JP2016004331A - Medal identification device and medal identification method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a medal identification device that can accurately identify a medal even when the medal has an individual difference.SOLUTION: A medal identification device 1 has: a medal region detection unit 61 for detecting, from an image generated by an image-capturing unit 4, a medal region in which the medal to be identified flowing down in a passage 6 is seen and the center of the medal region; an edge detection unit 62 for detecting an edge image in which the edge of the surface pattern of the medal to be identified is seen in the medal region; a similarity degree calculation unit 69 for selecting, for each partial region divided from the medal region, an edge image in which the ratio of the number of edge images to the area of the partial region is less than a prescribed ratio, aligning the center of a model region on a model image indicating the model of a prescribed medal and the center of the medal region, and calculating, on the basis of a match or mismatch of the selected edge image with the edge image in the model region, the similarity degree between the surface pattern of the model and the surface pattern of the medal to be identified; and a determination unit 70 for determining, in accordance with the similarity degree, whether or not the medal to be identified is a prescribed medal.

Description

  The present invention relates to a medal identification device and a medal identification method for identifying a medal using an image obtained by photographing a medal.

  Conventionally, when a medal (also called a coin) having a valuable value is inserted into a gaming machine such as a spinning machine, it is inserted to determine the number of times a player can play according to the medal. A medal identification device for identifying a medal is installed. In stores where such medals are used, medals having a store-specific pattern are used to prevent unauthorized use of medals from other stores. Therefore, the medal identification device is required to be able to identify a medal used in a store where a gaming machine equipped with the medal identification device is installed and a medal used in another store. In view of this, a coin identification device has been proposed that can determine that appropriate medals with slightly different surface shapes are appropriate (see, for example, Patent Document 1).

  The coin identification device disclosed in Patent Document 1 identifies an identification target coin based on imaging data obtained by imaging the surface of the identification target coin. At this time, the coin identifying device determines the feature part from the imaging data related to the first reference coin serving as a reference for identification, and then registers the first feature data related to the feature part, and the first reference coin and Registration of second feature data relating to a part corresponding to a feature part of imaging data relating to the second reference coin of the same type, identification data relating to a part corresponding to a feature part among imaging data relating to a coin to be identified; By comparing each of the second feature data, the identification target coin is identified.

JP 2010-160702 A

In the coin identifying device disclosed in Patent Document 1, two types of feature data are used for identifying medals. Therefore, if the identification target medal has a feature close to one of the two types of feature data, The coin identification device can accurately determine whether or not the identification target medal is appropriate.
However, the position, shape, and size of wrinkles, wear, or dirt generated on medals vary depending on the usage status of medals, and individual differences for each medal may increase. In such a case, there is a possibility that not all medals can be properly identified with the two types of feature data.

  Therefore, an object of the present invention is to provide a medal identification device and a medal identification method that can accurately identify medals even if medals have individual differences.

  As one aspect of the present invention, a medal identification device is provided. The medal identification device includes a passage forming unit that forms a passage through which a medal passes, an imaging unit that generates a part of the passage as a shooting range, and generates an image of the shooting range, and a model of a predetermined medal. The storage unit stores a model image including edge pixels corresponding to the pattern edges on the surface of the model in the model region, which is the region in which the model is shown, and the image generated by the imaging unit flows down in the passage A medal area that is an area in which the identification target medal is shown and a medal area detection unit that detects the center of the medal area, and an edge that detects an edge pixel in the medal area where the edge of the pattern of the identification target medal is reflected The detection unit and the medal area are divided into a plurality of partial areas, and for each of the partial areas, the number of edge pixels included in the partial area with respect to the area of the partial area Edge pixels are selected so that is less than a predetermined ratio, the center of the model area and the center of the medal area are aligned, and the selected edge pixels match one of the edge pixels in the model area The degree of similarity is calculated so that the similarity between the pattern on the surface of the model and the pattern on the surface of the identification medal increases as the number of is larger than the number of edge pixels that do not match any edge pixel in the model area And a determination unit that determines whether or not the identification target medal is a predetermined medal according to the similarity.

  In this medal identification device, the similarity calculation unit divides the medal region into a plurality of partial regions for each of a plurality of different division methods, and the number of edge pixels occupying the area of each partial region for each of the division methods. Edge pixels are selected so as to be less than a predetermined ratio, the similarity is calculated, the average value of the similarity for each division method is calculated, and the determination unit determines whether the identification target medal is a predetermined value according to the average value of the similarity It is preferable to determine whether or not it is a medal.

  In this case, the similarity calculation unit divides the medal area into a plurality of partial areas along the radial direction from the center of the medal area as one of a plurality of division methods, As another one, it is preferable to divide the medal area into a plurality of partial areas along the circumferential direction about the center of the medal area.

  According to another aspect of the present invention, there is provided a passage forming unit that forms a passage through which medals pass, and an imaging unit that is arranged so that a part of the passage is a shooting range and generates an image of the shooting range. A medal identification method in a medal identification device is provided. The medal identification method includes a step of detecting from the image a medal area that is an area where an identification target medal flowing down in a passage and a center of the medal area, and an edge of a pattern on the surface of the identification target medal in the medal area Detecting the edge pixel in which the image is shown, and dividing the medal area into a plurality of partial areas, and for each of the plurality of partial areas, the ratio of the number of edge pixels included in the partial area to the area of the partial area is The edge pixel is selected to be less than the predetermined ratio, and the center of the model area and the center of the medal area, which are the areas on the model image representing the model of the predetermined medal stored in the storage unit. , And the number of edge pixels that match one of the selected edge pixels in the model area is equal to any edge image in the model area. Calculating the similarity so that the similarity between the pattern on the surface of the model and the pattern on the surface of the identification target medal increases as the number of edge pixels that do not match each other increases, and the identification target medal according to the similarity Determining whether or not is a predetermined medal.

  The medal identification device and medal identification method according to the present invention have an effect of accurately identifying medals even if medals have individual differences.

1 is a schematic top view of a medal identification device according to one embodiment of the present invention. It is a schematic rear view of a medal identification device. FIG. 2 is a schematic cross-sectional view of the medal identification device when a cross section indicated by AA ′ in FIG. 1 is viewed from the direction of an arrow. FIG. 3 is a schematic cross-sectional view of the medal identification device when a cross-section indicated by BB ′ in FIG. 2 is viewed from the direction of an arrow. It is a circuit block diagram of a medal identification device. It is a functional block diagram of the control circuit of a medal identification device. It is a schematic diagram of a medal for samples. It is a figure which shows an example of edge distribution information. It is a figure which shows an example of the relationship between a rotation angle and a coincidence degree. It is a figure which shows an example of edge distribution information of a temporary model, and edge distribution information of the medal on the sample image to which attention is paid. It is an operation | movement flowchart of a rotation angle calculation process. It is a figure which shows an example of a model image. (A) And (b) is a figure which shows an example of the partial area | region respectively set to the medal area on a model image. It is an operation | movement flowchart of a model registration process. It is a figure which shows an example of the medal from which the position of the pattern was correct, and the medal from which the position of the pattern shifted | deviated. It is a figure which shows an example of the edge image by which the position correction was carried out. It is an operation | movement flowchart of a medal identification process. It is an operation | movement flowchart of a medal identification process. 1 is a schematic perspective view of a spinning machine including a medal identification device according to an embodiment of the present invention. 1 is a schematic perspective view of a spinning game machine including a medal identification device according to an embodiment of the present invention in a state where a front panel is opened. FIG. It is a circuit block diagram of a spinning machine. It is an external appearance perspective view of the medal counting device which is another form of a medal identification device. It is a perspective view of the state which removed the upper surface cover of the medal counting device. It is a three-dimensional view of a conveyance part. It is a top view of a conveyance part. It is a figure explaining the operation | movement at the time of conveying the medal in the guide vicinity. It is an exploded view of a conveyance part. It is explanatory drawing of a stopper pin. It is a circuit block diagram of a medal counting device.

