JP4832054B2 - Identification device - Google Patents

Description

  The present invention relates to a method and apparatus for reading the surface information of securities and other articles and determining their authenticity.

  Purchases of goods using electronic money and credit cards are becoming popular, but the purchase of goods using banknotes is still active. This is clear from the increase in the number of banknotes issued. On the other hand, in recent years, the number of unauthorized access to an automatic cash handling machine, for example, the use of counterfeit bills has also increased dramatically. And the method has become more sophisticated. For example, various types of counterfeit tickets and counterfeit bills are counterfeited and put into an automatic cash handling machine, and the counterfeit tickets and counterfeit bills received are examined. On the next occasion, there are problems such as taking counterfeit tickets of the accepted type, counterfeiting a large number of counterfeit bills and putting them into the automatic cash handling machine again to take money.

  In order to prevent such crimes, a method for enhancing the reliability of the automatic cash handling machine is regarded as important. For example, there is a method of trying to prevent forgery of a counterfeit ticket by moving a sensor when the banknote passes to change a pattern for capturing a bill to be used for discrimination (see, for example, Patent Document 1). The purpose of this technology is to prevent counterfeiters from forging only the information of a specific part of the banknote and deceiving the device. In addition, there is a technique for changing encrypted information in a timely manner when the encrypted information is stored in a ticket and the automatic ticket checker performs reading corresponding to the information (see, for example, Patent Document 2).

JP-A-8-249512

JP-A-8-305904

The method described above is characterized by (1) changing the information reading position of the object and (2) writing the encrypted information in the object at a suitable time (random time). However, it is necessary to consider the following points in adapting the authenticity of highly forged banknotes. (1) Most banknotes that are highly counterfeited are those in which the front of the banknote is imitated. For this reason, even if the information reading position of the object is changed, there is a high possibility that the information is forged. Moreover, in the case of the present banknote, since the storage device is not equipped, it is difficult to adaptively write the encryption information. Furthermore, changing the information reading position or changing the timing for changing the encrypted information at random has a low accuracy that can be generally determined.

  In order to solve such a problem, in the present invention, information on the entire paper sheet of securities such as banknotes is acquired by a sensor, and a plurality of prepared determination units having different characteristics are used to determine the respective information. Authenticity is determined. Then, one of a plurality of determination results is selected at a certain timing.

  According to the present invention, even if there is a weak part of a certain determination method, that is, a determination region that takes in a counterfeit note, the determination method is replaced at an appropriate timing, and the intention is to enter the determination region. Even if a large number of counterfeit bills are inserted, it can be rejected. Further, in the case of such a typical crime, it is considered that the criminal loses confidence in decoding the discrimination determination characteristic of the automatic cash handling machine he / she thinks, and as a result, the crime number is reduced.

  In the present invention, a method represented by a code that processes the physical information on a sensor such as a microcomputer and a sensor that takes various physical information from bills or the like is used as a specific form.

  Embodiments of the present invention will be described in detail with reference to the drawings. First, FIG. 3 is a diagram showing an overview of an automatic teller machine 31 which is an example to which the present invention is applied. A CPU 13 is provided in the casing 31. Reference numeral 311 is a deposit port for the user to deposit cash, 312 is a keyboard for inputting a personal identification number and amount necessary for deposit and withdrawal, and 313 and 314 are bank card slot and withdrawal port. Reference numeral 315 denotes a display device for displaying a message to the user.

