JP2006293651A - Coin discriminating device, coin discriminating method, and storage medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the coin discriminating accuracy of a coin discriminating device which discriminates coins on the basis of a gray level histogram obtained from a coin image. <P>SOLUTION: Digital multi-leveled image data of the coin photographed by an imaging unit is incorporated. The coin image is divided into a plurality of section images in a first region 29, a second region 30, and a third region 31. The gray level histogram is prepared on the basis of each section image, and a feature quantity such as a peak position is extracted from each gray level histogram. Then, the authenticity of the coin is determined by collating the feature quantity with the feature quantity of a preregistered master coin. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a coin discriminating apparatus and a coin discriminating method for discriminating whether a coin is true or false based on an image. The present invention also relates to a storage medium that stores a processing process for executing the coin discrimination method.

  Built-in in game consoles, vending machines, ticket machines, etc., and discriminating the authenticity of inserted coins (currency, coins for game consoles, etc.) and other store coins (coin for game consoles used in other stores) For this purpose, a coin discriminating device is used. Some of these coin discriminating devices discriminate coins using a magnetic sensor. However, in such a coin discrimination device, the accuracy of coin discrimination is very low.

  Therefore, a coin discriminating apparatus has been proposed that converts a photographed coin image into digital multilevel image data, generates a density histogram from the digital multilevel image data, and discriminates coins based on the density histogram pattern. . In this coin discriminating apparatus, the authenticity of a coin is discriminated by comparing a density histogram pattern of a coin image with a master coin registration pattern registered in advance.

  However, since sophisticated coins such as 500 yen coins have become a social problem in recent years, such coin discriminating devices have not been sufficiently accurate in discriminating currencies. Further, in the case of a coin for a gaming machine that is a currency other than a currency, the pattern attached to the coin is different, but the size and pattern of the coin are often confusing. Therefore, in order to deal with such a game machine coin, a coin discrimination device with high coin discrimination accuracy is required.

  The present invention has been made in view of the technical problems as described above, and an object of the present invention is to provide a coin discriminating apparatus for discriminating coins based on a density histogram obtained from a coin image, and to improve the coin discriminating accuracy. It is to improve.

  A coin discriminating apparatus according to claim 1 of the present invention is a coin discriminating apparatus for discriminating the authenticity of a target coin based on an image, and a storage means for storing a feature value obtained from a density histogram of a master coin. Imaging means for acquiring an image of the target coin, histogram generation means for generating a density histogram from the image, and feature amount extraction for extracting a feature amount of the target coin from the density histogram generated by the histogram generation means And means for determining the authenticity of the target coin by comparing the feature quantity of the master coin stored in the storage means with the feature quantity of the target coin extracted by the feature quantity extraction means. Is.

  In the coin discriminating apparatus according to claim 1 of the present invention, since the master coin and the target coin are collated using the feature amount extracted from the density histogram, the discrimination accuracy of the target coin can be improved, and the target coin can be improved. It is possible to more accurately determine the true / false of.

  According to a second aspect of the present invention, in the coin discriminating apparatus according to the first aspect, an image division processing unit that divides the image of the target coin acquired by the imaging unit into a plurality of annular or circular regions. The histogram generation means generates a density histogram for each divided image divided by the image division processing means, and the feature quantity extraction means extracts each feature quantity from the density histogram for each divided image, The determination unit compares each feature amount extracted from the density histogram of each divided image of the target coin with a corresponding feature amount of the master coin. As in this embodiment, the coin image is divided into a plurality of circular or circular regions, and the feature amount is extracted from the density histogram for each divided image and compared with the master coin, thereby obtaining the target coin. Can be improved, and the true / false of the target coin can be more accurately determined.

  According to a third aspect of the present invention, in the coin discriminating apparatus according to the second aspect, the image division processing unit displays an image of the target coin with a plurality of annular or circular shapes with a gap between the divided images. It is characterized by being divided into regions. If the coin image is divided with a gap between the divided images as in this embodiment, the coin discrimination accuracy can be further improved, and misleading coin discrimination can be performed more accurately.

  According to an embodiment of the present invention, the feature amount of the master coin and the feature amount of the target coin in the coin discriminating device according to claim 1 are the maximum number of pixels in the density histogram of each coin. It is characterized by the value of the gradation.

  In the embodiment according to claim 5 of the present invention, the feature amount of the master coin and the feature amount of the target coin in the coin discriminating device according to claim 1 are the maximum in the number of pixels in the density histogram of each coin. It is characterized in that it is the difference between the gradation value at the time of the above and the gradation value when the number of pixels is at a predetermined level.

  In the embodiment according to claim 6 of the present invention, the feature amount of the master coin and the feature amount of the target coin are deviations of peak positions of the respective histograms of the divided images.

  In the embodiment according to claim 7 of the present invention, the feature amount of the master coin and the feature amount of the target coin are deviations between gradations at which the number of pixels becomes a predetermined value in each histogram of the divided images. It is characterized by that.

  As the feature amount of the density histogram, the features as in claims 4 to 7 can be used, or a combination of these can be used, thereby improving the discrimination accuracy of the target coin. Thus, it is possible to more accurately determine the authenticity of the target coin.

  According to an eighth aspect of the present invention, there is provided a coin discriminating method in which a feature amount obtained from a density histogram of a master coin is stored in a storage unit in advance, and an image of a target coin is obtained using an imaging unit; Generating a density histogram from the steps, extracting a feature amount of the target coin from the density histogram, comparing a feature amount of the master coin stored in the storage means with a feature amount of the target coin, It is characterized by determining the authenticity of the target coin.

