JP3640219B2 - Banknote recognition device - Google Patents

Description

[0001]
[Industrial application fields]
The present invention relates to a banknote recognition apparatus, and more particularly to a banknote identification apparatus that optically detects micro characters printed on a banknote or the like and determines the authenticity of the banknote or the like.
[0002]
[Prior art]
In recent years, the progress of copying technology has been remarkable, and a problem has arisen in that securities such as banknotes and checks are misused by copying.
[0003]
Therefore, in order to avoid such misuse, research for identification has been proceeding rapidly in order to make it easy to identify a copy by being false.
[0004]
In recent years, a method has been proposed in which minute characters called micro characters are formed at the end of a banknote, and the micro characters are crushed when copied so that they can be easily identified.
[0005]
Although this method is considered to be a very effective method for detecting counterfeit bills, there has been no bill recognition device that effectively detects this. Further, on the linear image sensor, there is a problem that the direction of the banknote needs to be regulated with high accuracy in order to form an image by aligning micro character strings in the direction in which the pixels are arranged.
[0006]
[Problems to be solved by the invention]
Thus, in order to detect a micro character when detecting a counterfeit bill, it is necessary to limit the traveling direction of the bill with high accuracy, and it is necessary to use a precision machine as a peripheral device.
[0007]
The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a banknote discriminating apparatus capable of discriminating banknotes with high accuracy, focusing on the fact that micro characters have a characteristic spatial frequency.
[0008]
[Means for Solving the Problems]
In the banknote identification apparatus of the present invention is therefore, a conveying means for conveying the banknote in a direction perpendicular to the string of micro characters printed with grid pattern on the bill, the detection light to the bill conveyed by the conveying means And a peak interval obtained from the output of the reflected light detecting means, and a reflected light detecting means for detecting reflected light from the banknote. Using a histogram of the number of pixels, a feature amount creating unit that creates a micro character feature amount for a predetermined number of pixels in a peak interval, and using the micro character feature amount created by the feature amount creating unit, the micro character feature amount And a correlation distance calculating means for calculating a correlation distance between a grid pattern and a correlation distance between the micro character feature and the micro character pattern, and the correlation By the pattern matching of the correlation distance between the micro character feature amount calculated by the separation calculating means and the lattice pattern, and the correlation processing between the micro character feature amount and the lattice pattern measured in advance for the genuine note of the banknote, or By pattern matching between the correlation distance between the micro character feature amount calculated by the correlation distance calculation means and the micro character pattern, and the correlation distance between the micro character feature amount and the micro character pattern by measuring in advance the genuine note of the bill. Pattern matching means for determining whether or not it is the micro character, and determination means for determining whether the banknote is a genuine bill or a counterfeit bill according to a result of pattern matching by the pattern matching means. To do.
[0009]
[Action]
According to the above configuration, the spatial frequency is read from the output of the reflected light detection means, the spatial frequency distribution is created in the bill conveyance direction, and there is a characteristic pattern peak in the genuine note. Authenticity is discriminated based on whether or not exists.
[0010]
In the case of counterfeit bills, the micro characters are read without being resolved at the time of copying and are crushed, the micro characters are not detected, and there is no characteristic peak of the spatial frequency.
[0011]
Therefore, by such pattern matching, it is possible to efficiently eliminate counterfeit tickets by copying and counterfeit tickets by poor printing, and it is possible to eliminate old tickets by determining the presence or absence of micro characters.
[0012]
【Example】
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0013]
FIG. 1 is a schematic explanatory diagram of a banknote recognition apparatus according to an embodiment of the present invention.
The identification device includes a transport unit (not shown) that transports the banknote 1 in a direction orthogonal to the character string 2 of micro characters printed on the banknote, and a banknote transported by the transport unit. LED array 3 for irradiating detection light to the CCD, CCD linear sensor 5 arranged in parallel to the character string 2 so as to detect the reflected light from the banknote via the lens optical system 4, and the CCD linear sensor 5 From the output, the peak interval is measured as the number of pixels, the number of pixels is represented by a histogram, and a feature quantity creating means for creating a feature quantity distribution, and a micro that has been created in advance from the output of the feature quantity creating means. A signal processing circuit 6 including pattern matching means for determining whether or not a character is a micro character compared with a character feature amount is provided.
[0014]
As shown in the front sectional view of FIG. 2 (a) and the bottom view of FIG. 2 (b), this structure has a light source comprising an LED array 3 that illuminates bills traveling underneath, an aspheric resin lens 4 and the like. The CCD linear sensor 5 for imaging the light from the banknote through the aspherical resin lens 4 and reading the spatial frequency of the imaged micro character, and the spatial frequency of the micro character string 2 Is output as an electrical signal.