  Hereinafter, a medal identification device according to an embodiment of the present invention will be described with reference to the drawings. This medal identification device uses an image of a medal to be identified obtained by an imaging unit that captures a medal flowing down a passage in the medal identification device, and the medal is approved for use. It is determined whether the medal is a good medal. In other words, the medal identification device includes an edge distribution in an area where a medal is shown in an image (hereinafter referred to as a medal area) and a medal area (a medal area on a model image that represents a model of an appropriate medal registered in advance). In the following, the similarity between the pattern on the surface of the identification target medal and the pattern on the surface of the model is calculated from the edge distribution in the model area), and the identification target medal is an appropriate medal according to the similarity. It is determined whether or not. At this time, this medal identification device equalizes the number of edge pixels in the medal area by adjusting the number of edge pixels corresponding to the edge of the pattern for each of the plurality of partial areas obtained by dividing the medal area of the identification target image. After that, the similarity is calculated. As a result, even if a part of the surface of the identification target medal is detected as an edge pixel, this medal identification device suppresses the influence on the similarity due to individual differences of the medal such as the wrinkle or dirt. it can.

  In the present embodiment, the outer shape of the medal is assumed to be substantially circular. Further, in this specification, a medal that is permitted to be used in a device in which the medal identification device is mounted is referred to as a proper medal. On the other hand, a medal that is not allowed to be used in a device on which the medal identification device is mounted is called an inappropriate medal. Further, in this specification, unless otherwise specified, the sides on both sides of the medal are not particularly distinguished and both are referred to as surfaces.

FIG. 1 is a schematic plan view of a medal identification device 1 according to one embodiment of the present invention, and FIG. 2 is a schematic rear view of the medal identification device 1. FIG. 3 is a schematic cross-sectional view of the medal identification device 1 when the cross section indicated by AA ′ in FIG. 1 is viewed from the direction of the arrow. FIG. 4 is a schematic cross-sectional view of the medal identification device 1 when the cross-section indicated by BB ′ in FIG. 2 is viewed from the direction of the arrow. FIG. 5 is a circuit block diagram of the medal identification device 1.
The medal identification device 1 includes a housing 2, a light source 3, an imaging unit 4, and a control board 5. The light source 3 and the imaging unit 4 are accommodated in the housing 2 and are connected to a control substrate 5 disposed on a substrate (not shown) provided outside the housing 2 via a signal line. Yes. Note that the control board 5 may also be housed inside the housing 2. Further, the control board 5 is connected via a signal line to a host control device (not shown) such as a main control circuit of a gaming machine on which the medal identification device 1 is mounted, and whether or not the medal is a proper medal is connected to the host control device. A notification signal representing the determination result is output.

  The housing 2 has, for example, the direction of the u-axis shown in FIG. 1 substantially parallel to the front surface of the spinning machine, and the direction of the w-axis shown in FIG. It is attached to the spinning cylinder game machine so that it is substantially parallel to. Therefore, in the following, for convenience, the u direction is referred to as the width direction and the w direction is referred to as the height direction. Further, the v-axis direction shown in FIG. 1 is referred to as a depth direction.

  As shown in FIG. 1, a medal slot 2 a is formed on the top surface of the housing 2. The medal slot 2a has a length along the width direction of the housing 2 plus a predetermined offset (for example, 1 mm to 2 mm) to the diameter of the medal, and the thickness of the medal in the depth direction of the housing 2 And a predetermined offset (for example, 0.5 mm to 1 mm).

  As illustrated in FIG. 3, the housing 2 is an example of a passage forming unit, and a passage 6 through which medals flow down is formed therein, and the light source 3 and the imaging unit 4 are accommodated. Therefore, the housing | casing 2 is comprised by combining the some components formed so that it might become a desired shape by shape | molding resin, for example. Or the housing | casing 2 may be formed with metals, such as iron, aluminum, copper, or the components formed with the metal and the components formed with resin may be combined.

  The passage 6 formed in the housing 2 is formed along the height direction on the upper end side of the housing 2, and is curved along the width direction of the housing 2 in the middle, and then obliquely toward the lower right. Is formed. The upper end of the passage 6 is directly connected to the medal slot 2 a, while the lower end of the passage 6 is directly connected to the medal outlet 2 b formed at the right end of the housing 2. Therefore, medals inserted from the medal insertion slot 2a flow down in the casing 2 along the passage 6 as indicated by an arrow, and then the medal identification apparatus 1 of the medal identification apparatus 1 from the medal discharge outlet 2b on the right side of the casing 2. Discharged outside. The flow direction of the medals indicated by the arrows is hereinafter referred to as the forward direction for convenience, and the direction from the medal discharge port 2b toward the medal insertion port 2a is referred to as the reverse direction.

  The passage 6 has a length obtained by adding a predetermined offset (for example, 1 mm to 2 mm) to the diameter of the medal along the width direction and the height direction of the housing 2 so that the medal can flow smoothly. A length obtained by adding a predetermined offset (for example, 0.5 mm to 1 mm) to the thickness of the medal along the depth direction of the housing 2.

  In addition, in order to reduce the contact area with the medal, a plurality of grooves may be formed on the side surface of the passage 6 along the flow direction of the medal. In addition, the member forming the side wall of the passage 6 is preferably formed of a material having a color different from the medal or having a different reflectance so that it can be easily distinguished from the medal on the image obtained by the imaging unit 4. . In general, since the medal is made of metal and has high light reflectance, the luminance value of the pixel included in the medal area where the medal is shown on the image is the luminance value of the pixel included in the background area where the medal is not shown. Higher than. Therefore, the member that forms the side wall of the passage 6 is configured to reduce the wavelength of light emitted from the light source 3 so that the difference between the luminance value of the pixel included in the medal area and the luminance value of the pixel included in the background area becomes large. It is preferable that a paint having absorbency is applied.

  As shown in FIGS. 3 and 4, below the curved portion of the passage 6, as will be described later, the housing 2 is more than the passage 6 so that the imaging unit 4 can photograph the medal M passing through the passage 6. There is a gap on the back side of the. Therefore, in order to prevent the medal M from deviating from the passage 6 in this gap, below the curved portion of the passage 6, the upper end and the lower end of the passage 6 are respectively closer to the rear side of the housing 2 than the passage 6. Further, guide members 7a and 7b for restricting the movement of medals in the depth direction are arranged. The distance between the lower end of the guide member 7a and the upper end of the guide member 7b is slightly smaller than the diameter of the medal M, for example, so that the imaging unit 4 can photograph most of the medal M. It is the value which subtracted 1mm-2mm from.

  The guide members 7a and 7b may be formed of a member that is transparent to the light from the light source 3, such as an optical resin or glass, so that the imaging unit 4 can capture the entire medal M.

As shown in FIG. 4, the light source 3 illuminates the medal M passing through the shooting range 4 a of the imaging unit 4 set in a section sandwiched between the guide members 7 a and 7 b of the passage 6, and It is arranged on the same side as the imaging unit 4 with respect to the passage 6 in the vicinity of the section so that the imaging unit 4 can capture the pattern of the medal M. Furthermore, the light source 3 includes, for example, a light emitting element such as one or a plurality of light emitting diodes arranged around the imaging unit 4 so that the photographing range 4a can be illuminated uniformly. Furthermore, a plate-like diffusion sheet that diffuses and emits light from the light emitting element toward the passage 6 may be provided on the front surface of the light emitting element.
The light source 3 illuminates the passage 6 while the medal identification device 1 is operating. Alternatively, the light source 3 may be controlled by the control board 5 so as to emit light only at the timing when the imaging unit 4 captures an image.