  The basic principle of the present invention will be described with reference to FIG. In general, the counter's counterfeit bill capture rate (the rate at which it is mistakenly imported into the device as a genuine note) and the genuine note rejection rate (the rate at which a genuine note is erroneously rejected as being a counterfeit or other paper sheet) Is generally a small value but has a statistical value other than zero. FIG. 2 shows examples of some determination methods. 2 of FIG. 2 shows the true distribution of bills in the feature amount space obtained from the information on the entire paper sheet. However, since the true shape of this distribution is generally unknown, it is currently estimated using a large number of samples. Reference numeral 22 in FIG. 2 indicates a determination boundary estimated by a certain estimation method 1 by a dotted line. In general, since the bias and number of collected samples are finite, there are an area 221 where a counterfeit note is taken in and an area 222 where a genuine note is rejected. The criminal tries to make a fake ticket that enters this area 221, but it is quite difficult to use a general technique. For this reason, recently, there are many cases where a counterfeit method is used to search for a fake ticket in this area 221. That is, a variety of types of counterfeit bills are prepared in advance, and the types of counterfeit bills that are loaded and taken into the automatic cash handling machine at a time are examined. Then, a large number of counterfeit bills to be taken in next are prepared and put into an automatic cash handling machine. Usually, after that, the transaction is canceled and a genuine ticket is refunded for the number of counterfeit tickets. As a method of preventing such crimes or eliminating the power of the criminal, in the present invention, a plurality of determination units having different characteristics are prepared, and a result from the determination unit is selected at a certain appropriate timing. . Reference numeral 23 in FIG. 2 indicates a determination boundary estimated by a certain estimation method 2 by a dotted line. In this case as well, there are an area 231 where a counterfeit note is taken in and an area 232 where a genuine note is rejected. Further, 24 in FIG. 2 indicates a determination boundary estimated by a certain estimation method 3 by a dotted line. In this case as well, there are an area 241 where a counterfeit note is taken in and an area 242 where a genuine note is rejected. In general, the area where a counterfeit note is taken in and the area where a genuine note is rejected are usually different depending on the distribution estimation method. Therefore, even if the region 221 where the counterfeit ticket is taken in is recognized, the determination method (method for determining authenticity inside and outside the dotted line obtained by the estimation method) is changed at the timing of receiving the next insertion. It is possible to cope with the use of counterfeit bills as described above. This is apparent from the fact that the regions 221, 231, and 241 do not overlap. In the present application, an effect can be obtained by using a combination of determination methods having slightly different characteristics in authentic bill determination. Considering the typical crimes mentioned above, the timing for switching between multiple stored methods should be set at the beginning of the next transaction after a large number of rejections, or every transaction when the transaction is sparse, such as night work. Is effective.

  Next, the overall processing flow will be described with reference to a block diagram illustrating an example of the configuration of the authenticity determination apparatus in the present embodiment in FIG. 1 and the flowchart in FIG. 4. In the present invention, the banknote deposited from the deposit port 311 is conveyed to the discrimination unit, and the sensor unit 111 acquires physical information of the entire paper sheet such as the banknote at 411. Here, the physical information is, for example, color image information. Also, the entire banknote is a banknote not to be overlooked even if it is crafted only at a certain specific place without subjecting only a specific partial area set in advance to information acquisition. This means that almost all of the information is targeted for information acquisition. This is because only a specific counterfeit banknote can be handled as a matter of course when only a specific partial area is targeted. Thereafter, in the feature value generation 416, the feature value generation unit 116 generates a feature value from the acquired physical information. In the feature quantity generation unit 116, a feature quantity that combines physical quantities, for example, a feature quantity expressed by components such as principal component analysis is generated. These feature amounts are usually generated by software processing because complicated processing is required. Next, a plurality of determination units 112, 113, and 114 perform authenticity determinations 412, 413, and 414 using the feature amounts. These authenticity determinations are usually processed by software in a different manner while using feature quantities using the same physical quantities obtained from the sensor 1. Therefore, it is possible to easily execute complicated processing and to provide various processing means. Also, in the processing by hardware, problems such as failure occur due to physical burdens, and the reliability and processing speed are excellent. As described above, the present application is characterized in that a plurality of determination processes having different characteristics for authenticity determination are provided.

