JPH09114920A - Automatic reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently read a code described in commodity by image recognition by reading the identification code from a position corresponding to an image that is picked up in accordance with position information of priority. SOLUTION: A data processing computer 3 instructs to start the processing of code detection while the commodity 10 is being transported after the image is image-picked up by a camera 1a, and an image recognition device searches an ISBN code. Then, image data of a divided screen where the ISBN code is mentioned is transmitted to a character recognition device 2, and the character of a part following the head character string of the character code is recognized. The data processing computer 3 to which the code that was read, has been transmitted collates the received code with product information registered in the master. If there is a master information matching with the code, the product is judged to be a commodity being a transaction object, and product information is given. Then, the information is kept. Thus, the commodity 10, product information or error information are transferred to a commodity distribution device in a postage.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic reading device, and more particularly to a device for reading an identification code written on products such as books, magazines, videos, CDs, cassette tapes and the like.

[0002]

2. Description of the Related Art As a code for identifying a book, a magazine, a video, a CD, a cassette tape, etc. (hereinafter, the book is generically referred to as a product), a code mainly used in books is used in Japan. Book code (ISBN code), JAN code used for books, videos, CDs, cassette tapes, etc. and common magazine code (T code) used for magazines (hereinafter, the ISBN code etc. When collectively referred to, it is described as a code.) These are generally displayed on the back cover, belt, cover, book case, and the like. The above-mentioned code has been conventionally displayed only with OCR characters, but recently, since the reading time is fast and accurate and easy to handle, for the magazines, the T-code converted into a bar code and the OCR characters are written together. Things make up most of the distribution. For this reason, some magazine sorting devices include an automatic recognition device having a barcode reading means to read required information from the barcode.

On the other hand, for general books other than magazines,
The number of items that have been issued in recent years with the addition of a bar code is increasing, but items with an old issue year are also shipped by inventory or resale at the publisher. Therefore, a large number of books without a bar code are distributed, and the code displayed with only OCR characters is 5 to 7 with respect to the total volume of books distributed.
It accounts for a percentage.

For this reason, a character recognition device for recognizing OCR characters is also used to read information on all products. As a character recognition device, there is an automatic reading device provided with an image recognition device that recognizes a screen imaged by a camera or a screen input by a line sensor.

Incidentally, Japanese Utility Model Laid-Open No. 5-1728 discloses a technique for sorting magazines on the basis of data read by the automatic reader, which is equipped with an automatic reader for reading a common magazine code.

[0006]

The code of a product is often written on the back cover or the like parallel to the short side. However, the position is not constant and varies from product to product. For example, when reading an ISBN code of a book with a conventional automatic reader for image recognition, ISBN
The back cover page of the book containing the code is captured by the camera and the image is captured, and the image data for one page is divided into a size that can be processed at one time by pattern matching in parallel with the short side, and one page is divided. The image is sequentially divided from one end to the other end, the screen is taken out, and it is checked whether or not the divided screen has a character string of ISBN at the beginning of the character string in the ISBN code. Thus, the OCR character code is searched for. Repeatedly take out the split screen according to the above sequence until the character of ISBN is found, and then,
The required character code is obtained by reading the character string following the found character string.

The positions where these codes are described occupy most of the upper end and the lower end, but since the design etc. changes depending on the handling time of the product, the description position changes,
Also, the position of the description differs depending on the type of product and the publisher.

Therefore, if the reading is sequentially performed from one end side,
When the code is written on the other end side or when the code writing position of the product changes, the processing time for searching for the identification code becomes long, and the number of products that can be processed within a fixed time decreases.

It is an object of the present invention to provide an automatic reading device for efficiently reading a code written on a product such as a book, a magazine, a video, a CD, a cassette tape by image recognition.

[0010]

The automatic reading apparatus of the present invention is an automatic reading apparatus equipped with an image recognition apparatus for picking up an image of an identification code written on a product and reading the image by image recognition. Means for storing position information prioritized for image division of
It is characterized in that a means is provided for reading the identification code from the corresponding position of the image picked up according to the position information of the priority order obtained from the storage means.

