JP2001291137A - Device for recognizing coin and medal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a coin and medal recognizing device which has a simple configuration and is capable of performing fast processing and is excellent in discrimination performance. SOLUTION: This coin and medal recognizing device 1 is provided with an image sensor 5 which picks up the images of the front and rear surfaces of a coin or a medal 2 based on a light flux LS and outputs image data DS as variable density data of each pixel, a preprocessing means 6 for mainly performing shading correction of irregular lighting of the image data DS, a feature extracting means 7 for fetching a plurality of pieces of gradation data corresponding to the variable density change of a stamped pattern for one cycle on a plurality of concentric circles from the center of the coin or the medal 2 as the cycle data of a time domain on the basis of image data DG and converting the gradation data into the one-dimensional spectrum of a frequency domain, and a discriminating means 8 for comparing the spectrum data SH of the coin of a discrimination object with reference spectrum data SK to discriminate the correctness of the coin of the discrimination object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coin and medal recognition device for discriminating pseudo coins and pseudo medals, and more particularly to recognition of the engraved pattern of coins and medals corresponding to one cycle of shading on a concentric circle. The present invention relates to a coin and medal recognition device which is executed by detecting a one-dimensional spectrum to identify a pseudo coin or a pseudo medal.

[0002]

2. Description of the Related Art Recently, a coin similar to a 500 yen coin,
There are many crimes of stealing vending machine change using foreign coins with a lower monetary value than 0 yen coins, and a huge social problem in Japan where a huge amount of vending machines are installed unprecedented in other countries It is becoming. Such crimes against vending machines are in a situation where it is difficult to establish a monitoring system due to the inherent characteristics of vending machines, which are located in a wide variety of places and have an unspecified number of users. Therefore, as crimes against vending machines increase, not only will the installer suffer enormous financial damage, but also the provision of the 500-yen coin as a provisional measure for the installer will cause inconvenience to users. The service is degraded. There are various types of foreign coins used as pseudo coins.However, in cases where abuse has been confirmed in large quantities, use a tool such as a drill or reamer to cut a part of the front or back of the coin into a conical shape. There is a Korean coin, 500 Won, whose size is equivalent to a 500 yen coin. Also,
It is expected that the use of pseudo medals will increase as the value of medals increases in game centers and amusement arcades where games are played using medals. Conventional coin and medal recognition devices for identifying pseudo coins and medals detect eddy currents generated when pseudo coins and pseudo medals pass through an electromagnetic field, and based on the detected eddy current, A device that detects the material, weight, and outer diameter of a pseudo medal is known. In addition, conventional coin and medal recognition devices include:
2. Description of the Related Art There is known a fake coin or a medal which detects a counterfeit product by obtaining a waveform corresponding to unevenness formed by a drill or a reamer using a magnetic field of a relatively high frequency or a photosensor on the surface of the pseudo coin. Furthermore, the conventional coin and medal recognition device takes a two-dimensional image of the imprinted image of a pseudo coin or a pseudo medal, and matches a large number of the taken digital data with a reference pattern of a normal coin or medal by template matching. A device for performing collation is known.

[0003]

A conventional coin and medal recognition device for detecting the material, weight, and outer diameter of a pseudo coin or a pseudo medal based on an eddy current is disclosed in Japanese Patent No. There is a problem that there is no discriminating ability for foreign falsified coins that are very similar to medals. In addition, even if a foreign forged coin is not used, there is no discriminating ability for a forged one whose material, weight, and outer diameter are equivalent to those of a Japanese coin or medal. In addition, conventional coin and medal recognition devices that detect authenticity by obtaining a waveform corresponding to the irregularities on the coin surface using a relatively high-frequency magnetic field or a photo sensor, for example. There is a problem that is not suitable for detecting pseudo coins and pseudo medals processed on flat surfaces. Further, in a conventional coin and medal recognition device that captures an engraved image two-dimensionally and performs matching by template matching, coins and medals are circular, and coins and medals are not necessarily oriented in a certain direction at the time of imaging. Therefore, in order to perform template matching, it is necessary to execute matching verification while rotating a captured image or a reference pattern, and there is a problem that the processing time for this becomes extremely long. The present invention has been made to solve such a problem, and an object of the present invention is to provide a coin and medal recognition device capable of accurately determining the authenticity of coins and medals with a simple configuration and a short processing time. It is in.

