JP2003108915A - Bar code recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To precisely recognize a bar code in an image with a low resolution photographed by a small-size image sensor built in a portable terminal. SOLUTION: After binarization, from a labeled input image, an adjacency relation of a connected region is analyzed on the basis of characteristics of a bar code structure, and a bar code area is extracted. From the width of the connected region of a black pixel inside the extracted bar code area, a unit width serving as a module width of the bar code is decided. When a defined bar code pattern and the inputted width pattern are collated with each other according to order of magnification of the unit width, the bar code is recognized. In collation, an allowance range is arranged for each magnification of the width collation, so that influence of noise and the like is reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a bar code recognizing device, a recognizing method, a program and the like for recognizing a bar code by extracting a bar code area from input image data.

[0002]

2. Description of the Related Art In recent years, with the development of a small image sensor with low power consumption, a camera can be built in a portable device such as a mobile phone, and an image taken by the built-in camera It is possible to send it immediately by email. These built-in cameras are prioritized for their small size, and have lower resolution than general digital cameras.

Further, recent mobile phones have a function of connecting to the Internet. In order to connect to the Internet, it is necessary to input the URL from the key, and with the key of the mobile phone, it takes time to enter a long URL. Therefore, manage the URL with a unique number,
It has been attempted to save the labor of inputting a URL and improve the usability of the user by reading a printed barcode corresponding to the number with a barcode reader. In such a device, it is necessary to separately connect a barcode reader to the mobile phone.

Therefore, in a portable device having a built-in camera,
If the barcode can be recognized with respect to the barcode image input by the built-in camera, such a service can be used without separately preparing a barcode reader. However, the resolution of the current built-in camera is lower than that of the image sensor used in the barcode reader, and it is difficult to recognize the barcode with high accuracy.

Another cause of difficulty in recognizing the bar code is that the width of the bar photographed in the input image is not constant. This is because the distance relationship between the barcode and the camera changes each time it is input. With a close-up type scanner, the bar code can always be input in a certain size, so the width of the bar can be set in advance.However, when shooting a bar code with a camera held by hand, a certain bar code can be used. The width cannot be predetermined.

There has been proposed a method of recognizing a bar code by determining the width of the bar from a bar code image input by an image scanner. For example, JP-A-6-32519
In Japanese Patent Publication No. 7, the width of the bar is determined by counting the frequency of the width of the pixel for each scanning line. The principle of determining the width will be described.

FIG. 19 is a diagram showing the relationship between the detection frequency and the width. As shown in FIG. 19, the width detection frequency is extracted for the scanning line. As a result, a region having a frequency equal to or higher than a certain threshold is detected as an effective width. In the example of FIG. 19, sections 1, 2, and 3 are effective widths. Furthermore, two long sections of this effective width are selected to be the width of the bar. In this method, from the two most frequent widths,
Two types of width, “thick”, can be detected. However, in a low-resolution image, there is a problem that the interval of frequency peaks becomes short and the detection becomes unstable. Also, it cannot support barcodes with multiple widths.

Further, when the bar code is photographed by the camera, the input image also includes areas around the bar code such as characters and patterns other than the bar code. Therefore, it is necessary to extract the barcode area from the input image. In Japanese Patent Application Laid-Open No. 9-16701 (FIG. 19), the bar code area is extracted on condition that the bar interval is within a certain threshold value. This is effective for a system that can obtain an input image relatively stably, such as a scanner. However, since the bar code is input while holding the camera with the hand, the width of the bar varies greatly depending on the distance between the subject and the camera, and there is a problem in that it cannot be extracted well with a fixed threshold. In order to solve such a problem, Japanese Patent Laid-Open No. 9-22437 attempts to improve the extraction accuracy by examining a pattern peculiar to the barcode when extracting the barcode area.

FIG. 18 is an explanatory diagram of the structure of the bar code. As shown in the figure, the barcode 1 includes a left guard bar 2 (or start bar) peculiar to the barcode, a center bar 3, and a right guard bar 4 (or end bar). In this case, the left guard bar 2 is provided on the starting point side (left side) of the barcode 1, and is configured by a pattern in which black bars and white bars are alternately arranged, and the information of "101" (start information of the barcode 1 ( This is the pattern for which (module expression) is set. The center bar 3 is provided at the center position of the barcode 1, and is composed of a pattern in which black bars and white bars are alternately arranged.
10 is a pattern in which information (module expression) is set. The light guard bar 4 is provided on the end point side of the barcode 1, and is composed of a pattern in which black bars and white bars are alternately arranged, and the information (module expression) of “101” is set as the end information of the barcode 1. It is a pattern.

Between the left guard bar 2 and the center bar 3, a 6-digit left data character 5 is displayed.
A 6-digit right-side data character 6 and a check digit 7 are arranged between and the light guard bar 4. In addition,
The numerical value at the left end below the barcode 1 represents the prefix digit 8.