  In the coin discriminating method according to claim 8 of the present invention, since the master coin and the target coin are collated using the feature amount extracted from the density histogram, the discrimination accuracy of the target coin can be improved. It is possible to more accurately determine the true / false of.

  A storage medium according to a ninth aspect of the present invention stores a feature amount obtained from a density histogram of a master coin, a step of acquiring an image of a target coin using an imaging unit, and a density histogram is generated from the image Recorded are a step of extracting the feature amount of the target coin from the density histogram, and a step of comparing the feature amount of the master coin stored in the storage unit with the feature amount of the target coin. It is. By executing such a storage medium by a computer, it is possible to more accurately determine the authenticity of the target coin.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. However, it is needless to say that the present invention is not limited to the following examples.

  First, a coin discriminating apparatus according to Embodiment 1 of the present invention will be described with reference to FIGS. FIG. 1 is a perspective view showing a part of a coin passage 1 provided inside a game machine, a vending machine, a ticket machine or the like. The coin passage 1 has a narrow rolling path 1a and a side wall surface 1b located above the rolling path 1a. The rolling path 1a is inclined along the coin passing direction, and the inserted coin 2 rolls on the rolling path 1a. Further, the side wall surface 1b is slightly inclined with respect to the vertical surface, and the undeformed coin 2 rolls without falling off the rolling path 1a, but the deformed coin 2 drops off from the rolling path 1a. It is discharged to the return port.

  FIG. 2 is a side view showing the illumination device 3 and the imaging device 4 of the coin discriminating device arranged facing the coin passage 1. 3 is a perspective view showing the illumination device 3 and the imaging device 4, and FIG. 4 is a sectional view thereof. The illuminating device 3 has a structure in which a plurality of light sources 6 are provided on the outer peripheral surface of the cylindrical body 5. A plurality of through holes 7 are opened on the outer peripheral surface of the cylindrical body 5 at equal intervals. The light source 6 is configured by a white LED (light emitting element) or the like, and is inserted into the through hole 7 and fixed. Each light source 6 is disposed so as to emit light toward the inside of the cylinder 5, and the emission direction of the principal ray is inclined obliquely with respect to the direction parallel to the opening surface of the cylinder 5. Thus, the light B emitted from each light source 6 is uniformly irradiated so as to overlap on the passing surface (side wall surface 1b) of the coin 2.

  The imaging device 4 is provided on a substrate 9 fixed to the end surface of the cylinder 5. A window 11 is opened at the center of the substrate 9, and the imaging lens 10 is held at the tip of the lens barrel 12 that extends from the edge of the window 11 toward the cylinder 5. An imaging device 8 such as a CCD camera is fixed at a position facing the window 11.

  The substrate 9 is provided with a light projecting element 14 such as an LED for coin detection or an LD (laser element). The light projecting element 14 is located in the coin passing direction with respect to the imaging element 8. As shown in FIG. 5, a cavity 15 is opened in the coin passage 1 so as to face the light projecting element 14, and light reception for receiving the detection light D emitted from the light projecting element 14 is received in the cavity 15. Element 13 is arranged.

  Therefore, in this coin discriminating apparatus, an image of the passing coin 2 is taken as shown in FIG. When the coin 2 has not passed, the light source 6 is turned off. On the other hand, the light projecting element 14 is turned on, and the detection light D emitted from the light projecting element 14 is received by the light receiving element 13, and the light receiving element 13 is turned on (light receiving state).

  At this time, when the coin 2 rolled along the rolling path 1a reaches a position facing the image sensor 8, the detection light D emitted from the light projecting element 14 is blocked by the edge of the coin 2, and the light receiving element 13 Turns off (non-light receiving state). At the same time as the light receiving element 13 is turned off, all the light sources 6 instantaneously emit a flash B to illuminate the coin 2 and the image sensor 8 acquires a still image of the passing coin 2. Thus, the image of the coin 2 photographed by the image sensor 8 is sent out from the image sensor 8 to the coin discrimination processing unit as image data (digital data).

  Note that a coin image may be captured by a method different from the above embodiment. In a different method, the lighting device 3 is always turned on, and is blocked (masked) so that image data is not sent from the image sensor 8 to the coin discrimination processing unit. When the light receiving element 13 is turned off and the coin 2 is detected, the image data of the image acquired by the imaging device 4 at that moment is sent from the imaging element 8 to the coin discrimination processing unit, as will be described later. Whether the coin is true or false is determined. When the instantaneous image data is sent to the coin discrimination processing unit, data transmission from the image sensor 8 to the coin discrimination processing unit is again cut off. According to this method, since the shutter of the imaging device 4 can be configured by an electric signal, the shutter speed can be increased and a still image of a coin can be easily obtained.

  FIG. 6 is a block diagram showing a configuration of the coin discrimination processing unit 16. The coin discrimination processing unit 16 is configured by a microcomputer, an electronic circuit, a computer, software, and the like. The coin discrimination processing unit 16 separates the digital multivalued image data received from the image sensor 8 into a coin image and a background image and extracts a coin shape, and detects the center from the extracted coin shape. A coin center detection unit 18 that performs the image shift unit 19 that moves the coin image to the center of the screen so that the center of the coin shape is at the center of the screen, and image division that divides the coin image into circular and concentric annular partial images. A processing unit 20, a histogram generation unit 21 that generates a density histogram for each partial image, a feature amount extraction unit 22 that extracts a feature amount of each histogram, and feature amounts on both sides or one side of the master coin are stored in advance. The master coin data storage unit 23 to be placed, and the feature amount of the target coin and the master coin stored in the master coin data storage unit 23 A determination processing section 24 the authenticity of the subject coin by comparing the feature amount, and an output unit 25 for outputting a determination result.

