JPH09212569A - Method and device for processing bar code - Google Patents

Abstract

PROBLEM TO BE SOLVED: To decode a bar code capable of securing high accuracy and reliability. SOLUTION: Disappearing correction form is obtained at the time of using a fixed weight code when errors are in the same direction, a decoding part 6 by a parity can correct the number of characters twice as many as the case of unfixed weight code and narrowing the object candidate of decoding by the combination of the content of the parity and the restoring position of a bar in an erroneous bar code with respect to the parity having a certain number of characters. However, the processing mentioned above follows an encoding system in which decoding is limited within the range of the shape of the bar code, satisfactory performance is not attained. Then, a part 7 decoding by means of a rule and a part 8 decoding by means of a data base respectively refer to rules depending on the object of coding, a new zip code book, etc., stored in a data base 8a to narrow the candidate of an original character string.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is, for example, a bar code code representing a postal code and a numeric part of an address, which is printed on a postal matter such as a letter or postcard and is used later for sorting the postal matter by destination. The present invention relates to a bar code processing method and a bar code processing system that perform encoding and decoding processing.

[0002]

2. Description of the Related Art Conventionally, there is a three-digit or five-digit postal code represented by characters as a code on a postal matter, and characters are recognized based on an image optically read by a postal code automatic reading / sorting machine. As a result, the zip code was read. Further, the address and the address were read by character recognition based on the image optically read by the automatic mail address sorting machine. Then, according to the information of the read zip code and address, the mail is classified according to the destination.

As described above, the conventional technique is based on the so-called character recognition technique of reading the zip code and address written in characters. However, the character recognition performance, speed, and cost of machinery required were not satisfactory for use in sorting mail by destination.

Therefore, in recent years, it has been studied to construct a mail sorting system with higher performance, higher speed and lower cost by encoding characters into a bar code. Specifically, at the Central Office, new postal codes (destination place name information, etc.), address display numbers (destination place number information, etc.) different from the current ones were used in accordance with the dedicated bar code system on the address display surface of postcards, envelopes, etc. It has been considered to improve the efficiency of work by printing as a bar code and reading the bar code at a sending station or a delivery station to automate sorting of mail items by destination.

The character string encoded in the bar code is, for example, a new postal code consisting of 7 digits, a chome,
It is an address display number consisting of a number, a number, etc., and is expressed as 1-2-3-A-405 in the room No. 405 of the apartment building No. 1-3 of 1-chome, and the new postal code part of the barcode and the address. A plurality of parities such as a general-purpose Reed-Solomon code are added to each display number portion to perform code error correction.

[0006]

The mail and the bar code indicating the address display number are printed at a predetermined position on the address display surface, but the position of the letter such as the address written by the sender of the mail is not correct. It's a rule. Therefore, especially in the case of barcodes printed with special ink such as fluorescent light (in-office barcodes) in order to reduce contamination of the mail in the post office, overlap with the address letter, background color of the mail, pattern or picture When the barcode is read, the bar may fall off due to a decrease in the amount of light emitted from the ink.

In this case, when error correction is performed on a code coded by a general coding method, it is usually possible to correct an error of only one character with respect to a parity code of two characters. There is a problem that the accuracy and reliability of the code decoding process are lacking.

Therefore, according to the present invention, a constant weight code suitable for erasure correction in the case where there is directionality in the error such that the bar is dropped (or the bar is added due to ink rubbing, scattering, etc.). The character of the bar is composed by the bar, and the bar that can be corrected as much as possible at the time of decoding the bar code by using the rule and the database for the erasure correction so that the erasure correction can be easily performed for the error exceeding the conventional parity. An object is to provide a code processing method and a bar code processing system.

[0009]

According to the bar code processing method of the present invention, an error correction code is added to a character string, and a code word obtained by encoding into a constant weight code that produces a directional error is obtained. In the barcode processing method of converting to a barcode and printing on a predetermined print object, reading the barcode from the print object, and restoring the original character string, the barcode read from the print object Is converted into a code word, erasure correction is performed based on the error correction code, and when erasure correction cannot be performed based on the error correction code, the weight is determined based on a predetermined condition depending on the original character string. By detecting the original character string from the character string candidates of the original character string extracted based on the characteristics of a constant code and the error correction code, it is possible to decode a barcode that can ensure high accuracy and reliability. To become.

Further, in the bar code processing method of the present invention, an error correction code is added to a character string, and a code word obtained by encoding into a constant weight code that generates a directional error is converted into a bar code. Then, the barcode is printed on a predetermined print object, the barcode is read from the print object, and in the barcode processing method of restoring the original character string, the barcode read from the print object is used as a code word. When the erasure correction is performed by performing the conversion and the erasure correction cannot be performed based on the error correction code, the characteristics of the constant weight code and the error are determined based on a predetermined rule of the original character string. The original character string is detected from the character string candidates of the original character string extracted based on the correction code, and when the original character string cannot be detected, the character string limited based on the rule of the original character string From the accessory, by detecting a character string within the representation range of the source string, high accuracy and reliability it becomes possible to decode the bar code can be ensured.

Further, the bar code processing system of the present invention converts a code word obtained by adding an error correction code to a character string and encoding it into a constant weight code which produces a directional error into a bar code. And print on the specified print target,
In a barcode processing system that reads a barcode from the print object and restores the original character string, the barcode read from the print object is converted into a code word, and erasure correction is performed based on the error correction code. When the erasure correction cannot be performed by the correction means and the correction means to be performed, the element extracted based on the characteristics of the constant weight code and the error correction code is based on a predetermined condition depending on the original character string. By including the detection unit that detects the original character string from the character string candidates of the character string of 1, the barcode can be decoded with high accuracy and reliability.