FIG. 20 is a flow chart showing the flow of conventional bar code extraction processing. First, as described above, a portion where the bar interval is within a certain threshold is extracted as a complex area (step 401). The complex area can be said to be an area where the bar code is supposed to exist.

Next, the presence or absence of the center bar, right guard bar and left guard bar is determined from the extracted complex area, and it is determined whether the extracted complex area is a bar code (steps 402 to 404). .
Then, each character of the barcode is recognized in the area determined to be the barcode (step 40).
5).

Here, when determining the presence / absence of the center bar, the right guard bar, and the left guard bar, the pattern of the module representation of the bar structure is collated based on the extracted bar width and a predetermined bar module width. ing. The module width depends on the resolution of the scanner.
Predetermined. For example, when a form is read by a scanner having a resolution of 200 dpi, each module is represented by 4 pixels, and thus the module width is
It is defined as 4 pixels.

However, as described above, since the bar code is input while holding the camera by hand, the width of the bar is
Since there are many variations depending on the distance between the subject and the camera, there is a problem that the module width cannot be specified in advance. In addition, the bar width of an image input by a low-resolution camera is about 2 pixels, and with a bar code that uses four types of width, a bar with a width completely equal to the module width will not be displayed due to noise during input. There is a problem that a thick bar may be erroneously recognized as a thin bar and the collation cannot be performed correctly by simply comparing the widths that do not exist.

In order to solve this problem, Japanese Patent Laid-Open No. 4-2
In Japanese Patent No. 63381, when comparing the width of a bar in a prescribed pattern with the width of an input bar, it is determined that the bar widths are equal when they are within the range with a certain width. . For example, the condition that the bar width a and the bar width b are equal is 1.25a ≧ b ≧ 0.75a. However, when a barcode is photographed by a low resolution digital camera, the minimum width of the bar in the image is only about 2 pixels. In the case of a width of 2 pixels, the width of 2 pixels may actually become 1 pixel width or 3 pixels due to an error in input processing or the like. In this case, the fluctuation of the width is 50% of the actual width, and the fluctuation is large. When the width is large, 4 pixels are often 5 pixels and 3 pixels. In this case, the width of fluctuation is 25% of the actual width. As described above, when the widths of all the bars are compared, if the width ranges are all determined to be the same, there is a problem that the judgment cannot be performed correctly when the variation with respect to the bar width is large.

[0016]

The problem to be solved by the present invention is to recognize a bar code with higher accuracy than a bar code image taken using a low resolution image sensor. For this purpose, it is necessary to accurately extract the barcode area from the input image. Further, it is necessary to extract the width of the bar from the input image without specifying the width of the bar in advance so that the bar can be correctly recognized even if the width of the bar in the input image varies. Also, by using a low-resolution image sensor, the width of the narrowest bar is about 2 pixels, but even at this time, the widths of multiple bars can be accurately determined and recognized without being affected by noise or the like. It is necessary to. It is also necessary to be able to handle barcodes with multiple widths. The present invention has been made in view of the above problems, and provides a barcode recognition device capable of accurately recognizing a barcode even in a low-resolution image taken by a small image sensor built in a mobile terminal. The purpose is to

[0017]

A bar code recognition apparatus according to the present invention comprises preprocessing means for preprocessing an input image,
A binarization processing unit that binarizes the preprocessed input image, a labeling unit that labels the binarized input image, and a barcode from the labeled input image. It is characterized by comprising a barcode area extraction means for extracting an area and a barcode recognition means for recognizing a barcode from the extracted barcode area.

Further, in the barcode recognition apparatus of the present invention, it is desirable that the preprocessing means performs histogram conversion. Further, it is preferable that the binarization processing means uses a discriminant analysis method as a method of determining a threshold value in binarizing an image.

In the bar code recognition apparatus of the present invention, it is desirable that the labeling means label the input image by assigning an individual numerical name to each connected pattern. Further, it is desirable to define the width of the bar from the input image labeled by the labeling means by the number of black pixels of the label of the bar / height in the vertical direction. Further, when collating the widths of the bars or spaces, it is desirable to set an allowable range of the width according to the width.

Further, in the barcode recognition apparatus of the present invention, the barcode area extracting means extracts the adjacency relationship between the bars, determines the left end of the bar and the right end of the bar from the adjacency relationship, and determines the left end of the bar and the bar. It is desirable to extract the bar code area by matching the number of bars between the right edge of and the specified value.

Further, the adjacency relationship between the bars is that there is a scanning line shared by the two bars, that there is a difference in height between the two bars, and that there is an interval between the two bars. It is desirable to determine that they are adjacent to each other when all of the above are satisfied, and to adaptively obtain the range of the difference in bar height and the range of bar interval from the height and width of the bar as a reference.