  7, 8 and 9 are flowcharts showing a procedure for determining the authenticity of the target coin by the coin determination processing unit. 10 to 13 are explanatory diagrams thereof. Hereinafter, the coin discrimination method will be described in detail with reference to FIGS. However, the following processing is processed as data, and FIGS. 10 to 13 illustrate the principle for easy understanding of the processing contents. When the image of the coin 2 is taken by the imaging device 4 as described above, the coin discrimination processing unit 16 reads the digital multi-value image data (for example, image data of 256 gradations) (step S1 in FIG. 7). ). The coin image read in this way is preferably at the center of the screen (screen for separating the images). However, as shown in FIG. 10A, the coin image 26 is not necessarily located at the center of the screen 27 due to a difference in the timing of photographing the passing coin 2, and is normally biased from the center of the screen 27.

  Therefore, in order to move the coin image 26 to the center of the screen 27, first, the center of the coin shape is detected (step S2). That is, the coin shape extraction unit 17 extracts the coin image 26 by separating the coin image 26 and the background image 28 in the read image data. Next, the center of the extracted coin image 26 is detected by the coin center detection unit 18. There are various methods for detecting the center of the coin shape. For example, as shown in FIG. 10A, a vertical line passing through the center in the width direction of the coin image 26 and a horizontal line passing through the center in the height direction. It can be obtained as an intersection with.

  When the center of the coin image 26 is detected in this manner, the image shift unit 19 moves the coin image 26 on the screen 27 so that the center of the coin image 26 coincides with the center of the screen 27 as shown in FIG. Move (step S3).

  When the coin image 26 finishes moving to the center of the screen 27, the image division processing unit 20 cuts the coin image 26 into a plurality of areas (step S4), and stores each divided image in the memory (step S5). There are various methods for dividing the coin image 26, but it is desirable to divide the coin image 26 into a circular region and a circular region whose centers coincide with each other. In FIG. 11, the first region 29 having a circular shape and the second and third regions 30 and 31 having a circular shape are separated.

  Next, the partial image of the first area 29 is read from the memory (step S6), and the histogram generation unit 21 performs a histogram process on the partial image of the first area 29 to generate a density histogram. Save in the memory (step S7).

  When the first to third regions 29 to 31 are divided into a circular region and an annular region that coincide with the center of the coin image 26, when a density histogram is created from each partial image, the density histogram of each region However, since the same in principle regardless of the rotation angle of the coin image 26 (the rotation angle of the coin 2 when the coin 2 is photographed), the process of rotating the image to a predetermined angle is not necessary. .

FIG. 12 is a diagram for explaining an actual method of dividing the coin image 26 into partial images and performing histogram processing. As shown in FIG. 12A, the screen 27 is divided into, for example, 2 ⁶ = 64 and is divided into 4096 pixels, and a number is assigned to each pixel. In FIG. 12A, the pixel numbers “1”, “2”, “3”, and “0” are assigned, but the regions with the same pixel number are as shown in FIG. It corresponds to a circular first area 29A ("1"), an annular second area 30A ("2"), an annular third area 31A ("3"), and a background area ("0"). . Therefore, if the density (gradation degree) of each pixel in the coin image 26 is grouped in association with each pixel number, and the statistical processing is performed on the pixel gradation degree for each group, each area divided as shown in FIG. Density histograms 29, 30, and 31 are obtained at a time. The density histogram of each area created in this way is temporarily stored in the memory.

  Thereafter, the feature amount extraction unit 22 extracts the feature amount of the density histogram according to the flowchart of FIG. 9 (step S8). For example, it is assumed that the density histogram 32 of the partial image is a curve as shown by a broken line in FIG. In FIGS. 13A and 13B, the horizontal axis represents the density (gradation degree) of each pixel expressed in 256 gradations (0 to 255), and the vertical axis represents the number of pixels in each gradation degree. The density histogram 32 is first reduced or enlarged in the vertical axis direction or translated in the vertical axis direction so that the peak value becomes a constant value (for example, 100 (arbitrary unit)), and a solid line in FIG. Normalization is performed as shown (step S21 in FIG. 9).

  Next, the density histogram 32 in the region where the number of pixels is smaller than the predetermined threshold Ns is discarded (step S22). For example, the threshold value Ns is a value when the number of pixels of the density histogram is 70% of the peak value, and the area below the threshold value is discarded to obtain a density histogram 32 as shown in FIG. The magnitude of the threshold value Ns is set to an appropriate level in advance.

  Thereafter, the gradation value at the peak position of the density histogram 32 (hereinafter referred to as peak position) Kp is detected (step S23), and the width of the density histogram 32 on the right side of the peak position (hereinafter referred to as high gradation width). WR and the width (hereinafter referred to as low gradation width) WL of the density histogram 32 on the left side of the peak position are obtained (steps S24 and S25), and the peak position Kp, high gradation width WR, and low gradation width WL are used as features in the memory. Remember.

  The master coin data storage unit 23 stores the peak position, the high gradation width, and the low gradation width for each of the first to third regions on both sides or one side of the master coin. Here, when the feature amounts of both sides of the master coin are stored in the master coin data storage unit 23, one side is referred to as a first side and the other side is referred to as a second side. Alternatively, when only the feature amount of one side of the master coin is stored, that side is referred to as a first side.