Further, in the bar code processing system of the present invention, an error correction code is added to a character string, and a code word obtained by encoding into a constant weight code which generates a directional error is converted into a bar code. And print on the specified print target,
In a barcode processing system that reads a barcode from the print object and restores the original character string, the barcode read from the print object is converted into a code word, and erasure correction is performed based on the error correction code. When the erasure correction cannot be performed by the correcting unit and the correcting unit that performs the correction, the original character string extracted based on the characteristics of the constant weight code and the error correction code is determined based on a predetermined rule of the original character string. First detecting means for detecting the original character string from character string candidates; and when the original character string cannot be detected by the first detecting means, the predetermined rule of the original character string And a second detecting means for detecting a character string within the expression range of the original character string from the character string candidates limited based on No. of it is possible.

Further, in the bar code processing method of the present invention, an error correction code is added to a character string, and a code word obtained by encoding into a constant weight code that generates a directional error is converted into a bar code. Then, the barcode is printed on a predetermined print object, the barcode is read from the print object, and in the barcode processing method of restoring the original character string, the barcode read from the print object is used as a code word. When the erasure correction is performed based on the error correction code after conversion, and the erasure correction cannot be performed based on the error correction code, the directionality of the error is determined based on the weight of the code word having the erasure error, and as a result, the code When it is determined that the direction of the error of the word is likely to be one direction, based on the predetermined constant character string dependent condition, based on the characteristics of the constant weight code and the error correction code. Extraction By detecting the original string from the candidate of the original string, high accuracy and reliability becomes possible to decode the bar code can be ensured. Furthermore, even in a system in which miscorrection has a significant effect, it is possible to reduce miscorrection by not performing decoding based on the condition depending on the original character string when an error having no directionality occurs, and thus, , System reliability is improved.

Further, in the bar code processing method of the present invention, an error correction code is added to a character string, and a code word obtained by encoding into a constant weight code that generates a directional error is converted into a bar code. Then, the barcode is printed on a predetermined print object, the barcode is read from the print object, and in the barcode processing method of restoring the original character string, the barcode read from the print object is used as a code word. When the erasure correction is performed based on the error correction code after conversion, and the erasure correction cannot be performed based on the error correction code, the directionality of the error is determined based on the weight of the code word having the erasure error, and as a result, the code When it is determined that there is a high possibility that the error direction of the word is one direction, it is extracted based on the characteristic of the constant weight code and the error correction code based on the rule of the original character string set in advance. The original character string is detected from the character string candidates of the original character string, and when the original character string cannot be detected, from the character string candidates limited based on the rule of the original character string, the original By detecting the character string within the expression range of the character string, it is possible to decode the barcode with high accuracy and reliability. Furthermore, even in a system in which miscorrection has a significant effect, it is possible to reduce miscorrection by not performing decoding based on the condition depending on the original character string when an error having no directionality occurs, and thus, , System reliability is improved.

Further, the bar code processing system of the present invention converts a code word obtained by adding an error correction code to a character string and coding it into a constant weight code which produces a directional error into a bar code. And print on the specified print target,
In a barcode processing system that reads a barcode from the print object and restores the original character string, the barcode read from the print object is converted into a code word, and erasure correction is performed based on the error correction code. Correcting means to be performed, judging means for judging the directionality of the error based on the weight of the code word having an erasure error when the erasure correction cannot be performed by the correcting means, and the direction of the error of the code word is one direction by this judging means. When it is determined that there is a high possibility that the original character extracted based on the characteristics of the constant weight code and the error correction code, based on a condition that depends on the predetermined original character string. By including the detection unit that detects the original character string from the character string candidates of the string, it becomes possible to decode the barcode with high accuracy and reliability. Furthermore, even in a system in which miscorrection has a significant effect, it is possible to reduce miscorrection by not performing decoding based on the condition depending on the original character string when an error having no directionality occurs, and thus, , System reliability is improved.

Further, the bar code processing system of the present invention converts a code word obtained by adding an error correction code to a character string and coding it into a constant weight code which produces a directional error into a bar code. And print on the specified print target,
In a barcode processing system that reads a barcode from the print object and restores the original character string, the barcode read from the print object is converted into a code word, and erasure correction is performed based on the error correction code. Correcting means to be performed, judging means for judging the directionality of the error based on the weight of the code word having an erasure error when the erasure correction cannot be performed by the correcting means, and the direction of the error of the code word is one direction by this judging means. When it is determined that there is a high possibility that the character string of the original character string extracted based on the characteristics of the constant weight code and the error correction code, based on a predetermined rule of the original character string. First detecting means for detecting the original character string from among the candidates, and based on the predetermined rule of the original character string when the original character string cannot be detected by the first detecting means. Decoding a barcode that can ensure high accuracy and reliability by including a second detection unit that detects a character string within the expression range of the original character string from the determined character string candidates. Is possible. Furthermore, even in a system in which miscorrection has a significant effect, it is possible to reduce miscorrection by not performing decoding based on the condition depending on the original character string when an error having no directionality occurs, and thus, , System reliability is improved.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. First, referring to FIG.
The constant weight code will be described. FIG. 2 (a) to (c)
Is an allocation example in the case where the number of code words is 15, in which one character is allocated to one code composed of three bars of 2 bits 4 states (upper, lower, upper, lower). Note that FIG.
In the figure, a bar code having a constant weight is shown.

2A shows the case of weight (the number of bits having a bar) of 4, and FIG. 2B shows the case of weight 3 of FIG.
(C) is a case of weight 2. The words (characters) to be encoded are numbers from 0 to 9, − (hyphen), and several control codes (for example, C1, C
2, C3, etc.) are assumed. The control code is used in combination with numbers to represent the English letters of the apartment building number in the address. For example, if "A" is represented by "C1" + "0", using C1, C2, and C3,
It will be able to represent 26 English characters. C4 is a control code for extension described later. That is, by using the extension control codes C4, C5 ..., More words can be encoded.

