JP2012226402A - Receipt data recognition device and program therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correctly and efficiently collect sales management data from an individual tenant shop and improve a speed and efficiency in sales management work at the tenant.SOLUTION: For a recognition character read from a receipt image, a receipt data extraction part 141 extracts a recognition character matching a character string of an item name registered in definition data different for each shop; a same-name determination part 140 obtains a position of a recognition character matching an anchor character or plural keywords, both of which are registered in the definition data, if an item name corresponding to the extracted recognition character is defined, in the definition data, that another character string having the same item name exists in the receipt; a receipt data identification part 144 extracts a recognition character representing a money-amount item using that position and the position of the extracted recognition characters; and a sale management data generation part 15 outputs, or stores, data of the recognition result based on a recognition character matching the item name and a recognition character representing the money-amount item.

Description

  The present invention relates to a receipt data recognition apparatus and a program thereof, and more particularly to a technique for accurately extracting necessary data from receipt data read by character recognition in order to manage sales at a plurality of stores such as a shopping mall. It is.

  The tenant management department that manages stores of tenants (such as companies that operate stores) that open stores in shopping centers, shopping malls, etc. collects daily sales for each store, and breakdowns of sales such as cash sales, coupon sales, etc. ( Sales management item). When creating the sales breakdown, payment receipts provided from each store are collected and the numerical value of the sales management item is calculated. However, since the format of the payment receipt provided from each store depends on the POS system installed in each store, it is currently created in various formats. Therefore, the payment receipt processed by the tenant management department has a wide variety of items, item names, print sizes, print fonts, and the like.

  For example, regarding the names of the items on the payment receipt, even though the total sales, total sales, total sales, total sales, etc. indicate the same contents, different names are printed depending on the POS system. In addition, since the tenant's store that opens a store is frequently replaced, a new payment receipt format must also be supported. For this reason, the tenant management department currently requires enormous labor for the work of totaling the payment receipts as unified sales data, and the human and time burden for that is enormous. is there.

  Note that Patent Document 1 discloses a document image recognition apparatus. In this document image recognition apparatus, a set of components constituting a character is extracted with reference to a document image, horizontal and vertical line candidates are extracted based on the set, and reliability is estimated for the line candidates. Thereafter, a set of lines with high accuracy is extracted based on the estimated line reliability, and the inclination is estimated using the arrangement of the character components of the set of lines with high accuracy based on the result. Then, a line set with a fixed line direction is extracted from the set of basic elements that make up the document image, and a line and a stage are extracted by associating the lines with each other and performing a line extraction with a stage as a constraint. To do.

  Patent Document 2 discloses a receipt processing apparatus. In this receipt processing apparatus, text data is generated by recognizing characters from the receipt image, a predetermined character string is selected from the text data, and a line position including the selected predetermined character string is detected. Then, line data in units of lines is sequentially selected from a predetermined line range on the text data determined based on the detected line position, and data relating to the item is extracted from the line data.

JP 11-219407 A JP 2004-164218 A

  In the tenant management department of a company (organization) that operates a shopping center or the like, the clerk has traditionally used to collect the payment receipts (hereinafter referred to as receipts) provided by each store such as a shopping center. The receipt was read visually, and the read data was manually input to the data processing device. The work amount of the sales totaling work is very heavy because the number of stores may reach 100 stores or more and must be performed every day.

  As a proposal for reducing this burden, it is conceivable to automatically collect sales data (receipt data) of each store from a terminal of each store via a network. However, in general, each store has a different corporate body, so it is necessary to obtain permission to open the network individually, and the type of network to be connected and the difficulty of adding / modifying applications differ depending on the installed system. There are difficult problems.

  Another way to reduce the above-mentioned work load by changing the existing system, especially information devices such as terminals in each store as much as possible, is to recognize character images instead of reading receipt data by humans. The reading method can be considered. That is, the receipt of each store is read by, for example, a reader or scanner having an OCR (Optical Character Recognition) function, and the necessary receipt items and their numerical data are extracted as sales management items, and the sales management data is extracted. This is a method to re-enter.

  However, even in this method, it is necessary to solve a problem that cannot be solved by the conventional recognition of character images. Next, this problem will be described.

  When a receipt is read by a reading device or the like having an OCR function, the correspondence between the receipt item printed on the receipt necessary for collecting sales management data and the sales management item is registered in advance and read from the reading device or the like. It is necessary to extract necessary receipt items from the receipt image. For example, problems that cannot be solved in that case will be described below with reference to FIGS.

  16 and 17 are diagrams for explaining an example of the correspondence between receipt items and sales management items. In particular, FIG. 16 shows an example in which the receipt item name does not have the same name in the receipt, and FIG. 17 shows that there are a plurality of receipt item names (receipt item names) in the receipt. An example of

  For example, as shown in FIG. 16, on the printed receipt 22u, the receipt item name “total sales” corresponding to “total sales” of the sales management item 1001u is printed only on the line Line # 101 in the receipt 22u. ing. Similarly, the receipt item name “gift certificate” corresponding to “gift certificate” of the sales management item 1001u is printed only on the line Line # 102 in the receipt 22u. As for other receipt item names corresponding to “credit sales” and “consumption tax” of the sales management item 1001u, only “line” # 103 and line # 104 in the receipt 22u are “credit” and “tax total”, respectively. "Is printed. That is, in the receipt 22u shown in FIG. 16, there is no receipt item name with the same name for the receipt item that is the target of the sales management item.

  On the other hand, in the receipt 22v shown in FIG. 17, the receipt item name “total sales” corresponding to “total sales” of the sales management item 1001v has a plurality of locations (line # 201, # 204, # 206 in the receipt 22v ( It is printed on a different line. In addition, the receipt item name “gift certificate stock” corresponding to “gift certificate” in the sales management item 1001v is printed only on the line Line # 202 in the receipt 22v. The “credit sales” of the sales management item 1001v is printed at a plurality of locations (three locations) with the name “credit amount” on the lines Line # 203, # 205, # 207 in the receipt 22v. That is, in the receipt 22v shown in FIG. 17, there are a plurality of receipt item names having the same name for the receipt item that is the target of the sales management item. Note that the receipt item name corresponding to “consumption tax” in the sales management item 1001v is not printed on the receipt 22v.

  In FIG. 17, the amount item corresponding to the receipt item name that should be extracted as the sales management item is the amount item in the line Line # 201 to 203 of the group G # 1. The “total sales” and “credit balance” in the line Line # 204 to 207 of the groups G # 2 and G # 3 indicate the breakdown for each person in charge group in the store. It is not a receipt item required as a sales management item. If there are a plurality of receipt item names having the same name and different monetary contents, for example, the receipt 22v shown in FIG. 17 is printed as a subtotal for each of the plurality of paragraphs. The receipt item name “total sales” corresponding to “sales” cannot be distinguished from the receipt item to be extracted only by character recognition by OCR processing. That is, it is not possible to determine whether or not the receipt data is to be extracted as a sales management item only by the receipt item name.

  In order to distinguish these, for example, it is conceivable to use position data based on the receipt print position at the upper end. However, when using position data with a fixed position as a reference, it is not possible to cope with a POS system in which position data changes such that a blank line or a plurality of receipt items increase or decrease depending on the printed contents of the receipt of the day. Actually, the payment receipt for each store often increases or decreases the number of blank items or a plurality of receipt items depending on the printing contents of the day.

  As described above, when a receipt item name printed on a receipt has a plurality of identical names, and the amount corresponding to each of the receipt item names is different, the receipt is read by a reading device or the like. When the receipt numerical data is subjected to OCR processing, a monetary item that does not correspond to the receipt item name that should be extracted may be extracted and processed, resulting in an aggregation error in the sales management data. In order to avoid this, the sales receipts for stores where there are receipt items with the same name as described above must be manually counted for sales management items. Even in this case, since the names are the same, the worker may misidentify the counting points. For this reason, it is necessary to check the settlement receipt, and the human burden is not reduced.

