JP6114320B2 - Code reading learning system - Google Patents

Description

  The present invention relates to a code reading learning system that learns about reading of a one-dimensional barcode.

  2. Description of the Related Art Conventionally, code readers that decode bar codes and obtain information are known. Bar codes are used to manage information on goods, for example, at retail, physical distribution, and production sites.

  Generally, a barcode is attached to an article by being printed on the surface of the article or by attaching a label or the like printed with a barcode to the article. The quality of the article or label on which the barcode is printed, Depending on the print quality of the printer that prints the bar code, the bar code shape may change or the bar code may become unclear, and the code reader may not be able to read it accurately. When the barcode cannot be read by the code reading device, the user needs to manually input the barcode information, and there is a problem that work efficiency is lowered.

  Therefore, if the code reader cannot read the target bar code information, the printer print head used to print the bar code is considered defective and the scanned bar code data And a manually input barcode data are identified, and a defective portion of the print head is identified, and the defective portion is stored for future barcode scanning correction (see Patent Document 1). ).

Patent No.4550262

  In a technique such as Patent Document 1, correction is performed on the assumption that there is a defect in the print head of the printer. However, the case where reading is not possible is not limited to such print quality, and a printing surface or label is attached. Various causes are conceivable depending on the usage environment, such as the surface being curved or dirty, and how the code reader is used by the user. However, if such various causes are individually corrected, a great deal of labor is required and the efficiency may deteriorate.

  The present invention has been made to solve such problems, and the object of the present invention is to provide a code reader that can easily improve the reading accuracy according to the use environment and can improve the working efficiency. To provide a learning system.

  In order to achieve the above object, a code reading learning system according to a first aspect of the invention analyzes a one-dimensional barcode detected by a detection unit that detects a one-dimensional barcode, and the one-dimensional barcode detected by the detection unit. A code extraction unit for extracting bar code space and bar data, a storage unit for storing a decoding threshold for decoding a one-dimensional bar code space and bar to be read, and a decoding stored in the storage unit The threshold value has a reference value. Using the reference value as teacher data, the teacher data is compared with the space and bar data extracted by the code extraction unit, and the difference between the corresponding data is calculated. A learning unit that updates a reference value of the decoding threshold value of the storage unit so as to reduce the difference.

  In the code reading learning system according to a second aspect of the present invention, in the first aspect, the teacher data is individually selected according to one or both of a space and a bar type and position of the one-dimensional bar code. It has a decoding threshold value.

  In the code reading learning system according to a third aspect of the present invention, in the first or second aspect of the invention, the learning unit is responsive to one or both of the space and bar type and position of the one-dimensional barcode. The degree to which the difference is decreased is individually set.

  A code reading learning system according to a fourth aspect of the present invention is the code reading learning system according to any one of the first to third aspects, further comprising a plurality of reading devices and a server that can communicate with each of the reading devices, Provided in the reading device, the code extraction unit is provided in the reading device or the server, the storage unit is provided in the reading device or the server, and the learning unit is provided in the server. It is characterized by that.

  According to the present invention using the above means, the difference between the teacher data and the space and bar data extracted by the code extraction unit is calculated using the reference value of the one-dimensional barcode space and the bar decoding threshold as teacher data. Then, by updating the decoding threshold reference value so as to reduce the difference, the decoding threshold is adapted to the actual use environment. As a result, the reading accuracy can be easily improved according to the use environment, and the working efficiency can be improved.

1 is an external view of the use of a code reading device that realizes a code reading learning system according to a first embodiment of the present invention. It is a block diagram which shows schematic structure of the code reading learning system which concerns on 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the code reader which concerns on 1st Embodiment of this invention. It is explanatory drawing about the decoding threshold value and reference value in 1st Embodiment. It is explanatory drawing of the learning process in 1st Embodiment. It is a block diagram which shows schematic structure of the code reading learning system which concerns on 2nd Embodiment of this invention. It is a flowchart which shows operation | movement of the code reader which concerns on 2nd Embodiment of this invention. It is explanatory drawing of (a) 1st modification of a learning process, and (b) 2nd modification.

  Hereinafter, a code reading learning system according to the present invention will be described with reference to the drawings.