When the error has directionality, the constant weight code is in the form of so-called erasure error correction (erasure correction) in which the error position is known, so that only the number of parities, that is, twice the number of ordinary errors is detected. There is an advantage that the character error of can be corrected.

Here, the character string to be encoded is a new zip code consisting of 7-character numbers, and as shown in FIG. 3, a 1-character parity is added to the 7-character new zip code for encoding. Shall be done. 2 bits 4
Since one character is represented by three bars in the state, 1
The character is composed of 6 bits, and the parity is also composed of 6 bits.

By using the constant weight code, if the erasure correction is performed, the error correction of one character can be performed by the parity of one character. For example, even if a 1-bit disturbance occurs in the entire bar code of 7 + 1 = 8 characters, at most, an erasure error of one character occurs, and the error is corrected regardless of the weight value.

Next, a case where a disturbance in the direction of dropping the 2-bit bar occurs will be described. In this case, the number of erroneous characters and the number of erroneous character error bits differ depending on whether or not the occurrence location is a bit with a bar. For example, for a code with a weight of 4, 48C2 (Note: 6 bits / character x 8 characters = 48 bits) = 112
Of the eight disturbances, if two characters are erroneous by one bit, then 8 C2 × 4 C1 × 4 C1 = 448.
The others are at most one character error and are corrected by the parity.

It is difficult to correct an error of 2 characters by a conventional mechanical method, but in the case of a code with a weight of 4, there are 3 candidates of bar restoration positions in one character, so there are 448 2 characters. 3 × for each of the errors
If it can be specified which of 3 = 9 ways, it can be corrected.

In the code of weight 3, two characters out of 1128 are erroneous by one bit because 8 C2 × 3 C1 × 3
With C1 = 252, there are 4 × 4 = 16 candidates for each.

With the code of weight 2, 2 out of 1128 ways
Characters are erroneous bit by bit 8 C2 × 2 C1 × 2
With C1 = 112 ways, there are 5 × 5 = 25 ways for each.

Next, a method of specifying restoration candidates will be described. Assuming that the new postal code is composed of numbers, in the case of a code with a constant weight of 15 code words as shown in FIG.
For example, if the parity is also a number, "-""C1""C
Since "2", "C3", and "C4" are omitted from the candidates, generally, 5/15 candidates for each character are excluded.
Further, since it is considered that the new postal code uses only part of the numbers expressed by 7 characters, if there is a new postal code list, it is possible to further narrow down the candidates by using it as a database. Depending on the degree of narrowing down the candidates, which of the code of weight 4 and the code of weight 2, and in some cases, the code of weight 3 has the best performance changes.

An effective narrowing-down means can be assumed for the address display number portion. You can use rules such as no hyphens at the beginning or the end, and no more than two hyphens in a row. If there is a database of address display numbers, there is a possibility that it is possible to use information such that the block number of a certain street is only 1 to 15.

The disturbance of 3 bits or more generally has a lower frequency of occurrence, and therefore its explanation is omitted, but the principle is that erasure correction is possible as in the case of 2 bits described above. In this way, when a one-way error occurs, by using a constant weight code, no character can be transformed into another character. Therefore, even if the number of character errors exceeds the parity correction capability, You can concentrate on the correction by narrowing down the candidates. Further, even if the correction cannot be made, no erroneous correction occurs and the entire code can be treated as a reject.

Next, the control code for extension will be described. In general, it is desirable in terms of performance that the number of code words is set to the minimum necessary. In the case of the present embodiment, if the number of words is changed from 15 to 14, it is considered that the candidates for the bar restoration position are 14/15 in total. However, the number of characters is set to 15 with the provision of C4 in consideration of future expansion such as support for numbers and characters other than English letters.

Which code is to be C4, in other words 1
Whether to omit the code word when the number of code words is limited to four may be selected from the viewpoint of performance. In the example of FIG.
The code having many bars in the upper half of the bar code is C4. This assumes the following:

FIG. 4 shows a specific example of the position where the bar code on the mail P is scheduled to be printed. Bar codes as shown in FIG. 2 are printed side by side at positions shown in FIG. 4 from left to right, and are printed near the lower end in parallel with the lower end. What is expected when evaluating the printing position is that the susceptibility to disturbance due to the letters of the address etc. differs depending on the position from the bottom edge. For example, at a position very close to the lower end, it is considered that the disturbance tends to be greater toward the inside of the mail P (in the direction away from the lower end). Therefore, it is considered that a code having many bars in the upper half of the bar code is more likely to be disturbed. Note that the code of weight 3 can be expressed up to 20 words, but in FIG. 2B, five codes including a code having a bar only in the upper half of the bar code are excluded. The codes of weight 2 and weight 4 can express exactly 15 words, respectively. Codes with a weight of 1 or less and 5 or more are possible, but they cannot be expressed because 14 words considered necessary here cannot be expressed. When the number of bars is four or more, the constant weight code can be similarly constructed. The possible weight values are different from those in the case of three bars, but the description is omitted.

The concept of the position susceptible to this disturbance is 14
It can also be used for intra-word assignment. That is, in FIG. 2, numbers, especially young numbers, are assigned to the upper half of them with fewer bars under the assumption that they will be used frequently. In addition to that, FIG. 2B is an example in which a bar-free code is assigned to the upper half under the assumption that the hyphen is used more frequently.

Another way of thinking is to reduce the number of consecutive bar-free positions when arranging bar codes. This is because the print position of the entire barcode and the extraction of the pitch may be adversely affected. In the case of FIG. 2A, when the bar codes corresponding to the numbers “7” and “9” are arranged from the left, two bar-free positions continue without any bar drop error. "9" to avoid this as much as possible
It is also conceivable to replace the signs of "C4" with ".

Note that the control code for expanding such character representation needs C1, C2, and C3 so that at least English letters can be represented. Therefore, the number of code words is at least 14 (hyphen (- )including)
Is required. That is, C4, C5, etc. are added in order to enable representation of characters other than numbers and English characters, and the number of code words must be 14 or more.

