JP2004118611A - Character reading system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a character reading system capable of making sure that correct character reading is under way by determining whether or not a threshold used in pattern matching is appropriate each time characters are read. <P>SOLUTION: The character reading system is provided with an image pickup means which picks up images of individual patterns from workpieces and outputs image pickup signals; a pattern collating means for obtaining binary image data from the image pickup signals and collating the data with a stored reference pattern; and a statistical means which performs a statistical process on the results of collation by the pattern collating means. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to reading an individual pattern of an article. In particular, the present invention relates to a character reading system capable of evaluating the validity of a threshold used for reading an individual pattern.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, on a production line or the like, an ID such as a lot number for specifying an article is assigned to an article, and management is performed so that a non-genuine product is not mixed.
[0003]
Further, in the case of products that need to specify each of the articles, an ID for identifying the individual is given to each product, and each product is managed. I have.
[0004]
Generally, the ID and the like are indicated by a method such as printing a bar code or an individual identification character on the product surface. The ID and the like are captured as image data by an imaging means such as a camera, and are stored in advance as data of a reference pattern. Recognition is performed automatically by pattern matching compared with.
[0005]
In pattern matching, the captured image data and the reference pattern are compared for each pixel of the image, the number of matching pixels is divided by the total number of pixels to be compared, multiplied by 100, and the degree of matching of the presence or absence of pixels (hereinafter, score Is calculated, and a match / mismatch is determined by comparing the score value with a preset threshold value.
[0006]
In the pattern matching, there is a problem that the image reading conditions such as the position of the article and the brightness of the illumination are not constant, so that it is difficult to accurately read the image data due to the influence of disturbance when the image data is imaged.
[0007]
In order to solve such a problem relating to imaging, there are an illumination device and a character reading device that read a character accurately by switching a plurality of illuminations by a switch and applying an appropriate illumination for each character (for example, see Patent Document 1). ).
[0008]
In addition, there is also a technique in which a newly generated reference pattern is registered based on the calculated stability of character reading and used for calculating a score value of subsequent imaging data (for example, Patent Document 2).
[0009]
However, in the related art, it cannot be evaluated that the threshold used for judging the match / mismatch is properly set. Therefore, when an inappropriate threshold is used, a match is determined to be a mismatch. In addition, a determination error may occur, for example, what is supposed to be mismatched is regarded as a match.
[0010]
Since the score value of each image data is a random variation, most of the score values are near the center, and the number of the score values decreases as the distance from the center decreases, resulting in a symmetric normal distribution. By inputting the score value into software for statistical analysis, calculating a statistical average value and standard deviation σ, and comparing the distribution with the threshold value, it can be determined whether the threshold value is appropriate.
[0011]
If it is determined that the set threshold value is inappropriate, it is necessary to re-determine the match / mismatch manually and find a threshold value at which the match / mismatch is clearly separated by trial and error.
[0012]
[Patent Document 1]
JP-A-8-161423 [Patent Document 2]
JP-A-11-306287
[Problems to be solved by the invention]
The present invention has been made in order to solve the above-described problem, and an object of the present invention is to determine whether or not a threshold used for pattern matching is appropriate each time a character is read, and to determine a correct character. An object of the present invention is to provide a character reading system capable of confirming that reading is being performed.
[0014]
[Means for Solving the Problems]
A first aspect of the present invention is a character reading system for recognizing an individual pattern of a work to identify a product and confirming whether or not the pattern is a correct individual pattern. An imager that outputs binary image data, a pattern matching unit that obtains binary image data from the image pickup signal and matches the reference pattern stored in advance, and a statistical unit that performs statistical processing of a result of matching in the pattern matching unit. It is provided.
[0015]
According to a second aspect of the present invention, in the character reading system according to the above-described first aspect, a work detection unit that instructs the imaging unit to perform an imaging timing of a work, an illumination device that illuminates the work at the time of imaging the work, It is provided with a brightness adjusting means for adjusting the brightness output of the lighting device, and an imaging device for imaging the work and outputting an imaging signal.