  First, the code reading learning system in the first embodiment will be described. 1 and 2 show the configuration of the code reading learning system according to the first embodiment of the present invention, which will be described below with reference to these drawings.

  As shown in FIG. 1, the code reading learning system of the first embodiment is applied to the code reading device 1 alone. Specifically, the code reading device 1 is a handy terminal code reading device used by a user. The code reader 1 reads the barcode 3 by taking an image of the slip 4 on which the barcode 3 shown in FIG. 1 is printed, for example. The type of the code reading device 1 is not limited to this, and may be a wearable type that can be fixed to the user's body or a fixed type. However, the barcode to be read in the present invention is a one-dimensional barcode having a plurality of spaces and bars as symbol characters.

  As shown in FIG. 2, the code reader 1 mainly includes a detection unit 11, an imaging memory 12, a binary image memory 13, a storage unit 14, a display unit 15, an operation unit 16, a trigger switch 17, an external communication unit 18, a power supply A unit 19 and a control unit 20 are provided. Each part except the power supply part 19 is connected so that data communication is possible.

  Specifically, the detection unit 11 includes an imaging camera 21 that captures an image and an illumination 22. The imaging camera 21 is a two-dimensional camera using a COMS sensor, for example, and images a rectangular imaging range from the code reading device 1 in one direction as indicated by a broken line in FIG. The illumination 22 illuminates the imaging range by projecting light in the same direction as the imaging direction of the imaging camera 21 so that the code symbol can be clearly captured by the imaging camera 21 even in a dark place.

  The imaging memory 12 is a part that stores multi-value image data captured by the imaging camera 21 of the detection unit 11. The binary image memory 13 is a part that stores binary image data obtained by binarizing the multi-value image data stored in the imaging memory 12. In the present embodiment, binary image data composed of binary values of a white value (value 0) and a black value (value 1) is used. The storage unit 14 is a part that stores various types of information. For example, a decoding threshold value used for decoding a barcode is stored. The decode threshold has a reference value, and the decode threshold is a value that defines a certain range from the reference value. For example, the initial value of the decode threshold is defined according to the barcode standard to be read, and the ideal barcode space and bar width (also referred to as symbol character width) defined in the standard. Use the reference value. The symbol character width is normalized, and decoding is performed based on the ratio of each symbol character.

  The display unit 15 is a monitor of the code reading device 1 as shown in FIG. 1 and uses the operation menu guidance of the code reading device 1, the state of the code reading device 1, the code reading result, etc., in characters, charts, etc. This is the part that is displayed so that it can be viewed by the user. The operation unit 16 includes a plurality of switches as shown in FIG. 1, and is a part where the user performs a selection instruction operation or inputs information according to the display on the display unit 15. The operation unit 16 is not limited to a switch and may be a touch panel. The trigger switch 17 is a switch for instructing the start of imaging in the detection unit 11. For example, in the present embodiment, as illustrated in FIG. 1, the trigger switch 17 is a push button type switch provided in the code reading device 1.

  The external communication unit 18 is for communicating between the code reading device 1 and a computer device such as an external management server. The communication method in the external communication unit 18 is not particularly limited, and may be wired, infrared, wireless, or the like. The power supply unit 19 includes a primary battery or a secondary battery, and is a power source for operating the code reading device 1. A configuration in which power is supplied from the outside may also be used.

  The control unit 20 includes a CPU that performs arithmetic processing in the code reader 1, and performs operation control according to programs of basic software and application software.

  The control unit 20 includes an extraction unit 23 (code extraction unit), a decoding unit 24, and a learning unit 25. The extraction unit 23 analyzes the one-dimensional barcode in the image captured by the detection unit 11, and extracts the space and bar data of the one-dimensional barcode. The space and bar data of the one-dimensional barcode includes information such as the symbol character width and the position of each symbol character.

  The decoding unit 24 decodes the barcode extracted by the extraction unit 23 using the decoding threshold stored in the storage unit 14 and displays the decoding result on the display unit 15 or the external communication unit 18 to an external computer.

  The learning unit 25 uses the reference value of the decoding threshold used in the decoding unit 24 as teacher data, compares the teacher data with each symbol character width extracted by the extraction unit 23, and determines the corresponding symbol character width. The difference is calculated, and the teacher data, that is, the decoding threshold reference value stored in the storage unit 14 is updated so as to decrease the difference.