  The imaging unit 4 shoots a medal M passing through the shooting range 4a set in a part of the passage 6, and generates an image in which the image of the medal M is reflected. Therefore, the imaging unit 4 is disposed on the back side of the passage 6 in the housing 2 so as to be positioned on the same side as the light source 3 with respect to the passage 6 below the curved portion of the passage 6. The imaging unit 4 includes, for example, an image sensor such as a CCD image sensor or a C-MOS image sensor arranged in a two-dimensional manner, a partial section of the passage 6 included in the photographing range 4a on the image sensor, and the section thereof. An imaging optical system that forms an image of the passing medal M. In addition, the imaging unit 4 has, for example, horizontal 320 × vertical 240 pixels so that the pattern of the surface of the medal M can be identified on the image. The luminance value of each pixel is represented by, for example, 0 to 255, and the larger the luminance value, the higher the luminance.

  The imaging unit 4 continuously photographs the imaging range 4a at predetermined time intervals while the medal identification device 1 is operating. The predetermined time interval is set to, for example, an interval of 50 msec to 100 msec so that at least one image is generated while the medal inserted into the medal identification device 1 passes through the shooting range 4a. . The imaging unit 4 outputs the generated image to the control board 5 every time it captures an image.

  Further, a sensor for detecting the passage of medals may be provided closer to the medal slot than the shooting range 4a. This sensor can be, for example, an optical sensor having a light source and a light receiving element arranged so as to face each other across a passage, or a magnet type proximity sensor. Then, the control board 5 causes the imaging unit 4 to perform shooting when a predetermined period has elapsed after receiving the detection signal from the sensor, so that the entire medal M is surely represented on one image. You may do it.

  As shown in FIG. 5, the control board 5 includes a control circuit 51, a memory 52, and a communication interface 53. The control circuit 51 includes, for example, a processor and its peripheral circuits. Then, the control circuit 51 analyzes the image received from the imaging unit 4 to determine whether or not the identification target medal passing through the passage 6 is an appropriate medal.

  The memory 52 is an example of a storage unit, and includes, for example, a nonvolatile read-only semiconductor memory and a volatile read / write semiconductor memory. The memory 52 stores a medal identification process program executed by the control circuit 51, parameters used in the medal identification process, various intermediate calculation values calculated during the medal identification process, and the like. Furthermore, the memory 52 stores a model image representing a model of a proper medal. Details of the model image will be described later. The memory 52 may store the latest plurality of images among the images received from the imaging unit 4.

  The communication interface 53 is an interface for connecting to a host control device (not shown), the light source 3 and the imaging unit 4, receives a control signal to the light source 3 or the imaging unit 4 from the control circuit 51, and receives the light source 3. Alternatively, the image is output to the imaging unit 4. The communication interface 53 receives various control signals from the host control device via the signal line and passes them to the control circuit 51. Further, the communication interface 53 receives a notification signal representing the medal identification result from the control circuit 51 and outputs the notification signal to the host control device.

  FIG. 6 is a functional block diagram of the control circuit 51. The control circuit 51 includes a medal area detection unit 61, an edge detection unit 62, a rotation angle calculation unit 63, a model generation unit 64, a model edge number adjustment unit 65, a shape correction unit 66, and an edge position correction unit 67. A direction coincidence calculation unit 68, a similarity calculation unit 69, and a determination unit 70. Each of these units included in the control circuit 51 is, for example, a functional module realized by a computer program executed on the control circuit 51.

  Among these units included in the control circuit 51, the medal area detection unit 61 and the edge detection unit 62 are used in the medal identification process for determining whether or not the identification target medal is an appropriate medal and the medal identification process. It is used in both model registration processes for generating a model image in which a model of a proper medal is represented. On the other hand, the rotation angle calculation unit 63, the model generation unit 64, and the model edge number adjustment unit 65 are used for model registration processing. The shape correction unit 66, the edge position correction unit 67, the orientation coincidence calculation unit 68, the similarity calculation unit 69, and the determination unit 70 are used in the medal identification process.

  Hereinafter, the process of each unit included in the control circuit 51 will be described separately for the model registration process and the medal identification process.

(Model registration process)
In the model registration process, a plurality of sample medals are used to generate a model image. In addition, in a store where a gaming machine equipped with the medal identification device 1 is installed, when a medal used by the store is an appropriate medal, a medal actually used in the store is used as a sample medal in the model registration process. Used. At this time, if the number of sample medals is large, improper medals that are not used in the own store, for example, medals used in other stores may be mixed into the sample medals. Therefore, in this embodiment, the medal identification device 1 generates a model image according to the following procedure, so that a small number of inappropriate medals are included among a plurality of sample medals used for generating a model image. An appropriate model image can be generated even if mixed.

  When starting the model registration process, for example, the control circuit 51 receives a control signal indicating that an image for model registration is to be captured from a higher-level control device (not shown). The camera enters a standby state for taking an image. When a medal is inserted into the medal identification device 1 in this state, the control circuit 51 stores the image obtained by the imaging unit 4 in the memory 52 as a sample image for model registration processing. Thereafter, when the control circuit 51 receives a control signal for starting the model registration process from the host control device, the control circuit 51 starts the model registration process.

Even with the same medal, if the orientation of the pattern on the surface of the medal with respect to the light source 3 and the imaging unit 4 is different, the illumination direction with respect to the pattern is different, so that the shading caused by the pattern on the surface of the medal is also different. For this reason, the appearance of medals displayed on the image is also different. Therefore, in this embodiment, each of the plurality of sample medals is inserted into the medal identification device 1 a plurality of times. Thereby, about each sample medal, the some image image | photographed in the state from which the direction of the pattern of the surface with respect to the light source 3 and the imaging part 4 differs can be obtained.
In the present embodiment, for example, each of hundreds to tens of thousands of sample medals is inserted into the medal identification device 1 about several to thirty times. Accordingly, the number of sample images obtained by multiplying the number of sample medals by the number of insertions of individual medals is obtained. The control circuit 51 executes a model registration process using these sample images.

  FIG. 7 is a schematic diagram of a sample medal. In the medal 700, a pattern 701 is engraved on the surface thereof. Therefore, a feature for identifying a medal is obtained using a shadow corresponding to the pattern 701 on the image. In the present embodiment, the control circuit 51 is characterized in that it detects edge pixels corresponding to shaded edges corresponding to the pattern on the medal surface and identifies the medal with the edge distribution.

The medal area detection unit 61 extracts a medal area where a medal is shown from each sample image, and obtains the coordinates and radius of the center of the medal area.
For example, the medal area detection unit 61 sets a plurality of scanning lines for the sample image of interest along the horizontal direction and the vertical direction on the image, respectively. Then, the medal area detection unit 61 calculates the luminance value of the pixel adjacent to the image end side with respect to the pixel of interest from the luminance value of the pixel of interest in order from the image end to the image center on each scanning line. Subtract the luminance difference value. The medal area detection unit 61 sets a pixel of interest when the luminance difference value first exceeds a predetermined threshold as a pixel candidate corresponding to the outer periphery of the medal (hereinafter referred to as an outer peripheral candidate pixel). It should be noted that the medal area detection unit 61 does not have to obtain the outer peripheral candidate pixels on a scanning line whose luminance difference value does not exceed a predetermined threshold value for any pixel on the scanning line.

