JPH0950481A - Two-dimensional code reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct reading without rearranging the recording carrier of original, etc., in which the symbol of a two-dimensional bar code is recorded when the symbol is placed obliquely to the scanning direction of a line sensor. SOLUTION: When the two-dimensional bar code of a PDF-417 is decoded by utilizing the scanner of a digital copying machine, either one of a starting pattern S or a stopping pattern T can be made to be detected on a scanning line by detecting the inclination for the scanning line of a symbol by a rotation processing means 6 when the both of the starting and stopping patterns do not exist and rotating symbol image data by every arbitrary angle θ on an image memory. Subsequently, a left reading or right reading T sequence is generated, and the T sequence to be read from the other pattern side is generated at a point of time when the other pattern is detected while the T sequence is generated. The right and left T sequences are stored in a total T sequence memory 8 in a format that the sequences coexist, stitchings are performed for these sequences and decoding processings are performed for the sequences.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention mainly uses a scanner of a digital image input / output device such as a digital copying machine, a facsimile device, a filing system, etc.
The present invention relates to a two-dimensional code reader for decoding a two-dimensional bar code of F-417, and more particularly to a diagonal reading correction of a two-dimensional code.

[0002]

2. Description of the Related Art Conventionally, as a prior art example proposed for the purpose of expressing data as a machine-readable graphic image as encoded information and recognizing it, Japanese Patent Laid-Open No. 5-29 is known.
No. 0204 and Japanese Patent Laid-Open No. 6-12515,
An apparatus is disclosed which optically reads a two-dimensional bar code recorded on a record carrier by a line sensor such as a handy scanner and executes the decoding.

By the way, a two-dimensional bar code is a symbol
A representative example is PDF-417 developed by Technologies. Hereinafter, description will be made assuming that PDF-417 is used as the two-dimensional barcode.
For details of PDF-417, refer to U on January 5, 1990.
It is described in SP07 / 461881, and its algorithm has already been published.

FIG. 7 illustrates different patterns of various sizes of PDF-417 symbols. As shown in FIGS. 7B and 7C, the PDF-417 symbol is 30 × 60 mils, 24
Two sizes of × 24mils are shown, but both have the same configuration. That is, as shown in FIG.
The DF-417 symbol is vertically erected by stacking bar coded data rows vertically, and each row in the symbol is a start pattern S, a left row indicator pattern RL, a code word data area D, and a right row in order from the left end. The marking pattern RR and the stop pattern T are constructed in series.

The start pattern S and stop pattern T identify the start and end points of each row of symbols. The left and right line indicator patterns RL and RR include the line number and the number of symbols, the number of data columns, or the security level of the symbol. Further, codewords are the basic units for encoding numbers, letters, other symbols or their associated values, and the codewords within each row collectively form a data string. C represents the codeword in a digitized form. The black part in the symbol is called a bar and the white part is called a space.

FIG. 8 shows a basic pattern of the symbol structure of the PDF-417 two-dimensional bar code. As is clear from the pattern of this figure, the cell a is a minimum unit block that constitutes a data string. Further, a portion in which each cell is filled with black is called a bar, and a portion that remains white is called a space, and data is coded by the interval between these bars and spaces.

The set of alternating bars and spaces within this codeword is defined as an X-sequence of codewords, and this X-sequence is generated when data is encoded. On the other hand, when decoding, a sequence for obtaining the bar data and space interval data of coded data is defined as a T sequence.

In the conventional reading device constructed for the PDF-417 two-dimensional bar code having the above-mentioned configuration, the P sensor is used by a line sensor such as a handy scanner.
When reading the DF-417 symbol, it is premised that the start pattern S and the stop pattern T are detected at the same time, and when the reading is started, the left and right row marking patterns adjacent to the start and stop patterns S and T Only by solving RL and RR, the decoding of the code word data area D can be executed.

[0009]

As described above, in the case of the conventional apparatus, the PDF-417 symbol is usually read by the handy scanner and its decoding is executed, and the conditions at that time are both start and stop. Since it is premised that the patterns S and T can be found, there is a problem that the symbol cannot be decoded when the symbol is read diagonally.

That is, when reading the two-dimensional code with the line sensor, if the symbol is placed in parallel with the scanning direction of the line sensor, the start pattern S and the stop pattern T are simultaneously detected from the first scanning line. Therefore, there is no problem that the decoding is disabled, but it is usually possible to set a symbol at a slight inclination angle with respect to the scanning direction of the line sensor on a daily basis.