Priority is assigned to the position information having the identification code in descending order of the probability that the identification code is described, and the position information and the priority are stored in the storage means in association with each other. The identification code is read from the corresponding position according to the priority order obtained from the storage means. The identification code can be efficiently read by preferentially reading from the position where the identification code is most likely to be written, so the image recognition processing time can be shortened, the processing speed can be improved, and a large number of products can be processed. Product information can be read.

When the long side direction of the product is the vertical direction, the position information is arranged on the basis of two types of distances from the upper end and the lower end, and the position information is positioned so as to gradually approach the center from the upper end and the lower end. When the information is encoded, the position information with the priority given to the end is obtained. Positions with the same distance from the end can be treated equally even for products of different sizes. Also, even if the product is conveyed with the top and bottom reversed,
It can be treated almost equally.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION An automatic reading apparatus according to an embodiment of the present invention will be specifically described below with reference to FIGS.

FIG. 1 is a schematic block diagram showing the main part of the apparatus, FIG. 2 is a diagram for explaining an example of image division processing, and FIG.
5 and 6 are flowcharts showing data processing in the main part of the apparatus, and FIG. 5 is a perspective view showing an example in which the automatic reading apparatus is connected to the article sorting apparatus in the subsequent stage. FIG. 1 corresponds to the main part of FIG.

As shown in FIG. 5, the product 10 is loaded into the automatic code reading device 100 through the loading slot 7. A conveyor 5 that conveys the product 10 is provided in the automatic reading device 100, and a camera 1 a that captures an image of the product 10 that is conveyed on the conveyor 5 is provided above the conveyor 5.

FIG. 1 is an enlarged view showing the periphery of the camera 1a.
The drive motor 11 of the conveyor 5 is controlled by the control device 6 according to a command from the data processing computer 3. If the product 10 conveyed on the conveyor 5 and stopped at a predetermined imaging position is a book, the camera 1a shoots one page of the cover. If the product 10 is a CD, a video, a cassette tape, etc., the outer box surface is imaged as one page corresponding to one page of the cover of the book.

The image picked up by the camera 1a is sent to the image recognition apparatus 1. The image recognition device 1 searches for a code in the image, sends the found code to the character recognition device 2 connected to the image recognition device 1, and performs character recognition there. The code read by the character recognition device 2 is sent to the data processing computer 3 connected to the character recognition device 2. The data processing computer 3 has a display 3a.
In addition to the keyboard 4 and the keyboard 4, it has a master in which necessary information (book name, publisher, price, etc.) of a code and a product corresponding to the code is registered.

As shown in FIG. 5, the automatic reading device 10
In the subsequent stage of 0, for example, a product sorting device 200 or the like is connected, where products are sorted for each shipping destination and forms are issued. The goods are supplied from the automatic reading device 100 together with the result of the code reading. Hereinafter, the information when the code can be read will be referred to as “code read data”, and the information when the code cannot be read will be referred to as “code unreadable data”.

The product sorting device 200 sorts the product conveyer 201 that carries the product 10 into the device, the product 10 having the code reading data into the good product conveyor 203, and the product 10 having the code unreadable data into the defective product conveyor 204. Handy reader 21 for commonly reading barcodes and character codes of products sent by sorting conveyor 202 and defective conveyor 204.
0, and handy reader 21 due to scratches on the code
A keyboard 211 for inputting a character code when 0 cannot be read, a processing device 212 that determines a sorting destination based on the read data of the product 10 having code read data, and a sorting result of the processing device 212. The printer 213 or the like prints a form of a product based on the product. An example of such a product sorting device is disclosed, for example, in the column of the embodiment of JP-A-7-33218.

Next, in the case of reading the ISBN code consisting of the OCR characters described on the back cover of the product (book) 10 in the automatic reading apparatus 100 having the above-mentioned configuration, as an example, referring to the flowchart of FIG. The operation will be described.

