JP3891072B2 - Bar code reading method and bar code reading apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a bar code reading method and a bar code reading apparatus, and more particularly to a black and white pattern which is determined with priority given to whether the center of a module is white or black.
[0002]
[Prior art]
The bar code reader captures a bar code with a camera unit consisting of a CCD sensor, etc., binarizes the image signal, discriminates the brightness of the bar code (monochrome), and detects the width of the white bar and black bar And decipher.
[0003]
[Problems to be solved by the invention]
When the barcode is taken by the camera unit, the edge of the bar is easily affected by noise, and it is difficult to accurately detect the boundary between the white bar and the black bar, so the bar width may not be detected accurately. There is a risk of misreading. In order to solve this problem, the barcode reader disclosed in Japanese Patent Application Laid-Open No. 61-125684 focuses on the fact that the center position of the bar does not change relative to the center positions of the start bar and end bar. Thus, stable reading can be performed by determining the white bar and the black bar by calculating the area of the logic 1 value near the center.
[0004]
However, in this white bar / black bar determination method (light / dark determination method), when the camera section is tilted with respect to the barcode, the photographed barcode is different in size due to the perspective difference. As a result, even with a bar having the same width, the area of the central portion of the bar changes, and the accuracy is lacking.
[0005]
The present invention has been made in view of the above circumstances, and the purpose of the present invention is to make the bar code image captured by the image detection means unclear at the edge portion of the bar, or the size of the captured image differs at the partial portion. It is an object of the present invention to provide a barcode reading method and a barcode reading apparatus which can be decoded more accurately even if they are.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, the inventions of claims 1 and 6 divide multi-level image data into light or dark and detect a light / dark boundary position.
(1) When a light / dark pattern obtained by dividing the multi-level image data into light or dark is used as a light / dark pattern at the time of capture, both sides of the light / dark pattern at the time of capture centered on a light / dark boundary in one codeword Set the predetermined length range of as the boundary error range,
(2) In the bright / dark pattern at the time of capture, the length of one code word composed of the number of bars determined for each type of barcode is calculated, and the length of the one code word and the number of modules of one code word are calculated. Calculate the length of one module from
(3) In the bright and dark pattern at the time of capturing, one codeword is divided for each module length, and the center position of each module is detected as an initial position;
(4) The number of modules whose center is located in the boundary error range is determined by moving the center of each module constituting one code word a plurality of times from the initial position to both sides so that the amount of movement differs each time. Counting for each moving amount including the initial position where the moving amount is 0,
(5) The center of each module at the amount of movement when the number of boundary modules satisfies a predetermined condition is compared with the light / dark pattern at the time of capture, and the brightness / darkness of each module is compared with the light / dark pattern at the time of capture at which the center is located. Create a normalized light and dark pattern that is the same as light and dark,
A decoding process is performed on the normalized light / dark pattern.
[0007]
According to this means, in the barcode image captured by the image detecting means, the predetermined length range on both sides of the light / dark boundary is set as the boundary error range where the light / dark is not clear, and each word constituting one word code is set. When the center of the module is moved from the initial position to both sides of the module, a normalized light / dark pattern is created with the number of boundary modules that fall within the boundary error range satisfying a predetermined condition. Therefore, even if the edge portion of the bar is unclear in the captured barcode image, the boundary position of the bar can be detected more accurately and the barcode can be decoded more accurately. Moreover, since the word code is decoded one by one, even if the size of the captured barcode image is different in the partial portion due to the perspective difference, the influence can be eliminated as much as possible and more accurate decoding processing can be performed.
[0008]
According to the second and seventh aspects of the present invention, the operations of (1) to (3) are performed after the type of barcode is determined based on a bright / dark pattern at the time of capture in which the multi-level image data is divided into light and dark. It is characterized by performing.
[0009]
According to the third and eighth aspects of the present invention, since the light / dark boundary of the light / dark pattern at the time of capturing is set to a position where the change in light / dark is maximized, even if there is a light / dark boundary between the pixels of the image detecting means, Can be detected.
[0010]
In the inventions of claims 4 and 9, the operation of (3) is stopped when the number of boundary modules satisfies a predetermined stop condition, so that a light / dark boundary can be detected to some extent and a normalized light / dark pattern can be created. In spite of being in a state, it is possible to solve the problem of continuing the operation for detecting a light / dark boundary.