  At the time of collation, first, the peak position Kp of the density histogram 32 in the first area 29 of the coin 2 is collated with the peak position of the density histogram in the first area of the first surface of the master coin (step S9). It is determined whether or not the peak position Kp matches the peak position of the master coin (step S10). Here, the coincidence between the peak position Kp of the coin 2 and the peak position of the master coin does not mean that they coincide completely, and the peak position Kp of the coin 2 is a predetermined centered on the peak position of the master coin. As long as it is within the range. This is because it is necessary to consider changes in illuminance over time of the lighting device 3.

  When the peak position Kp in the first area 29 does not coincide with the peak position of the first face of the master coin, it is checked whether or not the data of the master coin exists for two faces in the master coin data storage unit 23 ( Step S14) If there is no data on the two sides (NO in Step S14), it is determined that the coin 2 does not match the master coin, and the coin 2 is discharged to the return slot (Step S17).

  Further, when master coin data exists for two sides (YES in step S14), the second side is also checked. That is, the peak position Kp in the first area 29 of the coin 2 is collated with the peak position in the first area on the second surface of the master coin stored in the master coin data storage unit 23 (step S15). It is determined whether or not the peak position Kp of the coin 2 matches the peak position of the master coin (step S16). Again, the coin 2 peak position Kp and the master coin peak position coincide with each other as long as the coin 2 peak position Kp falls within a predetermined range centered on the master coin peak position.

  If the peak position Kp of the first area 29 does not coincide with the peak position in the first area of the second surface of the master coin (NO in step S16), it is determined that the coin 2 is not a genuine coin, The coin 2 is discharged to the return slot (step S17).

  On the other hand, when the peak position Kp of the coin 2 coincides with the peak position of the first or second surface of the master coin in step S10 or S16, the low gradation width WL in the first region 29 of the coin 2 is further increased. The high gradation width WR is compared with the low gradation width and the high gradation width in the first area of the master coin stored in the master coin data storage unit 23 (step S11). However, if the peak position of the coin 2 coincides with the peak position of the first surface of the master coin in step S10, the low gradation width and high gradation width of the master coin to be collated in step S11 are also on the first surface. is there. If the peak position of the coin 2 coincides with the peak position of the second surface of the master coin in step S16, the low gradation width and high gradation width of the master coin to be collated in step S11 are also on the second surface. is there. Also, here, if the gradation widths WL and WR of the coin 2 are included in a certain range centering on each gradation width of the master coin, they are assumed to match.

  As a result of the collation in step S11, if the low gradation width or the high gradation width does not match (NO in step S12), it is determined that the coin 2 does not match the master coin and is not a genuine coin and is discharged from the return slot. (Step S17).

  On the other hand, as a result of the collation in step S11, when the low gradation width and the high gradation width match (in the case of YES in step S12), the collation of the first area 29 of the coin 2 is finished and the second area 30 of the coin 2 is completed. Are collated (step S13). That is, the partial image in the second area 30 is read from the memory, and the processes in steps S6 to S17 are performed on the partial image in the second area 30 as in the case of the first area 29. Therefore, when the feature amount obtained from the density histogram 32 of the second region 30 does not match the feature amount in the second region of the master coin, the coin 2 is discharged to the return slot.

  Further, when the peak position, the low gradation width, and the high gradation width obtained from the density histogram 32 of the second area 30 coincide with the feature amount of the first or second surface of the master coin, the partial image of the third area 31 Are read from the memory, and the processes in steps S6 to S17 are performed on the partial image in the third area 31 as in the case of the first area 29. Therefore, when the feature amount obtained from the density histogram 32 of the third region 31 does not match the feature amount in the third region of the master coin, the coin 2 is discharged to the return slot.

  In this way, when the feature amount of the coin 2 matches the master coin in all the areas 29 to 31 and the coin 2 is successfully verified (steps S13 and S18), the coin 2 is determined to be a genuine coin, A signal indicating that the coin is a genuine coin is output from the output unit 25.

  Therefore, according to the coin discriminating apparatus of the present invention, since the comparison is made with the master coin using the feature amount of the target coin extracted from the density histogram, the authenticity of the coin can be determined with high accuracy. Furthermore, since the coin image is divided into a plurality of partial images, the feature amount of the coin is extracted for each partial image, and compared with the master coin, the amount of information for coin comparison increases, and the authenticity of the coin Can be determined with higher accuracy.

  In the above description, the feature amount of the master coin is stored in the master coin data storage unit 23 in advance, but the feature amount of the master coin is registered in advance according to the flowchart of FIG. When the coin discriminating apparatus is set to the master coin registration mode and a master coin (authentic coin) is inserted into the coin insertion slot, an image of the master coin is captured by the illumination device 3 and the imaging device 4 (step S31). Subsequent processes in steps S32 to S38 are the same as those in steps S2 to S8 in FIGS. 7 and 8, and the processing content of step S38 is as shown in FIG. Therefore, after the coin image of the inserted master coin is divided into a plurality of circular or annular partial images, each partial image is converted into a density histogram, and the feature amount of each density histogram is extracted. And the extracted feature-value is memorize | stored in the master coin data storage part 23 for every partial image of a 1st surface (step S39, S40).

  If there are patterns on both sides of the master coin, the master coin is placed face down and inserted into the coin slot, and the feature amount on the back side (second side) is also stored in the master coin data storage unit 23.

  As the feature amount of the coin, the peak position Kp, the low gradation width WL, and the high gradation width WR are used, but some of these feature amounts may be used.