  Here, referring to FIG. 3, an operation flow of barcode reading by the code reading device 1 is shown. Hereinafter, the operation procedure of the code reading device 1 of the present embodiment will be described along the same flow.

  First, in step S1, the control unit 20 determines whether or not the code reader 1 has been activated for the first time after factory shipment. When the determination result is true (Yes), that is, when it is activated for the first time after factory shipment (hereinafter referred to as initial activation), the process proceeds to step S2.

  In step S2, since the code reading device 1 is in the initial activation state, the control unit 20 sets no decoding threshold value for decoding the barcode, so that the storage unit 14 is used as the initial value of the decoding threshold value. The decoding threshold value in accordance with the barcode standard to be read is acquired.

  On the other hand, if it is not the first activation in step S1, that is, if a learning process described later is performed and the decoding threshold is updated, the process proceeds to step S3. In step S <b> 3, the control unit 20 acquires the updated decoding threshold value from the storage unit 14.

  After acquiring the decoding threshold value in step S2 or S3, the control unit 20 captures an image with the imaging camera 21 of the detection unit 11 in accordance with the user's operation of the trigger switch 17 in step S4. The captured image data is first stored in the imaging memory 12, binarized, and then stored in the binary image memory 13.

  In subsequent step S5, the extraction unit 23 of the control unit 20 determines whether or not the barcode can be extracted from the binary image data. If the determination result is false (No), that is, if the barcode cannot be extracted from the binary image data, the process returns to step S4 and waits for the trigger switch 17 to be operated again. On the other hand, if the determination result is true (Yes), that is, if the extraction unit 23 can extract the barcode, the process proceeds to the next step S6.

  In step S6, the decoding unit 24 of the control unit 20 determines whether or not the extracted barcode can be decoded using the decoding threshold acquired in step S2 or S3. If the determination result is false (No), that is, if there is a symbol character outside the range of the decoding threshold, it is determined that decoding has failed, and the process proceeds to step S7.

  In step S7, the control unit 20 outputs an error display to the display unit 15, for example, prompts the user for manual input, and proceeds to step S8 after manual input by the user is completed.

  In addition, if the determination result in step S6 is true (Yes), that is, if the extracted barcode is successfully decoded, the process proceeds to step S8.

  In step S8, the control unit 20 outputs the decoding result or the manually input result to the display unit 15 or the external communication unit 18, and proceeds to step S9.

  In step S9, the learning unit 25 of the control unit 20 performs a learning process using the reference value of the decoding threshold used in step S6 as teacher data, and returns the routine.

  Here, referring to FIGS. 4 and 5, FIG. 4 shows an explanatory diagram regarding the decoding threshold value and the reference value in the present embodiment, and FIG. 5 shows an explanatory diagram of learning processing in the present embodiment. Based on these drawings, the learning process performed by the learning unit 25 in step S9 will be described in more detail.

  First, the bar code 3a shown in FIG. 4 includes four types of symbol characters, a narrow space, a thick space, a thin bar, and a thick bar, and a left margin and a right margin (and a margin margin) arranged at both ends for barcode identification. It is also a bar code standard having The barcode standard is not limited to such a format. For example, the present invention can be applied to a barcode standard having three or more types of spaces and bars. 4 and 5, in order to simplify the description, the number and width of the symbol characters do not always conform to the existing standard.

The bar code 3a illustrated in FIG. 4 is an ideal bar code whose symbol character widths are all in accordance with the standard. That is, the symbol character width and margin margin width of the barcode 3a are the reference values of the decoding threshold value according to the standard. In the following description, the left margin width is α _l , the right margin width is α _r , the narrow space width is β _s , the thin bar width is β _b , the thick space width is γ _s , and the thick bar width is γ _b. And Each reference value in accordance with this standard is stored in advance in the storage unit 14, and the control unit 20 acquires the reference value as an initial value of the teacher data at the time of initial activation as in step S2.

The decoding threshold is set so as to provide a certain range with respect to the reference value of each symbol character. For example, as shown in FIG. 4, the left margin decoding threshold α _lTH is an arbitrary value α _lTH (α _lTH > α _l ) greater than the reference value α _l , and the right margin decoding threshold is an arbitrary value greater than the reference value αr. Α _rTH (α _rTH > α _r ). The extraction unit 23 identifies those having a margin margin width within the decoding threshold of these margin margins as barcodes to be read, and extracts the data of each symbol character between these margins.