[0011]
According to the fifth and tenth aspects of the present invention, when there are a plurality of cases where the number of the boundary modules satisfies a predetermined condition, the light and dark pattern is determined by adopting the smaller moving amount.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows an electrical configuration of a handy type bar code reader, and includes a camera unit 1, an amplifier circuit 2, an A / D converter 3, a clock generator circuit 4, an address generator circuit 5, It comprises an input / output circuit 8 for transmitting / receiving data to / from an image memory 6, a liquid crystal display 7, a host computer, various switches 9, a control circuit 10 as a control means, and the like.
[0013]
The camera unit 1 includes a CCD area sensor 11 as a two-dimensional image detection unit configured by arranging CCDs as light receiving elements vertically and horizontally, an LED (light emitting diode) 12 as an illuminating unit that illuminates a reading target unit, The lens 13 is configured to image the reflected light from the reading target portion on the CCD area sensor 12.
[0014]
When detecting a two-dimensional image, the CCD area sensor 11 outputs the two-dimensional image to the A / D converter 3 through the amplifier circuit 2 as a horizontal scanning line signal. The A / D converter 3 converts the analog scanning line signal output from the CCD area sensor 11 and amplified by the amplifier circuit 2 into a digital signal having multiple levels.
[0015]
On the other hand, the clock generation circuit 4 outputs a clock pulse sufficiently finer than the pulse of the scanning line signal output from the CCD area sensor 11 according to the synchronization pulse output from the CCD area sensor 11. The address generation circuit 5 counts the pulses output from the clock generation circuit 4 and generates an address for the image memory 6. Then, the control circuit 10 stores the multi-level scanning line signal at the address of the image memory 6 determined by the address generation circuit 5.
The control circuit 10 is composed of a microcomputer system including a CPU, RAM, ROM and the like.
[0016]
Here, the bar code is configured by alternately arranging black bars and white bars having different widths, and the minimum width of the bar is referred to as one module. In addition, a start code and a stop code are provided on both the left and right sides of the bar code, and a data code including a plurality of code words is provided between the start code and the stop code. The number of bars and modules that make up the start code and stop code vary depending on the type of barcode, and the number of bars that make up one code word and the number of modules that make up one code word also depend on the type of barcode. Different.
[0017]
Next, the operation of the barcode reading apparatus having the above configuration will be described with reference to the flowcharts of FIGS. In this description of the operation, it is assumed that the barcode reader is configured exclusively for reading a specific type of barcode, for example, PDF 417 shown in FIG.
[0018]
The PDF 417 is configured by overlapping a plurality of data code patterns DT in which white bars and black bars are arranged between the start code SR and the stop code ST on both the left and right sides. In the case of PDF417, one code word constituting a data code is composed of 8 bars as shown in FIG. 5, and the total module of these 8 bars is 17 modules.
[0019]
Now, under the control of the control circuit 10, when the CCD area sensor 11 detects the barcode of the PDF 417 (step S1), the image signal output from the CCD area sensor 11 is an image with multiple levels, for example, 256 gradations. The data is stored in the image memory 6 (step S2).
[0020]
Next, the control circuit 10 sets the number counter n of the data code of the PDF417 barcode having a plurality of stages to 1 (step S3), and the light / dark boundary position for the data code pattern of the first stage (the uppermost stage in FIG. 8). (Brightness / darkness boundary position detection means), and the bright / dark pattern is captured and stored in a storage device such as a RAM (not shown) as a bright / dark pattern (lightness / darkness pattern detection means) (step S4). A part of the bright / dark pattern at the time of capture is illustrated in FIG.
[0021]
In this case, the binary boundary position between light and dark is detected as follows. That is, it is assumed that the stored multi-level image data is as shown in FIG. In FIG. 6A, the plotted points are the brightness (digital) detected by the pixels of the CCD area sensor. In order to obtain the light / dark boundary position, the light / dark level difference is obtained by subtracting the light / dark level on the right side from the light / dark level of the left pixel. The primary difference of the light and dark levels shown is obtained. Next, for each point plotted in the primary difference data of FIG. 6B, the value of the right point is subtracted from the value of the left point for each point adjacent to each other. The secondary difference shown in (c) is obtained. Then, when each point plotted in the secondary difference data is connected by an envelope, the position of the point where the envelope intersects the zero level is defined as the boundary between light (white bar) and dark (black bar) (Fig. 6 (d)).