First, the product 10 is loaded on the conveyor 5 with its back cover facing up (step S100). The loaded product 10 is automatically read by the conveyor 5 by the automatic reading device 1
The inside of 00 is moved toward the imaging unit of the camera 1a (step S110). The control device 6 stops the conveyor 5 when the product 10 is conveyed to the imaging unit of the camera 1a (step S120). Camera 1 for the stopped product 10
a captures an image of one page of the back cover (step S130). After the image capturing by the camera 1a is completed, the control device 6 restarts the conveyor 5. While carrying the product 10 on the conveyor 5, the data processing computer 3 instructs the image recognition apparatus 1 to start the code detection processing according to the image division and the setting order in step S140, and the image recognition apparatus searches for the ISBN code.

The process of step S140 will be described with reference to the flowchart of FIG. The image recognition device 1 takes in the image data for one page of the back cover into its memory (step S
141). The image recognition apparatus 1 divides the screen into a size that can be processed at one time by pattern recognition, and the divided screens are extracted in the image recognition apparatus 1 in the order registered in advance according to an instruction from the data processing computer 3.

FIG. 2 shows a modified version of A4 (310 mm × 249 m).
An example in which the image data of the back cover of the book of m) is divided is shown. I
Since the SBN code is written in the direction perpendicular to the binding side of the book, that is, parallel to the short side of the book, the screen is divided in parallel with the short side of the book from the top of the book and the bottom of the book as the starting points. To be done.

As an example, the width of the divided screen is 20 m
m. Further, the boundaries of the divided screens are overlapped with each other, and the overlapping size is, for example, 5 mm. This is so that when the ISBN code is printed at the boundary of the divided screens, it can be read by any of the adjacent divided screens. Since the size of the OCR character of the ISBN code is 2.4 mm in height, even if the character exists at the boundary, it can be read on any of the divided screens.

The order of extracting the divided screens in the memory of the image recognition apparatus 1 is set based on the order of extracting the divided screens stored in the memory by the data processing computer 3. This order is set in order from the highest probability that the code is described, and the divided screens are numbered from the upper end and the lower end of the product 10 as shown in FIG.

As an example, the order of extracting the divided screens is set in the order of the above numbers. That is, the divided screens are sequentially extracted from the upper end portion and the lower end portion of the product alternately toward the center. At this time, the image recognition device 1 is set to extract the screens in the order of the pattern numbers from the split screen of the pattern 1 according to the order set in the memory (step S14).
2) Next, the split screen of pattern 1 is extracted (step S143), and the leading character string of the character code (in the case of ISBN code, the leading character string is ISBN) is present in the extracted split screen of pattern 1. It is searched whether or not (step S144). When the first character string is detected on the split screen of pattern 1, the data processing for finding the ISBN code is ended, and the process proceeds to step S150.

If the first character string cannot be detected in the split screen of pattern 1, it is determined whether or not there is a split screen that has not been searched (step S145). If there is a split screen that has not been searched, the next priority order is given. The image setting is performed so as to extract the split screen (pattern 2) (step S14).
2).

Hereinafter, the split screen is extracted (step S1).
43), the leading character string is searched (character code detection) for the extracted split screen (step S144), and if the character code cannot be detected, it is determined whether there is a remaining split screen (all patterns have been searched). (Step S)
145), the operation of setting the screen to be searched next (step S142) is repeated, the divided screens are taken out in the set priority order, and the search is repeated until the character code is detected.

The image recognition device, when judging that the code cannot be detected (step S145) even if all the divided screens are searched, determines that the code data cannot be detected in the data processing computer 3 when the code is not detected (step S145). It is transmitted (step S146). This error data occurs, for example, when the book 10 is conveyed with its front cover facing up.

When the code can be retrieved in the process of step S144, the image data of the split screen in which the ISBN code is described is sent to the character recognition device 2, and the character of the portion following the first character string of the character code is sent. The ISBN code is read by recognizing (step S15).
0). The read ISBN codes are arranged according to a certain rule. The character recognition device 2 uses the check digit here to determine whether or not the reading is normally performed.