  When the medal area detection unit 61 obtains outer periphery candidate pixels for each scanning line, the medal region detection unit 61 selects any three outer periphery candidate pixels from the plurality of outer periphery candidate pixels and sets them as a set of outer periphery candidate pixels. The medal area detection unit 61 obtains a predetermined number of sets of outer periphery candidate pixels by changing the outer periphery candidate pixels to be selected. The predetermined number is set to an integer of 2 or more, for example, 12. For each set of outer periphery candidate pixels, the medal area detection unit 61 calculates an equation of a circle that passes through the three outer periphery candidate pixels included in the set, and determines the center of the circle represented by the circle equation. The medal area detection unit 61 obtains the center having the largest number of matches among the circle centers of each group as a medal center candidate. If the distance between the two centers of interest is less than one pixel, the medal area detection unit 61 determines that the two centers match. The medal area detection unit 61 determines the number of coincident circle centers in the medal center candidates and the outer circumference of the medal candidates obtained from the average value of the circle radii centered on the medal center candidates and coincident with the medal center candidates. The number of outer peripheral candidate pixels located at is counted. The medal area detection unit 61 stores in the memory 52 the coordinates of the medal center candidate, the number of coincident circle centers in the medal center candidate, and the number of outer peripheral candidate pixels on the outer periphery of the medal candidate counted for the medal center candidate. To do.

  The medal region detection unit 61 repeats a plurality of times by changing the value of the predetermined threshold for detecting the outer periphery candidate pixels and executing the same processing as described above. For example, the predetermined threshold value is first set to 50, and thereafter, the threshold value is incremented by five. The medal area detection unit 61 then sets the number of coincidence circle centers on the outer circumference of the medal candidates with respect to the total number of outer peripheral candidate pixels among medal center candidates whose number is equal to or greater than a predetermined value (for example, half of the set of outer peripheral candidate pixels). The medal center candidate having the largest ratio of the number of outer peripheral candidate pixels located at is set as the center of the medal area, and the radius obtained for the medal center candidate is set as the radius of the medal area.

  The medal area detection unit 61 is not limited to the above example, and extracts the medal area from the sample image by using any of various processes for extracting a predetermined object shown on the image from the image. May be. For example, on the sample image, when the difference between the brightness of the area where the medal is shown and the brightness of the background area where the medal is not shown is large, the medal area detection unit 61 determines the brightness value of each pixel of the sample image as a predetermined value. A set of pixels larger than the predetermined threshold may be detected as a medal area. Then, the medal area detection unit 61 may use the center of gravity of the detected medal area as the center of the medal area and the average value of the distance from the center of the medal area to the outer periphery of the medal area as the radius of the medal area.

  The medal area detection unit 61 stores the coordinates of the center of the medal area and the radius of the medal area obtained for each sample image in the memory 52 in association with the sample image.

  The edge detection unit 62 detects, for each sample image, an edge pixel in the medal area that represents a shadowed edge caused by the pattern on the medal surface.

  In the present embodiment, the edge detection unit 62 calculates an edge strength by applying an edge detection filter to each pixel in the medal area, and detects a pixel whose edge strength is equal to or greater than a predetermined edge threshold as an edge pixel. To do. The edge detection unit 62 preferably uses a second-order differential filter such as a Laplacian filter as the edge detection filter. In this case, the edge threshold is set to 200, for example. As a result, even when the luminance change changes across a plurality of pixels due to a gentle change in the pattern on the surface of the medal, the edge detection unit 62 selects pixels having a large difference in luminance change between neighboring pixels among the plurality of pixels. It can be detected as an edge pixel.

  For each sample image, the edge detector 62 sets the edge pixel value to '1', the pixel values other than the edge pixel to '0', polar coordinates with the center of the medal area as the origin, and the clockwise direction as positive. Edge distribution information representing each edge pixel is generated. Then, the edge detection unit 62 stores the generated edge distribution information in the memory 52 in association with the original sample image.

  FIG. 8 is a diagram illustrating an example of edge distribution information. In the edge distribution information 800, the coordinates of each pixel are represented by an angle θ with respect to a reference direction and a distance L from the center of the medal area. The value corresponding to the edge pixel is “1”, and the coordinates corresponding to the other pixels are “0”. Note that the reference direction may be arbitrarily determined, and may be, for example, the horizontal direction or the vertical direction of the sample image.

  Based on the edge distribution information of each sample image, the rotation angle calculation unit 63 determines, for each sample image, the orientation of the pattern on the surface of the medal shown in the sample image with respect to the orientation of the surface pattern in the temporary model of the medal. Calculate the rotation angle. The rotation angle is used to align each sample image in the rotation direction with the center of the medal area as the rotation center.

ここで、NumCは、回転角θだけ着目するサンプル画像のエッジ分布情報を回転させたときの、サンプル画像のエッジ画素のうち、仮モデルのエッジ画素と一致するエッジ画素の数であり、NumEは、着目するサンプル画像のエッジ画素の総数である。すなわち、NumEは、着目するサンプル画像のエッジ画素のうちで仮モデルの何れのエッジ画素とも一致しないエッジ画素の総数NumDとNumCの和である。エッジ分布情報に含まれる各エッジ画素の座標が極座標系で表されているので、サンプル画像に表されたメダルの各エッジ画素を回転角θだけ回転させるには、回転角算出部６３は、そのサンプル画像についてのエッジ分布情報に含まれる各エッジ画素の座標の角度値を回転角θだけずらせばよい。 In the present embodiment, the rotation angle calculation unit 63 uses a medal shown in one of a plurality of sample images as a temporary model. Then, the rotation angle calculation unit 63 sets the other sample images one by one as the sample image of interest. The rotation angle calculation unit 63 sets the relative rotation angle of the edge distribution information to a predetermined angle with respect to the edge distribution information obtained from the sample image corresponding to the temporary model, with the center of the medal region of the sample image of interest as the rotation center. The degree of coincidence representing the degree of coincidence of the edge distribution is calculated while being changed (for example, once). In the present embodiment, the rotation angle calculation unit 63 calculates the degree of coincidence C (θ) at the rotation angle θ according to the following equation.

Here, NumC is the number of edge pixels that match the edge pixels of the temporary model among the edge pixels of the sample image when the edge distribution information of the sample image of interest is rotated by the rotation angle θ, and NumE is , The total number of edge pixels of the sample image of interest. That is, NumE is the sum of the total number of edge pixels NumD and NumC that do not match any of the edge pixels of the temporary model among the edge pixels of the sample image of interest. Since the coordinates of each edge pixel included in the edge distribution information are expressed in the polar coordinate system, the rotation angle calculation unit 63 is used to rotate each edge pixel of the medal represented in the sample image by the rotation angle θ. The angle value of the coordinates of each edge pixel included in the edge distribution information for the sample image may be shifted by the rotation angle θ.

  FIG. 9 is a diagram illustrating an example of the relationship between the rotation angle and the degree of coincidence. In FIG. 9, the horizontal axis represents the rotation angle, and the vertical axis represents the degree of coincidence. A graph 901 shown on the upper side represents the relationship between the rotation angle and the degree of coincidence when the medal pattern shown in the sample image of interest and the medal pattern represented by the temporary model are the same. On the other hand, the graph 902 represents the relationship between the rotation angle and the degree of coincidence when the medal pattern shown in the sample image of interest differs from the medal pattern represented by the temporary model.

  If the medal pattern shown in the sample image of interest and the medal pattern represented by the temporary model are the same, the position of each edge pixel on the sample image at the rotation angle where the orientation of the pattern represented by the medal matches And the position of each edge pixel in the provisional model almost coincide. Therefore, the degree of coincidence C (θ) is a high value. On the other hand, at other rotation angles, as long as the pattern shown on the medal is not rotationally symmetric, there are few edge pixels on the sample image at the same position as any edge pixel in the temporary model. Therefore, the degree of coincidence C (θ) is uniformly low regardless of the rotation angle θ. On the other hand, if the medal pattern shown in the sample image of interest differs from the medal pattern represented by the temporary model, any edge in the temporary model among the edge pixels on the sample image at any rotation angle. Since there are few edge pixels at the same position as the pixels, the degree of coincidence C (θ) is uniformly low regardless of the rotation angle θ.