Next, using the constant weight code as described above, an encoding process of encoding a new postal code and an address display number and printing a bar code on a mail piece, and a printing process from the mail piece A bar code processing system that performs a decoding process of reading a bar code, decoding with a parity, decoding according to a predetermined rule with a weight value, and decoding based on the contents stored in a database will be described.

FIG. 1 shows a first embodiment of the present invention.
For example, it schematically shows the configuration of a bar code processing system used for mail processing. In FIG. 1A, the character string generating unit 1 is, for example, a postal code entered in a postal code entry field of a postal matter (bar code print target) such as a postcard or a letter of mail, and an address display number in the addressee. Is read by a scanner, character recognition processing is performed based on the obtained image signal, and character information is extracted. Then, from the extracted character information, if it is an address display number, 1-2 chome
No. 405, Apartment A Building No. 405 is “1-2-3-A-4
For example, in the case of a zip code, for example, in the case of a zip code, the character string is expanded (changed) by a system agreement expressed as "05" The character string on the system in which the postal code "123-4567" is "1234567" is expanded (changed).

The parity adding section 2 adds a parity corresponding to the character string output from the character string generating section 1 and outputs it to the bar code and printing section 3. In the bar code conversion / printing unit 3, the input character string is converted into each code of constant weight code for each character (see FIG. 3), converted into a bar code pattern, and then a corresponding mail is sent according to the bar code pattern. A bar code representing the postal code and address display number is placed at a predetermined position on the address display surface of the object.
Printing is performed, for example, with fluorescent ink together with a stop detection bar.

On the other hand, in FIG. 1 (b), the bar code reading section 5 reads the bar code printed on the postal matter with a scanner and converts it into a code for each character based on the read bar code information. Then, when the dropping of the bar is recognized, the dropping bar is grasped (for example, the position and the number of the dropping bar are grasped), and based on the feature of the constant weight code, the original character string Extract candidates.

The parity decoding unit 6 decodes the original character string based on the information of the barcode read by the barcode reading unit 5 by the erasure correction and the candidate of the restored character (hereinafter also referred to as the restoration candidate). Turn into.

The decoding unit 7 according to the rule narrows down the candidates for the restored character and decodes the bar code information into the original character string, for example, according to the rule that "the original character string is composed of only numbers". To do.

The database decoding unit 8 further narrows down the restored character candidates by referring to the database 8a of the new postal code book and address display number book prepared in advance, and decodes the bar code information into the original character string. To do.

Next, with reference to the flow chart shown in FIG. 5, the procedure of decoding the barcode (new zip code portion) in the decoding unit 6 by the parity of the barcode processing system of FIG. 1 will be described.

Since the constant weight code is used here, the parity decoding unit 6 can decode the number of characters twice the normal number by erasure correction. Further, it is possible to narrow down the candidates for the restoration position of the bar of the character, that is, the restoration candidates of the code word to some extent. For example, "227-
It is assumed that there is a new postal code "0048". 7 digits excluding hyphens are encoded into a bar code, but with the addition of one parity modulo 15 with a code of weight 4 (for example, if the total sum of parity is 5), In this case, the parity value is 12 and the corresponding word is C2).

The first 2 of the 448 errors described above
3 out of 4 out of 16 inaccurate digits
Each of the restoration candidates of “0470048C” in addition to the correct candidate “2270048C2”.
2 "," 4070048C2 "," 8C170048C
2 "," C1870048C2 "wrong candidates remain.

Now, in the flow chart shown in FIG. 5, one of the 16 wrong candidates "04700
The decoding process by parity will be described by taking the disturbance that causes "48" as an example. From the barcode information read by the barcode reading unit 5, a parity check is performed on the 8-digit read character (steps S10 to S11), and if they match, the read content is output as a result (step S15). Since the parity for one character is used here, if there is a one-character loss error, the parity is used to correct it and the result is output (steps S12 and S15). The erasure correction process using this parity is a known technique. If there is an erasure error of two or more characters (step S12), code words satisfying the parity are selected from the combinations of the bar restoration position candidates of each character, and all are output (steps S14 and S15). In this example, the erasure error is 2 characters. When the upper left of the leftmost digit and the lower left of the second digit drop out from the code representing the numeral 2, three restoration candidates of 0, 1, 2 and 2, 4, 6 are generated respectively. When a combination that satisfies the parity is examined among the 3 × 3 candidates, “0470048” and “227004”
8 ”(correct answer). These are output as a result.

With only the above-mentioned decoding by parity,
Up to this point, we have to reject the code. Next, when the rule-based decryption unit 6 follows the rule that only numbers are used,
In the case of 6 cases, "0
It can be narrowed down to two, "470048" and "4070048".

If there is no new postal code book database, narrowing down is done up to this point in this example, but in general, the possibility that the wrong candidates will be eliminated by the narrowing down and decoding is possible is higher in principle than only decoding by parity. If the parity is applied to the parity at the time of adding the parity in the parity adding unit 2, and the parity is modulo “10”, for example, the decoding performance by the parity is generally improved.

Furthermore, it is highly possible that the remaining two erroneous candidates can be eliminated by the database of the new postal code book in the database decoding unit 8. That is, the new postal code is a 7-digit number string, that is, "00000
Of the 10 million sets from "00" to "9999999", only those actually used are stored in advance in the database 8a as a new postal code list, and the database decoding unit 8 uses this database. A process of deleting from the candidates a number string that is not stored in 8a is performed.

FIG. 6 shows an example of storage of the new postal code book in the database 8a. The new postal code is stored in association with the town name (not including the chome portion) in a certain administrative district. Here, it is assumed that the town names are arranged in the order of the Japanese syllabary, but they may be sorted using the new postal code as a key in order to speed up the search.