If the reading is performed normally, the read code is transmitted to the data processing computer 3, and if the reading is abnormal, the abnormal reading data is transmitted to the data processing computer 3 (step S160). The data processing computer 3 collates the received code with the product information registered in the master.

That is, the read content is a character string of the ISBN code and is not required information such as a product name, a publisher, a price, a return deadline, and so the code read by the data processing computer 3 to obtain the product information. Is compared with the master information (step S170), and if there is a code having the same code as the master information, it is determined to be an item to be handled, the product information is given, and the information is stored (step S180).

When there is no code having the same code as the master information, the data processing computer 3 judges that the product 10 is not a product to be handled and stores the error information. Error information is also stored for those that are determined to be errors in step S140 or step S160 (step S
190).

Thereafter, the product 10 and the product information or error information are passed to the product sorting device 200 at the subsequent stage (step S200). The product sorting device 200 sorts the products 10 for which a code reading error has occurred into the defective product conveyor 204, and inputs information using the handy reader 210 or the keyboard 211.

The extraction of the divided screens is carried out in order from the upper end and the lower end where the code is likely to appear, toward the center, so that the code can be efficiently read.

The procedure for extracting the split screen may be arbitrarily set in accordance with the property of the product other than the above. For example, the order of extraction can be changed according to the product, such as changing the dividing method such as dividing the image parallel to the long side instead of the short side, or prioritizing a position at a certain distance from the end. .

Further, the data processing computer 3 may change the order of extracting the divided screens. At this time, when the code can be detected in the process of step S144 of the above-described embodiment, the number of the divided screen (pattern number) read as the position data of the divided screen searched is transmitted to the data processing computer 3. To be configured.

In parallel with the processing of steps S100 to S200 of the above-described embodiment, the data processing computer 3 sets data for changing the order in which the image recognition apparatus 1 searches for split screens based on the received split screen numbers. Perform processing.

This data processing will be described with reference to the flowchart of FIG. First, the number of the split screen in which the code is written is fetched (step S151), and then the number of readings is totaled for each split screen number (step S15).
2). This data is, for example, pattern 1 (0 to 20 from the top).
It is stored in the form of 100 times of code reading in mm) and 90 times of code reading in pattern 3 (corresponding to 15 mm to 35 mm from the top). As an example, the maximum size of the product 10 to be handled is A4 modified version 310 m.
In the device for m × 249 mm, the number of data divisions is 21.

Next, the total number of times of reading is summed, and the total number is counted until a desired amount is obtained (step S1).
53). When the total number of readings reaches a desired amount, the data processing computer 3 ranks the divided screens in the descending order of probability of a code based on the number of readings totaled for each divided screen number (step S154).

Next, the data processing computer 3 switches the order (priority order) for extracting the divided screens registered in the image recognition apparatus 1 based on the above-mentioned ranking (step S).
155). Next, the data totalized for each divided screen number is erased (step S156), the screen numbers at the new code entry positions are totaled again, and the ranking operation is repeated.

That is, the screen numbers that have been read are fetched, the number of times is summed up for each divided screen number, and when the total number of readings exceeds a certain number, the divided screen code retrieval is ranked, and based on the ranking. Image recognition device 1
Reset the search order for split screens registered in.
After that, the same operation is repeated.

As described above, the number of readings is totaled for each screen number divided as the position data in which the code is written, and the order of sequential reading is changed based on the totaled result, so the coded position according to the fashion etc. The reading order can be automatically set according to the change of, and the reading process can be performed efficiently.

In the above embodiment, the code read is O.
Although the CR character is used, it may be a bar code.

In the embodiment shown in FIG. 6, the product 10
Although the data is collectively stored and ranked regardless of the size of the item, the data may be stored and ranked according to the size of the product 10. At this time, for example, the product 10 to be screen-divided is divided by the long-side dimension. For example, B7 size or less (128 mm or less), B7 size to B6 size (128 mm to 182 mm or less) are classified. The data processing computer 3 has position data by the number of the divided screen for each of the product size divisions described above, takes in the code description position data and the product size data, and counts the number of data (the number of times of taking in). As a method of obtaining the dimension data of the product 10, there is a method of measuring the dimension when the camera 1a is imaged. Products such as paperback books, new books, and weekly magazines can be identified almost according to size, so
By accumulating the code description position data for each product size in this way, the scanning order can be automatically set according to the tendency of the code description position at that time, corresponding to the type of product, and the reading process can be performed more efficiently. .