  Authenticity is determined based on whether or not the feature amount is included in the genuine note determination distributions 22, 23, and 24 indicated by the dotted lines of the determination methods 1, 2, and 3 shown in FIG. For example, the genuine note determination distributions 22, 23, and 24 are regions determined by learning with a layered neural network using each feature amount set obtained from a large number of genuine notes as an input. At this time, genuine note determination distributions 22, 23, and 24 are obtained by preparing and learning three neural networks having different numbers and structures of neurons. Alternatively, three different regions may be obtained by changing the initial value during learning with the same number of neurons and structure. Also, without using three neural networks, one using a neural network, one using a Mahalanobis distance from a genuine sample distribution, and one using the magnitude of the shortest Euclidean distance from a genuine sample. Different determination methods may be used as in the case of determination. As for such a determination method, for example, the University of Tokyo Press “Image Analysis Handbook” (published on January 17, 1991, supervised by Mikio Takagi and Yoshihisa Shimoda “Classification” page 641, “Neural Network” page 190) The existing method described in the above. Then, the selection unit 115 (415 in FIG. 4) selects one of a plurality of determination results based on the selection signal controlled by the CPU 13. Here, the number of determination methods is not limited to three, and may be two or more. It should be noted that any one of a plurality of determination units prepared may be selected based on the selection signal, and only the processing in the determination unit may be performed.

  The identified banknotes are processed based on the determination result. That is, in the case of a genuine note, it is counted as a bill inserted for deposit or remittance, stored in a collection box, the amount of money inserted is displayed on the display device 315, and deposit or remittance processing is performed. On the other hand, if it is determined that it is not a genuine note, it is generally rejected and returned to the deposit / withdrawal port 311 in FIG. When the number of rejects is large, a warning message notifying the abnormality of the device is displayed on the display device 315 and information is transmitted to the staff.

  In this embodiment, all three determination methods are performed and the result is selected. However, even if one determination method is selected and the result is used, the same result can be obtained. The selection signal opens and closes the gate 122 at the timing described as a rule in the selection timing generation unit controlled by the CPU 13. The rule can be, for example, when the next transaction starts when a large amount of rejects exceeds the number of processed sheets. As described in the background art, this utilizes a phenomenon in which a large number of rejects are first generated when various types of counterfeit bills that are not rejected are checked by inserting various types of counterfeit bills. Therefore, a reference value for the number of rejects within a predetermined time is provided, and the timing for switching is determined by comparison with the reference value. Further, it may be at the start of individual transactions when transactions at night are sparse. This is because counterfeit banknotes are often aimed at nighttime when there are few people. Alternatively, the change timing may be determined by a random number without providing a specific rule. The timing determination can be realized by putting reject information and transaction information on the control signal from the CPU 13.

  As a specific example of the rule, while executing one transaction, the determination result from the determination unit 11 is accumulated, and the number of sheets determined to be rejected in a predetermined time is counted. And when the value becomes more than a certain threshold value, the determination method of the next transaction is changed. Also, in the second rule, the time during which transactions are not performed is counted, and when the value exceeds a certain threshold, or the number of securities determined within a certain time, or transactions If the number becomes smaller than a certain threshold value, the determination method in the next transaction is changed. These rules can be implemented in the above procedure.

  On the other hand, the selection method signal generation 123 that generates a selection signal instructing which determination result to select does not select the same method result from one of the three determination results by using a control signal from the CPU 13 as a trigger. Randomly decide as follows, or select cyclically in order. Alternatively, the following method may be used. The number of times the determination is adopted is measured for each determination method. Then, the determination result with the smallest number of times is selected. By doing so, specific determination results are not selected unevenly.

  In this embodiment, three methods are prepared. However, if there are a plurality of methods, other methods may be used. Further, the determination unit 11 may prepare a plurality of sets after the feature amount generation unit. That is, a plurality of feature quantity generation units for generating different feature quantities from sensor physical quantities are prepared. Processing may be performed by a determination unit associated with each feature amount generation unit in advance, or each of the feature amount generation unit and the determination unit may be selected by a selection circuit, and a plurality of feature amounts may be determined by a plurality of determination methods. It is good also as judging authenticity. Then, one of the overall determination results is selected. With this configuration, it is possible to make a determination with different characteristics, and the effect is enhanced.

  In addition, the present invention can be widely applied without being limited to the implementation object, that is, the securities such as banknotes. As long as information on the surface and shape of the object can be acquired, it is possible to determine the authenticity of a brand product, for example, with the same configuration of the present invention.