[0022]
When the light / dark boundary is defined in this way, even if there is a light / dark boundary between pixels, the intermediate position between the pixels is recognized as the boundary and conforms to the actual barcode. Can be defined as a boundary. For example, as shown in FIG. 7, the primary difference between the primary difference between the pixels 1 and 2 and the secondary difference a between the primary differences between the pixels 2 and 3 is +1, and the primary difference between the pixels 2 and 3 is between the pixels 3 and 4. When the secondary difference b point with respect to the primary difference is −4, a light / dark boundary exists between the pixel 2 and the pixel 3. Therefore, since the envelope S between the secondary difference points a and b can be regarded as a straight line, it is assumed that the zero-cross point is at a position 1/5 of the distance R between the pixels 2 and the pixels. As described above, according to the boundary detection method of this embodiment, even when a light / dark boundary is in the middle of two pixels, this can be detected.
[0023]
After detecting the light / dark boundary position as described above, the control circuit 10 calculates the widths of a number of white bars and black bars constituting the first stage data code pattern from the detected boundary position (bars). Width detection means, step S5). Next, for the first code word of the first-stage data code pattern, for example, the leftmost one code word (FIG. 5B), the control circuit 10 includes a plurality of bars constituting the code word. And the length of one code word is calculated (code word length detecting means, step S6). Then, the control circuit 10 divides the length of one codeword by the module (17) of one codeword, and calculates the length of one module in the codeword (module length detection means, step S7). As shown in FIG. 5C, the center position of each module is calculated as the initial position (module position detecting means, step S8).
[0024]
Thereafter, the control circuit 10 sets a predetermined range, for example, 1/16 of the module width M, on the left and right sides (both sides in the length direction of the codeword) from the light / dark boundary position. d)) (step S9), and as shown in FIG. 5E, at the initial position, the number of modules whose center is located in the boundary error range is counted, and this is moved as the number of boundary modules. It is stored in a storage device such as a ROM (not shown) in association with the amount 0 (initial position) (step S11).
[0025]
Next, the control circuit 10 sets a movement number counter m to 0 (step S12), and then adds 1 to the counter m (step S13). Then, the control circuit 10 moves the center of each module to the left by 1/16 of the length M of one module from the initial position (step S14), and the number of modules whose center is in the boundary error range at this movement position. Is counted as the number of boundary modules (step S15), and the number of boundary modules is associated with the amount of movement (including the direction) and stored in the storage device (step S16).
[0026]
Further, the control circuit 10 moves the center of each module to the right from the initial position by 1/16 of the length M of one module (step S17), and the number of modules whose center is in the boundary error range at this movement position. Is counted as the number of boundary modules (step S18), and the number of boundary modules is associated with the amount of movement (including the direction) and stored in the storage device (step S19).
Next, the control circuit 10 determines whether m is 8 (step S21). When m is less than 8 ("NO" in step S21), whether the number of boundary modules counted next is 1 or less. (Step S21), if it exceeds 1, "NO" is determined in the step S21, the process returns to the step S13, 1 is added to the counter m, and m is set to 2.
[0027]
Then, the control circuit 10 moves each module to the left and right from the initial position by 2/16 of the module length M, and stores the number of modules whose center is in the boundary error range in association with the amount of movement. The operation is performed (boundary module number detection means) until m reaches 8, that is, the amount of movement from the initial position to the left and right is increased by M / 16, and each amount of movement reaches M / 2.
[0028]
When m reaches 8, the control circuit 10 determines “YES” in step S20, and then detects the amount of movement in which the number of boundary modules satisfies the predetermined condition, that is, the amount of movement in which the number of boundary modules is minimum in this embodiment. (Step S22) After that, the number of movement amounts which is the minimum number of boundary modules is detected, and it is determined whether or not the number is plural (Step S23). Then, when there is only one movement amount that is the minimum boundary module number, the control circuit 10 becomes “NO” in step S23 and proceeds to the next step S25, and the movement amount that is the minimum boundary module number. If there is a plurality, the answer is “YES” in step S23, and the process proceeds to step S24. Here, the movement amount closest to the center position of the original module and farthest from the light / dark boundary position is set as the minimum boundary module. The movement amount is set to a number and the process proceeds to step S25.
[0029]
Then, in step S25, the control circuit 10 moves the center of each module to the movement amount when the number of modules is the minimum boundary module, as shown in FIG. 5 (i). Bright and dark patterns belonging to the light and dark pattern are set, and this is used as a normalized light and dark pattern (normalized pattern creating means). In the example of FIG. 5I, as shown in FIG. 5J, the fifth module from the left is dark (black), the next one module is bright (white), the next two modules are dark, the next two modules Module is light, next 2 modules are dark, next 2 modules are light, next 2 modules are dark, last 1 module is light. Then, the control circuit 10 decodes the code word as one code word as this normalized light / dark pattern (decoding means, step S26).