[0016]
According to a third aspect of the present invention, in the character reading system according to the first or second aspect described above, a reference pattern storage unit that stores a character pattern for each character in advance and the imaging signal are input to the pattern matching unit. Binarizing means for obtaining binarized binary image data, threshold setting means for setting a threshold for evaluating a comparison result between the binary image data and the reference pattern, It is provided with score value calculating means for comparing a reference pattern to calculate a score value, and determining means for comparing the score value with the threshold value to determine an individual pattern of the work.
[0017]
According to a fourth aspect of the present invention, in the character reading system according to the first to third aspects described above, the pattern matching unit includes a binary image storage unit that stores the binary image data; And a binary image display means for displaying image data.
[0018]
According to a fifth aspect of the present invention, in the character reading system according to the above-described first to fourth aspects, the statistical means includes the latest fixed number of constants each time the pattern matching means reads an individual pattern of one work. A partial statistic means for calculating a variation distribution including an average value, a standard deviation, and a discrimination value of the score values using the score values is provided.
[0019]
According to a sixth aspect of the present invention, in the character reading system according to the first to fifth aspects described above, the partial statistics means is provided with the latest constant number each time the pattern collating means reads an individual pattern of one work. A partial statistical processing means for calculating a variation distribution of score values using the score values of the partial statistical storage means for storing a variation distribution calculated using the fixed number of score values; and And partial statistical display means for displaying a variation distribution calculated by the calculation.
[0020]
According to a seventh aspect of the present invention, in the character reading system according to the first to sixth aspects, a sample number setting means for setting a fixed number of score values used for calculating a variation distribution in the partial statistical processing means. And, a probability accuracy setting means for setting how many times the standard deviation is to be calculated in the variation distribution, and a warning zone for instructing re-determination if the discrimination value falls in a specific area in the determination of the individual pattern in advance. A warning zone setting means for setting the width of the warning zone, and a partial statistic calculating means for calculating a variation distribution using the set fixed number of score values.
[0021]
According to an eighth aspect of the present invention, in the character reading system according to any one of the first to seventh aspects, the partial statistics display means displays the variation distribution immediately after the partial statistics calculation means calculates the variation distribution. It is provided with latest display means and history display means for calling and displaying the variation distribution stored by the partial statistical analysis storage means.
[0022]
According to a ninth aspect of the present invention, in the character reading system according to the first to eighth aspects described above, the statistical means is provided with an overall statistical means for calculating a variation distribution by using score values of all the captured works. It is something.
[0023]
According to a tenth aspect of the present invention, in the character reading system according to the first to ninth aspects, the overall statistic means includes a variance distribution calculating means for calculating a variation distribution by using score values of all the captured works. And an overall statistical display means for displaying a variation distribution using the score values of all the images of the workpieces.
[0024]
An eleventh aspect of the present invention is the character reading system according to the first to tenth aspects, wherein the partial statistic display means and the overall statistic display means calculate an average value, It is constituted by a graph display means for displaying in a graph format comprising a standard deviation, a discrimination value, a threshold value and a warning zone.
[0025]
According to a twelfth aspect of the present invention, in the character reading system according to the first to eleventh aspects, the partial statistic display means and the overall statistic display means include a comment area for displaying a comment on a variation distribution. The statistical means is provided with comment display means for displaying a comment when the discrimination value for the comment area is within the range of the warning zone.
[0026]
According to a thirteenth aspect of the present invention, in the character reading system according to the first to twelfth aspects, the statistical means monitors a variation distribution calculated by the partial statistical processing, and manually or when exceeding a permissible range. An image pickup system adjusting means for automatically adjusting the image pickup means and resetting the threshold value is provided.
[0027]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, the present invention will be described based on an illustrated embodiment. FIG. 1 is a configuration diagram showing one embodiment of the character reading system of the present invention.
[0028]
The character reading system includes an imaging unit 2, a pattern matching unit 3, and a statistical unit 4, reads an individual pattern printed on the work 1, and checks whether the individual pattern is correct. .
[0029]
The work 1 (product, that is, the object to be read) is, for example, a print of an individual pattern including a plurality of alphanumeric characters.