  In view of the above, the present invention collects sales management data accurately and efficiently from the receipts of each tenant's store without significantly changing the current system of each store's information device, etc. The purpose is to improve efficiency and efficiency.

  In the present invention, for example, in order to recalculate receipts printed for each store into a unified sales management item, recognition characters obtained by reading receipt images are registered in different definition data for each store that issued the receipt. If there is a character string of the item name in the received receipt, information on whether or not the same name as the item name to be read exists, and another character string with the same item name in the receipt, Extracted as a sales management item when there are multiple item names that are the same in the item name in the receipt from the position information of the item name and the character string of the other item name associated with the receipt print position. Judgment is made on receipt data, and necessary recognition characters are extracted, and from the extracted recognition character position data, the recognition character and the authorization item representing the monetary item from the position data are extracted. It extracts character, from the recognized character representing the recognized character and amount fields that match the item name, and outputs or stores the data of the recognition result.

  In other words, when the receipt image is read by a reader, the items and the amount in the receipt can be read accurately. From the receipts printed in a plurality of different formats, the name of the item to be read Even if there are multiple items with the same item name, only the items necessary for sales management are accurately extracted, and the items and amounts in the receipt necessary for sales management are output or stored as recognition result data. To be able to. Specifically, the present invention includes the following means.

  (1) The present invention provides an image data storage unit for storing a recognized character obtained by character recognition of a receipt image and its position data, and a character string of an item name printed on the receipt. Information indicating whether there is another character string having the same item name in the receipt, and the item name and receipt when there is another character string having the same item name in the receipt. Define anchor characters, which are character strings of other item names associated with print positions, and information on the distance between the character string of the item name and the anchor character for each item name to be read in a receipt. A definition data storage unit for storing the definition data, and a recognition character that matches the character string of the item name defined in the definition data from the recognition character stored in the image data storage unit. Corresponds when the receipt data extraction unit to be issued and the item name corresponding to the extracted recognized character are defined in the definition data that there is another character string having the same item name in the receipt The same name determination unit that obtains the position of the recognized character that matches the anchor character, calculates the distance between the position and the position of the extracted recognized character, and the item name corresponding to the extracted recognized character is: When it is defined that another character string having the same item name exists in the receipt in the definition data, the calculated distance is compared with the distance information defined in the definition data. , Identify the closest recognized character as the item name to be read, and the item name corresponding to the extracted recognized character is the same item name in the receipt in the definition data. If the other character string is defined not to exist, the receipt data specifying unit for specifying the extracted recognized character as the item name to be read, the item name to be read, and the item name A data generation unit that generates recognition result data from the recognized character representing the money item corresponding to the matching recognized character and outputs or stores the recognition result data is provided.

  Thus, for a receipt image from which a receipt can be read, for each item name to be read, a character string of the item name in the receipt and information on whether or not the same name as the item name to be read exists. When there are multiple item names that are the same among the item names in the receipt, the definition data includes the association of position information with the anchor character that is a unique item name associated with the item name in the receipt in the definition data. Even if there are multiple item names that are the same among the item names in the receipt, by extracting the recognized characters that match the anchor characters for the recognized characters obtained by reading the receipt image , Using the relationship between the extracted recognized character and the position data of the recognized character that matches the item name to be read. It can be of the item name and recognized character representing the amount fields corresponding thereto accurately extracted, and outputs or stores the data of the recognition result.

  (2) Further, according to the present invention, the receipt data recognition apparatus stores an image data storage unit for storing a recognized character obtained by character recognition of a receipt image and its position data, and an item name printed on the receipt. If a character string, information indicating whether there is another character string having the same item name in the receipt, and another character string having the same item name in the receipt, the item name A definition data storage unit for storing definition data defined for each item name to be read in a receipt, and information on a plurality of keywords that are character strings of other item names associated with receipt print positions; A receipt data extracting unit that extracts a recognized character that matches the character string of the item name defined in the definition data from the recognized character stored in the image data storage unit; If the item name corresponding to the recognized character is defined in the definition data as another character string having the same item name in the receipt, the plurality of items defined for the item name A plurality of recognitions that match the plurality of keywords when there are a plurality of recognition characters that match the plurality of keywords and the same name determination unit that determines whether or not there are a plurality of recognition characters that match the keywords A recognized character that matches the character string of the item name defined in the definition data in the character is identified as the item name to be read, and the item name corresponding to the extracted recognized character is in the definition data. A receipt that identifies the extracted recognized character as the item name to be read when it is defined that no other character string having the same item name exists in the receipt A recognition result data is generated from the data identification unit, the item name to be read, and the recognition character representing the monetary item corresponding to the recognition character matching the item name, and the recognition result data is output or stored And a data generation unit.

  As a result, when there are multiple item names that are the same in the item name character string in the receipt, information on whether or not the same name as the item name to be read exists, and the item name in the receipt, When there is more than one item name that is the same in the item name in the receipt, because the definition data has registered in the definition data the positional information association with a group of keywords associated with the item name in the receipt Even so, by extracting the recognized characters that match a plurality of keywords for the recognized characters obtained by reading the receipt image, the position data of the recognized characters that match the extracted recognized characters and the item name to be read Using the relationship, the recognition character representing the item name to be read and the corresponding monetary item is extracted accurately, and the recognition result data is output or It can be stored.

  (3) Further, according to the present invention, in any one of the above inventions, the receipt is a receipt printed in a different format for each store, and the definition data storage unit stores the definition data for each store. The receipt data extraction unit, the same name determination unit, the receipt data identification unit, and the data generation unit execute processing using different definition data for each store that issued the receipt. And

  As a result, even if the item name and print position of the receipt printed at each store are different, the item name and the relative position of the corresponding amount field are registered in the definition data for each store. The necessary receipt data can be read from the receipt image regardless of the difference between the receipt item name and the position of the corresponding money amount column.

  (4) Further, according to the present invention, in the above invention, the definition data stored in the definition data storage unit corresponds to the character string of the item name printed on the receipt and is used for sales management of a plurality of stores. The data generation unit includes data of a management item name used in common, and the data generation unit generates recognition result data obtained by converting a recognition character matching the item name into a corresponding management item name defined in the definition data. It is characterized by that.

  As a result, even if the item name and print position of the receipt printed at each store are different, the item name and the corresponding amount for each receipt item to be converted into the sales management item input in advance. Regardless of the difference between the receipt item name for each store and the position of the corresponding monetary amount column, the necessary receipt data is read from the receipt image and unified sales management data is generated to register the relative position of the column in the definition data. can do.

The effects of the present invention are as follows.
(1) By registering the receipt items necessary for sales management items from the receipt in advance in the definition data, the necessary receipt data can be automatically extracted using them. And speeding up can be achieved.
(2) Receipt items required for sales management items from receipts When there are multiple item names with the same name, the relative position of the amount required for the sales management item is changed to another unique item name or multiple items. Since it can be registered in the definition data in association with a group of position information of the item name as a keyword, it is possible to avoid erroneous recognition of receipt data extraction and to extract correct receipt data. For this reason, labor and speed of the check work can be achieved.
(3) Using the definition data registered and set in advance for each store, what kind of receipt it has, what item name it has, and information such as the relative position with respect to the amount of money is used for image analysis Reading accuracy, including character recognition.
(4) Further, basic data required by the sales management department can be uniformly collected from the receipt data of each store having different receipt items, and stored in a sales management database or the like.

It is a figure which shows the apparatus structural example of embodiment of this invention. It is a figure explaining the sheet | seat for receipt clamping. It is a figure which shows the example of the correspondence of a sales management item and a receipt item. It is a process flowchart of the receipt data recognition apparatus which concerns on embodiment of this invention. It is a figure which shows the example of a receipt image. It is a figure which shows an example of recognition character data. It is a figure which shows the example of a recognition character table. It is a figure explaining a definition data table. It is a flowchart of a receipt data process. It is a flowchart of the same name determination process. It is a flowchart of the same name determination process. It is a figure which shows an example of a receipt and the definition data corresponding to it. It is a figure which shows an example of a receipt and the definition data corresponding to it. It is a flowchart of a sales management data generation process. It is a figure which shows the example of sales management DB. It is a figure explaining the subject of this invention. It is a figure explaining the subject of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following, an example of an apparatus for totalizing store sales will be described. However, the present invention is not limited to sales totalization, and totals actual data printed on purchase total data or other slips. The present invention can be applied to a device for the same.