[0004]

In order to solve the above-mentioned problems, a coin and medal recognition device according to the present invention is based on image data from an image pickup means, and the shading of a stamp pattern on a plurality of concentric circles from the center of the coin or medal. Changes of 1
A feature extracting means for fetching the period data in the time domain of the circumference and converting the periodic data into a one-dimensional spectrum in the frequency domain; and a one-dimensional reference in the frequency domain for one circumference corresponding to the reference coin or medal in advance. Storage means for storing the spectrum; and identification means for comparing the one-dimensional spectrum from the feature extraction means with the one-dimensional reference spectrum from the storage means to identify the correctness of the coin or medal to be recognized. I do. The feature extracting means of the coin and medal recognition device according to the present invention, for image data of an engraved pattern of a coin or a medal to be identified, changes the shading of the engraved pattern on a plurality of concentric circles in a time domain of one circle. Since it is taken in as the cycle data of
Even if you start taking data from any position on one circumference,
The data for the circumference can be processed as a periodic function in the time domain. Subsequently, since a one-dimensional Fourier transform is performed on the periodic function in the time domain for one circle and stored as a one-dimensional spectrum in the frequency domain consisting of the fundamental component and the harmonic component, the shading of the engraved pattern on the concentric circles Irrespective of where the data corresponding to the change is taken from anywhere on one circle, the same spectrum can always be obtained. Since a one-dimensional Fourier transform is performed on the periodic function in the time domain for one circle,
The data amount can be reduced, and the data calculation processing time can be shortened. The storage means stores the one-dimensional reference spectrum of the frequency domain corresponding to one circumference of the reference coin or the medal in advance using the feature extraction means. The identification means corresponds to the one-dimensional reference spectrum of the frequency domain corresponding to one circumference corresponding to the reference coin or medal stored in the storage means, and the coin or medal to be identified supplied from the feature extraction means. By comparing the one-dimensional spectrum in the frequency domain for one circumference, the validity of the coin or medal to be identified can be identified.

[0005] The feature extracting means of the coin and medal recognition device according to the present invention may be arranged such that the periodic data of the time domain corresponding to one circle are respectively provided at a predetermined time interval and a plurality of times corresponding to the change in shading of the engraved pattern. It is characterized by comprising sampling means for converting sampling data into gradation data, and Fourier transform means for performing high-speed Fourier transform on the sampling data. The characteristic extracting means includes sampling means for converting the periodic data of the time region corresponding to one circle into sampling data at predetermined time intervals and at a plurality of gradations corresponding to the change in shading of the engraved pattern; And a Fourier transform means for performing a high-speed Fourier transform, so that regardless of the length of the circumference of the concentric circle, the cycle data in the time domain is the same, and detailed information corresponding to the change in shading of the engraved pattern Sampling data is obtained, and a high-speed Fourier transform is performed on the sampling data, so that a feature of a stamp pattern of a coin or a medal can be specified as a fundamental component or a harmonic component of a spectrum in a frequency domain. Further, the feature extracting means according to the present invention provides a predetermined allowable range for each of the plurality of concentric circles, and sets an average value of a change in shading within a predetermined allowable range from the center of the concentric circles as periodic data of a time domain for one circumference. It is characterized in that it takes in and converts it into a one-dimensional spectrum in the frequency domain. The feature extracting means according to the present invention provides a plurality of concentric circles each having a predetermined allowable range, and captures an average value of a change in shading within a predetermined allowable range from the center of the concentric circles as period data of a time region for one circle. Since the data is converted into a one-dimensional spectrum in the frequency domain, even if there is an error in determining the center of the image data of the coin or medal engraving pattern, the center error is absorbed to make the feature of the coin or medal engraving pattern one-dimensional in the frequency domain. It can be represented by a fundamental component or a harmonic component of the spectrum.