In the barcode recognition apparatus of the present invention, the barcode recognition is performed by using the minimum width of the bar in the barcode area extracted by the barcode area extraction means as a unit width. It is desirable to perform this by collating the arrangement of the patterns of the width of the bar and the space, which is an integral multiple of the unit width, with the arrangement of the patterns of the width of the bar and the space defined in advance in the area. In addition, it is desirable that the barcode recognition be repeated by changing the unit width.

According to the bar code recognition device of the present invention, the bar code can be recognized more accurately than the bar code image taken by the camera used in the small image sensor incorporated in the portable terminal. . Further, according to the present invention, the barcode can be recognized even when a low-resolution camera is used, so that the barcode can be read without adding a barcode reader.

A mobile phone according to the present invention is characterized by including the barcode recognition device of the present invention. By incorporating a bar code recognition device in a mobile phone, it becomes possible to easily recognize a bar code anywhere, and it is possible to immediately transmit a bar code recognized by the bar code recognition device.

Further, in the barcode recognition method according to the present invention, the step of preprocessing the input image, the step of binarizing the preprocessed input image, and the step of performing the binarization processing And a step of extracting a barcode area from the labeled input image, and a step of recognizing a barcode from the extracted barcode area. . Further, the present invention can be realized as a program that causes a computer to function as a barcode recognition device, and also as a recording medium that records the program.

[0026]

BEST MODE FOR CARRYING OUT THE INVENTION In the present invention, after performing preprocessing of histogram conversion on an input image,
Binarization is performed, and labeling processing is further performed. From the labeled image, based on the features of the barcode structure,
Calculate the adjacency of bars. The adjacency relationship is, for example,
All connected regions of the image are examined with the connected regions of the bar candidates satisfying the following conditions as adjacent bars. (1) As shown in FIG. 21, there is a horizontal scanning line 9 shared by two bars. (2) As shown in FIG. 22, it is within a range in which there is a difference in bar height between adjacent bars. Here, it is assumed that the height is within 20% of the bar height. (3) As shown in FIG. 23, the difference in position coordinates in the X direction is within a range. Here, it is assumed to be within 6 times the bar width.

Next, the number of bars sandwiched between the left end and the right end is counted, and if the count value is a specified number, it is extracted as a bar code area. The unit width which is the module width of the barcode is determined from the width of the connected area of the black pixels in the extracted barcode area. The barcode is recognized by collating the defined barcode pattern with the pattern in the extracted barcode area in the arrangement of the patterns of the width of the bar and the space which is an integral multiple of the unit width. For recognition, the pattern of the left guard of the bar code is first matched, and if successful, the pattern of the left 6 digits, the prefix digit, the center bar, and the right 6 digits are sequentially matched. At the time of the pattern matching, the influence of noise or the like in the low resolution image is reduced by providing an allowable range according to the width of the bar and the space, that is, for each magnification. According to the present invention, a barcode can be accurately recognized even in a low-resolution image taken by a small image sensor built in a mobile terminal.
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a flow chart showing the flow of processing for barcode recognition. In FIG. 1, in step 1, histogram conversion is performed on the input image as preprocessing. When the minimum value of the brightness of the input image is Yi and the maximum value is Yj, the converted brightness Y ′ for the brightness Y is obtained by the following formula (1).

[0029]

[Equation 1]

First, a differential value of the brightness Y'is obtained, and a pixel having a differential value equal to or more than a threshold value is set as an edge portion, and a brightness histogram for this edge portion is created. In step 2, the image processed in step 1 is binarized. The binarization process stabilizes the calculation of the differential value of the brightness and improves the contrast for an image with low contrast. The binarization threshold is determined from the histogram obtained above, and the image is binarized. As a method of determining the threshold value, for example, a discriminant analysis method may be used. The discriminant analysis method is performed as follows.

When an image having a brightness range of "0 to D" is binarized with a threshold value t, the average brightness of pixels having a brightness of "0 to t-1" is f ₀ and the brightness is "t to D". Let f _{1 be} the average luminance of pixels, f be the average luminance of the entire image, and n _k be the number of pixels having luminance k, then the interclass variance σ _B ² is given by the following equation (2), and the intraclass variance σ _I ² becomes the following formula (3).

[0032]

[Equation 2]

[0033]

[Equation 3] Since the dispersion ratio at this time is as shown in the following formula (4), t that maximizes this F (t) is set as the threshold value.

[0034]

[Equation 4]

In step 3, labeling processing is performed.
As shown in FIG. 2, the labeling process is to give a unique label to all connected black pixels with respect to the black pixels (pixel value is 1) of the binarized image. The area occupied by the connected black pixels is called a connected area. There are various labeling methods, but one example is as follows.

(1) Scan the image from the upper left to the lower right, find a pixel P that has a pixel value of 1 and is not labeled, and attach a new label. (2) For the pixel P, all connected pixels in the image are labeled with the same label (10 to 12 in the figure are labels). (3) Returning to (1), if a pixel that is not yet labeled is found, a new label is attached and the process of (2) is performed. (4) When the scanning of the entire image is completed, the processing ends.