In the above embodiment, the ranking for searching the split screens is reset according to the number of data storages, but the split screen search order may be reset depending on the data storage period. By changing the setting in the data accumulation period in this way, the setting can be changed according to the latest change in the code description position even if the product handling amount changes.

Further, in the above embodiment, all the accumulated data was deleted when resetting the ranking for retrieving the divided screens. However, the data is further divided and accumulated in the order in which it was acquired, and statistical processing is performed. At this time, the data replacement process may be performed so that only a certain section is deleted from the old data accumulated and the data is accumulated again.

That is, when resetting the ranking by the total number of data, the ranking is reset when the total number of data reaches a desired number, and at this time, only the fixed number of data from the oldest one is set. Erase. For example, when the number of data becomes 10,000, the data is sorted and stored in units of 2000 for resetting the ranking, and after resetting, the oldest data of 2000 is deleted and the data is stored again. Then, 2000 pieces of data are newly taken in and resetting is performed when the total number of data pieces reaches 10000 pieces.

When resetting the ranking during the data accumulation period, the ranking is reset when the data accumulation period reaches a desired period, and at this time, the oldest data is stored for a certain period. Erase. For example, when accumulating data for 4 weeks and resetting the data, the data is divided and stored for each week, the data of the oldest week is deleted after resetting, and the data is aggregated again. Reconfigure after a week. By performing the data processing in this way, it is possible to deal with more detailed changes in the code description position.

In the above embodiment, the resetting timing of the split screen reading order is based on the amount of data such as the number of data and the accumulation period, but the split screen reading is performed based on the average number of split screen reading times. The ranking may be reset. At this time, when the position data of the code is fetched into the data processing computer 3, information about how many times the divided screen has been searched and the code can be retrieved is fetched, and this average value is constantly calculated.

When the average value exceeds the set value, the reading order of the divided screens is reset. The time required to retrieve the split screen and retrieve the code requires a time that is the number of times the retrieved screen is times the retrieval time per screen. Since the average value of the data processing time can be suppressed by performing the data processing in this way, a stable processing speed can be obtained and the number of products that can be processed can be increased.

As the average value of the number of readings in the above-mentioned data processing, a fixed amount may be averaged from the latest data. By performing the data processing in this way, it is possible to detect a change in the code description position that occurs when the product to be processed changes, for example.

In the above embodiment, the reset order of the reading order of the split screens is performed by the data processing computer 3 side, but the reset order may be performed at any time.
For example, on the display 3a, desired information such as the average value of the current reading time, the statistics for each type of the read product 10, the position distribution of the code, etc. may be displayed and the information on the display 3a may be monitored. If a change in the tendency appears due to, it is possible to input the resetting of the reading order of the split screen from the keyboard 4 of the data processing computer 3. With such a configuration, it is possible to cope with minute changes in the code description position.

Further, the reading order may be reset by observing the processing status of the automatic reading apparatus 100 without using the display 3a. For example, when the reading speed of the code is reduced, the product 10 is placed on the conveyor 5.
Begins to stagnate. Observing this situation, the read order of the split screens is reset using the keyboard 4. By doing so, the processing load on the data processing computer 3 can be reduced.

When the products sent from the publisher are used for reading the product information when sorting by delivery destination (bookstore, etc.), the data storage classification is further subdivided to store data for each publisher. You can go. That is, the product has a characteristic at the position where the code is described by the publisher. Since the sorting of products is processed for each package sent from the publisher, it is only necessary to provide a publisher information input means so as to provide the publisher information of the products when the products are put in. For example, the keyboard 4 of the data processing computer 3 is used as the input means. By performing data processing in this way, it is possible to more accurately respond to changes.

