JP2003248852A - Genuineness discrimination device for paper sheet - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the precision of genuineness discrimination of a paper sheet by intaglio printing. <P>SOLUTION: When the paper sheet is carried, an unevenness pattern generated by ink in printing is detected. The unevenness pattern is obtained by extracting a high-frequency component by a high-pass filter from the output of a displacement meter in contact with a paper sheet surface. The unevenness pattern includes a component due to unevenness of the ink of the intaglio printing. As to a specified calculation section, the time integral Ve of the output absolute value of the unevenness pattern is found. The time integral Ve is used as an evaluation parameter to compare a reference value obtained from a genuine ticket with a value obtained from the paper sheet to be discriminated, thereby precisely discriminating whether the paper sheet is genuine. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to authenticity discrimination of intaglio printed sheets.

[0002]

2. Description of the Related Art Conventionally, bills have been verified by various methods. As one of them, there is a technique for performing authenticity discrimination based on the presence / absence of unevenness on the surface of a banknote caused by ink during printing. For example, JP-A-54-5999
1 discloses a technique in which a stylus is brought into contact with a printing portion on the paper surface to detect irregularities of special printing on a bill to determine the authenticity of the bill. In Japanese Patent Laid-Open No. 63-247895, a roller is brought into contact with a printing portion on a paper surface to detect unevenness of a bill,
A technique is disclosed in which when the absolute value of the output signal exceeds the range determined by the upper threshold value and the lower threshold value, it is determined as a genuine bill. Japanese Unexamined Patent Application Publication No. 9-62893 discloses a technique for discriminating a bill from a genuine bill when minute irregularities due to ink exist in a region of the bill surface called a detection zone.

Generally, banknotes are often manufactured by intaglio printing, and unevenness due to ink is generated on the surface thereof, whereas such unevenness cannot be generated in a fake bill manufactured by a color copy or a printer. . Considering a general manufacturing method of counterfeit bills, authenticity discrimination based on unevenness due to ink is useful.

[0004]

However, bills often have various wrinkles due to their use. In addition, the thickness may differ depending on the denomination. With conventional technology,
In some cases, the presence or absence of wrinkles and the difference in thickness caused false identification of authenticity. This problem is not limited to paper money, but is a common problem for authenticating paper sheets. The present invention has been made to solve such a problem, and an object of the present invention is to improve the accuracy of authenticating a paper sheet by intaglio printing.

[0005]

In order to solve the above problems, the present invention utilizes an uneven pattern formed on the surface of a paper sheet by printing. The concavo-convex pattern means the state of concavo-convex obtained over a certain area, not the presence or absence of concavity and convexity due to printing ink. The concavo-convex pattern is obtained as a concavo-convex waveform over a certain area, and contains the concavo-convex component caused by the ink. Therefore, by analyzing the concavo-convex pattern, it becomes possible to distinguish the concavo-convex caused by printing from the concavo-convex caused by other factors, and it is possible to avoid the influence of wrinkles, noises, etc. and perform accurate authenticity discrimination. .

In the present invention, the detection of the uneven pattern is performed by
It is preferable to carry out in the sheet conveying direction. By doing so, there is an advantage that the uneven pattern can be detected while conveying the paper sheet. Of course, the uneven pattern may be detected in a direction intersecting the transport direction,
It may be two-dimensionally detected on the surface of the paper sheet.

In the present invention, it is preferable to mainly use a concavo-convex pattern of a short wavelength component in which the concavo-convex wavelength is a predetermined value or less. When the uneven pattern is viewed as a waveform, the unevenness caused by printing generally has a shorter wavelength than the unevenness caused by wrinkles or the like. Therefore, by using the short wavelength component, it is possible to avoid the influence of wrinkles and improve the accuracy of the authenticity discrimination. The predetermined value can be set experimentally or analytically according to the wavelength range of the uneven waveform generated by wrinkles or the like.

When detecting an uneven pattern in the sheet conveying direction, an uneven waveform is obtained as a function of time. In such a case, the short wavelength component may be restated as a high frequency component. However, the frequency included in the uneven waveform varies depending on the transport speed. Therefore, it is preferable to set the predetermined value as the extraction reference of the high frequency component according to the transport speed.

The high frequency component can be extracted by, for example, a high pass filter circuit. Alternatively, the extraction may be performed based on various frequency analyzes such as Fourier analysis.

In true / false discrimination, for example, a concavo-convex pattern obtained from a paper sheet to be discriminated is directly compared with a concavo-convex pattern obtained from a genuine paper sheet, and based on whether or not they coincide with each other. It can be carried out.