Further, FIG. 6 shows the part of the new postal code mentioned in the above-mentioned example.
If "270048" and "0470048" remain, the database compounding unit 8 performs a process of deleting the number string "0470048" from the candidate because the database 8a does not have a number string "0470048".

The parity decoding unit 6 of FIG.
Decoding section 7 by rule, decoding section 8 by database
At each processing step, if only one candidate is narrowed down and decoded, the subsequent processing is terminated.
Also, if the decryption is not performed even in the final stage decryption (decoding section 8 by the database in FIG. 1), it is rejected.

The database decryption unit 8 can also be moved to an earlier stage, for example, before the rule decryption unit 7. Further, the database 8a may store not only the new postal code book and the address display number book but also a predetermined rule referred to by the rule decryption unit 7. Further, when the database 8a can be used for referring to the new postal code book, if there is no problem in the access speed, or if the decryption by the rule alone is not possible in many cases, the decryption by the decryption unit 7 by the rule can be omitted. Is.

Although the new zip code portion has been described above, the address display number portion is also decrypted using the parity, the rule and the database in the same manner as described with reference to FIGS.

FIG. 7 shows an example of storage of the address display number book in the database 8a. For example, in the case of the numerical values of the chome, address, and number of the xxx ward of Tokyo, as shown in FIG. 7, for example, the chome is from 1 to b1 and its i (1≤i≤b1) The number says that there are from No. 1 to No. g2. Although it may not always start from No. 1 or there may be a missing number, it will not be described in detail because it is not essential as it is devised appropriately. Also, in the area where the house display is not implemented, the arrangement of numbers in the address is more complicated, but the same idea may be used.

In reality, the bar code may be decoded in a device having a character recognition function. In that case, even if the bar code processing system as shown in FIG. 1 is subjected to the decoding process and a rejection occurs, the candidates can be expected to be reduced by matching with the reading result of the characters of the address and the new postal code.

As described above, if the constant weight code is used when the error is unidirectional, erasure correction is performed. In the parity decoding unit 6, the constant weight code is applied to a certain number of characters of parity. If not, usually a maximum of 2
Double the number of characters can be corrected, and the candidates for decoding can be narrowed down by combining the content of parity and the restoration position of the bar of the wrong barcode. However, the decoding up to this point is limited to the encoding system and the range of the shape of the barcode, and it cannot be said that sufficient performance is obtained. Therefore, each of the rule-based decoding unit 7 and the database-based decoding unit 8 uses a rule (for example, may be stored in advance in the database 8a) or a database 8a depending on the target of encoding. Only then can high performance decoding be achieved easily.

If a code other than the constant weight code is used, it is impossible to narrow down the candidates by applying a simple rule or database due to the possibility of garbled characters, and it is difficult to effectively use the rule or database. Is.

Now, when the cause of the dropout of the bar and the cause of the addition of the bar occur at the same time, or the cause of some non-unidirectional error occurs, an error that is not completely unidirectional can occur. At this time, as described above,
When error correction is performed in the parity decoding unit 6, the rule decoding unit 7, and the database decoding unit 8 according to the condition depending on the original character string, an error corresponding to the number of parities, which is the range of the code logic, is generated. The result may be worse than if only correction was made.

In a system where erroneous correction has a significant effect, it is better that erroneous correction is small even if the number of rejects increases. Therefore, in such a system, it is effective to judge whether or not the correction is limited to the range guaranteed by the code theory, depending on the degree of the error directionality not being one direction.
For example, in the case of postal mail, if an error occurs in the postal code, the postal matter will be sent to a post office far away from the destination, causing a delay in delivery.

Therefore, in such a case, the bar code processing system having the decoding unit having the configuration shown in FIG. 8 is effective. FIG. 8 schematically shows the configuration of the barcode decoding unit of the barcode processing system according to the second embodiment of the present invention. When an error occurs when reading the barcode, The directionality is judged and the correction is limited to the range guaranteed by the code theory or the error correction is performed under the condition depending on the original character string. The same parts as those in FIG.
Only different parts will be described.

In FIG. 8, the functions and operations of the barcode reading unit 5, the parity decoding unit 6, the rule decoding unit 7, and the database decoding unit 8 are the same as those in FIG. That is, when the bar code reading unit 5 reads a bar code on a mail, for example, the first half of the parity decoding unit 6 performs error correction within the range of the number of error characters guaranteed by the conventional code theory. Decrypt. Then, in the latter half of the decoding unit 6 by parity, the decoding unit 10 by the rule, the decoding unit 8 by the database, and the decoding unit 10 which depends on the original character string composed of the database 8a, According to the output from the sex determination unit 11,
The error which could not be corrected in the first half of the parity decoding unit 6 is corrected or not corrected.

An example of a code as shown in FIG. 9 which is composed of three vertically divided bars and a character having a weight of 6 bits (number of bits) 3 will be described below.
Here, in particular, it is assumed that a 4-bit disturbance occurs in a code composed of 48 bits, which is a total of 8 digits including a new zip code 7 digits and a parity 1 digit.

If the disturbance is either the dropout or the addition of the bar, that is, the error direction is one direction, then 48C4 =
128 out of 194580 patterns
There are 562 erasure errors of two or more characters (errors whose position is known), and these codes must be treated as rejects within the scope of code theory. However, many of them can be corrected by using the barcode processing method of the first embodiment described above.

On the other hand, when the disturbance is not unidirectional, for example, when the bar is missing 3 bits and additional 1 bit, better results may not be obtained by the same method. For example, there is an error of 1 character (garbled character) and a deletion error of 2 characters. If error correction is performed using a condition depending on the original character string in order to correct only the erasure error of two characters, the correction result will be selected from the erroneous character string candidates. As a result, erroneous correction may occur. They can be rejected by correction up to the range guaranteed by the code theory. Further, it is difficult to set a character string candidate on the assumption that one or more characters are garbled, because the number of candidates is significantly increased.