[0030]
By the way, during the execution of the operation of detecting the number of boundary modules by moving the module center left and right while increasing the module length by 1/16, the number of boundary modules becomes 1 or less (predetermined stop condition). There is a case. In this case, the control circuit 10 becomes “YES” in Step S21, and proceeds to Step S25. This means that even if m does not become 8 (before moving the center of the module from the initial position to 8M / 16), the operation of detecting the number of boundary modules is stopped.
[0031]
Thereafter, the control circuit 10 determines “NO” in step S27 to determine whether or not all the first-stage code words have been decoded, and proceeds to step S28 to set the next code word as a decoding target. Returning to step S5, the decoding operation of the second code word is started. Then, the second code word is decoded in the same manner as described above, and then the third and subsequent code words are sequentially decoded.
[0032]
When all the code words of the data code pattern of the first stage are decoded (“YES” in step S27), the control circuit 10 determines whether or not the decoding is finished up to the next data code word of the final stage. In S29, “NO” is set, and then 1 is added to n, and the process proceeds to decoding of the second-stage data code pattern. Then, the decoding operation of the data code patterns of all the stages is started. When the decoding of the data code word patterns of all the stages is finished (“YES” in step S29), the decoding operation of the PDF417 barcode is finished.
[0033]
As described above, according to the present embodiment, since the barcode is decoded for each codeword, the camera unit 1 is inclined with respect to the barcode, and the barcode has a partial size due to a perspective difference. Even if they are different, they can be deciphered while eliminating the influence of the difference in size as much as possible.
[0034]
In addition, when the range where light and dark boundaries are unclear and wins is defined as a boundary error range, and the center of each module constituting the codeword is moved to both sides, when the number of modules whose center falls within the boundary error range is the smallest, That is, even if the position of the boundary is uncertain, the case where the number of modules without error is the largest is determined as a normalized light / dark pattern and decoded, so that the barcode can be more accurately Can be deciphered.
[0035]
In this embodiment, when there are a plurality of movement amounts that are the minimum boundary module number, the movement amount that is the farthest from the module center and the farthest from the light / dark boundary position is set as the movement amount that is the minimum boundary module number. Since it is set and a normalized light / dark pattern is created, a more accurate light / dark pattern can be created.
[0036]
The present invention is not limited to the embodiments described above and shown in the drawings, and the following modifications or expansions are possible.
In claim 1, the order of the operations (1) to (5) is not limited to ascending order.
The bar code reader may be provided with a function for determining the type of bar code. In this way, the barcode can be decoded even if the type of barcode to be read is not known in advance, and the types of barcodes that can be read are expanded. The control circuit 10 determines the bar code type at this time. However, since the number of bars and the number of modules constituting the start code or the stop code are different for each type, the control circuit 10 as a type detecting means is used. First, assuming that it is an A type barcode, it is determined from one side of the barcode whether the light / dark pattern of the number of bars constituting the start code of the A type barcode matches the start code pattern of the A type barcode. If they match, it is determined that the barcode is A type. If it is not possible to determine that it is a type A barcode, it is possible that the next barcode image may have been captured in the opposite direction, so the number of bars that make up the end code of the type A barcode from one side of the barcode It is determined whether the pattern matches the light / dark pattern of the end code of the A type barcode, and if it matches, it is determined as the A type barcode. If it is not possible to determine that it is an A type barcode end code, then assume that it is a B type barcode and perform the same operation as above, then repeat the same operation to determine the type of barcode. It is.
As a method of dividing one codeword into light and dark, a distribution of lightness and darkness of multi-level levels of image data may be obtained, and this may be divided into light and dark as a threshold value.
The light / dark boundary position may be detected collectively for the entire barcode, or may be performed for each code word.
[Brief description of the drawings]
FIG. 1 is an electrical configuration diagram of a barcode reader showing an embodiment of the present invention. FIG. 2 is a flowchart showing the contents of a barcode decoding operation.
Fig. 2
Fig. 4
FIG. 5 is a diagram showing the execution contents of main steps. FIG. 6 is an explanatory diagram of a method for obtaining a bar code light / dark boundary position. FIG. 7 is a detailed diagram. FIG. 8 is a diagram showing a barcode to be read. Explanation】
6 is an image memory, 9 is a control circuit (light / darkness detection means, boundary error range setting means, module position detection means, boundary module number detection means, normalization pattern creation means), and 10 is a CCD area sensor (image detection means). .