As another method, based on the uneven pattern,
The authenticity may be discriminated based on whether or not the value of the predetermined quantitatively determined evaluation parameter is within the range allowed for the genuine note. By doing so, objective discrimination can be performed relatively easily.

As the evaluation parameter, for example, a parameter that quantitatively represents the wavelength distribution of the concavo-convex pattern can be used. For example, the average value of wavelength distribution, standard deviation, wavelength having maximum amplitude, maximum wavelength, minimum wavelength, and the like can be used. These may be used alone or in combination.

The evaluation parameter may be the integral of the absolute output values of the concavo-convex pattern. The integral is defined by the detection pattern of the uneven pattern. If the concavo-convex pattern is obtained as a function of position, it is line integral or area. When the uneven pattern is obtained as a function of time, it is time integration. Such an evaluation parameter has an advantage that the authenticity can be accurately discriminated by a relatively simple operation.

The integral may be obtained by targeting an absolute output value that exceeds a predetermined threshold value set in advance. In this case, the output below the threshold value is not reflected in the evaluation parameter. Depending on the setting of the threshold value, it is possible to avoid the influence of noise whose output absolute value is relatively small,
In some cases, it is possible to improve the detection accuracy.

The uneven pattern may be detected by a non-contact sensor or may be mechanically detected by a mechanism having a contact portion that contacts the surface of the paper sheet. According to the former, it is possible to suppress the influence of vibration and the like and to accurately detect the uneven pattern. According to the latter, the concavo-convex pattern can be detected with a relatively simple and low-cost mechanism.

When the uneven pattern is mechanically detected, it is preferable to perform the authenticity discrimination except at least one of a predetermined period from the start of detection and a predetermined period before the end of detection. This is because, in general, noise is likely to be included between the start of detection and the end of detection. By excluding such a period, the discrimination accuracy can be improved.

In the present invention, it is preferable to store a plurality of analysis results of the concavo-convex pattern to be obtained from a genuine note in association with at least one of the type of paper sheet and the conveyance posture. The accuracy of the authenticity discrimination can be improved by performing the discrimination by properly using the analysis result according to the type of the paper sheet and the conveyance posture.

The present invention can be configured in various ways.
For example, it may be configured as a genuine / counterfeit discriminating apparatus for discriminating the authenticity of a paper sheet, or may be configured as a genuine / counterfeit discriminating method. Further, it may be configured as a computer program for causing a computer to realize the above-mentioned authenticity discrimination, or may be configured as a computer-readable recording medium recording the computer program. INDUSTRIAL APPLICABILITY The present invention can be used for authenticating various kinds of intaglio printed sheets, and for example, can be used for authenticating banknotes.

As the storage medium, a flexible disk, a CD-ROM, a DVD, a magneto-optical disk, I
Various computer-readable media such as a C card, a ROM cartridge, a punched card, a printed matter on which codes such as a bar code are printed, an internal storage device (memory such as RAM and ROM) of the computer, and an external storage device can be used.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below in terms of the following items. A. Device configuration: B. Principle of authenticity discrimination: C. Authenticity discrimination processing: D. Modification of First Embodiment: E. Second Example: F. Modification:

A. Device Configuration: FIG. 1 is an explanatory diagram showing a configuration of a bill discriminating device as an embodiment. It is the top view which looked at the banknote discrimination device from the upper part. This bill discriminating apparatus is provided in, for example, an automatic teller machine of a financial institution, a so-called ATM, and discriminates authenticity of bills to be deposited and withdrawn. The bill discriminating device can be incorporated not only in the ATM but also in various devices such as an automatic vending machine, which require the discrimination of bills. Further, it may be configured so that it can operate independently.

The bill discriminating apparatus conveys the bill 10 at the conveying speed Vp in the direction indicated by the arrow in the drawing by the roller 62 attached to the rotary shaft 61. The image sensor 30 scans the image of the banknote 10 during the transportation process. This image is used to discriminate the authenticity of the banknote 10 and to determine the conveyance attitude. In the figure, two optical timing sensors 40 are provided at positions hidden by the lower surface of the bill. The timing sensor 40 detects the presence or absence of banknotes.

FIG. 2 is a side sectional view of the bill validator. The state of the AA cross section in FIG. 1 is schematically shown. The banknote 10 is conveyed by the roller 62 at the conveying speed Vp. A scraper 63 for removing dirt is attached to the roller 62.

A movable roller 72 is provided at a position facing the roller 62 with the banknote 10 interposed therebetween. The movable roller 72 rotates around the rotation shaft 71 and conveys the banknote 10. The rotary shaft 71 is attached to an arm 75 that is rotatable around the rotary shaft 74. A spring 76 is attached to the arm 75 and urges the entire arm 75 in the arrow F direction, that is, in the direction in which the roller 72 is pressed against the banknote 10. A scraper 73 for removing dirt is attached to the roller 72.