[0015]
Here, in order to read the micro character 2 (see FIG. 4) under “10000” at the upper left of the banknote as shown in FIG. 3, this portion is irradiated with light from the LED array 3, and this CCD linear An image is formed on the sensor 5, and it is detected by this read signal whether it is a genuine note or a forged bill.
[0016]
Partial enlarged views of the micro characters are shown in FIGS. 5 (a) and 5 (b). Here, FIG. 5 (a) is an enlargement of a genuine note and incorporated into a computer. In the figure, the integration area is represented by V, which can be determined by the exposure time and the conveyance speed of the CCD linear sensor 5 . FIG. 5 (b) is a diagram showing the output intensity distribution from the CCD, and the spatial frequency increases as the number of narrower peak intervals increases.
[0017]
Utilizing this fact, the peak interval is measured by the number of pixels, and a histogram showing how many peak intervals exist in one scanning image of the CCD linear sensor 5 is obtained. FIG. 6 shows a histogram when the peak interval pixel number is measured for one of the micro characters. The number of pixels means the number of each pixel of the CCD linear sensor 5 . As is clear from this figure, the peak is at 6 pixels. The same was true for other micro characters. Using this fact, pattern matching is performed using a histogram.
[0018]
By the way, blurring due to copying may have a high spatial frequency, but this is because the peak is low, so that only the peak-to-peak height above a certain level is shown in FIGS. 7 (a) and (b). It can cut by counting.
[0019]
Next, the signal processing circuit of the banknote identification device that actually performs such processing will be described. As shown in the block diagram of FIG. 8, the signal processing circuit includes a banknote detection sensor 11 that detects the start of banknote travel, a CCD linear sensor 5 that detects micro characters, and a sensor that drives the CCD linear sensor 5. A drive circuit 12, an amplifier 13 that amplifies the output of the CCD linear sensor 5, an A / D converter 14 that digitally converts the output of the amplifier 13, a RAM 15 that stores this output, and an output of the banknote detection sensor 11 On the basis of this, the CPU 15 is configured to extract a desired signal from the RAM 15 and calculate it. The ROM 17 stores various processing programs, reference patterns, and the like. The CPU 16, RAM 15, and ROM 17 constitute feature quantity creating means and pattern matching means.
[0020]
The CPU executes a calculation as shown in the flowchart of FIG.
[0021]
First, it is determined whether or not a banknote is detected based on the output of the banknote detection sensor 11 (determination step 100). If it is determined that the banknote is detected, the output of the CCD linear sensor 5 is amplified by an amplifier 13 and A / D converted. Digital conversion is performed by the device 14 (step 101). Then, one line of CCD pixel data is stored in the RAM 15 (step 102), and a peak in which the output difference between peaks exceeds a certain value is detected from the stored data (step 103).
[0022]
Then, the number of pixels in the peak interval is counted, and the number of peaks within the predetermined number of pixels is set as the micro character feature amount (step 104). Thereafter, the micro character feature value for each line is stored in the memory (step 105). Here, a micro character feature amount distribution is created from the micro character feature amounts of about 500 lines of one bill.
[0023]
Then, it is determined whether or not the banknote detection sensor 11 has detected the trailing edge of the banknote (determination step 106). If it is determined that the trailing edge has been detected, the microcharacter pattern and the lattice pattern pattern are determined by the microcharacteristic feature distribution. Is calculated (step 107).
[0024]
It is determined whether or not there is only one pattern at a predetermined position where the distance between the lattice pattern and the micro character is a predetermined value or less (determination step 108).
[0025]
If it is determined that only one is present, it is determined that the bill is a bill with micro characters, that is, a genuine note (step 109).
[0026]
On the other hand, if it is determined that a pattern equal to or less than the predetermined value does not exist, is not in a predetermined position, or is not only one, it is determined that the banknote is not a micro-character banknote, that is, a genuine note but a false note (step 110). .
[0027]
In this way, micro characters can be detected very easily, and counterfeit bills can be easily detected.
[0028]
Here, the calculation of the micro character feature amount in step 104 will be described with reference to FIG. In this figure, the lattice pattern CCD output and its peak interval pixel number histogram, the micro character CCD output and its peak interval pixel number histogram, the grass pattern CCD output and its peak interval pixel number histogram are shown in order from the top. It can be seen that the number of pixels in the peak interval of CCD output for micro characters is 5 and peaks.