  Further, when the density histogram of a partial image of a coin (master coin) has a plurality of peaks as shown in FIG. 15, a set of feature amounts as shown in FIG. 15A depending on how the threshold value Ns is set. Can be extracted, and a plurality of sets of feature quantities can be extracted as shown in FIG. Of course, if a plurality of sets of feature quantities are extracted as shown in FIG. 15B, information for coin verification increases, and the coin discrimination system is improved.

  Further, when the coin image is divided into a plurality of partial images, in the above-described embodiment, as shown in FIG. However, as shown in FIG. 16, if the coin image is divided into a plurality of partial images with a gap, the accuracy of coin authenticity determination is further increased. That is, if the image of the shaded region α in FIG. 16 is discarded and a gap is provided between the first region 29, the second region 30, and the third region 31, which are partial images, the first region 29, the second region 30 are provided. And the image of the 3rd field 31 becomes discontinuous, and the difference in the pattern between the density histograms of each field 29-31 becomes remarkable. As a result, the feature amount of each basin is also greatly different, and the coin discrimination accuracy is improved.

  FIGS. 17 and 18 are diagrams showing the difference in density histogram between the case where no gap is provided between partial images and the case where a gap is provided. FIGS. 17A and 17C show a circular partial image (first region 29) and an annular partial image (second region 30) which are divided without a gap. When the image is divided in this way, the density histogram obtained from the circular partial image as shown in FIG. 17A is as shown in FIG. 17B, and the annular shape as shown in FIG. The density histogram obtained from the partial image is as shown in FIG.

  18A and 18C show a circular partial image (first region 29) and an annular partial image (second region 30) divided with a gap. When the image is divided in this way, the density histogram obtained from the circular partial image as shown in FIG. 18A is as shown in FIG. 18B, and the annular shape as shown in FIG. The density histogram obtained from the partial image is as shown in FIG.

  When the difference between the two density histograms in FIGS. 17B and 17D is compared with the difference between the two density histograms in FIGS. 18B and 18D, when a gap is formed between the partial images. The difference between the two density histograms becomes more remarkable. Therefore, by making a gap between the partial images, when comparing the target coin and the master coin by extracting the respective feature amounts of the partial images, the coin collation accuracy is further improved.

  Next, a coin discrimination device according to Embodiment 2 of the present invention will be described. In this case, instead of the high gradation width WR and the low gradation width WL, the gradation level when the number of pixels is smaller than the peak value of the histogram (for example, ½ times the peak value) is used as the feature amount. is there. In the coin discriminating apparatus according to the second embodiment, the structures of the coin passage 1, the illumination device 3, the imaging device 4, and the like are the same as those in the first embodiment. The configuration of the coin discrimination processing unit 16 is the same as that of the first embodiment except that the processing in the feature quantity extraction unit 22 and the feature quantity to be collated in the judgment processing unit 24 are different. Therefore, in the following description of the second embodiment, the same components as those in the first embodiment will not be described in detail by using the same reference numerals as those in the first embodiment, and only differences from the first embodiment will be described.

  19, 20 and 21 are flowcharts showing a procedure for determining the authenticity of the target coin in the second embodiment. FIG. 22 is an explanatory diagram thereof. Also in the coin discrimination device according to the second embodiment, the illumination device 3 illuminates the coin 2 and the imaging device 4 takes an image of the coin 2, and the image data is transmitted from the imaging element 8 to the coin discrimination processing unit 16. The multivalued image data of the coin 2 is read into the coin discrimination processing unit 16 (step S1). The coin multi-value image data of the coin 2 that has been read is separated into a coin image 26 and a background image 28 by the coin shape extraction unit 17, and the coin image 26 is extracted. Next, the center of the extracted coin image 26 is detected by the coin center detection unit 18 (step S2). When the center of the coin image 26 is detected, the image shift unit 19 moves the coin image 26 within the screen 27 so that the center of the coin image 26 coincides with the center of the screen 27 (step S3).

  When the coin image 26 moves to the center of the screen 27, the image division processing unit 20 divides the coin image 26 into a plurality of circular or annular regions (steps S4 to S6). Next, the divided partial images for each region are subjected to histogram processing by the histogram generation unit 21, and a density histogram of each partial image is generated (step S7). The processing so far is the same as that in the first embodiment.

  However, the feature amount extraction (step S41) by the next feature amount extraction unit 22 is different in the second embodiment from that in the first embodiment. FIG. 21 is a flowchart for explaining the feature amount extraction process in the second embodiment, and FIG. 22 is an explanatory diagram thereof. In the case of the second embodiment, the histogram is not normalized and the density histogram 32 obtained from the partial image is used as it is. Then, as shown in FIG. 22, the peak value Np of the density histogram 32 is determined, and the density (gradation degree) value Kp of the corresponding pixel is obtained (step S42). Further, as shown in FIG. 22, the density Kr of the right pixel and the density Kl of the left pixel when the peak value Np is ½ (half value) Np / 2 are obtained (steps S43 and S44). .

  In the master coin data storage unit 23, feature amounts Kp, Kr, and Kl of partial images for each area on both sides or one side of the master coin are stored in advance. And after calculating | requiring the feature-values Kp, Kr, and Kl of the coin 2 as mentioned above, it is substantially the same as that of Example 1. FIG. That is, the authenticity of the coin 2 is determined by comparing and comparing the feature values Kp, Kr, and Kl of the target coin 2 and the master coin according to the procedures of steps S9, S10, S45, S46, and S13 to S18 of FIG. Is done. The difference from the first embodiment is that in steps S45 and S46, the left and right half-value gradations Kr and Kl in the histogram of each partial image of the target coin and the master coins feature values Kr and Kl in accordance with the difference in feature amount. Are compared (step S45), and if any feature quantity does not match (NO in step S46), it is determined that the coin is not a genuine coin and the coin is discharged (step S17). If the amounts match (YES in steps S46 and S13), it is determined that the verification is successful and the coin is determined to be a genuine coin (step S18).