  FIG. 1 shows an apparatus configuration example for explaining the outline of the present invention. In FIG. 1, reference numeral 1 is composed of a CPU and a storage device such as a hard disk and a memory, etc., and a receipt data recognition device for executing processing relating to sales aggregation based on data read from a receipt, and 20 is associated with each store. This is a receipt holding sheet that is marked with the identification code 21 (shown in FIG. 2) and holds the receipt 22 of each store. The receipt 22 is a receipt printed in a different format for each store.

  The receipt 22 is printed for each tenant's store, for example, the result of totaling the sales of the store in one day using a cash register provided in each store. Further, these receipts 22 are shown as an example related to sales aggregation, and may be related to purchase aggregation and other form processing in addition to sales aggregation.

  First, the receipt clamping sheet 20 will be described. FIG. 2 is a diagram showing a configuration of one embodiment of the receipt clamping sheet 20. 2A is a top view of the receipt sandwiching sheet 20 as viewed from the transparent film 231 side, and FIG. 2B is a receipt sandwiching sheet 20 in which the receipt 22 is sandwiched by the transparent sheet 23. FIG. It is a perspective view explaining the previous state.

  As shown in FIG. 2A, for example, the receipt holding sheet 20 holds a receipt 22 of each store on a transparent sheet 23 with a mount. In the upper right corner of the transparent sheet 23, an identification code 21 (two-dimensional code such as QR code (registered trademark)) corresponding to the store is printed or marked with a seal or the like.

  As shown in FIG. 2B, the transparent sheet 23 is, for example, a transparent film (sheet) 231 and a mount 232 connected (adhesive or the like) at one end 233, and in a state before the receipt 22 is sandwiched, 23 is in a half-opened state. The receipt 22 is placed on the mount 232 of the transparent sheet 23 in the semi-opened state, and the receipt 22 is sandwiched, for example, by adhering the mount 232 and the transparent film 231 having a slightly adhesive surface. The above-described identification code 21 is marked by printing, sticking, or the like within a range positioned at the upper right corner of the mount 232, for example.

  Next, the correspondence relationship between the receipt items in the receipt 22 for each store and the sales management items counted by the tenant management department, which is processed by the receipt data recognition apparatus 1, will be described.

  FIG. 3 is a diagram illustrating an example of a correspondence relationship between sales management items and receipt items. In the example of FIG. 3, the sales summary data actually required by the tenant management department for store sales management is “No. No. 1 cash sales 11 sales management items up to 11 customers. However, these sales management items are not printed as a one-to-one relationship as receipt items of the receipt 22 of each store. Therefore, processing for calculating each of the eleven sales management items from the receipt item of each store is performed using correspondence information 1022 (1022a) for each store.

  In the case of the receipt 22 a shown in FIG. 3, the “cash sales amount” of the 14th receipt item is “No. Correspondence information 1022 (1022a) indicates that it corresponds to “cash sales” of one sales management item. Other sales management items are also associated with each other by the correspondence information 1022. The correspondence information 1022 (1022a) between the sales management item and the receipt item for each store as shown in FIG. 3 is registered and stored in advance in the definition data storage unit 102 (shown in FIG. 1).

  In FIG. 3, the receipt item shown in the receipt 22a is, for example, 14th “cash sales amount” or the like, but in actuality, the 14th (nth) row is fixed from the top position in the receipt. Is not in the position. Depending on the result of receipt collection on the day of the store, the number of print lines increases in the nth or earlier receipt items, or there are receipt items that are omitted during printing. Therefore, in the receipt data recognition apparatus 1, definition data including the receipt item name in the receipt item and the relative position (position of the relative amount column) of the corresponding amount item is registered in advance, and correspondence information 1022 (1022a) ) Is used to solve the above problems.

  The receipt data recognizing device 1 determines the “sales date”, the sales management item No., for each store code in accordance with the definition data and the correspondence information 1022 (1022a) as described above. 1 “Cash Sales”, No. 1 2 “Multiple sales”, No. 2 3 Amounts corresponding to items related to sales such as “credit sales”,... Are aggregated and stored in the sales management data storage unit 103. For this purpose, the receipt data recognition apparatus 1 executes processing described below.

  The receipt data recognition device 1 reads a receipt 22 from a receipt holding sheet 20. The receipt data recognition device 1 uses the definition data for the receipt 22 to extract receipt data including the item name and amount in the receipt item. The receipt data recognition apparatus 1 converts the extracted receipt data item into sales management data using the correspondence information 1022 (1022a) between the sales management item and the receipt item. That is, receipt items necessary for the sales management data are extracted, and the amounts of the receipt items are aggregated as sales management data.

  As described above, the receipt data recognition device 1 executes the receipt data recognition process using the definition data for each store that issued the receipt 22.

  The configuration of the receipt data recognition apparatus 1 shown in FIG. 1 for executing the receipt data recognition process described above will be specifically described below.

  The image data storage unit 101 stores the image data of the receipt sandwiching sheet 20 read by the reader 10. The image data is data including a recognized character and its coordinates (position data) obtained by performing OCR processing on the image of the receipt 22. Furthermore, the image data may include an identification code 21. The image data storage unit 101 may store an image file of the receipt sandwiching sheet 20 such as a bitmap format file in association with the image data.

  The definition data storage unit 102 is a storage unit that stores definition data used in the receipt data recognition apparatus 1. The definition data includes information such as the receipt item of the receipt 22 and the position of the amount for each store.

  Specifically, in the definition data, sales management item names that are commonly used for sales management of a plurality of stores are registered corresponding to the character string of the receipt item name printed on the receipt 22. For example, the definition data includes a receipt item name character string printed on the receipt 22 and a relative position data of a monetary item corresponding to the receipt item name from the position in the receipt of the receipt item name. It is defined for each item name (receipt item name) of a receipt item to be read. Such definition data is created in advance for each receipt 22 of the store to be tabulated and stored in the definition data storage unit 102. Details of the definition data will be described later with reference to FIG.

  Further, the definition data includes a character string of the receipt item name printed on the receipt 22, information indicating whether another character string having the same receipt item name exists in the receipt 22, and the receipt 22 If there is another character string with the same receipt item name, the anchor character that is the character string of the other receipt item name associated with the receipt item name and the receipt print position, and the character string of the receipt item name And information on the distance between anchor characters are registered for each receipt item name to be read in the receipt.

  Alternatively, the definition data includes a character string of the receipt item name printed on the receipt 22, information indicating whether another character string having the same receipt item name exists in the receipt 22, and the receipt 22 If there is another character string having the same receipt item name in the receipt 22, the receipt item name and the information of a plurality of keywords that are character strings of other receipt item names associated with the receipt print position are the receipt 22. Is registered for each receipt item name to be read.

  The sales management data storage unit 103 is a storage unit that stores sales management data generated by the receipt data recognition apparatus 1. The sales management data is data obtained as a result of collecting the amounts of the receipt items of each store in association with the sales management items in which the receipt items different for each store are unified. Details of the sales management data will be described later with reference to FIG.

  The reader 10 is an apparatus that reads an image on the receipt sandwiching sheet 20 by using an optical method or the like for image scanning. When the receipt holding sheet 20 is input, the reader 10 converts the receipt holding sheet 20 into image data. The image data includes, for example, an image file and character data recognized by OCR. The reader 10 stores the converted image data in the image data storage unit 101. For example, the reader 10 is based on a known technique such as a scanner or an OCR processing device.

  The image data reading unit 11 reads the image data stored in the image data storage unit 101. Note that the image data reading unit 11 may input image data from an input device external to the receipt data recognition device 1. The external input device is a reading device, a network device, or the like. In this case, an input device (not shown) inputs the image data of the receipt sandwiching sheet 20 input by a reading device installed at a location other than the installation location of the receipt data recognition device 1 via a network or the like. And stored in the image data storage unit 101.