The fine space decoding threshold β _sTH is a range where the lower limit is 0.5β _s and the upper limit is β _s +0.5 (γ _s −β _s ), and the thin bar decoding threshold β _bTH is 0.5β _b, the range of the a _{β b +0.5 (γ b -β b} ) upper limit, the decoding thresholds gamma _STH lower limit _{β s +0.5 (γ s -β s} ) thick space, and the upper limit 1.5Ganma _s range gamma _STH that, thick bar decode threshold gamma _BTH is the lower limit _{β b +0.5 (γ b -β b} ), the range of the upper limit and 1.5γ _b. The decoding unit 24 performs decoding by determining a symbol character within the decoding threshold value in the extracted barcode.

  FIG. 5 shows an example of binarized data of teacher data and extracted data. As shown in FIG. 5, the teacher data is the same as the binarized data of the barcode according to the standard shown in FIG. 4, and the extracted data has a margin margin and the width of each symbol character with respect to the teacher data. Error. When this error is within the range of each decoding threshold, decoding by the decoding unit 24 is possible.

  In the learning process by the learning unit 25 in the present embodiment, the difference between the teacher data and the extracted data is calculated for each margin and each symbol character. For a plurality of symbol characters, an average value is calculated by adding all the differences and dividing by the number of the same kind of symbol characters. The teacher data is updated by adding a value obtained by multiplying the average value by an arbitrary convergence coefficient K to the teacher data used this time. The convergence coefficient K is arbitrarily set in the range of 0 <K ≦ 1, and the greater the value, the greater the degree of decreasing the difference, that is, the degree of learning reflection.

Specifically, in the extracted data shown in FIG. 5, there are five fine bars β _b1 to β _b5 , and the difference from the corresponding teacher data β _b is calculated for each of these, and converged to the average value The teacher bar data is updated by adding the value multiplied by the coefficient K to the teacher data β _b (β _b ← β _b + K × {(β _b −β _b1 ) + (β _b −β _b2 ) + ([beta] _{b- [} beta] _b3 ) + ([beta] _{b- [} beta] _b4 ) + ([beta] _{b- [} beta] _b5 )}} / 5). The same applies to other symbol characters, and a description thereof will be omitted.

If the decoding from the extracted data fails in step S6 described above, the teacher data is updated based on the data manually input in step S7. For example, in FIG. 5, when the first thin bar β _b1 is outside the range of the thin bar decoding threshold β _bTH and the symbol character cannot be decoded, the first thin bar β _b1 is _calculated from the value input manually. The width of one thin bar β _b1 is corrected, and the teacher data is updated using the corrected value of the first thin bar β _b1 .

  By repeatedly updating the teacher data in this manner, the difference between each symbol character of the extracted barcode and the teacher data decreases. For example, if there is a problem with the print quality of the printer, the printed surface is uneven, curved, or dirty, or if there is a flaw in usage by the user, some symbols Even if the width of the character tends to be narrower or wider than the prescribed width, the teacher data becomes a value suitable for the use environment by performing the learning process as described above. Thereby, the reading accuracy of the bar code can be improved and the frequency of manual input can be reduced regardless of the cause that the decoding becomes impossible.

  Therefore, according to the code reading learning system in the present embodiment, the code reading device 1 capable of executing the learning process as described above can easily improve the reading accuracy according to the use environment, and improve the work efficiency. Can be made.

  Next, a second embodiment of the present invention will be described. Referring to FIG. 6, there is shown a schematic configuration diagram of a code reading learning system according to the second embodiment, which will be described below with reference to FIG.

  Although the code reading learning system in the first embodiment described above performs from barcode reading to learning processing by the code reading device 1 alone, the code reading learning system in the second embodiment directly reads the barcode. The apparatus 30 includes a management server 50 that receives information from the code reading apparatus 30 and performs a learning process.

  As in the first embodiment, the code reader 30 includes a detection unit 31, an imaging memory 32, a binary image memory 33, a storage unit 34, a display unit 35, an operation unit 36, a trigger switch 37, an external communication unit 38, a power supply The unit 39 and the control unit 40 are provided. Note that the imaging camera and the proof of the detection unit 31 are not shown.

