block, and the third block as described above are extracted. The processed Kanji character code and the data of the fourth block are transferred to the kicking WS7.

As a result, the WS block follows the format of the transfer request form l, determines the code data for the data of the first block and the kanji character code, and displays the corresponding character font on the display, and also displays the data of the second block and the 4th block. After the data is expanded at a specific data expansion rate by a data expansion section, it is displayed in a specific area of the display as image data.

That is, the characters in the areas marked with ○ and Δ shown in FIG.
The built-in character font of S7 is displayed, and the characters in the area marked with an x are displayed as image data of the handwriting written by Yorito.

Here, the operator at the district center or office center looks at the displayed image data and uses the keyboard to perform supplementary input operations such as characters in the X-marked area and unrecognized characters.
Complete the data entry for the transfer request form l.

This completed transfer request form data is sent again from the WS 7 to the exchange center processing device 6 and stored in the transaction data file of the file section 11 through processing by the CPU 9.
At this time, a serial number is also given for communicating with the sending bank center.

Thereafter, the transfer request form data is read out from the transaction data file by the verification WS of the executive (not shown), and after being verified by the executive, it is sent to the sending bank center via the communication section 15, and then sent to the gold and silver center. will be sent as a money order notice.

This money transfer notice is then processed at the gold and silver center and then notified to the other party's business branch, that is, the destination store, and the destination store deposits the transferred funds into the recipient's savings account based on the received money order notice. .

By the way, the transfer request form l is normally read by the image reader section 2a of the 0CR2 in one transaction process and the data is sent to the exchange center processing device 6, after which it is stored at the branch and is used for transactions again. It is never used for processing. However, the transfer request form 1 whose data has been sent to the exchange center processing device 6 gets mixed up with the unsent transfer request form 1 due to a clerical error, etc.
If the R2 reading process is performed, a situation of double transfer will occur.

In addition, due to reasons such as low density of characters written by the customer on the transfer request form, character recognition and image reading are insufficient.
The data is sent to the exchange center processing device 6, and as a result of the above-mentioned processing being performed in the exchange center processing device 6, the WS7
If there is any ambiguity in the content displayed on the display, the operator operating the WS7 may make an inquiry to the branch. In this case, it is necessary to accurately report to the branch which transfer request form 1 applies, and to quickly process the subsequent process, but the reality is that no effective measures have been taken at present. be.

Next, a second embodiment of the present invention which is effective against these disadvantages will be explained with reference to FIGS. 5 to 7.

Fig. 5 is a block diagram showing the functions of the main devices in the second embodiment, Fig. 6 is a format diagram showing the transfer request form used in the second embodiment, and Fig. 7 is a diagram showing the characteristic features of the second embodiment. 2 is a flowchart showing the operation.

In the system used in this embodiment, as shown in FIG. 5, the OCR control unit 2g of the 0CR2 functions as a determination unit that determines whether or not a management code is printed on the transfer request form l, which will be described later. A printing section 2h controlled by the OCR control section 2g is added, the operation section 3d of the OCR adapter 3 has a mode setting switch for mode setting, which will be described later, and the OCR adapter control section 3f is used as a management code. As shown in FIG. 6, there is a management code area 14 for printing the serial number as the management code on the transfer request form l.
The configuration differs from that of the system used in the first embodiment in that the system is provided with , but the other configurations are the same.

Next, the operation of the second embodiment will be explained with reference to FIG. Note that the following explanation will be made corresponding to the step indicated by "S" in FIG.

First, when the transfer request form 1 is set in 0CR2, the -th transfer request form 1 is sucked in, and the management code area 14 of the transfer request form 1 is read by the image league section 2a (Sl).

Based on the reading result, the OCR control unit 2a determines whether or not a serial number is printed in the management code area 14 (S2), and if it is determined that it is not printed, the image reader unit 2a subsequently requests a transfer. The data in all areas of the ticket 1 is read (S3), and then the serial number sent from the OCR adapter control section 3f to the OCR2 is printed in the management code area 14 by the printing section 2h (S4).

The OCR adapter control unit 3f has one port, for example, “
It manages a series of numbers from 0001" to 9999", and each time all the data on the transfer request form is received from 0CR2, one number is transferred to 0CR2, and based on this, the transfer request form is sent as described above. A serial number is printed in the management code area 14 of l.