In the above embodiment, the read order of the divided screens was reset by the data processing computer 3, but the read order may be reset by the keyboard 4 or the like in any order. The data processing computer 3 stores the data at the code entry position, and when a certain amount of data is stored, the data processing is performed and the data processing result is displayed on the display 3a. Since the worker can make the setting based on the experience of the worker while referring to the data processing result, the setting can be changed according to the tendency of the code description position peculiar to the product.

Further, the code position data is not stored in the data processing computer 3, and the automatic reading device 100 is used.
Alternatively, the reading order may be reset based on the experience of the operator while observing the processing status of (1) and looking at, for example, the product 10 to be loaded. Even in this case, the display 3
Desired information such as the reading order of the split screen currently being performed may be displayed on a. With this configuration, the data processing computer 3 does not need to totalize the position information data, so that the processing load can be reduced.

The present invention has been described above with reference to the embodiments.
The present invention is not limited to these. For example,
It will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

[0059]

As described above, according to the present invention, it is possible to obtain an automatic reading device for efficiently reading a code described in a product such as a book, a magazine, a video, a CD, a cassette tape, etc. by image recognition efficiently. it can.

[Brief description of the drawings]

FIG. 1 is a schematic perspective view showing a main part of an automatic reading device according to an embodiment of the present invention.

FIG. 2 is a plan view illustrating an example of image division in the device shown in FIG.

FIG. 3 is a flowchart showing an image recognition process of the apparatus shown in FIG.

FIG. 4 is a flowchart showing partial details of the image recognition processing of FIG.

5 is a perspective view showing an example in which the apparatus shown in FIG. 1 and a product sorting apparatus are connected.

FIG. 6 is a flowchart in the case where the apparatus shown in FIG. 1 further performs an order setting process for searching a divided screen.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Image recognition device, 1a ... Camera, 2 ... Character recognition device 3 ... Data processing computer, 3a ... Display 4 ... Keyboard, 5 ... Conveyor, 6 ... Control device 7 ... Input port, 10 ... Product (book, magazine, video) , CD,
11 ... Drive motor 100 ... Automatic reading device 200 ... Commodity sorting device 201 ... Carry-in conveyor, 202 ... Sorting conveyor 203 ... Good conveyor, 204 ... Defective conveyor 210 ... Handy reader, 211 ... Keyboard 212 ... Processing device, 213 ... printer

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yoshio Kanesei 7-23-5 Minamisenju, Arakawa-ku, Tokyo Hitachi Techno Engineering Co., Ltd. Information Systems Department (72) Inventor Kenji Suzuki Hajime Nakase, Mihama-ku, Chiba-shi, Chiba 3-chome Kawatetsu Information Systems Co., Ltd. (72) Inventor Masahiro Kurakake 1-3c Shimaya, Konohana-ku, Osaka Prefecture Sumitomo Electric Industries, Ltd.

Claims (4)

[Claims] 1. An automatic reading device equipped with an image recognition device for picking up an identification code described on a product and reading the image by image recognition, wherein priorities are given in advance for image division for confirmation of the identification code. An automatic reading apparatus comprising: a unit for storing a series of position information; and a unit for reading an identification code from a position corresponding to an image picked up according to the priority position information obtained from the storage unit. 2. The automatic reading device according to claim 1, further comprising an input means for changing the priority of the position information stored in the storage means. 3. The apparatus according to claim 1, further comprising: storing the number of times the identification code is taken in for each position read from the result of reading the identification code by the means for reading the identification code, Data processing means for performing statistical processing when a desired amount is provided is provided, and the data processing means replaces the priority order of the position information in which the identification code stored in the storage means is described based on the statistical processing result. Automatic reading device characterized by. 4. The device according to claim 1, further comprising: storing the number of times the identification code is taken in for each position read from the result of the identification code read by the means for reading the identification code. A data processing unit that performs statistical processing when the desired amount is reached, a unit that displays the statistical processing result by the data processing unit, and a priority order of the position information in which the identification code stored in the storage unit is described. An automatic reading device provided with an input means for exchanging.