[0006]

Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a schematic block diagram of a coin and medal recognition device according to the present invention. In FIG. 1, a coin and medal recognition device 1 includes illuminations 3A and 3B for irradiating light LI from a diagonally downward direction on the front or back of a coin or medal 2 to be recognized, and reflection of light LI reflected on the coin or medal 2. Light beam LS converging light LR
A lens 4, an image sensor 5 for imaging the front or back of the coin or medal 2 based on the light beam LS, and outputting image data DS as grayscale data for each pixel, a preprocessing unit 6, a feature extracting unit 7, Means 8 and storage means 9. The pre-processing means 6 has a shading correction function, a calculation function, and a memory, and after performing shading correction of mainly illumination unevenness of the image data DS supplied from the image sensor 5, the preprocessing means 6 removes the coin or the medal 2 from the image data DS. The coordinates corresponding to the diameter and the center are calculated and stored as new image data DG. The feature extracting means 7 includes a pre-processing means 6
A plurality of gradation data corresponding to a change in the shading of the engraved pattern for one cycle on a plurality of concentric circles from the center of the coin or the medal 2 is taken in as the cycle data of the time domain based on the image data DG stored in. Further, the feature extracting means 7 performs a fast Fourier transform (for example, FFT) on the periodic data, converts it into a one-dimensional spectrum in the frequency domain, stores it, and supplies the one-dimensional spectrum data SH to the identifying means 8. FIG. 2 is a block diagram of a main part of the feature extracting means according to the present invention. In FIG. 2, the feature extraction means 7
The apparatus includes a sampling unit 10, a sampling data storage unit 11, and a Fourier transform unit 12. Based on the image data DG stored in the preprocessing means 6, the sampling means 10 converts the periodic data of the time domain of one circle on a plurality of concentric circles from the data corresponding to the center of the coin or the medal 2 for a predetermined time. Intervals (for example, one circle is equally spaced 51
A plurality of gray scales (for example, 256 gray scales) corresponding to changes in shading of the engraved pattern with two data at a cycle of 10 ms)
Is used as sampling data, and sampling data S
D is supplied to the sampling data storage means 11.

FIG. 3 is an image diagram of acquiring sampling data for one circumference on a concentric circle according to the present invention. FIG.
In this case, 512 pieces of sampling data corresponding to one circumference of each of the radius r1 and the radius r2 from the center O of the circle are sequentially taken from an arbitrary position 0 on the circumference in the order of, for example, 10 ms at a period of 10 ms. The 512 pieces of sampling data for one circumference have a value corresponding to the change in shading of the engraved pattern having a plurality of (for example, 256) gradations. With the data of 256 gradations, it is possible to accurately reproduce the change in shading of the engraved pattern of the coin or the medal 2. in this way,
Since the sampling data for one circumference of 512 cycles of 10 ms and 256 gradations is a periodic function in the time domain,
The same periodic function is obtained regardless of the data taken from any position on the same circumference. Alternatively, the sampling means 10 may provide a predetermined allowable range for each of the plurality of concentric circles, and take in the average value of the change in shading within the predetermined allowable range from the center of the concentric circles as period data of a time region for one circumference. FIG.
FIG. 3 is an image diagram of acquiring sampling data for one circumference on a concentric circle in which an allowable range is set according to the present invention. In FIG. 4, two circumferences (a radius r1 −Δr and a circumference r1 + Δr) in which a permissible range of ± Δr is set on the circumference of the radius r1.
And the intersection Y between an arbitrary position X on the circumference of the radius r1 and the straight line from the position X to the center O of the circle and the circumference of the radius r1−Δr, and the straight line from the position X to the center O of the circle The average value of the three data points at the intersection Z between the extension line and the circumference of the radius r1 + Δr is used as sampling data for starting the data acquisition of the radius r1. Similarly, with respect to the three circles, the 512 average values of the 1st,..., 511th are taken in as sampling data at a period of 10 ms. in this way,
A predetermined allowable range is provided for each of the plurality of concentric circles, and an average value of a change in shading within a predetermined allowable range from the center of the concentric circle is taken in as a period data of a time region for one circumference, and image data of a stamp pattern of a coin or medal is taken. Even if there is an error in the determination of the center, the center error can be absorbed. FIG.
In this example, the average value of the data on a total of three circumferences, that is, the circumference of the radius r1 and the two circumferences within the allowable range of ± Δr, is used.
The average value of data on the circumference of the radius r1 and a plurality of circumferences within the allowable range of ± Δr may be used. FIG. 5 is an image diagram in which two concentric circles are set on the back surface of a 500 yen coin.
In FIG. 5, the circumference (L1) of 30% of the radius of the coin from the center of the back side of the 500 yen coin (500 display is the front side)
And the circumference (L2) of 50% of the radius.