[0030]
The imaging unit 2 includes a work detection unit 5, an illumination device 6, a brightness adjustment unit 7, and an imaging device 8.
[0031]
The work detection means 5 is a sensor, a switch, or the like, and detects passage of the conveyed work 1 in order to determine an imaging timing.
[0032]
The illumination device 6 is a light source that illuminates the individual identification pattern printed on the work 1.
[0033]
The brightness adjustment means 7 is a power supply having an output adjustment function for adjusting the brightness of the lighting device 6. It is preferable to use a digital power supply capable of accurate output adjustment.
[0034]
The imaging device 8 is actually a CCD camera or the like, and captures an individual pattern printed on the work 1 and outputs an imaging signal to the pattern matching unit 3.
[0035]
The pattern matching means 3 and the statistical means 4 are actually software operating on a computer system or dedicated devices equipped with equivalent functions. Each of them can be mounted on a separate computer system or dedicated device and operated cooperatively by communication, or can be mounted on one computer system or dedicated device.
[0036]
FIG. 2 shows the configuration of the pattern matching means 3. The pattern matching unit 3 includes a reference pattern storage unit 9, a threshold setting unit 10, a binarization unit 11, a score value calculation unit 12, a determination unit 13, a binary image storage unit 14, and a binary image data display. Means 15 for obtaining binary image data from an image pickup signal, performing matching with a reference pattern, and reading an individual pattern.
[0037]
The reference pattern storage means 9 stores in advance a reference pattern to be compared with binary image data obtained from an imaging signal. The reference pattern is composed of a plurality of character patterns, and is, for example, binary image data for each character.
[0038]
The threshold setting unit 10 determines whether or not the binary image data matches the character of the reference pattern, that is, determines what the character of the individual pattern is based on the score value comparing the binary image data with the reference pattern. The threshold is set in advance. The threshold is set, for example, to 90%.
[0039]
The binarizing means 11 converts the imaging signal into binary digital data, and generates binary image data of the individual identification pattern.
[0040]
The score value calculating means 12 compares the binary image data with the reference pattern and calculates a score value.
[0041]
Generally, there is a slight displacement between the binary image data and the reference pattern. For the purpose of correcting the positional deviation, a relative position change from the reference position is performed within a predetermined range (for example, a range of three pixels on each side of the top and bottom).
[0042]
At each of the reference position and the relative position change, the pixel value of the corresponding pixel is compared, a score value is calculated, and the maximum score value is set as the score value between the binary image data and the one character in the reference pattern.
[0043]
Such a calculation of the score value is performed for the binary image data and all the character patterns in the reference pattern, and the character pattern that gives the maximum score value among the score values for all the character patterns is calculated as the binary image data. The character pattern matches the data.
[0044]
The determination means 13 compares the score value calculated by the score value calculation means 12 with a threshold value set in advance by the threshold value setting means 11, and determines whether or not a correct character has been read.
[0045]
For example, assuming that the threshold value is 90%, if the score value is 90% or more, it is determined that reading is OK, and if it is less than 90%, it is determined that reading is NG.
[0046]
The binary image storage means 14 stores the binary image data.
[0047]
The binary image display means 15 is for displaying the binary image stored by the binary image storage means 14. In particular, as shown in the example of FIG. 5, it can be used to confirm an individual pattern determined to be read NG, a position of the read NG, and the like.
[0048]
FIG. 3 is a configuration diagram of the statistical means 4. The statistic unit 4 includes a partial statistic unit 16, an overall statistic unit 17, an imaging system adjustment unit 18, and a comment display unit 30, and performs statistical processing of the matching result of the pattern matching unit 3.
[0049]
The partial statistical unit 16 includes a partial statistical processing unit 19, a partial statistical storage unit 20, and a partial statistical display unit 21, and is set in advance each time the pattern matching unit 3 calculates a score value of one work. The statistical processing is performed using the score values of a fixed number.
[0050]
The partial statistical processing unit 19 includes a partial statistical calculating unit 22, a sample number setting unit 23, a probability accuracy setting unit 24, and a warning zone setting unit 25.