  The store specifying unit 12 specifies the store code (or store) by identifying the identification code 21 of the receipt sandwiching sheet 20. In the configuration example of FIG. 1, the identification code 21 for the store code is formed as a recognition mark on the receipt sandwiching sheet 20, but the method of reading the store code is not limited to this embodiment. For example, when reading the receipt 22, the user may input the store code via an external input device such as a numeric keypad or a card reader. In that case, the store specifying unit 12 receives a store code input from an external input device. Alternatively, the reader 10 may read the identification code 21 of the receipt holding sheet 20 and the store specifying unit 12 may specify the store code from the read identification code 21.

  When the definition data reading unit 13 receives the store code specified by the store specifying unit 12, the definition data reading unit 13 searches the definition data storage unit 102 for definition data associated with the store code. The definition data reading unit 13 reads the definition data retrieved from the definition data storage unit 102 and sends the read definition data to the receipt data processing unit 14.

  The receipt data processing unit 14 uses the definition data for each store code to extract receipt data including the receipt item name and its amount from the image data, and specifies the receipt data required for the sales management item from the extracted receipt data. To do. For this purpose, the receipt data processing unit 14 includes the following functional units as shown in FIG.

  The receipt data extraction unit 141 searches the recognized character data of the image data stored in the image data storage unit 101 for a recognized character that matches the character string of the receipt item name defined in the definition data, and uses the recognized character data. Extract. The recognized character data includes a value indicating a difference between a candidate character by a dictionary (not shown) provided in the receipt data recognition device 1 in advance and a recognized character by OCR processing, and a relative coordinate position (relative position) indicating a character range. )including. The recognized character data will be described later with reference to FIG.

  The same name determination unit 140 determines whether the corresponding receipt item name includes the item name of the same name in the receipt 22 for the recognized characters extracted from the image data by the receipt data extraction unit 141 using the definition data. To do.

  Specifically, the same name determination unit 140 is defined such that the receipt item name corresponding to the extracted recognized character exists in the definition data as another character string having the same receipt item name in the receipt 22. If there is a recognized character that matches the anchor character or a plurality of keywords defined for the receipt item name, for example.

  Here, the anchor character is a unique item name among other receipt item names associated with the position information such as the receipt printing position, etc., on the receipt item name to be the target of the sales management item. The plurality of keywords are a group of keywords including one or a plurality of other receipt item names that include a receipt item name to be a sales management item and are associated with position information such as a receipt printing position. For example, one group is continuous as a receipt line.

  The same name determination unit 140 includes an anchor character detection unit 142 and a keyword detection unit 143 in order to detect an anchor character and a plurality of keywords from image data. Hereinafter, the anchor character detection unit 142 and the keyword detection unit 143 will be described.

(1) Anchor character detection processing The anchor character detection unit 142 determines whether the item name of the receipt item to be read has the same item name in the receipt in the definition data for the definition data corresponding to the read receipt 22. When a character string is defined to exist and an anchor character is defined, anchor character detection processing is executed as follows.

  The anchor character detection unit 142 recognizes a recognition character that matches the set anchor character from the image data stored in the image data storage unit 101 when the anchor character is set in association with the read character of the definition data. And the distance between the position and the extracted recognized character position is calculated.

(2) Multiple Keyword Detection Processes The keyword detection unit 143 determines whether the item name of the receipt item to be read has the same item name in the receipt in the definition data for the definition data corresponding to the read receipt 22. When a character string is defined to exist and a plurality of keywords are defined, a plurality of keyword detection processes are executed as follows.

  When a plurality of keywords are set in association with the read characters of the definition data, the keyword detection unit 143 matches a plurality of set keywords from the image data stored in the image data storage unit 101. It is determined whether or not a recognized character exists.

(3) When the same name exists in the receipt 22 and an anchor character is detected The receipt data specifying unit 144 defines the calculated distance and definition when the anchor character is detected by the anchor character detecting unit 142 The distance information set in the data is compared, and the recognized character closest to the detected anchor character is identified as the item name (receipt item name) of the receipt item to be read.

(4) When the same name exists in the receipt 22 and a plurality of keywords are detected When the keyword detection unit 143 detects a plurality of keywords, the receipt data specifying unit 144 selects one of the keywords. A recognized character that matches the character string of one receipt item name is specified as the item name of the receipt item to be read.

(5) When the same name does not exist in the receipt 22 On the other hand, the receipt data specifying unit 144 has the same receipt item name corresponding to the recognized character extracted by the receipt data extracting unit 141 in the receipt in the definition data. When it is defined that there is no other character string having the receipt item name, the extracted recognized character is specified as the item name of the receipt item to be read.

(6) Specify the identified recognized character as receipt data The receipt data specifying unit 144 corresponds to the position data of the recognized character specified in any of (3) to (5) and the item name of the receipt item to be read. From the relative position data of the amount item to be recognized, a recognition character representing the amount item is extracted.

  As described above, the receipt data specifying unit 144 receives the relative position of the amount item corresponding to the receipt item name from the receipt data extraction unit 141 or the same name determination unit 140, and stores it in the image data storage unit 101. A recognition character corresponding to the relative position of the monetary item is extracted from the recognition characters. Thereby, the receipt data specifying unit 144 can accurately specify the amount item corresponding to the receipt item name. That is, it is possible to specify receipt data that includes a receipt item name character string and a corresponding money amount item character string in relation to the receipt item.

  As described above, even when there are a plurality of the same names in the receipt item names to be read printed on the receipt 22, these other items are determined by the anchor character or the plurality of keyword information set in the definition data. The amount item corresponding to the receipt item name to be read can be accurately read by determining the relative position to the item name.

  The sales management data generation unit 15 generates sales management data using the receipt data specified by the receipt data processing unit 14. Specifically, the sales management data generation unit 15 recognizes using the recognition character that matches the receipt item name and the correspondence information 1022 that is associated with the corresponding sales management item name defined in the definition data. Calculate (convert) the numerical data of the monetary item in the result to generate sales management data. The sales management data includes a store code, a sales management item, and an amount corresponding to the sales management item.

  For example, the sales management data generation unit 15 sets the “sales date”, the sales management item No., for each store code in accordance with the correspondence information 1022 between the sales management item and the receipt item. 1 “Cash Sales”, No. 1 2 “Multiple sales”, No. 2 3 Converts the amount corresponding to the item related to sales such as “credit sales”,... And stores them in the sales management data storage unit 103.

  In the following, the detailed functions of the receipt data recognition apparatus 1 shown in FIG. 1 will be described with reference to the processing flowcharts of FIGS. 4, 9, 10, 11, and 14, using other figures.

  FIG. 4 is a process flowchart of the receipt data recognition apparatus 1 according to the embodiment of the present invention. The processing executed by the receipt data recognition apparatus 1 will be specifically described with reference to FIG.

  The receipt data recognition apparatus 1 is activated and this process is started. When the receipt holding sheet 20 is input to the reader 10, the reader 10 reads the receipt holding sheet 20, and stores the read image data of the receipt holding sheet 20 in the image data storage unit 101 (step S11). ).

  For specific description, a character string that is OCR processed from receipt image data will be described with reference to FIGS.

  FIG. 5 shows an example of an image of the receipt 22b read by the reader 10. The image data generated by the reader 10 includes recognized character data by OCR processing. When the image of the receipt 22 (22b) is OCR processed, the reader 10 generates recognized character data and stores it in the image data reading unit 11. The recognized character data includes the number of characters of the character string composed of the recognized characters and the coordinates of the character string, the coordinates of the characters constituting the character string, and a recognized character list that is a candidate character from the dictionary. Recognized characters by the OCR process are character string # 1, character string # 2,..., Character string # 14,. For example, the character string # 14 is “total sales”.

  FIG. 6 is a diagram showing an example of recognized character data by OCR processing. In particular, FIG. 6 shows the recognized character data 110 of the character string # 14 “total sales” shown in FIG.