FIG. 6 shows that the circumference of 30% of the radius of FIG.
% Sampling data in a time domain for one cycle with an allowable range of%. In FIG. 6, the horizontal axis represents sampling numbers 1 to 512 (for example, a period of 10 ms), and the vertical axis represents sampling data (for example, a value corresponding to a change in density of 256 gradations: a density value). The average value of the sampling numbers 1 to 512 for one cycle of 30% of the radius and for one cycle having an allowable range of ± 10% is obtained by sampling the density for one cycle of 30% of the radius. Data and
The average value (DC component) of all 512 samples is removed from each sampled data.
The sampling data in FIG. 6 contains many one-cycle fundamental wave components (indicated by a broken line: 100 Hz in the case of a cycle of 10 ms). This means that about half of the concentric circle with a radius of 30% on the back of the 500 yen coin in Fig. 5 has a high density of bright pixels, and the other half has a high density of flat dark pixels of "Kiriha" with engraved pattern. by. The sampling data storage means 11 is composed of a memory such as a rewritable RAM or EEPROM, stores the sampling data SD supplied from the sampling means 10, and provides it to the Fourier transform means 12 as sampling data DM. The Fourier transform means 12 is constituted by an FFT (Fast Fourier Transformer), converts the sampling data DM supplied from the sampling data storage means 11 into a one-dimensional spectrum in the frequency domain, and converts the spectrum data SH into identification means. 7 Since the Fourier transform converts a periodic function in the time domain into a fundamental component and a harmonic component in the frequency domain, the sampling data SD for one cycle taken by the sampling means 10 can be started from anywhere on the circumference. The spectrum data SH is the same.
As described above, the Fourier transform is performed on the sampling data SD in the time domain to obtain the spectrum data SH in the frequency domain, so that the sampling data SD in the time domain is obtained.
Even if the spectrum data S is started from any position on the target circle without determining the start position of
H is always the same, so that it is possible to eliminate the inconvenience, increase in software and processing time required for determining the start position of the sampling data SD. Further, Fourier transform is performed on the sampling data SD in the time domain on one circle to obtain the spectrum data S in the one-dimensional frequency domain.
Since H is set, the amount of data to be handled is reduced, the software scale is reduced, and the processing time is shortened.