  Therefore, as shown in the graph 901, if the medal pattern shown in the sample image of interest and the medal pattern represented by the temporary model are the same, the maximum matching value C1 and the second largest matching score are obtained. The difference between the maximum values C2 becomes large. In such a case, the rotation angle θ1 corresponding to the degree of coincidence C1 corresponds to the direction difference of the pattern orientation on the surface of the medal represented in the sample image of interest with respect to the orientation of the pattern on the medal surface represented by the temporary model. Presumed.

  On the other hand, as shown in the graph 902, if the medal pattern shown in the sample image of interest is different from the medal pattern represented by the temporary model, the maximum matching value C1 and the second largest matching score are obtained. The difference between the maximum values C2 is small, and C1 / C2 is close to 1. For this reason, it is unclear which rotation angle corresponds to the direction difference between the orientation of the pattern on the surface of the medal represented in the sample image of interest and the orientation of the pattern on the surface of the medal represented by the temporary model.

Therefore, the rotation angle calculation unit 63 calculates the second rotation angle θ1 corresponding to the maximum value C1 of the coincidence among the maximum values of the coincidence C (θ) in the function of the coincidence C (θ) with respect to the rotation angle θ. And a rotation angle θ2 corresponding to the maximum value C2 of the degree of coincidence greater than. Then, the rotation angle calculation unit 63 compares the ratio C1 / C2 of the maximum coincidence value to the maximum value of the second largest coincidence degree with a predetermined coincidence ratio threshold value. If the ratio C1 / C2 is equal to or greater than the coincidence ratio threshold, the rotation angle calculation unit 63 sets the rotation angle θ1 corresponding to the maximum coincidence value C1 for the sample image to the pattern orientation on the surface of the temporary model. Is stored in the memory 52 in association with the sample image as the rotation angle of the pattern orientation of the surface of the medal on the sample image. Further, the rotation angle calculation unit 63 stores a flag indicating that the rotation angle has been calculated in the memory 52 in association with the sample image.
Further, the rotation angle calculation unit 63 rotates the edge distribution information about the sample image by the rotation angle θ1, and sets the value “1” of each edge pixel included in the rotated edge distribution information to the corresponding pixel of the temporary model. Add to the value.

FIG. 10 is a diagram illustrating an example of edge distribution information of a temporary model and edge distribution information of medals on a sample image of interest. As described above, in the edge distribution information, each edge pixel is displayed in a polar coordinate system, but here, in order to facilitate understanding, each edge pixel is represented in an orthogonal coordinate system.
In this example, edge distributions 1001 to 1004 in units of pixels when the sample image of interest is not rotated and when rotated by 90 degrees, 180 degrees, and 270 degrees are shown. In each edge distribution, pixels that are not 0 are edge pixels. In this example, when the sample image of interest is rotated by 90 degrees, the edge distribution and the edge distribution 1000 of the temporary model most closely match. Therefore, the rotation angle calculation unit 63 adds “1” to the pixel corresponding to each edge pixel when the sample image of interest is rotated 90 degrees in the edge distribution 1000 of the temporary model.

  On the other hand, if the ratio C1 / C2 for the sample image of interest is less than the coincidence ratio threshold, the rotation angle calculation unit 63 sets the coincidence maximum value C1 and a flag indicating that it has not been added to the provisional model, The sample image is stored in the memory 52 in association with the sample image.

  The rotation angle calculation unit 63 repeats the above processing for all sample images. As a result, the edge pixels of the sample image in which medals of the same type as the medals represented in the temporary model are sequentially added to the temporary model. Therefore, in the temporary model, a pixel having a higher possibility of having an edge has a larger value. Become.

  However, in general, even if the edge distribution information is rotated by the rotation angle θ1 and aligned with the temporary model for each sample image due to lighting conditions, the orientation of the medal at the time of shooting, or the scratch on the medal itself, the sample Not all edge pixels on the image match the edge pixels included in the temporary model. As a result, as the number of sample images added to the temporary model increases, the proportion of pixels having values other than “0” in the temporary model, that is, edge pixels increases. For example, referring to FIG. 10 again, when “1” is added to the pixel corresponding to each edge pixel when the sample image of interest is rotated 90 degrees in the edge distribution 1000 of the temporary model, The pixel value changes from “0” to “1”, and there is a pixel that becomes a new edge pixel. If the number of edge pixels included in the temporary model is too large, the degree of coincidence cannot be calculated accurately.

  Therefore, the rotation angle calculation unit 63 adds a predetermined number (for example, 10 to 100) of sample images to the temporary model so that the ratio of the edge pixels among the pixels included in the medal area of the temporary model does not become too high. Each time an edge pixel is added, a certain value is subtracted from each pixel of the temporary model. At this time, the rotation angle calculation unit 63 sets the pixel value to “0” for a pixel that has become a negative value due to the subtraction. Note that the fixed value can be, for example, half of the predetermined number.

  Alternatively, the rotation angle calculation unit 63 adjusts the threshold value so that the ratio of the edge pixels in the medal area of the temporary model is equal to or less than a predetermined value (for example, 0.4), and each threshold value in the medal area of the temporary model is adjusted with the threshold value. The pixel value may be binarized. At the time of binarization, the rotation angle calculation unit 63 sets a value representing an edge pixel, for example, “1” for a pixel having a pixel value equal to or greater than a threshold value. A pixel whose value is less than the threshold is set to “0” indicating that it is not an edge pixel.

  After executing the above processing for all the sample images, the rotation angle calculation unit 63 refers to the flag, and for the sample images in which the edge pixels are not added to the temporary model, from the one with the highest matching value C1 Sort in order. Then, the rotation angle calculation unit 63 repeats the above processing for those sample images. As a result, in the second and subsequent processes, there is a high possibility that the temporary model can more accurately represent the edge distribution of the appropriate medals. Therefore, in the previous process, the rotation angle for the temporary model was not determined. The possibility that the rotation angle with respect to the temporary model is determined for the image is also increased. At this time, the coincidence ratio threshold value may be reduced by a predetermined amount (for example, 0.02). Then, the rotation angle calculation unit 63 ends the process when the edge pixels are added to the temporary model for all the sample images or when the coincidence ratio threshold value decreases to the lower limit value.

  Note that when the sample image initially adopted as the temporary model is an image in which an inappropriate medal is captured, the number of sample images in which (C1 / C2) is equal to or greater than the coincidence ratio threshold is very small. Therefore, when the first rotation angle is calculated for all the sample images, the number of sample images for which the rotation angles are determined is equal to or less than a certain number (for example, 1/100 of the total number of sample images). The rotation angle calculation unit 63 may replace the sample image as the temporary model first and execute the series of processes again.

Since C1 / C2 has a minimum value of 1 from its definition, by setting the lower limit value of the coincidence ratio threshold value to a value somewhat larger than 1, for example, 1.5 to 2.0, the rotation angle calculation unit 63 If the sample medal includes an improper medal, or an appropriate medal with a high degree of wrinkle or wear, a sample image showing the improper medal or the appropriate medal with a high degree of wrinkle or wear is created from the model creation. Can be excluded.
On the other hand, the rotation angle calculation unit 63 can use almost all sample images for model creation by setting the lower limit of the coincidence ratio threshold to 1 or a value close to 1, for example, 1.0 to 1.1. Alternatively, even if the rotation angle calculation unit 63 reduces the coincidence ratio threshold to the lower limit value, the rotation angle calculation unit 63 also performs s coincidence for the sample image in which (C1 / C2) is less than the coincidence ratio threshold. The rotation angle θ1 corresponding to the maximum degree C1 is stored in the memory 52 in association with the sample image as the rotation angle of the pattern orientation of the surface of the medal on the sample image with respect to the pattern orientation of the surface of the temporary model. May be.

FIG. 11 is an operation flowchart of the rotation angle calculation process.
The rotation angle calculation unit 63 changes the rotation angle of the pattern orientation on the surface of the medal on the sample image to be focused with respect to the pattern orientation on the surface of the temporary model, and the maximum value C1 of the edge distribution coincidence and the second maximum. A value C2 is calculated (step S101).