[0029]
FIG. 11 shows the CCD output and histogram of the color copy ticket for the grid pattern ((a), (b)) and micro characters ((c), (d)), respectively. As is clear from the comparison with FIG. 10, the frequency of the number of pixels in the peak interval varies greatly and there is no peak, so that they can be identified.
[0030]
In this example, as shown in FIG. 12, a value obtained by integrating a histogram from 2 pixels to 8 pixels is set as a micro character feature amount.
[0031]
In practice, as shown in FIG. 13, the bill reading area R in FIG. 13 (a) is measured by the CCD. Here, the bill conveyance speed was 2000 mm / sec, and the CCD 256 pixel video rate was 4 MHz. The micro character feature amount at this time is as shown in FIGS. 13 (b) and 13 (c) (FIG. 13 (c) is a partially enlarged view of FIG. 13 (b)). As is clear from the above, a characteristic pattern in the micro character feature amount distribution exists at a predetermined position.
[0032]
FIG. 14 is an explanatory diagram showing the correlation distance S (X) between the micro character feature value, the lattice pattern, and the micro character. The pattern of the lattice pattern portion in the micro character feature distribution is shown in the middle of FIG. 14 and the pattern of the micro character portion is shown in the lower, and the correlation distance between these and the upper distribution is calculated.
[0033]
According to the method of FIG. 14, the measurement result of the correlation distance between the micro character feature amount and the lattice pattern pattern with respect to the peak interval pixel number of 2 to 8 is measured for a genuine note of 10,000 yen, a color copy ticket, and three black and white copy tickets. They are shown in FIGS. 15 to 19, respectively. From these comparisons, the correlation distance is minimum at the lattice pattern portion in the genuine note, but the correlation distance is not minimum at the lattice pattern portion in the color copy, as shown in FIGS. 15 and 16. Be identifiable. However, in the case of a black and white copy ticket, the minimum value of the correlation distance is larger than that of a genuine ticket, so that it can be identified.
[0034]
Further, in order to determine only the presence or absence of the micro character, the micro character feature amount is obtained as an integral value with respect to the peak interval pixel numbers 2 to 4, the feature amount not affected by the lattice pattern is obtained, and the micro character feature amount and the micro character are obtained. FIGS. 20 to 24 show the results of measuring the correlation distance with the pattern for a genuine ticket of 10,000 yen, a color copy ticket, and three black-and-white copy tickets, respectively. From these comparisons, the correlation distance is minimum in the micro character part in the genuine note, but in addition to the minimum correlation distance in the lattice pattern part in the color copy and black-and-white copy, there is also a difference in the minimum value. Therefore, it can be identified.
[0035]
FIG. 25 (a) shows the measurement result for a genuine note, and FIG. 25 (b) shows the measurement result for a copy ticket. That is, histograms are created from 100 samples for the minimum correlation distance with the lattice pattern shown in FIGS.
[0036]
In order to separate with higher accuracy, it is detected whether or not the place where the minimum value is generated matches the lattice pattern (the place where the micro characters are present).
[0037]
Note that in the determination step 108, it is determined whether or not there is only one pattern at a predetermined position where the distance between the lattice pattern and the micro character is equal to or smaller than a predetermined value. This is a case of 10,000 yen. In the case of 5,000 yen, it is determined by the distance between micro characters. The position of 10,000 yen and 5000 yen micro characters are the same, but in the case of 1000 yen, the position and direction are different (diagonal direction), so the same sensor as this application is used to detect 1000 yen micro characters. It is necessary to change the position separately.
[0038]
【The invention's effect】
As described above, according to the present invention, the micro character feature amount and the lattice pattern pattern with respect to a predetermined number of peak interval pixels, or the micro character feature amount and the micro character pattern based on the correlation distance. It is characterized by performing a process for determining whether or not it is a character string, and since it is fundamental to determine only the presence or absence of a micro character string in a banknote subject to authenticity determination by this process, whether or not the banknote is genuine It is not necessary to recognize each character to perform authenticity determination, and it is possible to perform highly accurate authenticity determination of banknotes without being regulated to carry a medium with very high accuracy. Since there is no need to regulate the conveyance of the medium with high accuracy, it is possible to use the banknote transport mechanism of the existing banknote identification device as it is, and the logic can be added to the existing device. Enough can be applied.