FIG. 7 is a frequency spectrum characteristic diagram obtained by performing a Fourier transform on the data of FIG. In FIG. 7, the horizontal axis represents the frequency of the fundamental component (order 1) to the harmonic component (order 41), and the horizontal axis represents the amplitude value (spectral value) corresponding to the density change of the engraved pattern. The data cycle of FIG.
If 0 ms, the frequency of the fundamental wave component (order 1) is 1
00 Hz, and the frequency of the harmonic component order 41 is 41
00 Hz. FIG. 7 shows, for comparison, the frequency spectrum characteristics of the front surface of a Greek 10-dragma coin and the front and back surfaces of a Korean 500-won coin. FIG.
In the frequency spectrum characteristic diagram of FIG. 5, the amplitude value corresponding to the back surface of the 500 yen coin is such that the fundamental component is significantly larger than the fundamental component of the comparative coin. Although not shown,
The fundamental wave component of the back of a 500-yen coin (a concentric circle with a 30% radius) shows an extremely large value compared to other foreign coins or forged foreign coins, and is a characteristic of the back of a 500-yen coin. ing. Fig. 8 shows the radius of 50 on the back of the 500 yen coin.
It is a frequency spectrum characteristic figure with respect to the circumference of% ± 10%. The horizontal axis and the vertical axis, and the types of foreign coins to be compared are the same as those in FIG. In FIG. 7, in the amplitude value corresponding to the back surface of the 500 yen coin, the harmonic component of order 6 is significantly larger than the harmonic component of order 6 of the foreign coin to be compared.
Therefore, the characteristic of the back surface of the 500 yen coin also appears remarkably in the frequency spectrum of a concentric circle having a radius of 50%. In this way, the reference coin (for example, a regular 500
Frequency spectra on a plurality of concentric circles, which are characteristic features of a yen coin, a 100 yen coin, etc., are stored in advance, and the correctness is easily identified by comparing with the frequency spectrum of a foreign or falsified coin to be recognized. be able to. The discriminating means 8 has a comparison function, and includes spectrum data SH of a coin to be discriminated supplied from the feature extracting means 7;
A reference coin (for example, regular 5
00 yen) and the reference spectrum data SK.
When the spectrum data SH matches the reference spectrum data SK, the coin to be identified is determined to be normal, and when the spectrum data SH does not match the reference spectrum data SK, the coin to be identified is determined to be abnormal. I do. The discriminating means 8 includes a spectrum data SH and a reference spectrum data S for a certain concentric circle.
If K is identical, the spectrum data SH of the other concentric circles is compared with the reference spectrum data SK. If the comparisons of a plurality of preset concentric circles all match, the target coin is May be determined to match the reference coin. Storage means 9
Is configured by a memory such as a ROM or a RAM, stores a characteristic frequency spectrum on a concentric circle of a reference coin in advance as reference spectrum data SK, and supplies the reference spectrum data SK to the identification means 8. It should be noted that the reference spectrum data SK of the reference coin is obtained as a frequency spectrum in which a coin or a medal 2 is used as a reference coin in FIG.
This spectrum data SH is stored as reference spectrum data.

Next, the operation of identifying coins or medals will be described with reference to an operation flowchart. FIG. 9 is an operation flowchart of the coin and medal recognition device according to the present invention. In step S1, the image sensor 5 captures an image of the front or back of a coin or medal. In step S2, shading correction for illumination unevenness is mainly performed.
In step 3, the diameter and center of the coin or medal are calculated from the image data. Steps S2 and S3 are executed by the preprocessing means 6. In step S4, the concentric circle 1
The shading data of the engraved pattern for the period is converted into a periodic function in the time domain. In step S5, the periodic function in the time domain is one-dimensionally Fourier-transformed and stored as a frequency spectrum in the frequency domain. Step S4 and step S4
Step 5 is executed by the feature extracting unit 7. In step S6, the frequency spectrum of step S5 is compared with the reference frequency spectrum. If the frequency spectrum and the reference frequency spectrum match, the comparison of the frequency spectrum with the reference frequency spectrum is repeated for other concentric circles, and the predetermined number of times is repeated. If the comparison results in a match, the flow shifts to step S7 to determine that the coin or medal to be identified is normal. On the other hand, if the frequency spectrum does not match the reference frequency spectrum in step S6, the process proceeds to step S8, and the coin or medal to be identified is determined to be abnormal. Steps S6 to S8 are executed by the identification unit 8.