Various irregularities are present on the surface of the bill 10 due to the printing ink 11. As shown by the two-dot chain line in the figure, the arm 75 rotates about the axis 74, so that the movable roller 72 acts as a contact sensor for detecting the unevenness of the banknote 10. As shown in the drawing, a displacement gauge 80 for detecting the displacement of the arm 75 is provided at one end of the arm 75. When the arm 75 rotates due to the unevenness of the banknote 10, the distance between the arm 75 and the displacement meter 80 changes like d1 and d2. The displacement meter 80 outputs this displacement to the high-pass filter 81 as an electric signal.

This electric signal has a waveform corresponding to the unevenness of the banknote 10 and the transport speed Vp. This waveform contains various frequencies. The high-pass filter 81 is a circuit that extracts only high-frequency components of a predetermined value or more from this. The range to be extracted will be described later.

The output of the high pass filter 81 is converted into a digital signal by the A / D converter 90 and input to the control unit 100. The detection signal from the timing sensor 40 is similarly converted into a digital signal by the A / D converter 90 and input to the control unit 100.

The control unit 100 is a microcomputer having a CPU and a memory inside. The control unit 100 controls the operation of the bill validator and also
Receiving signals from the D converter 90, the image sensor 30, and the like, the banknotes are subjected to authenticity discrimination processing.

In the figure, the functional blocks provided for authenticity discrimination in the control unit 100 are illustrated. In this embodiment, each of these functional blocks is provided as software, but it may be configured as hardware.

The image processing unit 105 determines the type and conveyance attitude of the bill 10 based on the image acquired from the image sensor 30. The types of banknotes mean passage types and amounts such as yen, dollar, and euro. The conveyance posture means a conveyance state such as front and back of bills.

The concavo-convex pattern acquisition unit 101 uses the ink 11
Get the uneven pattern of. The high frequency component detected by the movable roller 72 and extracted by the high pass filter 81 is input as a digital signal. Here, the concavo-convex pattern acquisition unit 101 does not simply input whether or not the concavities and convexities due to the ink 11 exist in the banknote 10, but the movable roller 72.
Input the output of as a series of waveforms. Hereinafter, such a series of waveforms will be referred to as an uneven pattern.

The evaluation parameter calculation unit 102 calculates a quantitative parameter for performing authenticity discrimination, that is, an evaluation parameter, according to the uneven pattern. The method of calculating the evaluation parameter will be described in detail later.

The authenticity discriminating unit 103 discriminates the authenticity of the bill 10 based on the evaluation parameter. The reference value storage unit 106 is referred to for discrimination. What is the reference value storage unit 106?
The reference value, that is, the evaluation parameter value to be obtained for the genuine note is stored. In the authenticity discriminating unit 103, the value of the evaluation parameter obtained from the bill 10 is the reference value storage unit 106.
It is possible to distinguish between true and false by determining whether or not the value matches the value stored in.

An image of the stored contents of the reference value storage unit 106 is shown in the figure. In the present embodiment, the table format is used for storage, but various modes such as a function format can be adopted. Since the evaluation parameters handled in the present embodiment differ depending on the bill conveyance state, this table stores reference values corresponding to the types and orientations of bills. The authenticity discrimination unit 103 acquires the type of banknote and the conveyance attitude from the image processing unit 105, and refers to the reference value storage unit 106 based on this information.

The result of the authenticity discrimination is output to the outside by the output unit 104. The output destination is determined according to the usage mode of the bill validator. For example, when the bill validator is incorporated in a part of the automated teller machine, the control unit of the automated teller machine is the output destination. When the bill discriminating device is configured as a single device, a display or the like for displaying the discrimination result is an output destination.

FIG. 3 is an explanatory view showing a banknote transporting posture. In this embodiment, the banknote transport attitude is specified by the following parameters. The first parameter is the front and back of the bill. In the figure, the state of the table is illustrated. The second parameter is the transport direction. There are forward and backward directions. In this embodiment, the arrow A1 direction in the drawing is defined as the forward direction and the arrow A2 direction is defined as the reverse direction. Even when the banknote is the back side, the forward direction and the backward direction can be defined. The third parameter is the shift amount. The shift means, as shown by a two-dot chain line in the figure,
It means a state in which the center of the bill is displaced from the center of the transport mechanism. The shift amount Ds at this time is the shift amount. In this embodiment, the shift amount to the right of the orthogonal direction in the transport direction is defined as the positive shift amount. The fourth parameter is the skew angle. The skew means a state in which the left-right symmetry axis of the bill is deviated from the conveyance direction, as indicated by a broken line in the figure. The deviation angle θ at this time is the skew angle. In this embodiment, the rotation direction counterclockwise from the transport direction is defined as the positive skew angle.