[Brief description of the drawings]
FIG. 1 is a diagram showing a banknote identification device according to an embodiment of the present invention. FIG. 2 is a diagram showing a sensor unit of the device. FIG. 3 is a diagram showing an example of a bill to be read by the device. Fig. 5 is an enlarged view of the micro character portion of the banknote. Fig. 6 is an enlarged view of the micro character portion of the bank note. Fig. 7 is a comparative diagram showing an example of reading by the apparatus. 8 is a block diagram of a signal processing circuit. FIG. 9 is a flowchart of true / false detection. FIG. 10 is a calculation explanatory diagram of a micro character feature amount histogram. FIG. FIG. 12 is a diagram showing a histogram from 2 pixels to 8 pixels. FIG. 13 is a micro character feature amount for a banknote reading region R and an enlarged explanatory view thereof. FIG. FIG. 15 is a diagram showing correlation distances S (X) with micro characters. FIG. 15 shows the result of measuring the correlation distance between the micro character feature quantity and the lattice pattern pattern with respect to the peak interval pixel numbers 2 to 8 according to the method of FIG. A figure showing (a genuine note of 10,000 yen),
FIG. 16 is a diagram showing a result of measuring a correlation distance between a micro character feature amount and its lattice pattern pattern for peak interval pixel numbers 2 to 8 according to the method of FIG. 14 (color copy ticket);
FIG. 17 is a diagram showing the result of measuring the correlation distance between the micro character feature quantity and its lattice pattern pattern with respect to the peak interval pixel number of 2 to 8 according to the method of FIG. 14 (monochrome copy ticket)
FIG. 18 is a diagram showing the result of measuring the correlation distance between the micro character feature quantity and its lattice pattern pattern with respect to the peak interval pixel number of 2 to 8 according to the method of FIG. 14 (monochrome copy ticket)
FIG. 19 is a diagram showing the result of measuring the correlation distance between the micro character feature quantity and its lattice pattern pattern with respect to the peak interval pixel number of 2 to 8 according to the method of FIG. 14 (monochrome copy ticket)
FIG. 20 is a diagram showing a result of measuring the correlation distance between the micro character feature value and the micro character pattern by obtaining the micro character feature value as an integrated value with respect to the peak interval pixel numbers of 2 to 4 (genuine bill of 10,000 yen).
FIG. 21 is a diagram (color copy ticket) showing the result of measuring the correlation distance between the micro character feature value and the micro character pattern by obtaining the micro character feature value as an integrated value with respect to the peak interval pixel numbers of 2 to 4;
FIG. 22 is a diagram (black and white copy ticket) showing the result of measuring the correlation distance between the micro character feature value and the micro character pattern by obtaining the micro character feature value as an integrated value with respect to the peak interval pixel number of 2 to 4;
FIG. 23 is a diagram (black and white copy ticket) showing the result of measuring the correlation distance between the micro character feature value and the micro character pattern by obtaining the micro character feature value as an integral value with respect to the peak interval pixel numbers 2 to 4;
FIG. 24 is a diagram (black and white copy ticket) showing the result of measuring the correlation distance between the micro character feature value and the micro character pattern by obtaining the micro character feature value as an integrated value with respect to the peak interval pixel numbers 2 to 4;
FIG. 25 is a histogram of the minimum correlation distance with the lattice pattern.
DESCRIPTION OF SYMBOLS 1 Banknote 2 Character string 3 of micro characters LED array 4 Lens optical system 5 CCD linear sensor 6 Signal processing circuit

Claims (2)

Conveying means for conveying the banknote in a direction orthogonal to a character string of micro characters printed on the banknote together with a lattice pattern ;
A detection light irradiating means for irradiating the detection light to the bill conveyed by the conveying means,
A reflected light detecting means comprising a sensor array arranged in parallel with the character string, and detecting reflected light from the banknote;
Feature quantity creating means for creating a micro character feature quantity for a predetermined peak interval pixel number using a histogram of the peak interval pixel number obtained from the output of the reflected light detection means ;
Correlation distance calculation for calculating a correlation distance between the micro character feature quantity and the lattice pattern pattern, or a correlation distance between the micro character feature quantity and the micro character pattern, using the micro character feature quantity created by the feature quantity creating means. Means,
By the pattern matching of the correlation processing between the micro character feature amount calculated by the correlation distance calculating means and the lattice pattern, and the correlation processing between the micro character feature amount and the lattice pattern measured in advance for the genuine note of the banknote, Alternatively, the correlation distance between the micro character feature amount and the micro character pattern calculated by the correlation distance calculating means, and the pattern of the correlation distance between the micro character feature amount and the micro character pattern measured in advance for the bill of the bill. Pattern matching means for determining whether or not the character is a micro character by matching; and
A bill discriminating apparatus comprising: a judging unit for judging whether the bill is a genuine bill or a forged bill according to a result of pattern matching by the pattern matching unit. The banknote identification device according to claim 1 , wherein the reflected light detection means is a CCD linear sensor .

Images