  In FIG. 6, the recognized character data 110 includes the number of characters 3 and the coordinates (x, y) − (x ′, y ′) of the character string # 14 “total sales”. The recognized character data 110 also includes the coordinates of the three characters of character # 141 “total”, # 142 “sell”, and # 143 “upper” that constitute character string # 14. For example, the coordinates of the character # 141 “total” are (x1, y1) − (x′1, y′1), and the other characters are as illustrated. Note that (x, y) − (x ′, y ′) and the like represent the coordinates of the illustrated rectangular frame, and (x, y) is the upper left as viewed from the paper, and (x ′, y ′). Is the coordinate corresponding to the lower right. For example, the upper left is the direction toward the coordinate origin, and the lower right is the direction away from the origin coordinates. The coordinate origin is, for example, a specific position of the image on the receipt holding sheet 20.

  Further, the recognized character data 110 includes a recognized character list of three characters of character # 141 “total”, # 142 “sell”, and # 143 “upper”. The recognized character list stores the recognized characters that are candidates and the degree of difference (or degree of coincidence) between the recognized characters and the read characters in accordance with the recognition order that is the candidate character from the dictionary. For example, in the recognized character list of the character # 141 “total”, the rank “1” is the recognized character “total” and the dissimilarity “1311”, the rank “2” is “fragile”, and the dissimilarity is “1581”. Yes, other rankings are shown in the same way. The degree of difference indicates that, for example, a numerical value closer to 0 indicates a higher degree of matching with a candidate character (not shown) provided in advance, and a higher value indicates a higher degree of difference from the dictionary candidate character. The recognized character lists of the characters # 142 “Sell” and # 143 “Up” are also as shown in FIG.

  FIG. 7 is a diagram illustrating an example of the recognized character table 1011 stored in the image data storage unit 101. The recognized character table 1011 is a table that stores recognized character data for each read image data. Further, the recognized character table 1011 includes recognized character data for each recognized character string of the receipt 22. The recognized character data is as described above.

  The recognized character table 1011 shown in FIG. 7 stores recognized character data such as character strings # 1 to # 14,... Shown in FIG. For example, the character string # 14 “total sales” and the characters # 141 “total”, # 142 “sell”, and # 143 “upper” constituting the character string # 14 are stored in the recognized character table 1011. Other character strings are stored in the same manner. Note that the definitions of items, settings, etc. in the recognized character table 1011 shown in FIG. 7 are the same as those of the recognized character data in FIG.

  Next, when the image data reading unit 11 reads the image data from the image data storage unit 101, the store specifying unit 12 determines the identification code 21 of the receipt sandwiching sheet 20 from the image data, and from the determined identification code 21. The store code is read (step S12).

  Specifically, the store specifying unit 12 decodes a QR code or the like, for example, and specifies a store code or an identification number associated with the store code from the decoded data. The QR code may be decoded by the reader 10 and the store code may be specified from the data decoded by the store specifying unit 12, and the specifying method is not limited.

  Next, upon receiving the store code specified by the store specifying unit 12, the definition data reading unit 13 acquires definition data associated with the store code from the definition data storage unit 102 (step S13).

  FIG. 8 is a diagram for explaining the definition data table 1021. Hereinafter, an example of the definition data acquired by the definition data reading unit 13 will be described with reference to FIG.

  A receipt 22t shown in FIG. 8 is a part of an image of a receipt 22 (referred to as a reference receipt) prepared for creating definition data of a certain store. In FIG. 8, a definition data table 1021 stores definition data created in advance and is stored in the definition data storage unit 102. The receipt 22t displays the line number, receipt item name, and amount. From this receipt 22t, it is assumed that “cash sales amount” on the 14th line (BX14) and “other voucher amount” on the 18th line (BX18) are the receipt items to be extracted.

  In the definition data table 1021, receipt items to be extracted are registered as definition data. In order to create the definition data in advance, a reference receipt (receipt 22t) used as a reference for each store is used, and definition data for the receipt 22 for each store is created based on the reference receipt.

  The definition data table 1021 stores “read item”, “item attribute”, “read character”, and “read position of the money amount column” for each target store code and receipt item to be extracted. “Store code” is an identification number for each store corresponding to the store of the target receipt. “Read item” is the name of a sales management item. “Item attribute” is an attribute of a receipt item name to be read. The attribute is, for example, a value of “1”, “2”, or “3”. “1” indicates that the receipt item name corresponding to the sales management item is not duplicated in the standard receipt (ordinary item), “2” indicates that there is an identical duplicate name and an anchor character is set, and “3” "" Is an attribute when there are duplicate names and multiple keywords are set. “Reading character” is a receipt item name printed on the reference receipt corresponding to the sales management item. The “reading position in the amount field” is a relative position of the amount (including numerical characters, currency symbols, etc.) of the receipt item from the reading character.

  In the receipt 22t that is the reference receipt shown in FIG. 8, the image position BX14 includes character strings Str # 1 and Dat # 1 that are “cash sales amount” and “¥ 139,639”. The relative distance between the character string Str # 1 and the character string Dat # 1 is set to Dis # 1. In the image position BX18, there are character strings Str # 2 and Dat # 2 that are “other money voucher amount” and “¥ 8,000”. The relative distance between the character string Str # 2 and the character Dat # 2 is set to Dis # 2.

  The definition data table 1021 shown in FIG. 8 stores definition data created from the receipt 22t shown in FIG. As shown in FIG. 8, “store code”: “100001” is set in the definition data table 1021 as a store code that is a target of the reference receipt. “Reading item # 1”: “Cash sales” is a sales management item No. This is the result of setting the item name “cash sales” of 1. “Item attribute # 1”: “1” is a result of setting the above-described normal item. “Reading character # 1”: “cash sales amount” is a result of setting the character string Str # 1 at the image position BX14 of the receipt 22t. “Money amount reading position # 1”: “(x11, y11) − (x′11, y′11)” is a character as a relative distance Dis # 1 between the character string Str # 1 and the character string Dat # 1. This is a result of setting the relative position of the position of the character string Dat # 1 from the position of the string Str # 1. As relative positions, the upper left coordinates (x11, y11) and the lower right coordinates (x'11, y'11) of the rectangular frame of the illustrated character string Dat # 1 are stored. These coordinates are relative coordinates. Other definition data is stored in the same manner. Note that #n in the definition data table 1021 shown in FIG. 8 is an identifier for distinguishing a plurality of definition data setting items, and n is an integer such as 1, 2,.

  The definition data table 1021 described above is referred to via the definition data reading unit 13 when the receipt data extraction unit 141 extracts receipt data. The receipt data extraction unit 141 extracts receipt items for the read image data (recognized characters of the receipt 22) based on the definition data stored in the definition data table 1021. That is, the receipt data extraction unit 141 can determine the receipt item to be extracted based on the definition data of the reference receipt.

  Actually, even when the payment receipt (receipt 22) is printed by the same POS system, some receipt items in the middle may be skipped (dropped) depending on the output at that time. is there. Accordingly, if the “reading position of the amount column” in the definition data table 1021 is created with fixed coordinates, it is not possible to cope with a case where such a part of receipt item is skipped and printed. Therefore, as in the definition data table 1021 shown in FIG. 8, by using the coordinate position of the relative position from the character string position of the receipt item name as the definition data, printing such a part of the receipt items is skipped. It can correspond to. In FIG. 8, the item name of the receipt item and the amount corresponding to the receipt item are described on the same line. However, the item name of the receipt item and the corresponding amount of money are different. May be. In that case, the relative position of the amount of money in the different rows is registered in the definition data.

  Next, when the receipt data processing unit 14 receives the definition data associated with the store code from the definition data reading unit 13, the receipt data processing unit 14 executes receipt data processing (step S14). A detailed flow of the receipt data processing will be described later with reference to FIG.

  After receipt data processing is performed by the receipt data processing unit 14, the sales management data generating unit 15 generates sales management data from the processed receipt data. The sales management data generation unit 15 stores the generated sales management data in the sales management data storage unit 103 (step S15). After storing the sales management data, the receipt data recognition apparatus 1 ends this processing. A detailed processing flow of step S15 will be described later with reference to FIG.