The reference value storage section 106 described above stores the reference values in accordance with these transporting postures. Regarding the shift amount and the skew angle, reference values are stored for some representative values. The conveyance postures described here are merely examples, and the present invention is not limited to these. The conveyance posture may be represented by more parameters, or the above-mentioned part may be omitted. In addition, in the present embodiment, the above-described transporting posture is obtained by image processing of banknotes, but a sensor for detecting the transporting posture may be separately provided.

B. Principle of authenticity discrimination: FIG. 4 is an explanatory diagram showing the principle of authenticity discrimination in the embodiment. The explanation will be given based on the experimental results of genuine bills.

The output of the displacement meter 80 is shown in the upper part of the figure. The output during this time is shown as the detection start time t0 when the leading edge of the banknote 10 reaches the movable roller 72 and the detection end te when the trailing edge of the banknote 10 arrives. Depending on the thickness of the banknote 10, the output shifts to the negative side as a whole,
An output corresponding to the unevenness of the ink 11 is obtained. The amplitude due to the unevenness of the ink is much smaller than the amplitude due to the thickness of the banknote, and has a high frequency. In the figure, an enlarged view of a part A of the output is shown. By enlarging the amplitude, the output due to unevenness is confirmed.

As described above, the high frequency component extracted by the high pass filter 81 is input to the control unit 100. The signal after passing through the filter is shown in the second row from the top in the figure. The output mainly due to the unevenness of the ink 11 is extracted. The predetermined period dt1 of detection start and the predetermined period dt2 of detection end include noise due to vibration of the movable roller 72 in addition to the uneven pattern of the ink 11. In this embodiment, the section of the uneven pattern excluding the predetermined period dt1 of detection start and the predetermined period dt2 of detection end is used as the calculation section of the evaluation parameter. dt1,
The interval of dt2 can be set arbitrarily. It is preferable to set it in consideration of both the securing of the time required for damping the vibration of the movable roller 72 and the securing of the calculation section necessary for the authenticity discrimination with sufficient accuracy. In this embodiment,
Although the predetermined period is excluded at both the start and the end of detection, only one of them may be excluded, or the process may be performed without excluding both.

The evaluation parameter is a numerical value that quantitatively represents the uneven pattern thus obtained. In this embodiment, the time integration of the absolute output value of the concavo-convex pattern is used as the evaluation parameter. In the third row in the figure, the output absolute value of the concavo-convex pattern after passing through the filter is shown for the calculation section. The evaluation parameter is obtained by time integration of this output absolute value. This value corresponds to the area between the absolute output value and the time axis. The change in the time integral value is shown in the lower part of the figure. In this example, the evaluation parameter is the value Ve. The value of this evaluation parameter is stored in the reference value storage unit 106 as a reference value for genuine bills.

The control unit 100 calculates the evaluation parameter for the bill to be discriminated by the procedure described above. When the value of the evaluation parameter thus obtained substantially matches the reference value, it is determined that the bill is a genuine note, and when it is different, it is determined that the bill is a counterfeit note.

FIG. 5 is an explanatory view showing the difference between the concavo-convex patterns of the genuine bill and the test bill. The test ticket is a banknote corresponding to a counterfeit note, and a pattern similar to a genuine banknote is printed by an inkjet printer or a laser printer instead of the intaglio printing. In this example, two types of test tickets, "wrinkle ticket" and "folded ticket", which have different damages, were used. A wrinkle ticket is a test ticket that has been subjected to several rounds of rounding into a ping-pong ball and returning to the original state. A fold ticket is a test ticket with a crease that is equivalent to a normal circulating bill. A genuine note has the same tendency to fold as a folded ticket.

FIG. 5 shows the frequency distribution included in the concavo-convex pattern, that is, the amplitude distribution included in each frequency.
The solid line is the experimental result for the genuine bill C1, the alternate long and short dash line is the experimental result for the wrinkle bill C2, and the broken line is the experimental result for the folded bill C3.
Since the frequency of the concavo-convex pattern depends on the transportation speed of the banknote, the horizontal axis represents the normalized frequency, that is, the value obtained by dividing the frequency by the transportation speed. As is clear from the dimensions of the frequency and the transport speed, the normalized frequency is a value indicating the cycle of unevenness contained in a unit length of a bill. This is a value corresponding to the wavelength when the uneven pattern is regarded as a waveform.
In this embodiment, each component included in the concavo-convex pattern is called a frequency, but a component called a high frequency can be read as a short wavelength.