For example, assume that characters are assigned to the barcode as shown in FIG. They are ten numbers (“0” to “9”), hyphens (“−”), and control codes (“C1” to “C5”) for expanding characters.

Here, as in the first embodiment, "2270" is set.
Take the example of the new postal code "048". For the numerical value of parity, for example, the checksum modulo 16 is "0".
Then, the value becomes "9", and the numerical value of each digit at that time is "32". Due to the disturbance, the addition and removal of the bar to the first 2 bits of the first character "2" causes the character to become "8". Further, when the bar of the bit having the first bar of the next character "2" and the bar of the next character "7" is dropped, it is determined that the 1-bit bar is dropped for both of them, resulting in an erasure error. For each, there are four possible bit restoration candidate positions.

In terms of characters, there are two of the four characters, "2" and "8", which are the restoration candidate characters when the bar of the character "2" is missing. In addition, the restoration candidate characters of the missing character "7" bar are "1", "3", "5",
It is "7". If the first character “2” is garbled to “8” and a candidate satisfying the parity condition is selected, the entire new postal code will be decoded into an incorrect number “8210048”.

If the correction is limited to the range guaranteed by the code theory, it can be seen that there are two characters with missing bars, so that the entire code can be rejected as being uncorrectable with one parity character.

Next, as a method of determining the strength of the error directionality in the error directionality determining unit 11 of FIG. 8, a case of determining the error direction of an erasure error will be described. If the weight of the code is w, normally w =
It is 3 and when it is not "3", it can be seen that an erasure error has occurred in the character. When an erasure error of w is "2" or less occurs, it is the dropout direction of the bar, and when an erasure error of w is "4" or more is the addition direction of the bar.
The strength of the direction can be determined by the number of letters lost.

For example, there is a three-character loss error, and the weight, that is, the number of bars, is "0",
Suppose they were "2" and "4". The average is "2". At this time, the error directionality determining unit 11 determines that the error in the dropping direction is strong if the average weight of the lost characters is “2” or less, and if the average weight is “4” or more, for example. , It is determined that the error in the addition direction is strong, and as described in the first embodiment, the decoding unit 10 that depends on the original character string.
Error correction. When the average of the weights is "3", it is determined that the directionality of the error is weak, and in this case, the correction is performed up to the range guaranteed by the code theory in the decoding unit 6 by the parity, and the range is exceeded. In case, it is possible to reject.

Next, with reference to the flow chart shown in FIG. 10, the procedure of decoding the barcode (new postal code portion) in the decoding section of the barcode processing system of FIG. 8 will be described. The same parts as those in FIG. 5 are designated by the same reference numerals, and different parts will be described.

First, a parity check is performed on the read character of, for example, 8 digits from the information of the bar code read by the bar code reading section 5 (step S1).
0 to step S11), and if they match, the read content is output as a result (step S25). Since the parity for one character is used here, if there is a one-character loss error, the error is corrected using the parity, the result is output, and the process ends (steps S20 and S25).

If the correction corresponding to the number of parities cannot be made in step S21, the processes in and after step S22 are performed. In step S22, the error directionality determination unit 11 obtains the average of the weights of the lost characters by the method as described above, and calculates the average value and the threshold value that determines that the error directionality is strong (for example, , When the code weight w = 3,
The threshold value for judging that the error in the addition direction is strong is “4”, and the threshold value for judging that the error in the drop direction is strong is “2”), and the average value of the weights of the lost characters is added. It is checked whether it is equal to or more than a threshold value at which it is determined that the error in the direction is strong, or less than or equal to a threshold value at which the error in the drop direction is determined to be strong (step S23).

When it is judged that the direction of error is strong,
As described in the first embodiment, the decoding unit 10 that depends on the original character string performs error correction based on the condition that depends on the original character string, and outputs the result (step S24).
-Step S25).

In step S23, when it is determined that the error directionality is weak (for example, when the code weight w = 3, the threshold value determined to be strong in the addition direction error is "4", and the drop direction error is When the threshold value that is judged to be strong is “2”, and when the average value of the weights of the lost characters is “3”, decoding is performed by the parity decoding unit 6 by correction up to the range guaranteed by the code theory. The converted result is output or the barcode is rejected (step S25).

As another method for judging the strength of the error directionality in the error directionality judging section 11, for example, the following method can be considered. That is, the original weight w of the code is used as a reference value, and the result of cumulative addition of the difference between this reference value and the weight of the character with the erasure error is determined to have a strong predetermined error directionality. The directionality of the error is determined in the same manner as above by comparing with the threshold value.

Similarly to the above, the code weight w = 3, there is an erasure error in three characters, and the weights are "0", respectively.
A specific description will be given by taking the case of "2" and "4" as an example. That is, the error between the weight of the character having the erasure error and the reference value “3” is cumulatively added to obtain (3-0) + (3-2).
= 3. At this time, for example, if the threshold for determining that the error in the addition direction is strong is “−3” and the threshold for determining that the error in the drop direction is strong is “3”, it is determined that the error in the drop direction is strong. , As described in the first embodiment,
The decoding unit 10 depending on the original character string performs error correction.

Further, in step S23, the strength of the directionality of the error may be determined for each mail piece or for a plurality of mail pieces.
That is, when the mail pieces are continuously processed, the average of the weights of the characters having the erasure error or the cumulative addition value of the error with the reference weight w is calculated for each certain number of recent mails, and those values are calculated. It is possible to judge the direction of error based on. When the fluctuation is large, it is possible to switch the judgment frequently with a smaller number of copies. Then, in the case of collectively processing mail items with a certain property in advance,
Either condition that is appropriate only during that time, that is, one by one,
Alternatively, it is possible to set a certain number of copies.