  The rotation angle calculation unit 63 determines whether the ratio (C1 / C2) of the matching degree maximum value C1 to the second maximum value C2 is equal to or greater than the matching degree ratio threshold Th (step S102). When the ratio (C1 / C2) is equal to or greater than the coincidence ratio threshold Th (step S102—Yes), the rotation angle calculation unit 63 determines the orientation of the pattern on the surface of the medal on the sample image with respect to the orientation of the pattern on the surface of the temporary model. Is set to a rotation angle θ1 corresponding to C1, and each edge pixel on the sample image is rotated by the rotation angle θ1 to obtain a temporary model corresponding to each edge pixel on the sample image. '1' is added (step S103). Further, the rotation angle calculation unit 63 may adjust the number of edge pixels included in the temporary model.

  After step S103 or when the ratio (C1 / C2) is less than the coincidence ratio threshold value Th in step S102 (step S102-No), the rotation angle calculation unit 63 determines whether a sample image that has not been observed remains. It is determined whether or not (step S104). If an unfocused sample image remains (step S104-Yes), the rotation angle calculation unit 63 sets the unordered sample image in the next order as the focused sample image (step S105). And the rotation angle calculation part 63 repeats the process after step S101.

  On the other hand, if there is no unfocused sample image remaining (step S104-No), the rotation angle calculation unit 63 determines whether or not the coincidence ratio threshold value Th has reached the lower limit value Thmin (step S106). If the coincidence ratio threshold value Th has reached the lower limit value Thmin (step S106—Yes), the rotation angle calculation unit 63 ends the rotation angle calculation process even if an unadded sample image remains in the temporary model.

  On the other hand, if the coincidence ratio threshold value Th has not reached the lower limit value Thmin (No in step S106), the rotation angle calculation unit 63 decreases the coincidence degree ratio threshold value by a predetermined amount δ (step S107). Then, the rotation angle calculation unit 63 determines whether or not an unadded sample image remains in the temporary model (step S108).

  If unadded sample images remain (step S108—Yes), the rotation angle calculation unit 63 rearranges the unadded sample images in descending order of C1 (step S109). Then, the rotation angle calculation unit 63 repeats the processing from step S101 onwards, taking the sample image with the largest C1 among the sample images that have not been added as the sample image to which attention is paid and the rest as the sample images that have not been noticed.

  On the other hand, if no unadded sample image remains (step S108-No), the rotation angle calculation unit 63 ends the rotation angle calculation process.

  The model generation unit 64 generates a model image representing a medal model used for medal identification processing by cumulatively adding edge pixels in the medal area of each sample image for which the rotation angle θ1 is calculated.

The model generation unit 64 performs edge detection filter processing on each sample image for which the rotation angle θ1 is calculated, similarly to the edge detection unit 62, and calculates the edge intensity of each pixel in the medal area. Then, the model generation unit 64 performs binarization processing in which each pixel in the medal area has a pixel having an edge strength equal to or higher than the binarization threshold while changing the binarization threshold for the edge strength. The ratio of edge pixels included in the medal area is set to be less than a predetermined value (for example, 0, 4). For each sample image for which the rotation angle θ1 has been calculated by this binarization processing, the value of the edge pixel in the medal area is set to “1”, and the value of the pixel that is outside the medal area or is not an edge pixel in the medal area. A binary image with “0” is obtained.
The model generation unit 64 rotates each binary image by the rotation angle θ1 obtained for the corresponding sample image with the center of the medal area as the rotation center, and then aligns each binary image at the center of the medal area. By adding the pixel values, a model image representing a model of a proper medal is generated. Therefore, on the model image, the larger the number of sample images determined as edge pixels, the larger the value.

  Finally, the model generation unit 64 detects the maximum pixel value Pmax on the model image and multiplies the value of each pixel by (255 / Pmax) so that the maximum pixel value becomes 255. Then, the value of each pixel on the model image is normalized.

  FIG. 12 is a diagram illustrating an example of the obtained model image. In the model image 1200, a pixel determined as an edge pixel in many sample images has a larger value. Therefore, the individual characteristics of the medal for the sample shown in each sample image, for example, the edge due to the eyelid is not reflected in the model of the appropriate medal represented on the model image 1200. Edges with patterns common to each medal are emphasized. Further, even if some sample medals are contaminated and a part of the pattern is not visible, the model image 1200 can be used if an edge corresponding to a part of the pattern is detected in another sample medal. The edge is also expressed above.

The model generation unit 64 sets the center of the medal region in the sample image that is the reference of the two sample images that are initially aligned as the center of the medal region in the model image. Then, the model generation unit 64 sets the average value of the radius of the medal area in each sample image used for generating the model image as the radius of the medal area in the model image.
The model generation unit 64 stores the model image, the coordinates of the center of the medal area on the model image, and the radius of the medal area in the memory 52.

  The model edge number adjustment unit 65 adjusts the number of edge pixels so that the ratio of the number of edge pixels having one or more pixel values in the medal area of the appropriate model represented in the model image is equal to or less than a predetermined value. To do.

  In the present embodiment, the model edge number adjustment unit 65 converts the medal area into a circle using the center of the medal area on the model image as a reference point so that the number of edge pixels for each local area in the medal area is equalized. A predetermined number of partial areas are divided along the circumferential direction. As shown in FIG. 13A, each partial area 1301 set in the medal area 1300 is a fan-shaped area. The predetermined number is, for example, 36 (that is, the central angle of each partial region is 10 degrees). Then, the model edge number adjustment unit 65 calculates, for each partial region, the ratio of the number of edge pixels in the area of the partial region. When the ratio is 0.8 or more, the model edge number adjustment unit 65 subtracts 1 from the value of each edge pixel. Note that for a pixel having a pixel value of 0, the model edge number adjustment unit 65 excludes the pixel from the edge pixel. The model edge number adjustment unit 65 repeats the above processing for each partial region until the ratio of the number of edge pixels to the area of the partial region is less than 0.8.

When the above processing is completed for all the partial areas divided in the circumferential direction, the model edge number adjustment unit 65 converts the medal area into a plurality of partial areas along the radial direction from the center of the medal area on the model image. Divide into That is, as shown in FIG. 13B, each partial area 1302 set in the medal area 1300 is a ring-shaped area except for the most central circular partial area. Moreover, the width along the radial direction of each partial region is set to 2 pixels, for example.
The model edge number adjusting unit 65 performs the same process as that performed on the fan-shaped partial area for each partial area. At that time, the model edge number adjustment unit 65 reduces the value of the edge pixel until the ratio of the number of edge pixels in the partial region becomes less than 0.5.
As a result, the edge pixels are distributed almost uniformly throughout the medal area on the model image. Further, the ratio of the number of edge pixels to the area of the medal area is less than 0.4 (= 0.8 × 0.5).

  Note that the model edge number adjusting unit 65 may perform the adjustment processing of the number of edge pixels on each of the ring-shaped partial areas and then execute the process on each of the fan-shaped partial areas. In addition, the model edge number adjusting unit 65 may set the upper limit value of the ratio of the number of edge pixels in each partial region to a value other than 0.5 or 0.8. However, in the medal identification process described later, in order to appropriately compare the image showing the identification target medal with the model image, the edge pixel occupying the area of the medal region in the finally obtained model image. The number of edge pixels is preferably adjusted so that the ratio of the numbers is about 0.3 to 0.5.

  The model edge number adjustment unit 65 stores the model image in which the number of edge pixels is adjusted in the memory 52 as a model image used in the medal identification process. Further, the model edge number adjusting unit 65 generates edge distribution information of the model image in which the position of each edge pixel is expressed in a polar coordinate system with the center of the medal area as the origin, and stores the edge distribution information in the memory 52 as well. .