  In the second embodiment as well, when a coin image is divided into a plurality of partial images, if a gap is provided between each region, the coin discrimination accuracy is improved.

  Next, a coin discrimination device according to Embodiment 3 of the present invention will be described. This uses the deviation between the gradation levels at the peak position in the histogram of the partial image and the deviation between the gradation degrees at the position α times the peak value (α <1) as the feature amount. In the coin discriminating apparatus according to the third embodiment, the structures of the coin passage 1, the illumination device 3, the imaging device 4, and the like are the same as those in the first embodiment. Therefore, in the following description of the third embodiment, the same components as those in the first embodiment will be omitted by using the same reference numerals as those in the first embodiment, and only differences from the first embodiment will be described.

  23 to 26 are flowcharts showing a procedure for determining whether the target coin is true or false in the third embodiment. FIGS. 27 to 29 are explanatory diagrams thereof. Also in the coin discriminating apparatus of the third embodiment, the coin 2 is illuminated by the illumination device 3 and an image of the coin 2 is photographed by the imaging device 4, and the image data is transmitted from the image sensor 8 to the coin discrimination processing unit 16. The multivalued image data of the coin 2 is read into the coin discrimination processing unit 16 (step S1). The digital multi-valued image data of the coin 2 that has been read is separated into a coin image 26 and a background image 28 by the coin shape extraction unit 17, and the coin image 26 is extracted (step S1). Next, the center of the extracted coin image 26 is detected by the coin center detection unit 18 (step S2). When the center of the coin image 26 is detected, the image shift unit 19 moves the coin image 26 within the screen 27 so that the center of the coin image 26 coincides with the center of the screen 27 (step S3).

  When the coin image 26 moves to the center of the screen 27, the image division processing unit 20 divides the coin image 26 into a plurality of circular or annular regions (steps S4 to S6). For example, in FIG. 27, the coin image 26 is divided into one circular partial image 41 and five circular partial images 42 to 46. Next, the divided partial images 41 to 46 are processed into a histogram by the histogram generation unit 21, and density histograms of the partial images 41 to 46 are generated (step S7). The histograms of the partial images 41 to 46 thus formed into histograms are, for example, histograms 41A to 46A shown in FIG. The processing so far is the same as that in the first embodiment.

  Next, feature amounts are extracted by the feature amount extraction unit 22 (step S51). FIG. 26 is a flowchart for explaining the feature extraction process in the third embodiment. In the case of the third embodiment, the histograms 41A to 46A are first normalized (step S52). Then, any one of the normalized histograms is set as a reference histogram. FIG. 29 shows only two histograms 41A and 42A among the normalized histograms, and the histogram 41A obtained from the partial image 41 having a central annular shape is used as a reference histogram.

  Next, with respect to the histogram 41A, the gradation Kp at the peak position is detected (step S53), and the gradation Kr at the right position and the gradation Kl at the left position that are α times the height (number of pixels) of the peak value Np. (Steps S54 and S55). Similarly, with respect to the other histograms 42A to 46A, the gradation degrees Kp ′,... At the peak positions are detected (step S53), and the gradation degrees Kr ′,. And the gradation degree Kl ′,... At the left position are obtained (steps S54 and S55).

  Thus, for all the histograms 41A to 46A, the gradations Kp, Kp ′,... At the peak positions and the gradations Kr, Kr ′,... On the right and left sides at the positions where the height is α × Np (hereinafter referred to as a threshold). When K1, Kl ′,... Have been obtained (in the case of YES in step S56), as shown in FIG. 29, with reference to the histogram 41A, the peak position Kp of the other histogram 42A from the peak position Kp of the reference histogram 41A Deviation Gp = Kp′−Kp up to “′ (step S57), and then deviation Gr = Kr ′ from the right gradation Kr ′ at the threshold of the reference histogram 41A to the right gradation Kr ′ at the threshold of the other histogram 42A. -Kr is obtained (step S58), and another histogram is obtained from the left gradation degree K1 in the threshold value of the reference histogram 41A. A deviation Gl = Kl'-Kl to right gradient Kl' in ram 42A threshold (step S59).

  In this way, for all the histograms 42A to 46A other than the reference histogram 41A, when the deviations of the respective peak positions and the deviations of the right gradation and the left gradation in the threshold values have been obtained (step S60), the feature amount extraction processing is performed. finish.

  When the feature amount extraction is completed, the peak position Kp of the reference histogram 41A is compared with the peak position of the master coin reference histogram stored in the master coin data storage unit 23 (step S61). If the peak position Kp of the reference histogram 41A does not coincide with the peak position of the reference histogram of the master coin as a result of this collation (NO in step S62), the master coin data is stored in the master coin data storage unit 23. It is checked whether or not two sides exist (step S63). If the two side data does not exist (NO in step S63), it is determined that the coin 2 does not match the master coin, and the coin 2 Is discharged to the return port (step S78).

  Further, when master coin data exists for two sides (YES in step S63), the second side is also checked. That is, the peak position Kp of the reference histogram 41A is collated with the peak position of the reference histogram on the second surface of the master coin stored in the master coin data storage unit 23 (step S64), and the peak position Kp of the coin 2 is determined. And whether or not the peak positions of the master coins coincide with each other (step S65).