[0051]
The partial statistic calculation unit 22 calculates an average value, a standard deviation, and a discrimination value using the latest fixed number of score values, and obtains a normal distribution indicating a variation distribution of the score values.
[0052]
The above-mentioned latest fixed number of score values is, for example, when the fixed number is 500 and the work which has been read this time is the 600th, the 500th score value is retroactive from the latest 600th score value, that is, the 101st score value. Indicates that the score values from # to # 600 are used.
[0053]
The discrimination value is a value of the average value + 5σ or the average value −5σ in the variation distribution.
[0054]
The sample number setting means 23 presets a fixed number of score values used by the partial statistical calculation means 22 to obtain a normal distribution.
[0055]
The probability accuracy setting means 24 specifies how many times the standard deviation a score value variation distribution is to be viewed in the normal distribution. If necessary, 3σ (99.7% statistically appears in the range) to 5σ (99.9999994% statistically appear in the range) can be used properly.
[0056]
The warning zone setting means 25 sets the width of the warning zone on both sides of the threshold value, where a determination error may occur in the character reading determination.
[0057]
FIG. 4 is an explanatory diagram of the warning zone. In FIG. 4, the warning zone includes the threshold value−the margin value and the threshold value + the margin value sandwiching the set threshold value. The margin value is set arbitrarily. Hereinafter, a method of evaluating the discrimination value for the warning zone will be described.
[0058]
First, a character is read a plurality of times for the same character as the reference pattern, and a normal distribution of the score value is obtained. The average value at that time is X and the standard deviation is σx, and the determination value Hx = X−5σx is calculated.
[0059]
Next, character reading is performed a plurality of times for characters different from the reference pattern, and a normal distribution of the score values is obtained. The average value at that time is Y and the standard deviation is σy, and the determination value Hy = Y + 5σy is calculated.
[0060]
In the variation distribution using the latest fixed number of score values calculated each time the pattern matching unit 3 calculates the score value of one work, when the discrimination value H of the distribution falls in the warning zone, 1) the average value This means that the position is moved, 2) the variation is increased, and the likelihood of a determination error is increased.
[0061]
The partial statistics storage means 20 stores the average value, the standard deviation, and the discrimination value calculated by the partial statistics calculation means 22 each time the calculation is performed.
[0062]
The partial statistics display means 21 includes a latest display means 26 and a history display means 27. FIG. 5 shows a screen and a display example.
[0063]
The latest display means 26 displays a normal distribution, a discrimination value, and a warning zone as a graph based on the latest average value and standard deviation calculated by the partial statistics calculation means 22.
[0064]
The history display unit 27 reads an average value and a standard deviation at arbitrary times stored in the partial statistical storage unit 20 and displays a normal distribution. In FIG. 5, the selection of an arbitrary time is shown by a screen display called a tab. However, the selection method is arbitrary, such as displaying a list or displaying an icon indicating each time.
[0065]
The overall statistical unit 17 includes an overall statistical calculating unit 28 and an overall statistical display unit 29. The overall statistical unit 17 calculates an average value, a standard deviation, and a discrimination value by using the score values of all the captured workpieces. To display.
[0066]
The overall statistic calculation means 28 calculates an average value, a standard deviation, and a discrimination value by using all score values obtained by reading characters. As a result, the number of sampling score values is increased as compared with the partial statistical means 16, and the accuracy of the statistical processing can be increased. Not only score values for works of the same type but also scores for a plurality of different types of works may be used.
[0067]
The overall statistic display unit 29 displays a normal distribution, a discrimination value, and a warning zone as a graph based on the average value and the standard deviation of all the score values calculated by the overall statistic calculation unit 28.
[0068]
The imaging system adjustment unit 18 monitors the variation distribution obtained by the partial statistics unit 16 and adjusts the imaging system such as the aperture of the camera and resets the threshold value manually or automatically when the variation exceeds an allowable range. It is. It is also possible to periodically insert a sample work into the work 1 and make adjustments each time the sample work is read.
[0069]
The comment display means 30 displays a warning comment in the comment area on the screen shown in FIG. 5 when the discrimination value enters the warning zone.