Claims (10)

The image of the barcode to be read is captured by the image detection means, the image signal output as a multi-level signal from the image detection means is stored as image data, and the barcode is based on the stored image data. In the barcode reading method for decoding,
After the multi-level image data is divided into light and dark to detect a light / dark boundary position, the following operations (1) to (5) are performed for each of a plurality of code words constituting the barcode. A bar code reading method comprising: normalizing light and dark of modules constituting each code word and performing decoding processing on the normalized light and dark pattern.
(1) When a light / dark pattern obtained by dividing the multi-level image data into light or dark is used as a light / dark pattern at the time of capture, both sides of the light / dark pattern at the time of capture centered on a light / dark boundary in one codeword Is set as the boundary error range.
(2) In the bright / dark pattern at the time of capture, the length of one code word composed of the number of bars determined for each type of barcode is calculated, and the length of the one code word and the number of modules of one code word are calculated. From this, the length of one module is calculated.
(3) In the bright / dark pattern at the time of capturing, one code word is divided for each module length, and the center position of each module is detected as the initial position.
(4) The number of modules whose center is located in the boundary error range is determined by moving the center of each module constituting one code word a plurality of times from the initial position to both sides so that the amount of movement differs each time. And counting for each moving amount including the initial position where the moving amount is zero.
(5) The center of each module at the amount of movement when the number of boundary modules satisfies a predetermined condition is compared with the light / dark pattern at the time of capture, and the brightness / darkness of each module is compared with the light / dark pattern at the time of capture at which the center is located. Create a normalized light and dark pattern that is the same as light and dark. The operations (1) to (5) are executed after the type of bar code is determined based on the captured light / dark pattern obtained by dividing the multi-level image data into light or dark. Item 2. The barcode reading method according to Item 1. 3. The bar code reading method according to claim 1, wherein a light / dark boundary of the light / dark pattern at the time of capturing is set at a position where a change in light / dark becomes maximum. The center of each module constituting one code word is moved a plurality of times from the initial position to both sides so that the amount of movement is different each time, and the initial position is determined by using the number of modules whose center is located in the boundary error range as the number of boundary modules. 4. The barcode reading according to claim 1, wherein the operation of (3) including counting for each movement amount is stopped when the number of boundary modules satisfies a predetermined stop condition. 5. Method. 5. The operation according to claim 1, wherein when there are a plurality of movement amounts in which the number of boundary modules satisfies a predetermined condition, the operation of (5) is performed by adopting the smaller movement amount. Bar code reading method. The image of the barcode to be read is captured by the image detection means, the image signal output as a multi-level signal from the image detection means is stored as image data, and the barcode is based on the stored image data. In the barcode reader to decode,
Light / dark detection means for detecting the boundary position of light / dark by dividing the image data of the multi-stage level into light or dark,
When a light / dark pattern obtained by dividing the multi-level image data into light or dark is used as a light / dark pattern at the time of capturing, a predetermined length on both sides of the light / dark pattern at the time of capturing at the boundary of light / dark in one codeword Boundary error range setting means for setting the range as a boundary error range;
In the light / dark pattern at the time of capture, the length of one code word composed of the number of light / dark bars determined for each type of barcode is calculated, and the length of the one code word and the number of modules of one code word are calculated. Module length calculating means for calculating the length of one module;
Module position detecting means for detecting one central position of each module as an initial position by dividing one code word for each module length in the bright and dark pattern at the time of capture;
The center of each module constituting one code word is moved a plurality of times from the initial position to both sides of the module so that the amount of movement is different each time, and the number of modules whose center is located in the boundary error range is defined as the number of boundary modules. Boundary module number detection means for counting each movement amount including the initial position of 0,
The center of each module at the amount of movement where the number of boundary modules satisfies a predetermined condition is compared with the light / dark pattern at the time of capture, and the light / dark of each module is determined to be the same as the light / dark of the light / dark pattern at the time of capture. And a normalized light / dark pattern creating means for creating a normalized light / dark pattern,
A bar code reader, wherein the normalized light / dark pattern is decoded. 7. The bar code reader according to claim 6, further comprising a type discriminating unit for discriminating a bar code type based on the captured light / dark pattern. 8. The bar code reading apparatus according to claim 6, wherein the light / darkness detecting means determines a light / dark boundary of the light / dark pattern at the time of capture at a position where the change in light / darkness is maximized. 9. The module position detection unit, when the number of boundary modules satisfies a predetermined stop condition, stops the operation of counting the number of boundary modules for each movement amount. Bar code reader. The first-time normalized light / dark pattern creating means creates the normalized light / dark pattern by adopting the smaller moving amount when there are a plurality of moving amounts satisfying the predetermined number of boundary modules. The barcode reader according to any one of claims 6 to 9.

Images