  If the peak position Kp of the reference histogram 41A does not match the peak position of the reference histogram on the second surface of the master coin (NO in step S65), it is determined that the coin 2 is not a genuine coin, and the coin 2 is It discharges to the return port (step S78).

  On the other hand, in step S62 or S65, when the peak position Kp of the reference histogram 41A coincides with the peak position of the reference histogram on the first or second surface of the master coin, further on the right side of the threshold value of the reference histogram 41A. Is compared with the gradation degree on the right side of the threshold value of the reference histogram of the master coin stored in the master coin data storage unit 23 (step S66). Further, the gradation degree K1 on the left side of the threshold value of the reference histogram 41A is compared with the gradation degree on the left side of the threshold value of the reference histogram of the master coin stored in the master coin data storage unit 23 (step S68). However, if the peak position of the reference histogram 41A coincides with the peak position of the first face of the master coin in step S62, the gradation levels on the right and left sides of the threshold value of the master coin to be collated in steps S66 and S68 are also the same. In the first aspect. In addition, when the peak position of the reference histogram 41A matches the peak position of the second face of the master coin in step S65, the gradation levels on the right and left sides of the threshold value of the master coin to be collated in steps S66 and S68 are also set. This is in the second aspect.

  As a result of the collation in step S66, the gradation degree Kr on the right side of the threshold value does not match (in the case of NO in step S67), or as a result of the collation in step S68, the gradation degree Kl on the left side of the threshold value does not match (step In the case of NO in S69, the coin 2 does not match the master coin and is determined not to be a genuine coin, and is discharged from the return slot (step S78).

  On the other hand, as a result of the collation in steps S66 and S68, when the gradient on the right side and the gradient on the left side match in the threshold value (YES in steps S67 and S69), the peak based on the histogram 41A is used. The deviation of the position and the like and the deviation of the peak position and the like based on the reference histogram in the master coin are compared and collated (steps S70 to S76). That is, the peak position deviation Gp (hereinafter referred to as peak position deviation) between the histogram 42A obtained in step S51 and the reference histogram 41A is compared with the peak position deviation of the histogram 42A in the master coin (step S70). Further, the deviation Gr of the left gradation in the threshold values of the histogram 42A and the reference histogram 41A obtained in step S51 (hereinafter referred to as the left deviation) is compared with the left deviation of the histogram 42A in the master coin (step S72). Further, the deviation Gl of the right gradation (hereinafter referred to as the right deviation) in the threshold values of the histogram 42A and the reference histogram 41A obtained in step S51 is compared with the right deviation of the histogram 42A in the master coin (step S74).

  Similarly, with respect to the other histograms 43A to 46A, the peak position deviation, the right side deviation, and the left side deviation are compared with the peak position deviation, right side deviation, and left side deviation of the master coin, respectively (steps S70, S72, S74).

  Thus, as a result of comparing and comparing the peak position deviation, right side deviation, and left side deviation of the histograms 42A to 46A with the reference histogram 41A as the reference in step S70, S72, and S74, respectively, one of these feature quantities is the master. If it does not match the feature amount of the coin (NO in any of steps S71, S73, or S75), the coin 2 is discharged into the return slot (step S78).

  In addition, as a result of comparison and collation for all the histograms 42A to 46A (after YES in step S76), the feature amount of any histogram 42A to 46A matches the feature amount of the master coin (step S71). , S73 and S75 are YES), the coin 2 is determined to be a genuine coin, and a signal indicating that it is a genuine coin is output from the output unit 25.

  It should be noted that the fact that the feature values such as the peak position and the deviation match in the comparison and collation in steps S62, S66, S68, S70, S72, and S74 does not mean that the feature values of the coin 2 match. It suffices if it falls within a predetermined range centering on the feature amount of the master coin. This is because it is necessary to consider changes in illuminance over time of the lighting device 3. The feature amount of the master coin stored in the master coin data storage unit 23 may be generated from the feature amount acquired from a plurality of master coins. For example, the average value of the right side deviation and left side deviation obtained from multiple master coins may be used as the right side deviation and left side deviation of the master coin, or the maximum value of right side deviation and left side deviation obtained from multiple master coins. Alternatively, the minimum value may be the right deviation or left deviation of the master coin.

  If the feature amount as in the third embodiment is used, it is possible to further improve the accuracy of coin authenticity by the coin discriminating apparatus. Also in this embodiment, the discrimination accuracy can be further improved by providing a gap between the partial images.

  Further, by combining the coin discrimination methods described in the first to third embodiments, the coin discrimination accuracy can be further increased.