Thereafter, the compression process of the data read from the transfer request form 1 is performed in 0CR2 as in the first embodiment, and
The data is transferred to the CR adapter 3 (S5).

At the same time as this data transfer, the ocRffilI control unit 2a determines whether or not there is the next transfer request form 1 based on the detection data of a sensor (not shown) provided in the hopper unit 2c (36). When it is determined, an instruction to read the next transfer request form 1 is sent from the OCR control section 2a to the image reader section 2a (S7), and the processing from S1 is thereby repeated.

The data transferred to the OCR adapter 3 undergoes the same processing as in the first embodiment and is communicated to the exchange center processing device 6. After further processing by the exchange center processing device 6, the data is transferred to the WS7.
In this series of processing, the serial number is included in the data of the first block in the format shown in FIG. will be displayed on the display when doing so.

On the other hand, in S2, when the OC'R control section 2a determines that a serial number is printed in the management code area 14, the OCR adapter control section 3f then changes the mode of the mode setting switch (not shown) to retransmission mode. As a result, if it is determined that the retransmission gear is not set, error processing such as ejecting the transfer request form 1 from 0CR2 is performed, and it is determined that the retransmission mode is set. When this happens, data in all areas of the transfer request form 1 is read by the image reader section 2a as in S3 (S9), and then the process moves to S5 and S6.

In other words, if the content displayed on the display is unclear when performing supplementary input on WS7, the operator calls the branch with the serial number of the transfer request form 1 whose contents are unclear, and then the operator at the branch is OCR adapter 3
In this case, even if it is determined in s2 that a serial number is printed, it will not be treated as an error and the transfer request form will be processed. 1 data is read again and the read data is retransmitted.

This process prevents double transfers, and also allows the WS7 operator to reliably request the branch to resend the data or inquire about the details if the displayed content is unclear. .

In this second embodiment, the serial number or management code of the transfer request form 1 is read using the existing image reader section 2a.
However, it is possible to provide a separate reading section exclusively for management codes without charge.

Further, in the first and second embodiments described above, the knowledge processing of unrecognized characters that could not be recognized by the branch terminal device was performed by the exchange center processing device 6, but it is assumed that the knowledge processing of unrecognized characters that could not be recognized by the branch terminal device is performed by the branch terminal device. Furthermore, in the first and second embodiments described above, both the exchange center processing device 6 and the WS 7 decompress the compressed data. It is also possible to do so.

〔Effect of the invention 〕

As explained above, the present invention reads the data of the transfer request slip with the branch terminal device, transfers the read data to the WS through processing by the exchange center processing device, and transfers the read data to the WS through processing by the exchange center processing device.
In S, the operator performs supplementary input to complete the transaction data on the transfer request form, and also prints a control code on the transfer request form so that this control code can be checked on the branch terminal. The following effects can be obtained.

First, without using handwritten kanji OCR, the data of the transfer request form read by the branch terminal is transferred to the exchange center processing equipment, and supplementary input is performed by the operator at the workstation under the control of the exchange center processing equipment. Since the terminal equipment at the branch is low-cost, the investment amount for the entire system can be kept low, and the account number in the conventional FAX10CR system can be reduced. This has the effect of preventing erroneous input of information such as transfer amount, etc., and improving reliability.

In addition, since the transaction data for the transfer request form is completed on the workstation as described above, the only operation required by the operator when inputting the transfer request form on the terminal device at the branch is to set the transfer request form to the OCR. In addition to being highly flexible, it can significantly reduce the workload of branch operators, greatly contributing to overall rationalization, and also having the effect of being able to fully handle foreign exchange transactions on peak days. It will be done.

In addition, most of the supplementary input by the operator at the workstation is unrecognized characters and the names (addresses) of the transfer requestor and recipient.
It is possible to prevent erroneous inputs due to large numbers of kana characters being corrected, which improves the reliability of the data, and in conjunction with the fact that it is possible to continuously process transfer request forms on the terminal device at the branch. This system has the effect of increasing processing capacity, and also has the advantage of allowing a flexible system configuration because it is possible to increase the number of exchange center processing devices and WSs depending on the amount of exchange transactions handled.