In step 4, from the labeling result, from the pattern information composed of the height and width of the connected black pixel,
An area in which connected areas (bars) of black pixels, which are barcodes, are arranged under a predetermined condition is extracted as a barcode area.
In step 5, the arrangement of the widths of the bars and spaces in the extracted barcode area is checked to recognize the barcode. The extraction of the barcode area in step 4 is labeled,
It is performed by examining the adjacency relationship of each connected region. Hereinafter, the connected area labeled will be referred to as a bar.

FIGS. 3 and 4 are flowcharts showing the flow of the search processing for adjacent bars. The adjacency relationship is established by using the width, height, and position information of the bars. Obtain the bar width and bar height as follows. Bar height: height of a rectangle surrounding the connected area Bar width: area of the connected area / bar height The upper left coordinates of the rectangle surrounding the connected area are the position coordinates of the bar. Taking the condition of adjacency as an example, it is defined as follows. (1) There is a horizontal scan line shared by two bars. (2) The difference between adjacent bar heights is within 20% of the bar height. (3) The difference in position coordinates in the X direction is within 6 times the bar width.

A bar satisfying the above conditions and having a position coordinate closest to the position coordinate of the target bar is set as a bar adjacent to the bar. Adjacent bars can be on the left and right sides respectively. When the adjacency relations are obtained for all the bars, it is inspected whether they are in the bar code area or not by the following procedure.

The flow of the adjacent bar search processing will be described with reference to FIGS. 3 and 4. In FIG. 3 and FIG. 4, the following definitions are made. Bar (L): Connected area (bar) labeled L Bar height (L): Height of bar labeled L (bar height) Bar width (L): Bar labeled L Width (bar width) Bar position_x (L): x coordinate of the position coordinate of the bar with the label L MAX_LABEL: Maximum value of the label MIN_LABEL: Minimum value of the label Left distance: Horizontal of the bar on the left side of the bar Minimum distance right distance: minimum horizontal distance of bars on the right side of the bar bar distance (L, L2): horizontal distance between the bar labeled L and the bar labeled L2 (bar height difference ( L, L2): Height difference between bar with label L and bar with label L2 In step 101, first, the label indicating the bar of interest to be the reference of the search is substituted into the variable L. The label sequentially increases from the minimum value of the label (LABEL_MIN) to the maximum value of the label (LABEL_MAX).

In step 102, a label indicating a bar that is an adjacent candidate to the target bar is assigned to the variable L2.
The label sequentially increases from the minimum value of the label (LABEL_MIN) to the maximum value of the label (LABEL_MAX). In step 103, when the values of the variable L and the variable L2 are the same, that is, the same bar is shown, step 1
Go to 16. Otherwise, go to step 104.
In step 104, the condition shown in the above condition (1) is tested. If the condition (1) is satisfied, step 10
Go to 5. If not satisfied, the process proceeds to step 116.

In step 105, the bar height difference between the labeled bar indicated by the variable L and the labeled bar height indicated by the variable L2 is calculated. In step 106, the condition shown in condition (2) above is tested. If the condition (2) is satisfied, the process proceeds to step 107. If not satisfied, the process proceeds to step 116.

In step 107, the horizontal distance between the bar position of the labeled bar indicated by the variable L and the bar position of the labeled bar indicated by the variable L2 is calculated. In step 108, the condition shown in condition (3) above is tested. If the condition (2) is satisfied, the process proceeds to step 109. If not satisfied, the process proceeds to step 116. Since the condition (3) uses the relative value with respect to the width of the bar as the threshold value, it is possible to correctly determine even if the width of the bar of the input image is not constant.

In step 109, it is determined whether the bar labeled with the variable L2 is on the right side or the left side with respect to the bar labeled with the variable L, based on the bar position coordinates. . If it is determined to be on the right side, the process proceeds to step 110, and if it is determined to be on the left side, the process proceeds to step 113.

In step 110, the variable "right distance", which stores the minimum distance to the right bar found so far, is compared with the current right bar, and if the distance is shorter, step 111 Proceed to. If this is not the case, then this bar is not considered to be on the right and the process proceeds to step 116.

In step 111, assuming that the bar labeled with the variable L2 is on the right side, the value of the variable L2 is stored. In step 112, the variable "right distance" is updated. In step 113, the variable "left distance", which stores the minimum distance to the left bar found up to now, is compared with the current left bar, and if the distance is shorter, the process proceeds to step 114. Otherwise, the bar is not considered to be on the left, and the process proceeds to step 116.

In step 114, assuming that the bar labeled with the variable L2 is on the left side, the value of the variable L2 is stored. In step 115, the variable "left distance" is updated. In step 116, the variable L2 is updated to the next smaller label. In step 117, if there is a label that has not been searched as an adjacent candidate by the variable L2,
The process proceeds to step 103 and the process is repeated. When the search is completed for all labels, the process proceeds to step 118.