[0011]

As described above, the coin and medal recognition device according to the present invention performs a one-dimensional Fourier transform on a periodic function in the time domain of one circle, so that the data amount can be reduced and the data amount can be reduced. The calculation processing time can also be shortened. In addition, detailed sampling data corresponding to the change in shading of the engraved pattern is obtained, and a high-speed Fourier transform is performed on the sampled data to apply the characteristic of the engraved pattern of the coin or the medal to the fundamental wave component or the spectrum of the frequency domain spectrum. Since it can be specified as a harmonic component, it is possible to start sampling data at an arbitrary position on a concentric circle, thereby eliminating the troublesome positioning and shortening the processing time. Further, since the sampling data is formed with a plurality of gradations corresponding to the change in the density of the engraved pattern, the change in the density of the engraved pattern can be accurately reproduced. In addition, a predetermined allowable range is provided for each of the plurality of concentric circles, and an average value of a change in shading within a predetermined allowable range from the center of the concentric circles is taken as period data in a time domain for one circumference, and is included in a one-dimensional spectrum in a frequency domain. Since the conversion is performed, even if there is an error in determining the center of the image data of the coin or medal stamp pattern, the center error can be absorbed. Therefore, it is possible to provide a coin and medal recognition device having a simple configuration and capable of performing high-speed processing and having excellent discrimination performance.

[Brief description of the drawings]

FIG. 1 is a schematic block configuration diagram of a coin and medal recognition device according to the present invention.

FIG. 2 is a block diagram of a main part of a feature extracting unit according to the present invention;

FIG. 3 is an image diagram of acquiring sampling data for one circumference on a concentric circle according to the present invention;

FIG. 4 is an image diagram of acquiring sampling data for one circumference on a concentric circle in which an allowable range is set according to the present invention;

FIG. 5 is an image diagram in which two concentric circles are set on the back of a 500 yen coin

6 is sampling data in a time domain for one cycle in which a permissible range of ± 10% is provided on a circumference of 30% of the radius of FIG. 5;

7 is a frequency spectrum characteristic diagram obtained by performing a Fourier transform on the data of FIG. 6;

FIG. 8 is a frequency spectrum characteristic diagram for a circumference of 50% ± 10% of the radius of the back side of a 500 yen coin.

FIG. 9 is an operation flowchart of the coin and medal recognition device according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Coin and medal recognition device, 2 ... Coin or medal, 3A, 3B ... Lighting, 4 ... Lens, 5 ... Image sensor, 6 ... Pre-processing means, 7 ... Feature extraction means, 8 ... Identification means, 9 ... Storage means Reference numeral 10: sampling means, 11: sampling data storage means, 12: Fourier transform means.

Claims (3)

[Claims] 1. A coin comprising: illumination means for irradiating light to the front or back surface of a coin or medal to be recognized; and imaging means for detecting reflected light from said coin or medal to image an engraved pattern. In the medal recognition device, based on the image data from the imaging unit, changes in shading of the engraved pattern on a plurality of concentric circles from the center of the coin or medal are respectively taken as period data of a time region for one circle, and A feature extraction unit for converting the periodic data into a one-dimensional spectrum in the frequency domain, a storage unit for storing a one-dimensional reference spectrum in the frequency domain corresponding to one circumference of the reference coin or the medal in advance, and a feature extraction unit. Of the coin or medal to be recognized by comparing the one-dimensional spectrum of Coins and medals recognition apparatus characterized by comprising a identification means for identifying whether. 2. The feature extracting means converts the periodic data in the time domain of one circle into sampling data at predetermined time intervals and at a plurality of gradations corresponding to the change in shading of the engraved pattern. 2. The coin and medal recognition device according to claim 1, further comprising sampling means, and Fourier transform means for performing a high-speed Fourier transform on the sampled data. 3. The feature extracting means sets a predetermined permissible range in each of a plurality of concentric circles, and calculates an average value of a change in shading within a predetermined permissible range from the center of the concentric circles as periodic data of a time domain for one circumference. 3. The coin and medal recognition device according to claim 1, wherein the coin and the medal are taken in and converted into a one-dimensional spectrum in a frequency domain.