Furthermore, in the present invention, a management code can be printed on the transfer request form as described above, and this management code can be checked on the branch terminal, thereby preventing double transfers. If the content displayed on the display is unclear, the workstation operator can use the management code to reliably request the branch office to resend the data of the transfer request form whose content is unclear or inquire about the content. This effect can also be obtained.

[Brief explanation of the drawing]

Fig. 1 is a connection diagram of the system used in the first embodiment of the exchange processing method according to the present invention, Fig. 2 is a format diagram showing an example of the transfer request form in Fig. 1, and Fig. 3 is the main part in Fig. 1. A block diagram showing the functions of the device, FIG. 4 is an explanatory diagram showing an example of the format of data communicated by the OCR adapter in FIG. 3, FIG. 5 is a block diagram showing the functions of the main device in the second embodiment, FIG. 6 is a format diagram showing a transfer request form used in the second embodiment, and FIG. 7 is a flowchart showing characteristic operations in the second embodiment.
FIG. 8 is an explanatory diagram illustrating an example of a mechanism for transfer operations carried out at a bank. 1: Transfer request form 2: Optical character reader (OCR) 2a: Image reader section 2b: Character recognition section 2c: Hopper section 2d Nistats force section 2g: 0CRII
I? 11 part 3: OCR adapter 3b: file part 3c: data compression part 3f:
OCR adapter control unit 6: Exchange center processing device 7: Workstation (WS) 9: Central processing unit t! (CPU) lO: Memory 11: File section 12: Data expansion section 13: Kanji character recognition section 14: Management code area Patent applicant Oki Electric Industry Co., Ltd.

Claims (1)

[Scope of Claims] 1. An image reader unit installed in a branch of a financial institution that optically reads data on a transfer request form on which data for exchange transactions is handwritten, and data read by the image reader unit. A branch terminal device that has a character recognition unit that recognizes characters, a data transfer unit that transfers the character code and image data recognized by this character recognition unit, and a branch terminal device that is installed at a district center that handles administrative processing for multiple branches. is,
A money exchange center processing device having a data transmission/reception unit connected to the branch terminal device and a character recognition unit that recognizes characters from data transferred from the branch terminal device; He is equipped with a system consisting of a workstation having an input section and a display section for data supplementation, and reads the data entered in the first area of the transfer request form with the branch terminal device, recognizes the characters, and reads the data. The character recognition result and the read image data of the second area and the third area of the transfer request form are transferred to the money exchange center processing device, and the money transfer center processing device reads the characters based on the read image data of the second area. At the same time as performing the recognition, the character recognition results of the first and second areas and the image data of the third area are transferred to the workstation, and at the workstation, the image data of the third area and the character recognition are transferred to the workstation. A currency exchange processing method, characterized in that the input transaction data of the transfer request form is completed by displaying the result on a display section, and supplementing the input data and unrecognized characters in the third area using the input section. 2. An image reader unit installed at a financial institution's branch that optically reads data on transfer request slips on which data for foreign exchange transactions is handwritten, and whether or not a management code is printed on the transfer request slip. a determination unit that determines the management code; a printing unit that prints the management code on the transfer request form; a character recognition unit that recognizes the data read by the image reader unit;
A branch terminal device has a data transfer section for transferring character codes and image data recognized by this character recognition section, and is installed at a district center that handles administrative processing for multiple branches.
A money exchange center processing device having a data transmission/reception unit connected to the branch terminal device and a character recognition unit that recognizes characters from data transferred from the branch terminal device; The system is equipped with a workstation having an input section and a display section for data supplementation, and determines whether or not a management code is printed at a predetermined position on the transfer request form using the branch terminal device. However, if it is determined that the control code has already been printed, error handling is performed, and if it is determined that the control code has not been printed yet,
After reading the data entered in the first area of the transfer request form and recognizing the characters, and printing a management code at the position of the transfer request form using the printing section, the result of the character recognition and the second area of the transfer request form are read. The read image data of the area and the third area are transferred to the exchange center processing device, and the exchange center processor performs character recognition based on the read image data of the second area. Transferring the character recognition results of the second area and the image data of the third area to the workstation, and displaying the image data of the third area and the character recognition results on the display section of the workstation; A currency exchange processing method characterized in that the input transaction data of the transfer request form is completed by supplementary inputting the data entered in the third area and the unrecognizable characters using the input unit.