[0070]
The operation of the character reading system of the present invention will be described based on the flowchart shown in FIG.
[0071]
First, when the work 1 on which the individual pattern is printed is conveyed and reaches a predetermined position such as immediately below the imaging device 8, when the work detection means 5 detects the work 1, the imaging device 8 detects the individual pattern of the work 1. An image is taken, and an image pickup signal of the individual pattern is output to the pattern matching means 3 (steps S1 to S3).
[0072]
Next, the binarizing unit 11 binarizes the imaging signal to generate binary image data, and the score value calculating unit 12 compares the binary image data with a reference pattern to calculate a score value ( Step S4).
[0073]
The binary image storage means 14 stores the binary image data, and the binary image display means 15 displays the binary image data (step S5).
[0074]
The determination unit 13 compares the score value calculated by the score value calculation unit 12 with a threshold value set in advance by the threshold value setting unit 11, and determines whether the character matches the reference pattern, that is, whether the character is correctly read (step S1). S6).
[0075]
If it is determined that the characters have been correctly read, the partial statistical calculation means 22 calculates an average value, a standard deviation, and a discrimination value using the score values for the number of samples set by the number-of-samples setting means 23 (step S7). ).
[0076]
The partial statistic storage unit 20 stores the average value, the standard deviation, and the discrimination value calculated by the partial statistic calculation unit 22 together with, for example, an identification number indicating the processing of the work 1 (step S8).
[0077]
Further, the latest display means 25 creates a graph of the normal distribution from the average value and the standard deviation, and displays the latest normal distribution together with the discrimination value and the warning zone (step S9).
[0078]
When the calculated discrimination value enters the warning zone due to the movement of the average value or the increase or decrease of the standard deviation, a warning comment is displayed in the comment display means comment area (step S10).
[0079]
If the discrimination value is not acceptable and enters the warning zone, the imaging system adjustment unit 18 adjusts the imaging system such as the aperture of the camera and resets the threshold (step S11).
[0080]
As described above, the present invention has been described in detail based on the embodiment. However, the character reading system according to the present invention is not limited to the above embodiment at all, and does not depart from the gist of the present invention. Naturally, various modifications are possible.
[0081]
【The invention's effect】
As described above, according to the character reading system of the present invention, by calculating the variation distribution and displaying it along with the threshold value, it is possible to evaluate the threshold value on a statistical basis using the average value and the standard deviation. Become.
[0082]
Further, every time an individual pattern of a work is read, a variation distribution is calculated and displayed using a fixed number of score values from the latest, so that the state of the imaging system can be monitored in time series.
[0083]
In addition, every time an individual pattern of a workpiece is read, a variation distribution is calculated and displayed using a fixed number of score values from the latest, so that the imaging system adjustment unit can adjust the imaging system and reset the threshold as needed. It is possible to do.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a character reading system of the present invention.
FIG. 2 is a configuration diagram of a pattern matching unit of the present invention.
FIG. 3 is a configuration diagram of statistical means of the present invention.
FIG. 4 is an explanatory diagram of a warning zone according to the present invention.
FIG. 5 is a screen example of a partial statistics display unit of the present invention.
FIG. 6 is a flowchart illustrating an operation of the character reading system of the present invention.