However, in the conventional apparatus, when the symbols are set in such a tilted shape, a range in which only the start pattern S or only the stop pattern T can be read inevitably occurs in any of the scanning lines. . In this scan line range, both the start and stop patterns S and T are not detected at the same time, resulting in a decoding error, and it is necessary to correctly reposition the record carrier on which the symbols are recorded, for example, the document. Becomes

As described above, in the case of the conventional apparatus having the working condition that the symbols must be arranged parallel to the scanning direction of the line sensor, for example, in the case of a digital copying machine, the symbols must be arranged parallel to the edge of the contact glass surface. There was a disadvantage and trouble that it took a lot of trouble to the user's symbolic work.

The present invention has been made in order to solve the above problems, and even when a symbol of a two-dimensional bar code is placed obliquely to the scanning direction of the line sensor, It is an object of the present invention to provide a two-dimensional code reading device having a function of correcting reading without repositioning a record carrier such as a document on which the symbol is recorded.

[0014]

In order to achieve the above object, according to the present invention, a line sensor for optically reading a two-dimensional bar code on a record carrier and a symbol of the two-dimensional bar code read by the line sensor. An image memory for read data, which stores, and start pattern collating means for collating the presence or absence of a start pattern provided at the left end of the symbol for each scanning line stored in the image memory, and stored in the image memory. When both the start pattern and the stop pattern are not present in the stop pattern matching means for checking the presence or absence of the stop pattern provided at the right end of the symbol for each scanning line and the start pattern matching means and the stop pattern matching means The inclination of the symbol with respect to the scanning line is detected, and A rotation processing means for rotating the image data by an arbitrary angle and a T sequence of symbols to be read from one of the start pattern and the stop pattern, which is detected first by the collation, is generated for each scanning line, and T A T sequence generation buffer that generates a T sequence of symbols to be read from the other pattern side at the time when the other pattern is detected during sequence generation, and a scan line in which the left and right T sequences coexist. A decoding processing means for decoding the data is provided.

According to the above, first, the entire symbol placed obliquely with respect to the scanning line of the line sensor is read and stored in the image memory, and the start or stop pattern is collated for each scanning line. Due to this matching, there is a case where there is neither a start pattern nor a stop pattern in all lines. In that line, it is already known in advance which of the start pattern or the stop pattern was detected, and the direction of the inclination of the symbol is known.Therefore, the image is rotated by an arbitrary angle and the start pattern and the stop pattern are compared. The reading and correction can be performed by performing the above retry until either one is detected first.

When the start or stop pattern can be collated for each line, when the start pattern is first detected, first, the symbol is read from the left to generate the left T sequence. When the generation of the left T sequence for each scan line progresses and a stop pattern is detected, the symbol is read from the right at that time point and the right T sequence is generated. In this case, when the stop pattern comes first, the order is reversed. After that, codeword conversion is performed on the data of the scanning line in which the left and right T sequences are mixed, and the stitching process for rearranging the left and right T sequences to perform diagonal correction is performed to create a document code.

[0017]

1 to 6 show an embodiment of a two-dimensional code reading apparatus embodying the present invention. As shown in FIG. 1, this reading device includes a decoding device 1 which is configured in a central control unit of a digital copying machine, and reads a PDF-417 symbol using a scanner 2 attached to the copying machine. Is configured. Scanner 2 is CC
A two-dimensional bar code, which is composed of a D line sensor and is recorded on an original as a record carrier, is optically read,
The read data is output to the decoding apparatus 1.

The decoding apparatus 1 performs the skew correction of the symbol due to the inclination of the document with respect to the scanning direction of the scanner 2 when the T sequence is generated after the symbol of the two-dimensional bar code shown in FIG. 7 is read. Then, as shown in FIG. 2, an image memory 3 for read data, a start pattern matching means 4, a stop pattern matching means 5, a rotation processing means 6, a T sequence generation buffer 7, an entire T sequence memory 8 and rearrangement (stitching). Means 9), and its decoding (decoding) is performed in the order of symbol input → T sequence generation → code word conversion → stitching → document code output, and the diagonal correction is performed at the time of T sequence generation.

The image memory 3 stores the entire two-dimensional bar code symbol read by the scanner 2. The start pattern matching means 4 matches the presence or absence of the start pattern S provided at the left end of the symbol for each scanning line stored in the image memory 3. The stop pattern matching means 5 matches the presence or absence of the stop pattern T provided at the right end of the symbol for each scanning line stored in the image memory 3.

The rotation processing means 6 detects the inclination of the symbol with respect to the scanning line when there is no start pattern S or stop pattern S in both the start pattern matching means 4 and the stop pattern matching means 5, and also the image memory. 3, the symbol image data is rotated by an arbitrary angle.

The T sequence generation buffer 7 generates, for each scanning line, a T sequence of symbols to be read from one of the start pattern S and the stop pattern T, which is detected first by the collation, and the T sequence is generated. When the other pattern is detected during the generation of, the T sequence of symbols read from the other pattern side is generated for each scanning line.