  The processing flow described above is the overall processing flow executed by the receipt data recognition apparatus 1.

  FIG. 9 is a detailed process flowchart of the receipt data process in step S14 shown in FIG. The receipt data processing executed by the receipt data processing unit 14 will be specifically described with reference to FIG.

  The receipt data extraction unit 141 extracts only the item name of the line targeted for the receipt 22 from the image data (step S21).

  Specifically, upon receipt of image data (recognized character data) from the image data reading unit 11, the receipt data extraction unit 141 extracts only the item name from the first line of the receipt 22. For example, in the case of the image data of FIG. 5, the character string # 1 “Sunday, October 10, 2010 22:30 No: 0002” of the receipt 22b is extracted, and the character string # 2 “**”,. Character string # 7 “sales”, character string # 8 “297 points”, character string # 9 “¥ 226,701”,... Are extracted, and blank lines are not extracted. In step S21, the item name including the receipt item name, amount, symbol (for example, “**”) is extracted.

  Next, the receipt data extraction unit 141 refers to the definition data stored in the definition data table 1021 (step S22).

  Specifically, the definition data reading unit 13 reads definition data corresponding to the store code from the definition data table 1021 stored in the definition data storage unit 102. The definition data reading unit 13 sends the read definition data to the receipt data extraction unit 141. The receipt data extraction unit 141 uses the definition data sent from the definition data reading unit 13 in the processing from step S23. The receipt data extraction unit 141 stores the definition data sent from the definition data reading unit 13 in a work memory (not shown), and refers to the definition data on the work memory.

  Next, the receipt data extraction unit 141 checks whether or not the extracted item name is set in the definition data (step S23).

  Specifically, the receipt data extraction unit 141 determines whether there is a match among the read characters #n (n is an integer) in the definition data table 1021 for the character string # 14 “total sales” shown in FIG. To check.

  Next, when the extracted item name is set in the definition data (step S23: Yes), the same name determination unit 140 refers to the item attribute of the receipt item name to be read (step S24). On the other hand, if the extracted item name is not set in the definition data (step S23: No), the process proceeds to step S28.

  Next, the same name determination unit 140 checks whether or not it is defined that there is a receipt item name with the same name in the definition data (step S25). When it is defined that there is a receipt item name with the same name in the definition data (step S25: Yes), the same name determination unit 140 executes the same name determination process (step S26). On the other hand, when it is defined that there is no receipt item name with the same name in the definition data (step S25: No), the same name determination unit 140 moves the process to step S27.

  10 and 11 are detailed process flowcharts of the same name determination process in step S26. Hereinafter, the same name determination process executed by the same name determination unit 140 will be specifically described with reference to FIGS. 10 and 11.

  The anchor character detection unit 142 determines from the definition data whether the item attribute of the receipt item name to be read is an anchor character (step S261). If the item attribute is an anchor character (step S261: Yes), the process proceeds to step S262. On the other hand, if the item attribute is not an anchor character (step S261: No), that is, if the item attribute is a plurality of keywords, the process proceeds to step S267.

  When the process moves from step S261, the anchor character detection unit 142 acquires the position data of the recognized character that matches the receipt item name (step S262).

  For concrete explanation, FIG. 12 shows an example of a receipt 22c and a definition data table 1021c in which anchor characters corresponding to the receipt 22c are set. In the receipt 22c shown in FIG. 12, for example, a character string Str # 11, Str # 21, Str # 31 is present on the receipt 22c as a recognized character that matches the receipt item name “total sales”. Further, it is assumed that the money amount items (a character string of money amounts) Dat # 11 “¥ 168,735”, Dat # 21 “¥ 48,195”, and Dat # 31 “¥ 32,025” correspond to each. Hereinafter, the processing of steps S262 to S266 will be described with specific examples using FIG.

  The definition data table 1021c illustrated in FIG. 12 is a table in which definition data corresponding to, for example, the store code “100001” is stored. In the definition data table 1021 c shown in FIG. 12, the reading item # 1: “total sales” is the sales management item No. 5 corresponding to the receipt item name “total sales” of the receipt 22c. Item attribute # 1: “2” indicates that the receipt item name to be read has the same name and the anchor character is set as the read item attribute. In the reading character # 1: “total sales”, the receipt item name “total sales” of the receipt 22c is set as a reading character. The reading position # 1 in the amount column is (x11, y11) − (x′11, y′11), and on the receipt 22c, for example, the amount item Dat in the amount column from the character string Str # 11 “total sales”. The relative distance to # 11 “¥ 168,735” is shown.

  In the definition data table 1021c shown in FIG. 12, the anchor character is another receipt item name in the receipt 22c and is a unique character. For example, the anchor character “loan amount” is a unique character associated with the read character # 1: “total sales”. The distance of the anchor character is a distance based on the y coordinate from the character string that is the anchor character on the receipt 22 to the read character # 1 corresponding to the receipt item name on the receipt 22. For example, the anchor character distance “+100” is from the character string Str # 41 “loan amount” on the receipt 22c, which is the anchor character, to the character string Str # 11 “total sales” of the receipt item name on the receipt 22c. Is a distance “+100” based on the y coordinate of the character string Str # 11 “total sales”, and a relative value “ 100 ”.

  For example, in step S262 described above, the anchor character detection unit 142 uses the character strings Str # 11 and Str # 21 as recognition characters that match the receipt item name “total sales” from the image data of the receipt 22c shown in FIG. , Str # 31 position data is acquired.

  Next, the anchor character detection unit 142 acquires the position data of the recognized character that matches the anchor character from the image data (step S263).

  Specifically, the anchor character detection unit 142, from the image data of the receipt 22c stored in the image data storage unit 101 via the image data reading unit 11, reads the character string Str # 41 that matches the anchor character. The position data of the recognized character corresponding to “high” is acquired.

  Next, the anchor character detection unit 142 calculates each distance from the position of the recognized character that matches the receipt item name and the position of the recognized character that matches the anchor character (step S264).

  Specifically, as shown in FIG. 12, in the definition data table 1021c, the reading position # 1 in the amount column is (x11, y11) − (x′11, y′11), and the characters in the receipt 22c. The relative distance from the column Str # 11 “total sales” to the money amount item Dat # 11 is shown. The anchor character “loan amount” is a unique character associated with the read character # 1 “total sales”. The anchor character distance Ref # 1 is a distance based on the y coordinate to the read character # 1 corresponding to the receipt item name on the receipt 22c.

  For example, the distance Ref # 1 of the anchor character is a distance “based on the y coordinate from the character string Str # 41“ Lending ”on the receipt 22c to the character string Str # 11“ Total sales ”on the receipt 22c. +100 ". That is, it means that the character string Str # 11 “total sales” is a relative value “100” as a difference of the y-coordinates on the upper end side of the line from the character string Str # 41 “loan amount”. For other recognized characters, for example, the distance Ref # 2 between the character string Str # 21 “total sales” and the anchor character Str # 41 “loan amount”, and the character string Str # 31 “total sales” and the anchor This is the distance Ref # 3 between the character Str # 41 and the amount of lending. As shown in FIG. 12, the distances Ref # 2 and Ref # 3 are y coordinates on the lower end side of the line of the character string Str # 41 “loan amount”.

  Next, the anchor character detection unit 142 compares the calculated distance with the distance from the anchor character set in the definition data, and applies the one that has a small difference (subject to sales management items). It is determined as the item name of the item (step S265).

  Specifically, as shown in FIG. 12, the anchor character detection unit 142 determines that the distance Ref # 2 and the distance Ref # 3 are negative distances and are not close to “+100”. It is determined that Str # 21 “total sales” and Str # 31 “total sales” are not read objects of the read character # 1 “total sales”. On the other hand, since the distance Ref # 1 is the closest value to the distance “+100”, the anchor character detection unit 142 has the character string Str # 11 “total sales” as the reading target of the read character # 1 “total sales”. Judge. That is, the anchor character detection unit 142 determines that the character string Str # 11 “total sales” is the receipt item name to be read from the anchor character Str # 41 “rented amount” and the distance “+100” from the anchor character. Is detected.