[Explanation of symbols]
REFERENCE SIGNS LIST 1 work 2 imaging means 3 pattern matching means 4 statistical means 5 work detection means 6 lighting device 7 brightness adjustment means 8 imaging device 9 reference pattern storage means 10 threshold setting means 11 binarization means 12 score value calculation means 13 determination means 14 Binary image storage means 15 Binary image display means 16 Partial statistics means 17 Overall statistics means 18 Imaging system adjustment means 19 Partial statistics processing means 20 Partial statistics storage means 21 Partial statistics display means 22 Partial statistics calculation means 23 Sample number setting means 24 Probability accuracy setting means 25 Warning zone setting means 26 Latest display means 27 History display means 28 Overall statistics calculation means 29 Overall statistics display means 30 Comment display means

Claims (13)

A character reading system that recognizes an individual pattern of a workpiece in order to identify a product and checks whether the pattern is a correct individual pattern,
Imaging means for imaging the individual pattern from the work, and outputting an imaging signal;
Pattern matching means for obtaining binary image data from the imaging signal and matching with a reference pattern stored in advance;
In the pattern matching means, statistical means for performing statistical processing of the result of the matching,
A character reading system comprising: The character reading system according to claim 1,
The imaging means,
Work detection means for instructing the imaging timing of the work,
An illumination device that illuminates the work when the work is imaged;
Brightness adjustment means for adjusting the brightness output of the lighting device,
An imaging device that images the work and outputs an imaging signal;
A character reading system comprising: The character reading system according to claim 1 or 2,
The pattern matching means,
Reference pattern storage means for storing a character pattern for each character in advance;
A binarizing unit that receives the image signal and obtains binarized binary image data;
Threshold setting means for setting a threshold for evaluating a comparison result between the binary image data and the reference pattern;
A score value calculating unit that compares the binary image data with the reference pattern and calculates a score value;
Determining means for comparing the score value with the threshold value and determining an individual pattern of the work;
A character reading system comprising: The character reading system according to claim 1, wherein:
The pattern matching means,
A binary image storage means for storing the binary image data;
A binary image display means for displaying the stored binary image data;
A character reading system comprising: The character reading system according to any one of claims 1 to 4,
The statistical means,
Each time the pattern matching unit reads an individual pattern of one work, a partial statistical unit that calculates a variation distribution composed of an average value, a standard deviation, and a discrimination value of the score values using the latest fixed number of score values. ,
A character reading system comprising: The character reading system according to claim 1, wherein:
The partial statistical means,
Each time the pattern matching unit reads an individual pattern of one work, using a latest fixed number of score values, a partial statistical processing unit that calculates a variation distribution of the score values,
Partial statistical storage means for storing a variation distribution calculated using the certain number of score values,
Partial statistics display means for displaying a variation distribution calculated using the certain number of score values,
A character reading system comprising: The character reading system according to claim 1, wherein:
The partial statistical processing means,
Sample number setting means for setting a fixed number of score values used for calculating the variation distribution;
Probability accuracy setting means for setting how many times the standard deviation is to be calculated in the variation distribution,
In the determination of the individual pattern in advance, if the determination value falls within a specific area, a warning zone instructing re-determination,
Warning zone setting means for setting the width of the warning zone,
Using the set fixed number of score values, a partial statistical calculation means for calculating the variation distribution,
A character reading system comprising: The character reading system according to claim 1, wherein:
The partial statistics display means,
Immediately after the partial statistical calculation means calculates the variation distribution, a latest display means for displaying the variation distribution,
History display means for calling and displaying the variation distribution stored by the partial statistical analysis storage means,
A character reading system comprising: The character reading system according to claim 1, wherein:
The statistical means,
Overall statistical means for calculating a variation distribution using the score values of all of the captured workpieces,
A character reading system comprising: The character reading system according to claim 1, wherein:
The overall statistical means is:
Overall statistical calculation means for calculating the variation distribution using the score values of all the images of the workpiece,
Overall statistical display means for displaying a distribution of variations using the score values of all the workpieces imaged,
A character reading system comprising: The character reading system according to claim 1, wherein:
The partial statistics display means, and the overall statistics display means,
A graph display means for displaying each of the partial or entire variation distributions in a graph format including an average value, a standard deviation, a discrimination value, a threshold value, and a warning zone.
A character reading system, comprising: The character reading system according to claim 1, wherein:
The partial statistics display means, and the overall statistics display means,
The discrimination value, and based on the warning zone, comprising a comment area for displaying a comment on the variation distribution,
The statistical means,
Comment display means for displaying a comment when the determination value is in the range of the warning zone for the comment area;
A character reading system comprising: The character reading system according to claim 1, wherein:
The statistical means,
Monitoring the distribution of dispersion calculated by the partial statistical processing, manually, or automatically when exceeding the allowable range, adjustment of the imaging means, and an imaging system adjustment means for resetting the threshold,
A character reading system comprising:

Images