As shown, the normalized frequency has a predetermined value ft.
In the above range, a significant difference is confirmed between the genuine bill C1 and the folded bill C3. The predetermined value ft varies depending on the bill conveyance conditions and types. When targeting dollar bills, “ft = 15
0 Hz · s / m ”. The diameter of the movable roller 72 is 1
0 mm, the force F that presses the movable roller 72 against the bill is 2 to
It is a result when 3 N and the conveyance speed of a bill are 1.6 m / s. At the frequency before normalization, 240Hz or more (= 1
This means that a significant difference was obtained at a high frequency of 50 × 1.6). When converted to the wavelength of the uneven waveform, it means that a significant difference was obtained at a short wavelength of about 7 mm or less.

A predetermined value f is provided between the wrinkle bill C2 and the genuine bill C1.
A significant difference is confirmed in the range of 1 or more. This value f1 is higher than the predetermined value f for which a significant difference was confirmed with the folded sheet C3. When targeting dollar bills, “f1 = 1kHz · s /
m ”. 1.6 kHz at the frequency before normalization
This means that a significant difference was obtained at a high frequency equal to or higher than z (= 1 k × 1.6). When converted to the wavelength of the uneven waveform, it means that a significant difference was obtained at a short wavelength of about 1 mm or less. That is, in such a frequency band, it is possible to perform discrimination without being affected by wrinkles.

In order to discriminate between true and false with high accuracy, it is preferable to use the frequency region in which the difference between the genuine bill and the false bill is clearly shown. According to the experimental example of FIG. 5, the region B corresponds to this frequency region. In this experimental example, the normalized frequency is 1 to 10 KHz · s / m, and the frequency before normalization is 160
0 Hz to 16 KHz, 1 to 0.1 mm when converted to wavelength
Is the range.

FIG. 6 is an explanatory diagram showing the relationship between the cutoff frequency of the high pass filter 81 and the discrimination result. Dozens to several dozens for each of the nine cut-off frequencies shown in the figure
The experimental results of the discrimination using 0 test tickets and genuine tickets are shown. The discrimination was performed by the component in the band above the cutoff frequency. Genuine bills consist of uncirculated new bills and circulating bills, and the circulating bills generally include bills that are so damaged that they are treated as unfit bills. In the figure, the result of statistically evaluating the discrimination result between the folded sheet C3, the wrinkled sheet C2 and the genuine sheet is shown. "◎" means that it is substantially completely distinguishable, "○" means that it is almost distinguishable, that is, almost distinguishable, but sometimes an erroneous judgment occurs, and "△" means
It means to make mistakes quite often. In the experimental results, the folded sheet C3 was evaluated as “⊚” at a cutoff frequency of 150 Hz · s / m or higher. With wrinkle ticket C2,
The cut-off frequency was 1 KHz · s / m or more, and the evaluation was “⊚”. Also, the cutoff frequency is at least 3
0 KHz / s / m or more, 48 KHz at the frequency before normalization
When converted into the above range and wavelength, a rating of “◯” or higher was confirmed in the range of approximately 0.03 mm or less. This is a wrinkle ticket C
In the case of 2, it means that it is preferable to raise the cut-off frequency more than the folded sheet C3.

The frequency band extracted by the high-pass filter 81 described above is preferably set in consideration of the frequency at which a true / false difference appears in the bill to be discriminated. According to the experimental example of FIG. 5, it is preferable to set to extract a frequency of 480 Hz or higher in a device that conveys bills at 1.6 m / s. Furthermore, 1600Hz
It is more preferable to set to extract the above because the influence of wrinkles can be suppressed. In such a range, in the region where a significant difference between true and false appears in FIG. 5, the influence due to wrinkles hardly appears, and the influence due to the unevenness of the ink becomes dominant.

C. True / False Discrimination Processing: FIG. 7 is a flowchart of the true / false discrimination processing. This is a process executed by the CPU of the control unit 100. This processing is started when the timing sensor 40 detects that the bill has been conveyed into the apparatus.

When the processing is started, the CPU inputs the concave / convex pattern of the bill to be discriminated (hereinafter referred to as the target bill) and the image data (step S10). As described above, the concavo-convex pattern is data obtained through the displacement meter 80, the high pass filter 81, and the A / D converter 90, and the image data is the image sensor 30.
The data obtained through. Next, the image data is processed to determine the type of banknote and the conveyance attitude (step S11).

Next, the CPU calculates the evaluation parameter according to the following procedure based on the uneven pattern. First, in FIG.
As described above, the predetermined period dt for starting and ending the detection
Data of the calculation section is extracted by excluding 1 and dt2 (step S12). About the data of this calculation section,
The absolute value of the output is calculated (step S13), and the time integrated value is calculated (step S14). The time integration value becomes an evaluation parameter for the target bill.