In step 118, the variable L is sequentially updated to a larger label. At this time, the right adjacent bar and the left adjacent bar stored in steps 111 and 114 are confirmed as the right adjacent bar and the left adjacent bar of the bar indicated by the variable L, respectively. In step 119, if there is a label not searched by the variable L, the process proceeds to step 102 and the process is repeated. When the search is completed for all the labels, the process proceeds to step 120 and the process is completed.

When the adjacency relationship is obtained as described above, it is determined whether or not the area is the bar code area by the following procedure. First, the bar on the left is marked as the leftmost bar. Next, the bar on the right is searched sequentially from the left end bar. Mark the bar that does not have a bar to the right as the rightmost bar.

The number of bars from the left end bar to the right end bar is counted, and if the number of bars is a prescribed number, the bars from the left end bar to the right end bar and the area in which there is a space between these bars are counted. The code area. The specified number is determined by the bar code standard. For example, in the case of a JAN 13-digit bar code, the number of bars is 30. The bar code recognition processing is performed on the bar included in the extracted bar code area.

FIG. 5 is a flow chart showing the flow of processing for barcode recognition. In step 201, first,
Determine the unit width that is the standard for evaluating the width of bars and spaces. The minimum bar width within the bar code area is the unit width. The unit width corresponds to the module width of the barcode.

In step 202, the leftmost two bars are compared with the pattern of the left guard bar to test whether they are left guards. If it is determined to be the left guard, the process proceeds to step 203. Otherwise, proceed to step 212. In step 203, the left 6-digit numerical value is recognized for the next 12 bars. In step 204, it is determined whether the recognition of the left 6 digits is normally completed. Step 20 if completed normally
Go to 5. Otherwise, proceed to step 212.

In step 205, the prefix digit is recognized from the combination of the left and right 6-digit even and odd parity. In step 206, it is determined whether the prefix digits can be recognized without any contradiction. If it can be recognized without contradiction, the process proceeds to step 206. Otherwise, proceed to step 212. In step 207, the next two bars are compared with the pattern of the center bar to test whether they are the center bars. If it is determined to be the center bar, the process proceeds to step 208. Otherwise, proceed to step 212.

In step 208, the right six digits are recognized for the next twelve bars. In step 209, it is determined whether the recognition of the right six digits has been completed normally. If completed normally, the process proceeds to step 210. Otherwise, proceed to step 212. In step 210, the check digit is tested. If there is no contradiction in the check digit, the process proceeds to step 211. Otherwise, proceed to step 212.

The check digit is a numerical value calculated to check whether there is an error in reading.
Of the 6 digits on the right side, the last 1 digit is the check digit. The check digit is calculated from the 11 digits other than the check digit by the specified calculation method. The calculated check result and the read check digit are compared, and if they are equal, there is no contradiction. In the example of FIG. 18, the check digit is the numerical value 4 at the right end.

At step 211, the processing is terminated because the recognition is successful. In step 212, the unit width is increased by 1 pixel. In step 213, the increment of the unit width is 3
It is determined whether or not If it is 3 or less, the process proceeds to step 202 and the same recognition process is performed again. Otherwise, go to step 214. In step 214,
The process ends as recognition fails. In the left guard test, center bar test, and numerical value recognition, the bar and space widths are compared and evaluated with a predetermined pattern. The evaluation procedure will be described below.

FIG. 6 is a diagram for explaining the space width, and FIG. 7 is a diagram for explaining the allowable error rate.
The width of bars and spaces is an integral multiple of the unit width, and is evaluated for each magnification. The space width 15 is a width between two adjacent bars 13 and 14, as shown in FIG. The space width 15 is calculated as follows.

The adjacent bar (L) 13 and bar (L2) 1
For 4, the number of pixels between bars is calculated for each horizontal scanning line. The average number of pixels between the bars in all the scanning lines shared by the bar (L) 13 and the bar (L2) 14 is the space width 1
Set to 5. The bar code pattern is defined by a bar having a width that is an integral multiple of the reference module width (unit width) or a row of spaces. Width magnification is 1, respectively
Set to 2, 3 and 4. Here, the narrowest width is 1 and the widest width is 4.

The minimum magnification that satisfies the following condition is the magnification of the bar (space) width. That is, the condition is that the bar (space) width is within the range of magnification × unit width ± tolerance. Here, the allowable error is defined as allowable error = magnification × unit width × allowable error rate. The allowable error rate is specified for each magnification as shown in FIG. Here, the allowance is provided because the influence of pixel loss due to binarization is taken into consideration. When a barcode is shot with a low resolution digital camera, the unit width of the image is 2
There are only about pixels. When binarization is performed, 1 is displayed on the bar image.
There is a lack of pixels and noise of expansion. When compared as a bar having a width of 2 pixels without a tolerance, the 1-pixel width and the 3-pixel width are recognized as different widths. However, in reality, due to binarization noise, the 2-pixel width becomes 1-pixel width or 3-pixel width.
It may be the pixel width. By providing an allowable error, it is possible to recognize that the widths are the same in consideration of the variation in the width.