  As described above, according to the present invention, a plurality of methods for performing authenticity determination using physical information of the entire object are prepared, and a determination result of one method is selected at an appropriate timing. By doing in this way, even if the weakness of a certain judgment method is known by the introduction of many kinds of fake, it will change to another judgment method at the next timing, so there is an effect of preventing crime in advance . This invention can be applied to automatic cash handling machines, vending machines, and other automatic machine identification units that handle securities such as banknotes, and contributes greatly to further development of the industry, crime prevention, etc. is there.

It is a figure which shows an example of a structure of the authenticity determination apparatus in this embodiment. It is a figure for demonstrating the principle and effect of the discrimination method of FIG. It is the figure which showed the general view of the cash dispenser in this embodiment. It is a figure which shows an example of the processing flow of the authenticity determination apparatus in this embodiment.

Explanation of symbols

11: determination unit, 12: determination timing generation unit, 13: CPU, 111: sensor group, 112: determination method 1, 113: determination method 2, 114: determination method 3, 115: selection unit, 116: feature quantity generation unit 121: Selection timing generation unit, 122: Gate, 123: Selection method signal generation, 21: Genuine bill distribution, 22: Genuine bill judgment area of judgment method 1, 23: Genuine bill judgment region of judgment method 2 Judgment ticket acceptance area of determination method 1, 222: Genuine ticket rejection area of determination method 1, 231: Fake ticket acceptance area of determination method 2, 232: Genuine ticket rejection area of determination method 2, 241: Fake ticket of determination method 3 Accept area, 242: Genuine bill reject area of determination method 3. 31: Overview of an automatic teller machine, 311: Deposit port, 312: Keyboard for inputting a personal identification number and amount, 313: Bank card slot, 314: Withdrawal port, 315: Display device. 411: Physical quantity acquisition, 416: Feature quantity generation, 412: Determination method 1, 413: Determination method 2, 414: Determination method 3, 415: Selection of determination result

Claims (6)

A deposit department that accepts deposits of securities;
A sensor unit that scans the deposited securities and obtains physical information from almost the entire securities; and
A feature value generation unit that generates a feature value from the physical information;
A plurality of determination units that perform authenticity determination using a plurality of different determination methods prepared in advance using the feature amount;
A selection timing generation unit that generates a timing signal for selecting one of the determination results by the plurality of determination units;
The selection timing generation unit generates the timing signal when the determination unit rejects the counterfeit for a predetermined number of times within a predetermined period, and generates a timing signal from a determination result according to a certain determination method. Switch to the judgment result,
The discrimination apparatus characterized in that the plurality of determination methods use the same feature amount but have different authenticity determination characteristics. A deposit department that accepts deposits of securities;
A sensor unit that scans the deposited securities and obtains physical information from almost the entire securities; and
A feature value generation unit that generates a feature value from the physical information;
A plurality of determination units that perform authenticity determination using a plurality of different determination methods prepared in advance using the feature amount;
A selection timing generation unit that generates a timing signal for selecting one of the determination results by the plurality of determination units;
The determination timing generation unit generates the timing signal when a time during which a transaction is not performed exceeds a predetermined value, and switches from a determination result by a certain determination method to a determination result by another determination method,
The discrimination apparatus characterized in that the plurality of determination methods use the same feature amount but have different authenticity determination characteristics. A deposit department that accepts deposits of securities;
A sensor unit that scans the deposited securities and obtains physical information from almost the entire securities; and
A feature value generation unit that generates a feature value from the physical information;
A plurality of determination units that perform authenticity determination using a plurality of different determination methods prepared in advance using the feature amount;
A selection timing generation unit that generates a timing signal for selecting one of the determination results by the plurality of determination units;
The determination timing generation unit generates the timing signal based on the number of past selections of the plurality of determination methods, switches from a determination result according to one determination method to a determination result according to another determination method,
The discrimination apparatus characterized in that the plurality of determination methods use the same feature amount but have different authenticity determination characteristics. The discrimination apparatus according to any one of claims 1 to 3, wherein the determination timing generation unit switches the determination result so as to select one of the plurality of determination units and use the result. . The identification device according to claim 1, wherein the feature quantity generation unit and the determination unit are configured to perform software processing. Having a plurality of the feature quantity generation units,
The discrimination apparatus according to claim 1, wherein the determination timing generation unit further selects and switches any one of the plurality of feature quantity generation units.

Images