The CPU compares the evaluation parameter thus obtained with the reference value to make a true / false discrimination (step S15). As described above, the reference value is acquired from the reference value storage unit 106 according to the type of banknote and the conveyance attitude.

If the difference between the evaluation parameter and the reference value is within the permissible range (step S16), the discrimination result is a genuine note (step S17). If it exceeds the allowable range (step S16), the discrimination result is a fake bill (step S18). The allowable range can be appropriately set in consideration of noise included in the calculation of the evaluation parameter and the reference value so that sufficient discrimination accuracy can be obtained.

The CPU outputs the result obtained above (step S19) and completes the authenticity discrimination processing. As described above, the output destination of the result can be various modes depending on how the bill validator is used. Further, the output format of the result can be various modes. It may be output in the form of flag data, a message, or the like indicating the authenticity, or may be output in the form of a control signal for displaying the authenticity discrimination result.

According to the bill discriminating apparatus described above, it is possible to discriminate the authenticity of bills with high accuracy based on the uneven pattern of ink. In the bill discriminating apparatus of the present embodiment, not only the presence or absence of irregularities of a bill is detected, but also the concavo-convex pattern is analyzed to discriminate between true and false. Therefore, the influence of wrinkles can be avoided and discrimination can be performed accurately.

D. Modified Example of First Embodiment: FIG. 8 is an explanatory view showing the principle of the authenticity discrimination method as a modified example. The method of calculating the evaluation parameter is different from that of the example. The authenticity discrimination processing is the same as in the embodiment.

From the output of the displacement meter 80, the high pass filter 8
The process up to the step of extracting the high frequency component by 1 and obtaining the output absolute value is the same as in the embodiment (the graph in the third row from the top in the figure). In the embodiment, the absolute value of the output is directly integrated over time to obtain the evaluation parameter, whereas in the modified example, the time integration is obtained only for a portion exceeding the preset threshold Th. The absolute output values above the threshold Th are shown in the lower part of the figure. In the modification, the time integral value of this portion is used as an evaluation parameter.

Targeting a portion exceeding the threshold Th is equivalent to ignoring an output having a relatively small amplitude.
Therefore, by appropriately adjusting the threshold Th, it is possible to remove minute noise included in the output, such as minute vibration of the movable roller 72.

E. Second Example: In the first example and the modification, the case where the time integral value obtained from the concavo-convex pattern is used as the evaluation parameter is illustrated. As the evaluation parameter, various values that represent the features of the concavo-convex pattern can be used. In the second embodiment, as another example of the evaluation parameter, a case where the frequency distribution included in the uneven pattern is used will be illustrated.

FIG. 9 is an explanatory view showing the principle of authenticity discrimination in the second embodiment. For convenience of explanation, FIG.
It is shown that a part of the output after passing through the filter shown in Fig. 2 is expanded in both the time direction and the amplitude direction. The absolute value of this output is shown in the middle row. The processing up to this point is the same as that of the first embodiment.

In the second embodiment, this output absolute value is set to the threshold value T.
Binarize with h2. The binarized output is shown in the lower part of the figure. The binarization is a process in which a portion of the output absolute value that is equal to or greater than the threshold Th2 has a value of 1 and a portion that is smaller than the threshold Th2 has a value of 0. For example, since the range indicated by the two broken lines in the figure exceeds the threshold Th2, the value 1 is output.
By this processing, as shown in the lower part of the figure, pulses of various wavelengths are obtained.

The wavelength W of each pulse thus obtained has a value depending on the frequency of the output absolute value. A pulse having a long wavelength W corresponds to a low frequency because it is obtained when the output absolute value changes relatively slowly. Conversely, pulses with a short wavelength W correspond to high frequencies.

The wavelength of the obtained pulse differs depending on the threshold Th2. If the threshold value Th2 is set to a large value, pulses can be obtained only for outputs with large amplitude.
As a result, the output often corresponds to the low frequency component. On the contrary, if the threshold value Th2 is set small, the output often corresponds to the high frequency component. Threshold Th
It is preferable that 2 includes a frequency range required for authenticity discrimination and that the frequency distribution appears as a difference in pulse wavelength. As an example, the average value of the output absolute values can be used as the threshold Th2.

In the second embodiment, authenticity is discriminated based on the wavelength distribution of the pulse thus obtained. The wavelength distribution of the pulse can be evaluated by various methods. For example, the wavelength band with the largest number of pulses, the average value of the pulse wavelengths, and the standard deviation can be used as the evaluation parameters.
In this embodiment, the pulse is divided into three wavelength bands, a long wavelength band, a medium wavelength band, and a short wavelength band, and the number of pulses belonging to each of the categories is used as an evaluation parameter. The boundary value for dividing the wavelength band can be set arbitrarily.