The left guard test in step 202 is performed as follows. FIG. 8 is a diagram for explaining an arrangement pattern of left guard widths, and FIG. 9 is a diagram for explaining an arrangement pattern of left guard bars and spaces. The arrangement pattern of the widths of the left guard bars and spaces defined as shown in FIGS. 8 and 9, and the arrangement of the bars and spaces of the left guard width obtained based on the evaluation of the width in the extracted barcode area. The pattern is matched, and if the pattern matches, the bar is a bar that constitutes a valid left guard.

The recognition of the numerical values in steps 203 and 208 is performed as follows. FIG. 10 is a diagram showing the regulation of the arrangement pattern of the widths of the bars and spaces with respect to the numerical values,
11 is a diagram showing a combination of the width 1 to the width 4 shown in FIG. 10 being a bar width or a space width.
3 is a flowchart showing a flow of processing for recognizing a numerical value.

In FIG. 10, an odd parity and an even parity exist in the arrangement pattern of the width for the left 6 digits. The arrangement of widths shown in FIG. 10 is for odd parity. In the case of even parity, the same table is read from the opposite direction, that is, width 4, width 3, width 2, width 1. For example, with odd parity, the number 9 is a triple width space from the left, a unit width bar,
It is represented by a pattern of unit-width spaces and double-width bars. The number 1 for even parity is from the left, unit-width spaces, double-width bars, double-width spaces, and double-width bars. It is represented by a line-up pattern (refer to data characters 1 and 2 on the left side of FIG. 18).

In FIG. 11, the six digits on the left side and the six digits on the right side of the center bar have different patterns of arrangement of bars and spaces, so that two types of patterns are defined in FIG. For example, on the right side the number 6 with odd parity
Is represented by an array pattern of a unit width bar, a unit width space, a unit width bar, and a quadruple width space from the left (see the first data character on the right side of FIG. 18).

In the flowchart of FIG. 12, in step 301, the pattern of the width of the target bar and space is obtained for the left 6-digit data character by the same procedure as the left guard test. In step 302, the above-described width is evaluated using the matching patterns defined in FIGS. 10 and 11, and it is sequentially checked whether the patterns of the arrangement of bars and spaces match. In the left 6 digits, the width pattern of odd parity or the width pattern of even parity is also sequentially checked.

If it is determined in step 303 that the patterns match, the process proceeds to step 304. Otherwise, proceed to step 307. In step 304, the width difference between the actual width obtained from the input image and the pattern of the specified width is obtained. The width difference is the sum of the differences of the corresponding four widths.

In step 305, if the width difference is smaller than that found so far, step 306.
Proceed to. Otherwise, proceed to step 307. In step 306, the numerical value corresponding to the pattern is stored as the numerical value of the recognition candidate. At the same time, the type of odd parity or even parity is also stored. The initial value of the width difference is set to the maximum value.

In step 307, if the processing has been completed for all the matching patterns (numerical values 0 to 9) shown in FIG. 10, the process proceeds to step 308. If not, the process proceeds to step 302 and the process is repeated for the next matching pattern. In step 308, when the pattern matches the specified numerical value pattern, the numerical value is set as the recognition result, the process proceeds to step 309, and the process ends as recognition success.
Otherwise, the recognition is unsuccessful, and the process proceeds to step 310 and ends the process. The recognition of the prefix digit in step 205 is performed as follows.

FIG. 13 is a diagram showing a combination result (prefix digit) of odd parity and even parity, FIG. 14 is a diagram showing a pattern of center bar widths, and FIG. 15 is a bar of the center bar. It is a figure which shows the arrangement pattern of a space.

In step 205, the prefix digit is recognized from the recording of the odd parity and the even parity when the recognition of the left six digits is successful. The combination result of the odd parity and the even parity is as shown in FIG. 13, and the table is searched and the matched one is recognized as the prefix digit. When no matching combination is found, the recognition of the prefix digit is regarded as a failure and the processing is terminated. In the example of FIG. 18, the left 6-digit data character 5 has an odd parity of 9, an even parity of 1, an odd parity of 2, an odd parity of 3, an even parity of 4, an even parity of 5, and a combination of parity. Is 101100, the prefix digit 8 is 4 from FIG. 13 (displayed at the left end of FIG. 18). In the center bar test of step 207, pattern matching is performed as in the left guard test. The arrangement pattern of the widths of the bars and spaces used for the center bar test is shown in FIGS.
As shown in.

The present invention is not limited to the above-mentioned embodiment, and can be implemented with various modifications. Further, the barcode recognition device of the present invention can be incorporated in a mobile phone. By incorporating a bar code recognition device in a mobile phone, it becomes possible to easily recognize a bar code anywhere, and it is possible to immediately transmit a bar code recognized by the bar code recognition device.