  Next, the receipt data specifying unit 144 extracts the recognized character representing the money amount item from the image data using the position information of the receipt item name determined to be applicable (step S266).

  Specifically, the receipt data specifying unit 144 includes the position data of the receipt item name determined to be applicable by the anchor character detection unit 142, and the relative position data of the amount item corresponding to the receipt item name set in the definition data. From the image data, a recognition character representing a monetary item is extracted from the image data.

  For example, as shown in FIG. 12, in the definition data table 1021c, the reading position # 1 in the amount column is (x11, y11)-(x′11, y′11), and the character string Str # in the receipt 22c. 11 indicates the relative distance from “total sales” to the monetary item Dat # 11. The receipt data specifying unit 144 acquires this relative distance from the definition data table 1021c. Further, the receipt data specifying unit 144 acquires the position data of the recognized character corresponding to the character string Str # 11 “total sales” from the image data stored in the image data storage unit 101. As a result, the receipt data specifying unit 144 uses the relative distance and the position data of the recognized character corresponding to the character string Str # 11 “total sales” from the image data stored in the image data storage unit 101. , The amount item Dat # 11 “¥ 168,735” at the distance Dis # 11 from the character string Str # 11 “total sales” is extracted.

  After extracting the recognition character representing this monetary item, this process is terminated. That is, the process of step S26 (steps S261 to S2610) ends.

  On the other hand, when the process moves from step S261, the keyword detecting unit 143 acquires a plurality of keywords from the definition data setting (step S267).

  For concrete explanation, FIG. 13 shows an example of a receipt 22d and a definition data table 1021d in which a plurality of keywords corresponding to the receipt 22d are set. In the receipt 22d shown in FIG. 13, for example, a character string Str # 12, Str # 22, Str # 32 is present on the receipt 22d as a recognized character that matches the receipt item name “total sales”. Further, it is assumed that the money amount items (a character string of money amounts) Dat # 12 “¥ 168,735”, Dat # 22 “¥ 48,195”, and Dat # 32 “¥ 32,025” correspond to each. Hereinafter, the processing in steps S268 to S2610 will be described with specific examples using FIG.

  For example, as shown in FIG. 13, the keyword detection unit 143 has item attribute # 1: “3” in the definition data table 1021d, so that the read character # 1: “total sales”, “net sales”, “cash” A plurality of keywords Key # 1 including “sales amount” and “rental amount” are acquired.

  Next, the keyword detection unit 143 searches the image data for a group of recognized characters that match a plurality of keywords (step S268).

  Specifically, the keyword detection unit 143 receives a plurality of keyword keys # 1 “total sales”, “net sales” from the image data of the receipt 22d stored in the image data storage unit 101 via the image data reading unit 11. ”,“ Cash balance ”and“ lending balance ”are searched for a group of recognized characters. The keyword detection unit 143 acquires the retrieved recognized character data.

  Next, the keyword detecting unit 143 determines that the recognized character that matches the receipt item name is the item name of the corresponding receipt item from the group of recognized characters that match the searched keywords (step S269). .

  Specifically, as illustrated in FIG. 13, the keyword detection unit 143 includes the position data of the recognized characters that match the character string Str # 22 “total sales” and the positions of the recognized characters included in the plurality of keywords Key # 1. Based on the relationship with the data, it is determined that the character string Str # 22 “total sales” is not a line grouped with the plurality of keywords Key # 1. Similarly, it is determined that the character string Str # 32 “total sales” is not a grouped row.

  On the other hand, the keyword detection unit 143 determines the character string Str # from the relationship between the position data of the recognized character that matches the character string Str # 12 “total sales” and the position data of the recognized character included in the plurality of keywords Key # 1. 12 It is determined that “total sales” is in the position of a row included in a group of a plurality of keywords Key # 1. That is, the keyword detection unit 143 reads the character string Str # 12 “total sales” from the plurality of keywords Key # 1 “total sales”, “net sales”, “cash balance”, and “lending balance”. It is detected that it is the item name of the receipt item.

  The receipt 22d shown in FIG. 13 shows a case where there is no blank line between a group of keywords of a plurality of keywords Key # 1, but a blank line or a line in which no receipt item name exists between a group of keywords. May be included. In that case, a plurality of keywords Key # 1 may be regarded as a group of keywords except for blank lines and lines in which no receipt item name exists (not printed). In the definition data table 1021d shown in FIG. 13, an example in which a group of a plurality of keywords is set on the assumption that the reading character # 1: “total sales” is located in the upper row is shown. It may be the second row from the top or other rows. In that case, it is only necessary to have information indicating which item name among the plurality of keywords is the item name corresponding to the sales management item.

  Next, the receipt data specifying unit 144 extracts a recognition character representing the amount item from the image data using the position information of the receipt item name determined to be applicable (step S2610).

  Specifically, the receipt data specifying unit 144 uses the position data of the receipt item name determined to be applicable by the keyword detection unit 143 and the relative position data of the amount item corresponding to the receipt item name set in the definition data. , Recognize a character representing a monetary item from the image data.

  For example, as shown in FIG. 13, in the definition data table 1021d, the reading position # 1 in the amount column is (x11, y11) − (x′11, y′11), and the character string Str # in the receipt 22d. 12 indicates a relative distance from “total sales” to the monetary item Dat # 12. The receipt data specifying unit 144 acquires this relative distance from the definition data table 1021d. Further, the receipt data specifying unit 144 acquires the position data of the recognized character corresponding to the character string Str # 12 “total sales” from the image data stored in the image data storage unit 101. As a result, the receipt data specifying unit 144 uses the relative distance and the position data of the recognized character corresponding to the character string Str # 12 “total sales” from the image data stored in the image data storage unit 101. , The amount item Dat # 12 “¥ 168,735” at the distance Dis # 12 from the character string Str # 12 “total sales” is extracted.

  After extracting the recognition character representing this monetary item, this process is terminated. That is, the process of step S26 (steps S261 to S2610) ends.

  Next, the receipt data extraction unit 141 determines whether the processing of the last line of the receipt item has been completed (step S28). When the process of the last line of the receipt item is not completed (step S28: No), the process returns to step S21, and the processes of steps S21 to S27 from the line of the next receipt item are repeated. On the other hand, when the process on the last line of the receipt item is completed (step S28: Yes), this process is terminated.

  As shown in FIG. 12 or FIG. 13, the receipt data specifying unit 144 may specify, for example, the amount item “¥ 168,735” corresponding to the receipt item name “total sales” by the above-described identical name determination process or the like. it can.

  As described above, the receipt data processing unit 14 can extract necessary receipt data for the recognized character of the receipt item read from the receipt image using the registered definition data.

  The receipt data processing unit 14 can extract the receipt data necessary for the sales management data, and the receipt item name and the amount of the extracted receipt data are converted into the sales management data of the sales management item in a predetermined correspondence relationship. Can be converted (aggregated).

  Even if the receipt data recognition process described above has the same name as the item name of the receipt item to be read in the receipt 22, the unique anchor character associated with the item name is defined using the definition data. Alternatively, by retrieving a plurality of keywords, it is possible to accurately extract the item name of the receipt item necessary for the sales management item and the corresponding monetary item. Thereby, sales management data can be correctly totaled.

  FIG. 14 shows a detailed process flowchart of the sales management data generation process in step S15. The sales management data generation process executed by the sales management data generation unit 15 will be specifically described with reference to FIG.

  The sales management data generation unit 15 acquires receipt data necessary for sales aggregation from the receipt data specifying unit 144 (step S31).

  Specifically, the sales management data generation unit 15 receives data including a store code, data including a receipt item name and its amount specified from the recognized character data, and correspondence information 1022 (1022a) shown in FIG. Is acquired from the receipt data specifying unit 144.

  Next, the sales management data generation unit 15 aggregates the sales management data of the sales management items by using the correspondence relationship information 1022 (1022a) (step S32).