  The part different from the first embodiment is a control unit, and the control unit 40 of the second embodiment includes an extraction unit 41 and a decoding unit 42, but does not include a learning unit. The functions of the other parts are the same as in the first embodiment, and a description thereof will be omitted.

  On the other hand, the management server 50 includes a learning unit 51. The function of the learning unit 51 is the same as that of the first embodiment, and a description thereof will be omitted. The management server 50 includes an external communication unit 52, and is connected to the plurality of code reading devices 30 via the external communication unit 52 so as to be communicable.

  Referring now to FIG. 3, there is shown an operation flow of barcode reading by the code reading device 30 and the management server 50 of the second embodiment, and the code reading device 30 of the present embodiment and An operation procedure of the management server 50 will be described.

  First, in step S11, the control unit 40 of the code reading device 30 determines whether or not the code reading device 30 is in an initial activation state. If the determination result is true (Yes), the process proceeds to step S12.

  In step S <b> 12, the control unit 40 of the code reading device 30 acquires a decoding threshold value in accordance with the barcode standard to be read from the storage unit 34 as an initial value of the decoding threshold value.

  On the other hand, if it is not the first activation in step S11, the process proceeds to step S13. In step S <b> 13, the code reading device 30 acquires the updated decoding threshold value from the storage unit 34. The updated decoding threshold value here is the decoding threshold value updated in the learning unit 51 of the management server 50 and is stored in the storage unit 34 from the management server 50 via the external communication units 38 and 52. It is.

The barcode reading operation from image capturing to result display in steps S14 to S17 is the same as steps S4 to S7 in the first embodiment, and a description thereof will be omitted.
In step S <b> 18, the control unit 40 of the code reading device 30 outputs the decoding result or the result manually input to the display unit 35 and also to the management server 50 via the external communication unit 38.

  In step S19, the learning unit 51 of the management server 50 performs a learning process using the decoding threshold reference value as teacher data, and returns the routine. Here, the teacher data updated by the learning process, that is, the reference value of the decoding threshold value is distributed from the management server 50 to each code reading device 30 via the external communication units 52 and 38, and the storage unit of each code reading device 30 The reference value of the decoding threshold stored in 34 is updated. Note that the specific contents of the learning process are the same as those in the first embodiment, and a description thereof will be omitted.

  As described above, the code reading learning system according to the second embodiment performs the learning process on the management server 50 side. Thereby, the extracted data can be collected from each code reading device 30, and the reference value of the decoding threshold updated by the learning process can be reflected on each code reading device 30 in the same environment. Since a plurality of pieces of extracted data are collected in the management server 50 in a short time, the progress of the learning process is accelerated, the accuracy of each code reader 30 can be improved at an early stage, and the work efficiency can be further improved.

  This is the end of the description of the embodiment of the code reading learning system according to the present invention, but the embodiment is not limited to the above embodiment.

  In the learning process in each of the above-described embodiments, as shown in FIG. 5, the average value of the difference from the teacher data corresponding to the type is obtained for each symbol character, and the teacher data is updated by multiplying by the convergence coefficient K. However, the specific method of the learning process is not limited to this.

  Here, FIG. 8 shows (a) a first modified example and (b) a second modified example of the learning process, and another method of the learning process will be described with reference to FIG.

In the first modification shown in FIG. 8A, teacher data is provided for each type and position of each symbol character. For example, in FIG. 8A, there are five pieces of teacher data for the fine bars, β _ba , β _bb , β _bc , β _bd , and β _be in order from the left side to the right side in the figure.

The teacher data is updated by obtaining a difference between the teacher data and the extracted data corresponding to the type and position, and adding a value obtained by multiplying the difference by the convergence coefficient K to the teacher data. For example, if the teacher data of the leftmost thin bar is β _ba , the corresponding extracted data is β _b1 , and the teacher data to be updated is β _ba + K × (β _ba −β _b1 ).

  As described above, by performing the learning process for each type and position of the symbol character, it is possible to perform learning with higher accuracy.