FIG. 16 is a block diagram when the bar code recognition device of the present invention is incorporated in a camera-equipped mobile phone. The camera-equipped mobile phone controls an antenna 21 for transmitting and receiving radio waves and transmission and reception of wireless communication. A wireless unit 22, a key input unit 23 for inputting and generating key operation information from a user,
A display unit 24 for displaying character information and image information to the user, a speaker 25 for outputting a sound signal, a microphone 26 for inputting a sound signal from the user, character information received or input,
A storage memory 27 for storing image information and audio signals, a camera 28 for inputting target image information, and a control unit 29 which incorporates a barcode recognition device and controls each unit.

In the mobile phone with a built-in camera having the above structure, when the barcode is recognized from the image information (barcode) input from the camera 28, the user operates the key input unit 23 to select the camera operation. Control unit 29
Initializes the camera 28 based on the settings from the key input section,
Start importing image information (bar code). The image information (bar code) captured by the camera 28 is transferred to the storage memory 27 via the control unit 29. The control unit 29 transfers the image information (bar code) stored in the storage memory 27 to the display unit 24 and displays the image information (bar code). Further, the user can confirm the camera image as a moving image by continuously capturing and displaying the image information (bar code). On the other hand, the image information (barcode) stored in the storage memory 27 is transferred to the control unit 29 having a barcode recognition device, and barcode recognition is performed by the barcode recognition processing described in the present embodiment. When the recognition is successful, the recognition result is transferred to the storage memory 27 and stored in the storage memory 27 as bar code data.

Next, the operation of connecting the bar code data stored in the storage memory 27 to the destination by connecting to the wireless or Internet will be described with reference to FIG. The user operates the key input unit 23 to select wireless or Internet connection. The control unit 29 instructs the wireless unit 22 to connect to the wireless or Internet based on the setting from the key input unit. The wireless unit 22 starts wireless or Internet connection based on an instruction from the control unit 29. The user operates the key input unit 23 to select transmission of the barcode data stored in the storage memory 27. The control unit 29 transfers the barcode data stored in the storage memory 27 to the wireless unit 22 based on the setting from the key input unit 23, and instructs the wireless unit 22 to transmit the barcode data. The wireless unit 22 transmits the transferred barcode data wirelessly or to the Internet based on an instruction from the control unit 29. The barcode data transmitted from the wireless unit 22 is transmitted to the URL data server 32 via the wireless network 33. The URL data server 32 controls the operation of the wireless network 33 of the mobile phone terminal 34 and the gateway of the Internet 31. The URL data server 32 retrieves the URL of the destination from the received barcode data and the URL database, connects to the Internet 31, and transfers the barcode data to the destination. The result of the transfer completion is transmitted to the mobile terminal 34.

The present invention also provides a program for causing a computer to function as a bar code recognition device, and
It can also be realized as a recording medium recording the program. The electronic mail communication device of the present invention is also realized by a program for causing the electronic mail communication device to function. This program may be stored in a computer-readable recording medium.

As the recording medium, the ROM itself having the bar code recognition device may be the program medium, or the bar code recognition device may be connected to a program reading device such as a CD-ROM drive and the recording medium may be connected thereto. It may be a program medium such as a CD-ROM that can be read by inserting it. In any case, the stored program may be configured to be accessed and executed by the CPU, or the program may be read and the read program may be downloaded to a program storage area (not shown). The method in which the program is executed may be used. It is assumed that this download program is stored in the main body device in advance.

Here, the program medium is a recording medium which can be separated from the main body, and is a tape system such as a magnetic tape or a cassette tape, a magnetic disk such as a floppy (registered trademark) disk or a hard disk, or a CD-RO.
Disc system of optical disc such as M / MO / MD / DVD,
IC card (including memory card) / card system such as optical card, mask ROM, EPROM, EEPRO
It may be a medium that fixedly holds the program, including a semiconductor memory such as M and a flash ROM.

Further, the medium may be a medium that carries the program fluidly so that the program is downloaded from the communication network via the transmitting unit and the receiving unit of the mobile phone having the bar code recognition device. When the program is downloaded from the communication network as described above, the download program may be stored in the apparatus main body in advance, or may be installed from another recording medium. The content stored in the recording medium is not limited to the program and may be data.

[0078]

As described above, according to the present invention, it is possible to recognize a bar code more accurately than a bar code image taken by a camera used in a small image sensor incorporated in a portable terminal. You can Further, according to the present invention, the bar code can be recognized even when a low-resolution camera is used, so that the bar code can be read without adding a special bar code reader.

[Brief description of drawings]

FIG. 1 is a flowchart showing a flow of a barcode recognition process.

FIG. 2 is a diagram illustrating an example of a labeling process.

FIG. 3 is a flowchart (first half) showing a flow of a search process for adjacent bars.

FIG. 4 is a flowchart (second half) showing a flow of a search process for adjacent bars.