Next, referring to the flowchart of FIG. 3, the reading operation of the PDF-417 symbol by the reading device will be described. In step # 5, the reading operation of the PDF-417 symbol is started. Here, L indicates the scanning line number of the scanner 2. All T sequence memory 8 is
The generated left and right T sequences are stored. The rearrangement means 9 rearranges and combines the data of the scanning lines in which the left and right T sequences stored in the entire T sequence memory 8 are mixed, and creates correct read data.

Next, referring to the flowchart of FIG. 3, the reading operation of the PDF-417 symbol by the reading device will be described. In step # 5, the reading operation of the PDF-417 symbol is started. When the reading of this symbol is started, the read data image memory 3 stores the data of the entire read symbol. Then step #
In step 10, the T sequence at the left end of the symbol L line is calculated, and in step # 15 it is checked whether or not the result is start pattern data which is specific data.

FIG. 4 shows an example of how the T sequence is read when the symbol B is set obliquely to the scanning line, and FIG. 5 shows an example of how the T sequence is arranged in that case. As shown in FIG. 4, when the symbol B is arranged with a large inclination with respect to the scanning line,
As shown in FIG. 5, from the time of the first line (L = 1) when the scanning of the symbol B is started, immediately before the start pattern S appears (L
= N-1), there is no T sequence data, and it is recognized as a blank portion. In this example, scanning one cell corresponds to five scanning lines.

When the nth line in which the start pattern S appears appears, this is recognized in step # 15, and the left T sequence is generated in step # 20. In this case, when the start pattern S does not appear, the stop pattern T always appears at first because of the relation that the symbols B are arranged obliquely. In that case, the process proceeds to step # 25, the T sequence at the right end of the symbol L line is calculated, and it is checked in step # 30 whether or not the result is stop pattern data which is the specific data, and then step # 35. Generates a right T sequence. That is, the data is read backward.

At this time, the T sequence data is, for example, 92222247 422 ... As shown in FIG. As is clear from the contrast with the PDF-417 symbol of FIG. 7B, 922222 is the leftmost start pattern S, 422 ... Is the left row marking pattern RL, and the n-th line is the T sequence data. Then, it can be seen that only a part of the start pattern S and the left row marker pattern RL are present in the data.

Next, in the (n + 1) th line, 9222247 422
223 ..., and a part of the data area D of the codeword appears for the first time, so the left T sequence data is generated in step # 20 or the right T sequence is generated in step # 35. Start pattern S after (n + 2) line
While the state of being able to detect is generated, after the start pattern is detected, left T sequence data or right T sequence data is generated and stored in the entire T sequence memory 17.

By the way, as shown in FIG. 4, when the document set state is tilted to a certain limit, that is, beyond the diagonal angle of the symbol B, the start pattern S and the stop pattern T of the symbol B are not present in all lines. There are cases. In such a line, it is already known in advance which of the start pattern S or the stop pattern T is detected, and the direction of the inclination of the symbol B is known. Therefore, in step # 50, the rotation processing means 6 arbitrarily determines. The image can be read and corrected by rotating the image by an angle and retrying the collation of the start pattern S and the stop pattern T until either one is detected first.

That is, the rotation direction of the symbol B is determined from the appearance of the start pattern S or the stop pattern T, and the image of the symbol B is displayed on the image memory 3 by an arbitrary angle θ.
Rotate in CCW or CW direction. For the symbol B whose image has been rotated on the image memory 3, the same rotation process is repeated again until either the start pattern S or the stop pattern T can be detected, and the symbol image B having the original inclination angle shown by the thick solid line is displayed. To the symbol image B ′ indicated by the thin solid line.

In FIG. 4, a, b, c, and d indicate the original tilt angles, that is, the four corners in the state of the symbol image B read first. Also, a ', b',
c ′ and d ′ are symbol images B ′ after continuous rotation at an arbitrary angle θ
4 shows four corners in the state. As is clear from these, the rotation of the symbol B is caused by one corner a.
Is made mainly.

However, the arbitrary angle θ may be evenly distributed by a minute angle of about 2 to 3 °, or may be allocated by sequentially combining angles of different sizes. The rotation processing unit 6 performs affine transformation and interpolation processing to store binary density values.

Assuming that either the start pattern S or the stop pattern T can be detected at the m-th line shown in FIG. 5, at step # 30, the right T
The stop pattern T serving as the reference of the sequence appears, or the start pattern S serving as the reference of the left T sequence appears in step # 15.

Therefore, when it is determined in step # 30 that the stop pattern T appears, in step # 35, the generation of the T sequence is read backward from the line to the right T sequence, that is, from the stop pattern T. 8244 shown in FIG.
22 is T sequence data of the stop pattern T, 324 ... is T sequence data of the right row marker pattern RR, and 587 ... is T sequence data of the right end of the data area D. From this point, the T sequence data is in a form in which the left T sequence and the right T sequence are mixed.