Next, in the second modification shown in FIG. 8B, teacher data is provided for each type and range of symbol characters. For example, in FIG. 8B, the first range, the second range, and the third range are divided in order from the left side to the right side in the drawing, and teacher data is provided for each range. For example, the teacher data of the thin bar is β _ba for the first range, β _{bb for} the second range, and β _{bc for} the third range.

Then, the teacher data is updated by obtaining a difference between the teacher data in each range and the extracted data in the corresponding range, and adding a value obtained by multiplying the difference by the convergence coefficient K to the teacher data. . If there are a plurality of corresponding symbol characters in the range, the average value of the differences is multiplied by the convergence coefficient K. For example, since there are three β _b1 to β _b3 for the fine bars in the first range, the update of the teacher data is β _ba + K × {(β _ba −β _b1 ) + (β _ba −β _b2 ) + ( β _ba −β _b3 )} / 3. On the other hand, since each of the second and third ranges has one thin bar βb4 and βb5, updating of the teacher data in the range is performed by β _bb + K × (β _bb −β _b4 ), β _bc + K × (β _bc− β _b5 ).

  In this way, by performing the learning process for each symbol character type and for each range, it is possible to improve the learning accuracy while suppressing the amount of arithmetic processing as much as possible. This modification is effective when, for example, the printing surface of the barcode is curved, such as a can, and the error range is different between the central portion and the end portion.

  Further, in each of the above-described embodiments and modifications, the convergence coefficient K is an arbitrary constant value, but the convergence coefficient K may be individually set according to one or both of the type and position of the symbol character. For example, the symbol character at the position where decoding has failed can be dynamically varied so as to increase the value of the convergence coefficient K, thereby increasing the learning speed for the portion where decoding has failed.

  In the code readers 1 and 30 in the above embodiments, the barcode is detected by imaging the barcode with the imaging camera (CMOS sensor) of the detection unit 31, but the detection unit detects the one-dimensional barcode. Any means capable of detection may be used. For example, a laser or a CCD may be used.

DESCRIPTION OF SYMBOLS 1, 30 Code reader 11, 31 Detection part 12, 32 Imaging memory 13, 33 Binary image memory 14, 34 Storage part 20, 40 Control part 23, 41 Extraction part 24, 42 Decoding part 25, 51 Learning part 50 Management server

Claims (5)

A detection unit for detecting a one-dimensional barcode;
A code extraction unit that analyzes the one-dimensional barcode detected by the detection unit and extracts space and bar data of the one-dimensional barcode;
A storage unit for storing a decoding threshold for decoding a space and a bar of a one-dimensional barcode to be read;
The decoding threshold stored in the storage unit has a reference value, the reference value is used as teacher data, the teacher data is compared with the space and bar data extracted by the code extraction unit, A code reading learning system comprising: a learning unit that calculates a difference between corresponding data and updates a reference value of a decoding threshold value of the storage unit so as to reduce the difference. A detection unit for detecting a one-dimensional barcode;
A code extraction unit that analyzes the one-dimensional barcode detected by the detection unit and extracts space and bar data of the one-dimensional barcode;
A storage unit for storing a decoding threshold for decoding a space and a bar of a one-dimensional barcode to be read;
The decoding threshold stored in the storage unit has a reference value, the reference value is used as teacher data, the teacher data is compared with the space and bar data extracted by the code extraction unit, A learning unit that calculates a difference between corresponding data and updates a reference value of a decoding threshold value of the storage unit so as to reduce the difference,
The learning section, wherein in response to either one or both of the spaces and bars of the type and position of the one-dimensional bar code, and sets individually the degree of reducing the difference, code reading learning system. It further comprises an operation unit for the user to input information,
If the learning unit fails to decode the space and bar of the one-dimensional barcode to be read, the learning unit sets the reference value of the decoding threshold of the storage unit based on information input by the user using the operation unit. The code reading learning system according to claim 1, wherein the code reading learning system is updated. The said teacher data have the said decoding threshold value individually according to any one or both of the space of the said one-dimensional barcode, and the kind and position of a bar, The any one of Claims 1-3 characterized by the above-mentioned. The code reading learning system described in 1. A plurality of readers and a server capable of communicating with each of the readers;
The detection unit is provided in the reading device,
The code extraction unit is provided in the reading equipment,
The storage unit is provided in the reading equipment,
The learning unit, code reading learning system according to claim 1, characterized in that provided on the server.

Images