FIG. 5 is a flowchart showing a processing flow of barcode recognition.

FIG. 6 is a diagram for explaining a space width.

FIG. 7 is a diagram for explaining an allowable error rate.

FIG. 8 is a diagram for explaining an arrangement pattern of left guard widths.

FIG. 9 is a diagram for explaining an arrangement pattern of left guard bars and spaces.

FIG. 10 is a diagram showing a pattern of arrangement of widths with respect to numerical values.

FIG. 11 is a diagram showing a combination of widths 1 to 4 shown in FIG. 10, which are a bar width and a space width.

FIG. 12 is a flowchart showing a flow of processing for recognizing a numerical value.

FIG. 13 is a diagram showing a combination result (prefix digit) of odd parity and even parity.

FIG. 14 is a diagram showing an arrangement pattern of widths of center bars.

FIG. 15 is a view showing a bar and space arrangement pattern of a center bar.

FIG. 16 is an explanatory diagram of a configuration of a mobile phone with a built-in camera.

FIG. 17 is a diagram showing connection of a mobile phone terminal to the Internet.

FIG. 18 is an explanatory diagram of a barcode configuration.

FIG. 19 is a diagram showing the relationship between the detection frequency and the width.

FIG. 20 is a flowchart showing a flow of conventional barcode extraction processing.

FIG. 21 is a diagram for explaining the condition (1) of the adjacent bar.

FIG. 22 is a diagram for explaining the condition (2) of the adjacent bar.

FIG. 23 is a diagram for explaining the condition (3) of the adjacent bar.

[Explanation of symbols]

1 bar code 2 Left guard bar 3 Center bar 4 Light guard bar 5 Left data character 6 Right data character 7 check digit 8 prefix digit 9 horizontal scan lines 10-12 labels 13-14 bars 15 space width 21 antenna 22 Radio 23 Key input section 24 Display 25 speakers 26 microphone 27 memory 28 cameras 29 Control unit 31 Internet 32 URL data server 33 wireless network 34 Mobile phone terminals

Claims (14)

[Claims] 1. A preprocessing unit for preprocessing an input image, a binarization processing unit for binarizing the preprocessed input image, and an input image after the binarization process. A label area for extracting a barcode area from an input image after labeling, and a barcode recognition means for recognizing a barcode from the extracted barcode area. A barcode recognition device characterized by: 2. The bar code recognition device according to claim 1, wherein the preprocessing means performs histogram conversion. 3. The bar code recognition device according to claim 1, wherein the binarization processing means uses a discriminant analysis method as a method of determining a threshold value in binarizing an image. 4. The labeling means sets 1 for an input image.
4. The barcode recognition device according to claim 1, wherein labeling is performed by assigning an individual numerical name to each one of the connected patterns. 5. The input image labeled by the labeling means defines the width of the bar as the number of black pixels of the label associated with the bar / the height in the vertical direction. Barcode recognition device. 6. The bar code recognition device according to claim 5, wherein when the width of the bar or the space is collated, an allowable range of the width is set according to the width. 7. The bar code area extracting means extracts the adjacency relationship between the bars, determines the left end of the bar and the right end of the bar from the adjacency relationship, and determines the number of bars between the left end of the bar and the right end of the bar. 7. The barcode recognition device according to claim 1, wherein the barcode area is extracted by matching a certain specified value. 8. The adjacency relationship of bars is such that there is a scan line shared by two bars, that there is a difference in height between the two bars, and that there is a distance between the two bars. It is determined that they are adjacent to each other when all of the above are satisfied, and the range of the difference in bar height and the range of bar interval are adaptively determined from the height and width of the reference bar. Item 7. The barcode recognition device according to Item 7. 9. The barcode recognition uses a minimum width of a bar in the barcode area extracted by the barcode area extraction means as a unit width, and an integer multiple of this unit width in the extracted barcode area. 9. The bar code recognition device according to claim 7, wherein the bar code recognition is performed by checking the arrangement of the patterns of the widths of the bars and spaces against the arrangement of the patterns of the widths of the bars and spaces that are defined in advance. 10. The barcode recognition device according to claim 9, wherein the barcode recognition is repeated by changing the unit width. 11. A mobile phone comprising the barcode recognition device according to claim 1. 12. A step of pre-processing an input image, a step of binarizing the pre-processed input image,
A step of labeling the input image after the binarization processing, a step of extracting a barcode area from the labeled input image, and a step of recognizing a barcode from the extracted barcode area. A bar code recognition method comprising: 13. A preprocessing unit for preprocessing an input image, a binarization unit for binarizing the preprocessed input image, and an input after the binarization process. Labeling means for labeling the image, bar code area extracting means for extracting a bar code area from the labeled input image, and bar code recognizing means for recognizing a bar code from the extracted bar code area. A program for causing a barcode recognition device to function, comprising: 14. A computer-readable recording medium on which the program according to claim 13 is recorded.