FIG. 14 is an operation flowchart of the model registration process.
The medal area detection unit 61 detects a medal area from each sample image (step S201). Then, the edge detection unit 62 detects edge pixels in the medal area of each sample image (step S202).

  The rotation angle calculation unit 63 obtains the rotation angle of the medal on the sample image with respect to the temporary model of the appropriate medal for each sample image (step S203). Then, the model generation unit 64 rotates each edge pixel on the sample image by the corresponding rotation angle, aligns it at the center of the medal area, and sets “1” for each pixel that becomes the edge pixel. A model image representing a model of a proper medal is generated by adding and normalizing the value of each pixel so that the maximum pixel value becomes a predetermined value (step S204).

  After the model image is generated, the model edge number adjustment unit 65 divides the medal area on the model image into a plurality of partial areas along the circumferential direction with the center as a reference point, and is included in each partial area. The number of edge pixels is adjusted so that the ratio of the number of edge pixels to the area of the partial region is less than a predetermined ratio (step S205). Further, the model edge number adjustment unit 65 divides the medal region on the model image into a plurality of partial regions along the radial direction from the center, and determines the number of edge pixels included in each partial region as the area of the partial region. The ratio of the number of edge pixels occupying is adjusted to be less than a predetermined ratio (step S206). Then, the model edge number adjusting unit 65 stores the model image in which the number of edge pixels is adjusted in the memory 52, and ends the model registration process.

(Medal identification process)
Next, a medal identification process for determining whether or not an identification target medal is a proper medal will be described. The control circuit 51 of the medal identification device 1 identifies the inserted medal every time a medal is inserted into the medal identification device 1, for example, when a gaming machine equipped with the medal identification device 1 is operating. The medal identification process is executed as the target medal. At that time, the control circuit 51 identifies a medal using an image (hereinafter referred to as an identification target image) obtained by the imaging unit 4 shooting the medal while the inserted medal is in the shooting range of the imaging unit 4. Execute the process.

  The medal area detection unit 61 performs a process similar to the process performed on the sample image for the identification target image, and detects a medal area where the identification target medal is shown on the identification target image. Then, the medal area detection unit 61 passes the coordinates of the center of the detected medal area and the coordinates of each outer peripheral candidate pixel used when obtaining the coordinates of the center to the shape correction unit 66.

  Depending on the posture of the medal when the medal passes through the passage 6, the surface of the medal may be inclined without being orthogonal to the optical axis of the imaging unit 4. Further, depending on the width of the passage 6, the distance from the imaging unit 4 to the medal may not be constant. In such a case, the medal may be elliptical on the identification target image, or the size of the medal area may be different from the size of the medal area on the model image. Therefore, the shape correcting unit 66 corrects the distortion of the shape of the medal area on the identification target image. Further, the shape correcting unit 66 matches the size of the medal area on the model image with the size of the medal area on the identification target image.

  For this purpose, the shape correcting unit 66 sets an arbitrary straight line that passes through the center of the medal area and faces in various directions on the identification target image, and two opposing faces across the center of the medal area on the straight line. A peripheral candidate pixel is specified. Then, the shape correcting unit 66 sets the distance between the two outer peripheral candidate pixels as the diameter of the medal area in that direction. In this way, the shape correcting unit 66 obtains the diameter of the medal area along the straight line in each direction. Then, the shape correcting unit 66 obtains the longest diameter among the diameters in the respective directions as the major axis. The shape correcting unit 66 identifies two outer peripheral candidate pixels that are orthogonal to the direction corresponding to the major axis and that are opposed to each other across the center of the medal area on a straight line passing through the center of the medal area, and the two outer peripheral candidates. The distance between the pixels is the minor axis of the medal area. When the ratio of the minor axis to the major axis is a predetermined threshold (for example, 0.95) or less, or the difference between the major axis and the minor axis is a predetermined threshold (for example, three pixels) or more, the shape correcting unit 66 Interpolation processing is performed along the minor axis direction for each pixel in the medal area of the identification target image with the center of the medal area as the origin so that the diameter becomes equal to the major axis. Thereby, the shape of the medal area becomes substantially circular.

  Further, the shape correction unit 66 calculates the ratio of the major axis of the medal area on the identification target image to the radius of the medal area on the model image. If the ratio is larger than a first threshold value (for example, 1.05 to 1.1) that is larger than 1.0 by a predetermined amount, the shape correcting unit 66 determines that the medal area on the identification target image has a major axis of the medal area on the model image. Interpolation processing is performed along the horizontal and vertical directions for each pixel in the medal area of the identification target image with the center of the medal area as the origin so as to be equal to the radius of. The shape of the medal area on the identification target image with respect to the radius of the medal area on the model image is smaller than a second threshold (for example, 0.95 to 0.9) smaller than 1.0 by a predetermined amount. The correction unit 66 sets the center of the medal area as the origin and sets each pixel in the medal area of the identification target image so that the major axis of the medal area on the identification target image becomes equal to the radius of the medal area on the model image. Interpolation processing is performed along the horizontal and vertical directions.

  The shape correction unit 66 stores the shape-corrected identification target image in the memory 52 as an identification target image again.

  The edge detection unit 62 performs an edge detection filter process similar to the process of detecting edge pixels from the sample image for the identification target image, and represents an edge pixel in the medal area that represents a shadow edge caused by the pattern on the medal surface. Is detected. However, the edge detection unit 62 does not binarize the value of each edge pixel, and uses the edge intensity itself obtained by performing the edge detection filter processing for each pixel in the medal area as the value of the edge pixel. Further, the edge detection unit 62 sets pixels whose pixel values are not “0” as edge pixels. The edge detection unit 62 stores an edge image representing the edge strength of each pixel in the medal area in the memory 52.

  Note that the edge detection unit 62 may execute a smoothing filter process such as a Gaussian filter on the identification target image before executing the edge detection filter process so that the number of detected edge pixels does not become excessive. .

  Depending on the medal, the position of the pattern shown on the medal surface may shift due to the influence of the manufacturing process. FIG. 15 is a diagram illustrating an example of a medal with a correct pattern position and a medal with a shifted pattern position. In the medal 1500, a pattern 1510 on the surface is stamped at the center of the medal 1500. On the other hand, the position of the pattern 1511 on the surface of the medal 1501 is shifted from the center of the medal 1501 to the upper left. When a medal with such a pattern misalignment is an identification target medal, if the coordinates of the edge pixel detected by the edge detection unit 62 are used as they are for comparison with the model image, The position of the edge pixel on the identification target image indicating the same portion of the pattern on the medal surface is shifted from the position of the edge pixel on the model image, and the accuracy of calculating the coincidence is lowered. Therefore, the edge position correction unit 67 obtains the position where the medal area on the identification target image and the medal area on the model image most closely match, thereby correcting the positional deviation of the pattern on the surface of the identification target medal shown in the identification target image. to correct.

  In the present embodiment, the edge position correction unit 67 first aligns the center of the medal area of the edge image with the center of the medal area of the model image. From the aligned state, the edge position correction unit 67 determines the position of the center of the medal area of the edge image with respect to the center of the medal area of the model image by a predetermined search range (for example, the horizontal direction or the vertical direction). , ± 2 pixels), and the degree of coincidence between each edge pixel in the medal area of the edge image and each edge pixel in the medal area of the model image is calculated. Note that the degree of coincidence is calculated, for example, according to the above equation (1). Then, the edge position correcting unit 67 obtains the amount of positional deviation of the center of the medal area of the edge image with respect to the center of the medal area of the model image when the degree of coincidence is maximized.