FIG. 10 is a flow chart of the authenticity discrimination processing in the second embodiment. The calculation of the evaluation parameter is different from that of the first embodiment. The process up to the process of extracting the data of the calculation section from the concavo-convex pattern is the same as in the first embodiment (steps S10 to S12). In the second embodiment, the data in the calculation section is subjected to the binarization processing described in FIG. 9 (step S13).
A), the distribution of the pulse width is calculated (step S14)
A). As a result, three values of the number of pulses NL in the long wavelength band, the number of pulses NM in the medium wavelength band, and the number of pulses NS in the short wavelength band are obtained as evaluation parameters. The reference value is also expressed as a set of the above three values.

Similar to the first embodiment, the CPU compares the evaluation parameter thus obtained with the reference value, and discriminates based on the difference between the two (steps S15 to S19).

F. Modifications: The bill discriminating apparatus is not limited to the embodiments described above, and various modifications can be adopted.

For example, FIG. 11 is an explanatory view showing a modified example of the mechanism for detecting the uneven pattern of a bill. A mechanism that replaces the movable roller 72 in FIG. 2 has been illustrated. FIG. 11 (a)
Shows a modification of the contact type sensor. In this example, instead of the movable roller 72, a contact portion 72A for a bill is provided. The contact portion 72A is pressed against the surface of the banknote 10 by the leaf spring 76A supported by the fixing member 77A. The contact portion 72A slides on the surface of the banknote 10 and is displaced in the vertical direction in the figure due to the unevenness of the ink 11. The displacement meter 80A acquires the uneven pattern by detecting this displacement.

FIG. 11B shows an example provided with a non-contact sensor. The displacement meter 80B is a laser displacement meter and acquires the uneven pattern of the ink 11 in a non-contact manner. By using the non-contact sensor, there is an advantage that it is possible to acquire the uneven pattern with high accuracy. Further, when a non-contact sensor is used, it is not necessary to exclude a predetermined period of detection start and detection end as in the embodiment, so that a long calculation section can be secured and detection accuracy can be further improved. There is also an advantage.

In the embodiment, the case where the high-pass filter 81 extracts the high frequency component is illustrated. On the other hand, by analyzing the uneven pattern obtained from the displacement meter 80,
High frequency components may be extracted. Examples of the frequency analysis method include Fourier analysis. By performing Fourier analysis on the concavo-convex pattern obtained by A / D converting the output of the displacement meter 80, the frequency distribution (for example, the distribution shown in FIG. 5) included in the concavo-convex pattern can be acquired relatively easily. . The authenticity discrimination may be performed based on this distribution.

In the embodiment, the case where the evaluation parameter is calculated and the authenticity discrimination is performed is illustrated. It is not necessary to use a numerical value such as an evaluation parameter, but the concavo-convex pattern itself may be used for true / false discrimination. That is, the concavo-convex pattern obtained on the genuine note and the concavo-convex pattern obtained from the target bill may be directly superposed, and the difference between the two may be detected to discriminate the authenticity.

In the embodiment, the case where the reference value is selectively used according to the type of banknote and the conveyance attitude has been described, but a constant reference value may be used regardless of the conveyance attitude. Further, a common reference value may be used for different types of banknotes.

In the embodiment, the case where the concave-convex pattern is detected along the banknote transport direction has been illustrated. The concavo-convex pattern may be detected in a direction intersecting the transport direction. For example, a non-contact displacement gauge may be provided along the image sensor 30 to detect the concave / convex pattern in the direction orthogonal to the transport direction. Further, the uneven pattern on the surface of the bill may be two-dimensionally detected by detecting the uneven pattern in the direction intersecting the carrying direction while the bill is being carried. The two-dimensional detection can further improve the accuracy of the authenticity discrimination.

The discriminating apparatus of this embodiment can be used not only for bills but also for authenticating various paper sheets manufactured by intaglio printing.

[0076]

According to the bill discriminating apparatus of the present invention, the authenticity of bills can be discriminated with high accuracy based on the uneven pattern formed by the ink.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a configuration of a bill discriminating apparatus as an example.

FIG. 2 is a side sectional view of a bill validator.

FIG. 3 is an explanatory diagram showing a conveyance attitude of banknotes.

FIG. 4 is an explanatory diagram showing the principle of authenticity discrimination in the embodiment.

FIG. 5 is an explanatory diagram showing the difference between the concavo-convex patterns of a genuine note and a test note.

FIG. 6 is an explanatory diagram showing a relationship between a cutoff frequency of a high pass filter 81 and a discrimination result.