Further, there is a fixed format such as a postcard for contacting the electricity charges of local customers. It is also possible to empirically understand the nature of the error. Therefore, which method is used while processing the postcard, that is, whether the decoded result is output by the correction to the range guaranteed by the code theory in the decoding unit 6 by the parity, or It is also possible to determine in advance whether the decoding unit 10 that depends on the original character string outputs the result of performing error correction under the condition that depends on the original character string.

In the code system shown in FIG. 9, when the erasure error is matched, it may be possible to determine whether the erasure error occurs in both the bar drop direction and the addition direction at the same time. For example, when the original weight w of the bar is "3" and the weight of the character with the erasure error is "2", it is determined that one bar is not dropped (-1). What can be done is when copying a bar to any of the four lost recovery candidate positions in any of the remaining four bar patterns of 6 C3 = 20 unassigned codes. In this case, when determining the error direction, the average of the weights of the characters having the erasure error or the cumulative addition value of the error with the reference weight w is calculated in consideration of such a determination result. You may

As described above, according to the second embodiment, the barcode read by the barcode reading unit 5 is converted into a code word, and the parity decoding unit 6 makes an error. When the erasure correction is performed based on the correction code and the erasure correction cannot be performed based on the error correction code, the error directionality determination unit 11 determines the error directionality based on the weight of the codeword having the erasure error, and as a result, When it is determined that there is a high possibility that the error direction of the codeword is one direction, the decoding unit 10 that depends on the original character string uses a predetermined condition that depends on the original character string. , By detecting the original character string from the character string candidates of the original character string extracted based on the characteristics of the constant weight code and the error correction code, for example, in a system in which miscorrection has a significant effect,
When a non-directional error occurs, erroneous correction can be reduced by not performing the decoding based on the condition depending on the original character string, and therefore the reliability of the system is improved.

The second embodiment is premised on that it is disadvantageous to perform error correction using a condition depending on the original character string when there is a pair of errors (garbled characters) and an erasure error of two characters. It is based on the processing. On the other hand, for example, when there is a character error (garbled character) and one character disappearance error, a conventional decoding method causes a normal error. In order to prevent this, it may be advantageous to perform decoding based on the condition depending on the original character string even when the number of erasure errors is equal to the number of parity. The latter is not necessarily an error. It is effective to change the processing in this way depending on conditions such as error rate and burst property in addition to the directivity.

[0083]

As described above, according to the present invention,
A barcode processing method and a barcode processing system capable of decoding a barcode with high accuracy and reliability can be provided.

That is, by using the constant weight code when the error is unidirectional, the form of erasure correction can be achieved, and as many character errors as the number of parity, that is, twice the normal error can be corrected. It cannot be transformed into another character, and can be dedicated to the correction by the rules and the database for the number of character errors exceeding the parity correction ability. Moreover, even if the correction cannot be made, no erroneous correction occurs, and the barcode is The whole can be treated as a reject.

Further, in a system in which erroneous correction has a significant effect, it is better that erroneous correction is small even if the number of rejects increases. Therefore, in such a system, it is possible to improve the reliability of the system by reducing erroneous correction by determining whether or not the correction is limited to the range guaranteed by the code theory, depending on the degree to which the error directionality is not unidirectional.

[Brief description of drawings]

FIG. 1 schematically shows a configuration of a bar code processing system according to a first embodiment of the present invention, in which (a) is a coding section and (b) is a decoding section. It is a thing.

FIG. 2 is a diagram for explaining a specific example of a constant weight code, and 2 bits and 4 states (upper, lower,
This is an allocation example in the case where the number of codewords (codes) is 15, in which one code composed of three bars (upper, lower, none) is allocated. (A) is a case of weight (the number of bits having a bar) of 4, (b) is a case of weight 3, and (c) is a case of weight 2.

FIG. 3 is a diagram showing an example of a data configuration when a character string composed of 7-digit characters and a parity composed of 1-digit characters are encoded.

FIG. 4 is a diagram showing a specific example of a position where a bar code on a mail piece is scheduled to be printed.

5 is a flowchart for explaining a barcode decoding process procedure in a parity decoding unit of the barcode processing system of FIG. 1;

FIG. 6 is a diagram showing an example of storage of a new postal code book in a database.

FIG. 7 is a diagram showing a storage example of an address display number book of a database.

FIG. 8 is a diagram schematically showing a configuration of a decoding unit of the barcode processing system according to the second embodiment.

FIG. 9 is a diagram for explaining another example of the constant weight code, where 2 bits and 4 states (upper, lower,
This is an allocation example in the case where the number of codewords (codes) is 16, to which one code composed of three bars (upper, lower, none) is allocated.

10 is a flowchart for explaining the processing operation of the barcode processing system in FIG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Character string generation part, 2 ... Parity addition part, 3 ... Bar code conversion and printing part, 5 ... Bar code reading part, 6 ... Parity decoding part, 7 ... Rule decoding part, 8 ... Database decoding Database, 8a ... database, 10 ... decoding section depending on original character string, 11 ... error direction determining section.

Claims (30)