  Specifically, the sales management data generation unit 15 calculates the amount corresponding to the receipt item name from the correspondence information 1022 (1022a) based on the correspondence between the sales management item and the receipt item, and the sales management data of the sales management item. As As described above, the amount (numerical data) of the receipt item is converted into the numerical data of the sales management item.

  Next, the sales management data generation unit 15 stores the sales management data of the store code to be stored in the sales management DB 1031 of the sales management data storage unit 103 (step S33).

  For example, FIG. 15 shows an example of a sales management DB (database) 1031. In the sales management DB 1031, as shown in FIG. 15, “sales date”, sales management item No. 1 “Cash Sales”, No. 1 2 “Multiple sales”, No. 2 3 A numerical value (amount) corresponding to a settlement item relating to sales such as “credit sales”,... Is stored.

  As described above, in the sales management DB 1031, the receipt items of the receipt 22, which have a completely different format (sheet format) for each store, are unified, and the receipt item, name, Numerical values can be automatically collected without any manual input as sales management data based on unified sales management items. As a result, the person in charge of the tenant management department can automatically count the sales for each store, so that the store sales management business can be performed efficiently and speedily.

  In addition, since there is no need to change a receipt output device such as a cash register provided in each store, an extra expense can be suppressed for the tenant.

  The processing performed by the above receipt data recognition apparatus can be realized by a computer and a software program, and the program can be recorded on a computer-readable recording medium or provided through a network.

DESCRIPTION OF SYMBOLS 1 Receipt data recognition apparatus 10 Reader 11 Image data reading part 12 Store specification part 13 Definition data reading part 14 Receipt data processing part 15 Sales management data generation part 20 Receipt clamping sheet 22 Receipt 101 Image data storage part 102 Definition data storage part 103 Sales Management Data Storage Unit 140 Same Name Determination Unit 141 Receipt Data Extraction Unit 142 Anchor Character Detection Unit 143 Keyword Detection Unit 144 Receipt Data Identification Unit

Claims (6)

An image data storage unit for storing a recognized character obtained by character recognition of the receipt image and its position data;
The character string of the item name printed on the receipt, the information indicating whether there is another character string with the same item name in the receipt, and the other character string with the same item name in the receipt If present, an anchor character, which is a character string of another item name associated with the item name and the receipt print position, and information on the distance between the character string of the item name and the anchor character are displayed in the receipt. A definition data storage unit for storing definition data defined for each item name to be read;
A receipt data extracting unit that extracts a recognized character that matches the character string of the item name defined in the definition data from the recognized character stored in the image data storage unit;
A recognized character that matches the corresponding anchor character when the item name corresponding to the extracted recognized character is defined in the definition data as another character string having the same item name in the receipt The same name determination unit that calculates the distance between the position and the position of the extracted recognized character;
When the item name corresponding to the extracted recognized character is defined in the definition data as another character string having the same item name in the receipt, the calculated distance and the definition Compare the distance information defined in the data, identify the closest recognized character as the item name to be read, and the item name corresponding to the extracted recognized character is the receipt in the definition data A receipt data specifying unit for specifying the extracted recognized character as an item name to be read when it is defined that no other character string having the same item name exists in
A data generation unit that generates recognition result data from the item name to be read and a recognition character that represents a monetary item corresponding to the recognition character that matches the item name, and outputs or stores the recognition result data; A receipt data recognition device characterized by comprising: An image data storage unit for storing a recognized character obtained by character recognition of the receipt image and its position data;
The character string of the item name printed on the receipt, the information indicating whether there is another character string with the same item name in the receipt, and the other character string with the same item name in the receipt If there is, definition data that defines information about a plurality of keywords that are character strings of other item names associated with the item name and the receipt print position for each item name to be read in the receipt. A definition data storage unit to be stored;
A receipt data extracting unit that extracts a recognized character that matches the character string of the item name defined in the definition data from the recognized character stored in the image data storage unit;
The item name corresponding to the extracted recognized character is defined for the item name when it is defined in the definition data that there is another character string with the same item name in the receipt. The same name determination unit for determining whether or not there are a plurality of recognized characters that match the plurality of keywords;
When there are a plurality of recognized characters that match the plurality of keywords, a recognized character that matches the character string of the item name defined in the definition data among the plurality of recognized characters that match the plurality of keywords, When the item name corresponding to the extracted recognized character is defined as no other character string having the same item name in the receipt in the definition data, specified as the item name to be read A receipt data specifying unit for specifying the extracted recognized character as an item name to be read;
A data generation unit that generates recognition result data from the item name to be read and a recognition character that represents a monetary item corresponding to the recognition character that matches the item name, and outputs or stores the recognition result data; A receipt data recognition device characterized by comprising: In the receipt data recognition device according to claim 1 or 2,
The receipt is a receipt printed in a different format for each store,
The definition data storage unit stores the definition data for each store,
The processing by the receipt data extraction unit, the same name determination unit, the receipt data identification unit, and the data generation unit is executed using different definition data for each store that issued the receipt. Receipt data recognition device. In the receipt data recognition device according to claim 3,
The definition data stored in the definition data storage unit includes management item name data commonly used for sales management of a plurality of stores corresponding to the character string of the item name printed on the receipt,
The receipt data recognition apparatus, wherein the data generation unit generates recognition result data obtained by converting a recognition character that matches the item name into a corresponding management item name defined in the definition data. Computer
Image data storage means for storing a recognized character obtained by character recognition of a receipt image and its position data;
The character string of the item name printed on the receipt, the information indicating whether there is another character string with the same item name in the receipt, and the other character string with the same item name in the receipt If present, an anchor character, which is a character string of another item name associated with the item name and the receipt print position, and information on the distance between the character string of the item name and the anchor character are displayed in the receipt. Definition data storage means for storing definition data defined for each item name to be read;
Receipt data extracting means for extracting a recognized character that matches the character string of the item name defined in the definition data from the recognized character stored in the image data storage means;
A recognized character that matches the corresponding anchor character when the item name corresponding to the extracted recognized character is defined in the definition data as another character string having the same item name in the receipt And the same name determination means for calculating the distance between the position and the position of the extracted recognized character,
When the item name corresponding to the extracted recognized character is defined in the definition data as another character string having the same item name in the receipt, the calculated distance and the definition Compare the distance information defined in the data, identify the closest recognized character as the item name to be read, and the item name corresponding to the extracted recognized character is the receipt in the definition data Receipt data specifying means for specifying the extracted recognized character as an item name to be read when it is defined that no other character string having the same item name exists in
Function as data generation means for generating recognition result data from the item name to be read and a recognition character representing a monetary item corresponding to the recognition character matching the item name, and outputting or storing the recognition result data Receipt data recognition program. Computer
Image data storage means for storing a recognized character obtained by character recognition of a receipt image and its position data;
The character string of the item name printed on the receipt, the information indicating whether there is another character string with the same item name in the receipt, and the other character string with the same item name in the receipt If there is, definition data that defines information about a plurality of keywords that are character strings of other item names associated with the item name and the receipt print position for each item name to be read in the receipt. Definition data storage means for storing;
Receipt data extracting means for extracting a recognized character that matches the character string of the item name defined in the definition data from the recognized character stored in the image data storage means;
The item name corresponding to the extracted recognized character is defined for the item name when it is defined in the definition data that there is another character string with the same item name in the receipt. The same name determination means for determining whether or not there are a plurality of recognized characters that match the plurality of keywords;
When there are a plurality of recognized characters that match the plurality of keywords, a recognized character that matches the character string of the item name defined in the definition data among the plurality of recognized characters that match the plurality of keywords, When the item name corresponding to the extracted recognized character is defined as no other character string having the same item name in the receipt in the definition data, specified as the item name to be read Receipt data specifying means for specifying the extracted recognized character as an item name to be read;
Function as data generation means for generating recognition result data from the item name to be read and a recognition character representing a monetary item corresponding to the recognition character matching the item name, and outputting or storing the recognition result data Receipt data recognition program.

Images