FIG. 7 is a flowchart of authenticity discrimination processing.

FIG. 8 is an explanatory diagram showing the principle of the authenticity discrimination method as a modified example.

FIG. 9 is an explanatory diagram showing the principle of authenticity discrimination in the second embodiment.

FIG. 10 is a flowchart of authenticity discrimination processing according to the second embodiment.

FIG. 11 is an explanatory view showing a modified example of the mechanism for detecting the uneven pattern of a bill.

[Explanation of symbols]

10 ... Banknote 11 ... Ink 30 ... Image sensor 40 ... Timing sensor 61 ... Rotation axis 62 ... Laura 63 ... scraper 71 ... Rotating axis 72 ... Movable roller 72A ... Contact part 73 ... scraper 74 ... Rotary axis 75 ... Arm 76 ... Spring 76A ... Leaf spring 77A ... Fixing member 80, 80A, 80B ... Displacement meter 81 ... High-pass filter 82 ... A / D converter 100 ... Control unit 101 ... Concavo-convex pattern acquisition unit 102 ... Evaluation parameter calculation unit 103 ... Authenticity discrimination section 104 ... Output unit 105 ... Image processing unit 106 ... Reference value storage unit

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Muneharu Nakabayashi             1 Ikegami Stock Market, Haruoka Town, Owariasahi City, Aichi Prefecture             Hitachi, Ltd. Information Equipment Division (72) Inventor Mitsunari Kano             3-2-16 Omorikita, Ota-ku, Tokyo Stocks             Company Hitachi System and Service F-term (reference) 3E041 AA03 BA11 BB03 CA02 DB01                       EA03

Claims (10)

[Claims] 1. A genuine / counterfeit discrimination apparatus for discriminating the authenticity of a paper sheet, comprising: a conveyance section for conveying the paper sheet; a pattern detection section for detecting a concavo-convex pattern generated on the surface of the paper sheet by printing. A genuineness / counterfeit discriminating apparatus comprising: a discrimination unit for discriminating the genuineness of the paper sheet by an analysis mainly using a short wavelength component in which the wavelength of the irregularities in the concavo-convex pattern is equal to or less than a predetermined value. 2. The authenticity discriminating apparatus according to claim 1, further comprising: a calculation unit that calculates a predetermined evaluation parameter that is quantitatively determined based on the uneven pattern, and the discrimination unit includes the calculated evaluation. An authenticity discriminating apparatus for discriminating authenticity based on whether or not the value of a parameter is within a range allowed for a genuine note. 3. The authenticity discrimination device according to claim 2, wherein the evaluation parameter is a parameter that quantitatively represents the wavelength distribution of the concavo-convex pattern. 4. The authenticity discrimination device according to claim 2, wherein the evaluation parameter is an integral of an output absolute value of the concavo-convex pattern. 5. The authenticity discrimination device according to claim 1, wherein the pattern detection unit includes a contact unit that contacts the surface of the paper sheet and mechanically detects the uneven pattern. 6. The authenticity discriminating apparatus according to claim 5, wherein the discriminating unit performs the analysis except for at least one of a predetermined period from the start of detection of the uneven pattern and a predetermined period before the end of detection. A true / false discrimination device. 7. The authenticity discriminating apparatus according to claim 1, wherein a plurality of analysis results of the concavo-convex pattern to be obtained for the genuine bill are stored in association with at least one of the type of the paper sheet and the conveyance posture. An authenticity discrimination device comprising a holding unit, wherein the discrimination unit performs the discrimination by properly using the analysis result in accordance with the type and conveyance posture of the paper sheet. 8. A true / false discrimination method for discriminating the authenticity of a paper sheet, which comprises a step of conveying the paper sheet, a step of detecting a concavo-convex pattern generated on the surface of the paper sheet by printing, A method of discriminating the authenticity of the paper sheet by analysis. 9. A method for discriminating the authenticity of a paper sheet, comprising: (a) a step of conveying the paper sheet; and (b) an irregularity representing a state of irregularities generated by printing on the surface of the paper sheet. A step of detecting a pattern at the time of carrying, (c) a step of extracting a high-frequency component in which the frequency of the unevenness is a predetermined value or more according to the carrying speed from the uneven pattern, and (d) the high frequency as an evaluation parameter Calculating the time integral of the output absolute value of the concavo-convex pattern of the component, and (e) discriminating the authenticity of the paper sheet based on whether or not the value of the evaluation parameter is within the range allowed for a genuine note. A method for discriminating authenticity, which comprises the step of: 10. The authenticity discrimination method according to claim 9, wherein the step (c) extracts a high-frequency component in which the wavelength of the irregularities is 1 mm or less.