[Claims] 1. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code that produces a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing method of printing, reading a barcode from the print target, and restoring the original character string, converting the barcode read from the print target into a code word to obtain the error correction code. Erasure correction based on
When erasure correction cannot be performed based on this error correction code, the characteristics of the constant weight code and the original character string extracted based on the error correction code are determined based on the conditions that depend on the predetermined original character string. A method of processing a barcode, wherein the original character string is detected from the character string candidates. 2. A codeword obtained by adding an error correction code to a character string and encoding the code into a constant weight code that generates a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing method of printing, reading a barcode from the print target, and restoring the original character string, converting the barcode read from the print target into a code word to obtain the error correction code. Erasure correction based on
When erasure correction cannot be performed based on this error correction code, the character string of the original character string extracted based on the characteristics of the constant weight code and the error correction code based on a predetermined rule of the original character string. The original character string is detected from the candidates, and when the original character string cannot be detected, from among the character string candidates limited based on the rule of the original character string, within the expression range of the original character string A bar code processing method characterized by detecting a character string. 3. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code that generates a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing method of printing, reading a barcode from the print target, and restoring the original character string, converting the barcode read from the print target into a code word to obtain the error correction code. Erasure correction based on
When erasure correction cannot be performed based on this error correction code, the directionality of the error is determined based on the weight of the codeword having the erasure error, and as a result, the error direction of the codeword is highly likely to be unidirectional. When determined, based on a predetermined condition depending on the original character string, from the character string candidates of the original character string extracted based on the characteristics of the constant weight code and the error correction code to the original A barcode processing method characterized by detecting a character string of. 4. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code that produces a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing method of printing, reading a barcode from the print target, and restoring the original character string, converting the barcode read from the print target into a code word to obtain the error correction code. Erasure correction based on
When erasure correction cannot be performed based on this error correction code, the directionality of the error is determined based on the weight of the codeword having the erasure error, and as a result, the error direction of the codeword is highly likely to be unidirectional. When judged, the original character string is selected from the character string candidates of the original character string extracted based on the characteristic of the constant weight code and the error correction code based on a predetermined rule of the original character string. When the original character string cannot be detected, it is possible to detect a character string within the expression range of the original character string from the character string candidates limited based on the rule of the original character string. Characteristic barcode processing method. 5. The directionality of the error is determined by comparing an average of weights of code words having an erasure error with a predetermined weight for the constant weight code. 4. The barcode processing method according to any one of 4 above. 6. An error between a weight of a code word having an erasure error and a predetermined weight of the constant weight code is cumulatively added,
The bar code according to any one of claims 3 and 4, wherein the cumulative addition value is compared with a weight predetermined to the constant weight code to determine the directionality of the error. Processing method. 7. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 2 and a number of code words of 14. 8. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 3 and a code word number of 14. 9. The barcode processing method according to claim 1, wherein the constant weight code has a weight of 4 and a number of code words of 14. 10. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 2 and a number of code words of 15. 11. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 3 and a code word number of 15. 12. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 4 and a code word number of 15. 13. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 2 and the number of code words is 14 or more. 14. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 3 and the number of code words is 14 or more. 15. The bar code processing method according to claim 1, wherein the constant weight code has a weight of 4 and a code word number of 14 or more. 16. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code for generating a directional error, and converting the code word into a bar code on a predetermined print object. In a barcode processing system that prints, reads a barcode from this print object, and restores the original character string, converts the barcode read from the print object into a code word, and based on the error correction code When the erasure correction cannot be performed by the erasure correction and the erasure correction cannot be performed by the erasure correction, it is extracted based on the characteristic of the constant weight code and the error correction code based on a predetermined condition depending on the original character string. And a detection unit that detects the original character string from a character string candidate of the original character string. 17. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code that produces a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing system that prints, reads a barcode from this print object, and restores the original character string, converts the barcode read from the print object into a code word, and based on the error correction code When the erasure correction cannot be performed by the erasure correction, when the erasure correction cannot be performed by the erasure correction, based on a predetermined rule of the original character string, the characteristic of the constant weight code and the original extracted based on the error correction code First to detect the original character string from character string candidates of the character string
And the first detecting means cannot detect the original character string, the original character string is selected from character string candidates limited based on the predetermined rule of the original character string. And a second detecting means for detecting a character string within the expression range of the bar code processing system. 18. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code for generating a directional error, and converting the code word into a bar code on a predetermined print object. In a barcode processing system that prints, reads a barcode from this print object, and restores the original character string, converts the barcode read from the print object into a code word, and based on the error correction code A erasing correction means, a deciding means for deciding the directionality of the error based on the weight of the codeword having an erasure error when the erasure correction cannot be performed by the erasing correction means, and the deciding means for the error direction of the codeword. When it is determined that there is a high possibility that it is one-way, based on the condition depending on the predetermined original character string, the characteristics extracted by the constant weight code and the error correction code. Ex And a detection unit that detects the original character string from the character string candidates of the character string. 19. A code word obtained by adding an error correction code to a character string and encoding the code into a constant weight code that generates a directional error is converted into a bar code and printed on a predetermined print object. In a barcode processing system that prints, reads a barcode from this print object, and restores the original character string, converts the barcode read from the print object into a code word, and based on the error correction code A erasing correction means, a deciding means for deciding the directionality of the error based on the weight of the codeword having an erasure error when the erasure correction cannot be performed by the erasing correction means, and the deciding means for the error direction of the codeword. When it is determined that there is a high possibility of being unidirectional, the original character string extracted based on the characteristic of the constant weight code and the error correction code is determined based on the rule of the original character string determined in advance. First detecting means for detecting the original character string from among the character string candidates, and when the original character string cannot be detected by the first detecting means, the predetermined original character string A bar code processing system comprising: a second detection unit that detects a character string within the expression range of the original character string from character string candidates limited based on a rule. 20. The determination means compares the average of the weights of code words having an erasure error with a predetermined weight for the constant weight code to determine the directionality of the error. Item 20. The barcode processing system according to any one of items 16 to 19. 21. The determining means cumulatively adds an error between a weight of a code word having an erasure error and a predetermined weight of the constant weight code, and determines the cumulative addition value and the constant weight code in advance. The bar code processing system according to any one of claims 16 to 19, wherein the bar code processing system compares the weight with the determined weight to determine the direction of the error. 22. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 2 and the number of code words is 14. 23. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 3 and a number of code words of 14. 24. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 4 and a code word number of 14. 25. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 2 and the number of code words is 15. 26. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 3 and a number of code words of 15. 27. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 4 and a code word number of 15. 28. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 2 and the number of code words is 14 or more. 29. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 3 and the number of code words is 14 or more. 30. The bar code processing system according to claim 16, wherein the constant weight code has a weight of 4 and the number of code words is 14 or more.