Further, when the right T sequence is generated in the m-th line, the start pattern S which is the reference of the left T sequence appears together with the stop pattern T, so it is determined in step # 15 that the start pattern S has appeared. The left T sequence is generated in step # 20.

Then, in step # 40, it is determined whether or not the final line of the symbol has been reached, and if it has not reached, the above-mentioned steps # 5 to # 5 are performed for the next scanning line.
The process up to step # 40 is repeated, and thereafter, in the same manner, the above process is repeated for each scan line until the generation of the T sequence of the entire symbol is completed in step # 45.

In this way, when the reading up to the final line is completed, the T sequences of the respective rows output from the entire T sequence memory 17 are rearranged, that is, stitching is performed. FIG. 6 shows an example of the stitching process in this case.
In this figure, S is a start pattern, R is a left / right row marking pattern RL, RR, and T is a stop pattern.

The rearranging means 9 calculates a cluster or a codeword from each T sequence and performs diagonal correction from the attribute data. A cluster is a set of codewords divided into three mutually exclusive subsets.
For the F-417 symbol, each row encodes data using only one of the three clusters, with each cluster repeating in sequence every third row. Since any two adjacent rows use different clusters, the decoder can distinguish clusters from different rows within the same scan line.

In this diagonal correction, when a codeword matrix is created, the table is filled with row information such as [1] [2] as shown in FIG. 6 (A). When [S] indicating the start pattern S first exists, the data is filled from the left as it is, and when [T] indicating the stop pattern T exists, the data is sequentially filled from the right,
Based on the line information read diagonally by this,
As shown in FIG. 6 (B), the correction can be performed without omission.

In the above-described embodiment of the present invention, the configuration of the reading device for reading using the scanner of the digital copying machine has been described. In addition, in the present invention, the decoding device is installed independently of the copying machine. Of course, it is also possible to use a device such as a facsimile machine having a scanner other than the copying machine.

[0040]

As described above, according to the present invention, in the case of decoding a two-dimensional bar code of PDF-417 by using a scanner such as a digital copying machine, the entire symbol corresponds to the scanning line. If the start pattern and the stop pattern cannot be collated because they are scanned diagonally beyond the diagonal of the two-dimensional barcode symbol,
The image is rotated by an arbitrary angle in the opposite direction to the tilt direction of the symbol, the decoding is retried, and this retry is repeated until it can be resolved. Even when it is placed at an angle, it is possible to make an oblique correction without replacing the record carrier. Further, it becomes particularly effective when the tilt angle is large.

Further, it is possible to realize the reading of a document having a large inclination by a simple process of rotating the image data in the image memory without detecting the inclination of the entire image, which is unprecedented. Shows excellent effects.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing a reading device according to an embodiment of the present invention.

FIG. 2 is a schematic block diagram showing a configuration of a decoding apparatus.

FIG. 3 is a flowchart showing the operation.

FIG. 4 is a schematic diagram showing an example of how to read a T sequence.

FIG. 5 is a schematic diagram showing an example of how T sequences are arranged.

FIG. 6 is a schematic diagram showing an example of stitching processing.

FIG. 7 is a diagram showing patterns of four types of PDF-417 symbols having different sizes.

FIG. 8 is a diagram showing a pattern that is a basis of a symbol configuration of a PDF-417 two-dimensional barcode.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Decoding device 2 Line sensor 3 Image memory for read data 4 Start pattern matching means 5 Stop pattern matching means 6 Rotation processing means 7 T sequence generation buffer 8 All T sequence memory 9 Sorting means

Claims (1)

[Claims] 1. A line sensor for optically reading a two-dimensional bar code on a record carrier, an image memory for read data for storing a two-dimensional bar code symbol read by the line sensor, and an image memory for the image memory. Start pattern matching means for checking the presence / absence of a start pattern provided at the left end of the symbol for each stored scan line, and stop provided at the right end of the symbol for each scan line stored in the image memory The stop pattern matching means for matching the presence or absence of a pattern, and the start pattern matching means and the stop pattern matching means both detect the inclination of the symbol with respect to the scanning line when there is no start or stop pattern, and also on the image memory. And rotation processing means for rotating the symbol image data by arbitrary angles A time point at which a T sequence of symbols read from the side of one of the start pattern and the stop pattern, which is detected first by collation, is generated for each scanning line, and the other pattern is detected during the generation of the T sequence. And a T sequence generation buffer for generating a T sequence of symbols to be read from the other pattern side for each scanning line, and a decoding processing means for decoding the data of the scanning lines in which the left and right T sequences are mixed. A two-dimensional code reader characterized by.