It is a perspective view which shows a part of coin passage provided in the inside of a game machine, a vending machine, a ticket machine, etc. It is a side view which shows the illuminating device and imaging device of the coin discrimination device by one Example of this invention arrange | positioned facing a coin channel | path. It is a perspective view which shows an illuminating device and imaging device same as the above. It is sectional drawing of the illuminating device and imaging device of FIG. It is the schematic which shows a mode that the image of the coin which passes a coin passage is image | photographed with the illuminating device and imaging device of FIG. It is a block diagram of the coin discrimination | determination process part which comprises the coin discrimination | determination apparatus of this invention. It is a flowchart showing the procedure for discriminating the authenticity of the target coin by the coin discrimination processing unit. It is a flowchart showing the process following the process of the flowchart of FIG. It is a flowchart showing the process of extracting the feature-value of the coin image in FIG. (A) is a figure which shows the coin image which shifted | deviated from the center of the screen, (b) is a figure which shows the coin image moved to the center of a screen. It is a figure which shows the state which divided | segmented the coin image into the 1st area | region which carried out circular, and the 2nd and 3rd area | region which carried out the annular | circular shape. (A) (b) is a figure explaining the actual method which divides | segments a coin image into each partial image and carries out a histogram-ized process. (A) is a figure explaining the process which normalizes a density histogram, (b) is a figure explaining the process which truncates the density histogram below a fixed threshold value. It is a flowchart which shows the process for registering the feature-value of a master coin. (A) (b) is a figure explaining the relationship between the density histogram which has a some peak, and a threshold value. It is a figure explaining the method of different division | segmentation at the time of dividing | segmenting a coin image into a partial image. (A) is a schematic diagram showing a circular partial image, (b) is a diagram showing a density histogram obtained from the partial image of (a), and (c) is obtained without leaving a gap with the circular partial image. FIG. 4D is a schematic diagram showing an annular partial image, and FIG. 4D is a diagram showing a density histogram obtained from the partial image of FIG. (A) is a schematic diagram showing a circular partial image, (b) is a diagram showing a density histogram obtained from the partial image of (a), and (c) is obtained with a gap from the circular partial image. Schematic diagram showing an annular partial image, (d) is a diagram showing a density histogram obtained from the partial image of (c). In Example 2, it is the flowchart showing the procedure for discriminating the authenticity of a target coin by a coin discrimination | determination process part. FIG. 20 is a flowchart showing a process that follows the process of the flowchart of FIG. 19. FIG. 10 is a flowchart illustrating a procedure for extracting a feature amount of a partial image in the second embodiment. It is a figure explaining the feature-value same as the above. In Example 3, it is the flowchart showing the procedure for discriminating the authenticity of a target coin by a coin discrimination | determination process part. It is a flowchart showing the process following the process of the flowchart of FIG. It is a flowchart showing the process following the process of the flowchart of FIG. 10 is a flowchart illustrating a procedure for extracting a feature amount of a partial image in the third embodiment. It is explanatory drawing showing a mode that a coin image is divided | segmented into a some partial image. It is a figure showing the histogram of each partial image divided | segmented like FIG. It is explanatory drawing which shows the feature-value determined between the feature-value of the histogram used as a reference | standard, and the histogram used as a reference | standard, and a histogram other than a reference | standard.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Coin path rolling path 2 Coin 3 Illuminating device 4 Imaging device 6 Light source 8 Imaging element 10 Imaging lens 13 Light receiving element 14 Light projecting element 16 Coin discrimination | determination processing part 17 Coin shape extraction part 18 Coin center detection part 19 Image shift part 20 Image division processing unit 21 Histogram generation unit 22 Feature amount extraction unit 23 Master coin data storage unit 24 Judgment processing unit 25 Output unit 26 Coin image 27 Screen 28 Background image

Claims (9)

A coin discriminating device for discriminating the authenticity of a target coin based on an image,
Storage means for storing feature values obtained from the density histogram of the master coin;
Imaging means for acquiring an image of the target coin;
Histogram generating means for generating a density histogram from the image;
Feature amount extraction means for extracting the feature amount of the target coin from the density histogram generated by the histogram generation means;
Means for determining the authenticity of the target coin by comparing the feature quantity of the master coin stored in the storage means with the feature quantity of the target coin extracted by the feature quantity extraction means;
A coin discrimination device. Image division processing means for dividing the image of the target coin acquired by the imaging means into a plurality of annular or circular regions;
The histogram generation means generates a density histogram for each divided image divided by the image division processing means,
The feature amount extraction means extracts each feature amount from a density histogram for each divided image,
The coin determination apparatus according to claim 1, wherein the determination unit compares each feature amount extracted from the density histogram of each divided image of the target coin with a corresponding feature amount of the master coin.   3. The coin discriminating apparatus according to claim 2, wherein the image division processing unit divides the image of the target coin into a plurality of annular or circular regions with a gap between the divided images.   2. The coin discriminating apparatus according to claim 1, wherein the feature amount of the master coin and the feature amount of the target coin are values of gradation when the number of pixels in the density histogram of each coin is maximum. .   The feature amount of the master coin and the feature amount of the target coin are a gradation value when the number of pixels in the density histogram of each coin is maximum and a gradation value when the number of pixels is a predetermined level. The coin discriminating apparatus according to claim 1, wherein the coin discriminating apparatus is a difference.   3. The coin discriminating apparatus according to claim 2, wherein the feature amount of the master coin and the feature amount of the target coin are deviations of peak positions of respective histograms of the divided images.   The feature amount of the master coin and the feature amount of the target coin are deviations between gradations at which the number of pixels becomes a predetermined value in each histogram of the divided images. Coin discrimination device. The feature value obtained from the density histogram of the master coin is stored in advance in the storage means,
Acquiring an image of the target coin using an imaging means;
Generating a density histogram from the image;
Extracting a feature amount of the target coin from the density histogram;
Comparing the feature amount of the master coin stored in the storage means with the feature amount of the target coin;
A coin discriminating method comprising: determining whether the target coin is true or false. Storing feature values obtained from the density histogram of the master coin;
Acquiring an image of the target coin using an imaging means;
Generating a density histogram from the image;
Extracting a feature amount of the target coin from the density histogram;
Comparing the feature amount of the master coin stored in the storage means with the feature amount of the target coin;
The recorded storage medium.

Images