  The edge position correction unit 67 repeats the above processing while rotating the edge image by a predetermined angle (for example, 1 degree) with the center of the medal area as the rotation center. Then, the edge position correction unit 67 specifies the rotation angle and the center position shift amount when the degree of coincidence becomes maximum. The edge position correction unit 67 rotates each edge pixel on the edge image by the specified rotation angle with the center of the medal area as the rotation center, and further cancels the specified positional deviation amount. Is translated. Then, the edge position correcting unit 67 stores the edge image in which the position of the edge pixel is corrected in the memory 52. Further, the edge position correction unit 67 converts the coordinates of each edge pixel into polar coordinates with the center of the medal area as the origin and positive in the clockwise direction, and the coordinates of each edge pixel displayed in the polar coordinates are identified. The information is stored in the memory 52 as edge distribution information of the identification target medal of the image.

  FIG. 16 is a diagram illustrating an example of a position-corrected edge image. In the edge image 1600, it can be seen that the edge 1601 corresponding to the pattern on the surface is located at the approximate center of the medal area 1602.

  Note that, according to the modification, the edge position correction unit 67 rotates the identification target image by a predetermined angle (for example, 1 degree) around the center of the medal area as the rotation center, and the center of the medal area of the identification target image. And the center of the medal region of the model image may be aligned to calculate the degree of coincidence, and the rotation angle that maximizes the degree of coincidence may be obtained. Then, the edge position correcting unit 67 rotates each pixel of the identification target image by the rotation angle about the center of the medal area, and then determines the position of the center of the medal area of the identification target image within the search range. The degree of coincidence may be calculated while shifting in parallel with the center position of the medal area of the model image, and the maximum value of the degree of coincidence may be obtained. Then, the edge position correction unit 67 parallelizes the position of each edge pixel so as to cancel out the amount of positional deviation of the center of the medal area of the identification target image with respect to the center of the medal area of the model image corresponding to the maximum value of the matching degree. It may be moved.

  The orientation coincidence calculation unit 68 is an example of a correction rotation angle calculation unit, and matches the pattern direction in the model of the appropriate medal on the model image with the pattern direction of the identification target medal shown in the identification target image. In addition to calculating the correct rotation angle for calculating the orientation, the orientation coincidence, which is an index representing the difference between the coincidence when the identification target medal and the appropriate medal model are aligned with each other and the alignment when the alignment is not performed, is calculated. To do.

  The orientation coincidence calculation unit 68 performs a process similar to the process executed by the rotation angle calculation unit 63 to calculate the rotation angle and coincidence of medals on the sample image with respect to the temporary model, and the model of the appropriate medals on the model image. Then, the rotation angle of the orientation of the pattern on the surface of the identification medal is calculated with respect to the orientation of the pattern on the surface of the model of the appropriate medal on the model image. That is, the orientation coincidence calculation unit 68 uses the edge distribution information of the model image and the identification target image to set each edge pixel on the identification target image to a predetermined angle (for example, the rotation angle about the center of the medal area). The degree of coincidence between the edge distribution of the model and the edge distribution of the identification target medal is calculated while changing each time. In this example as well, the degree of coincidence is calculated by equation (1). Note that the orientation coincidence calculation unit 68 may calculate the degree of coincidence by rotating each edge pixel on the model image instead of rotating each edge pixel on the identification target image.

The orientation coincidence calculation unit 68 determines the rotation angle at which the degree of coincidence is maximum, the orientation of the pattern on the surface of the model of the appropriate medal on the model image, and the orientation of the pattern on the surface of the identification target medal shown in the identification target image. And a correction rotation angle for making them coincide with each other. Furthermore, the orientation coincidence calculation unit 68 calculates the average value of the coincidence degree for all rotation angles and the deviation value for the average value and the maximum value of the coincidence degree from the coincidence degree for each rotation angle. Then, the orientation coincidence calculation unit 68 sets the ratio of the deviation value of the maximum coincidence value to the deviation value of the average coincidence value as the orientation coincidence. Therefore, the degree of orientation matching is the same only when the orientation of the pattern on the surface of the model of the appropriate medal on the model image matches the orientation of the pattern on the surface of the identification medal that appears in the identification target image. Edge pixels that match when the orientation of the pattern on the surface of the model of the appropriate medal on the model image deviates from the orientation of the pattern on the surface of the identification medal that appears in the identification target image The smaller the number, the larger the value. That is, the lower the rotational symmetry of the pattern represented by the appropriate medal, the greater the degree of orientation matching.
This degree of orientation coincidence is less affected by individual differences between individual medals, and therefore accurately represents the degree of coincidence between the identification medal and the appropriate medal model even if the identification medal is scratched, worn, or dirty. be able to.

Note that, according to the modification, the direction coincidence degree calculation unit 68 has the maximum value of the coincidence degree (that is, the largest maximum value among the maximum values of the coincidence degree) out of the coincidence degrees obtained for each rotation angle C1. The difference between C2 and C2 (C1-C2), or the value obtained by normalizing the difference with C1 (C1-C2) / C1, is calculated as the degree of orientation coincidence. May be. Alternatively, the direction coincidence calculating unit 68 may calculate the ratio (C1 / C2) of the maximum value C1 of the coincidence to the maximum value C2 of the second largest coincidence as the direction coincidence.
The direction coincidence calculation unit 68 stores the corrected rotation angle and the direction coincidence in the memory 52.

  The similarity calculation unit 69 matches the center and pattern orientation of the appropriate medal model on the model image with the center and pattern orientation of the identification medal on the identification target image, and then the surface of the model of the appropriate medal model A similarity indicating the degree of similarity between the pattern and the pattern on the surface of the identification target medal is calculated.

  In the present embodiment, the similarity calculation unit 69 divides the medal area into a plurality of partial areas for each of the plurality of different division methods so that the similarity can be accurately evaluated, For each partial region, the edge pixels are selected so that the ratio of the number of edge pixels to the area of the partial region is less than a predetermined ratio. Then, the similarity calculation unit 69 calculates the similarity based on the coincidence / mismatch between the selected edge pixel and the edge pixel on the model image.

  In the present embodiment, the similarity calculation unit 69 applies the same processing as that of the model edge number adjustment unit 65 to the medal region of the edge image obtained from the identification target image in order to select the edge pixel. Adjust the number. That is, the similarity calculation unit 69 divides the medal area into a predetermined number of fan-shaped partial areas along the circumferential direction with the center of the medal area of the edge image as a reference point. Note that the similarity calculation unit 69 preferably makes each divided region larger than the sector-shaped partial region in the model edge number adjustment unit 65. Therefore, for example, the predetermined number is set to 18 (that is, the central angle of each partial region is 20 degrees). As a result, a portion where the pattern on the medal surface changes minutely is included in any of the partial regions, and the ratio of the number of edge pixels to the area of the partial region becomes very high. It is suppressed that the number of edge pixels to be reduced has to be reduced and the feature of the pattern is not accurately expressed.

The similarity calculation unit 69 calculates, for each partial region, the ratio of the number of edge pixels in the area of the partial region. When the ratio is equal to or greater than a predetermined ratio (for example, 0.4), the similarity calculation unit 69 subtracts 1 from the value of each edge pixel. For a pixel whose pixel value is less than the edge threshold used for determining whether or not the pixel is an edge pixel, the similarity calculation unit 69 excludes the pixel from the edge pixel. The value of the edge threshold value can be the same value as the edge threshold value used by the edge detection unit 62 in the model registration process. The similarity calculation unit 69 repeats the above processing for each partial region until the ratio of the number of edge pixels in the area of the partial region becomes less than a predetermined ratio. Conversely, the similarity calculation unit 69 may increase the edge threshold until the ratio of the number of edge pixels in the area of the partial region becomes less than a predetermined ratio.
Note that, when the ratio of the number of edge pixels in the area of the partial region is less than the predetermined ratio from the beginning, the similarity calculation unit 69 sets 1 to the value of each pixel in the partial area until the ratio reaches the predetermined ratio. May be added. In this case, the similarity calculation unit 69 adds the pixel to the edge pixel when the pixel value is equal to or greater than the edge threshold value.

ここで、NumCは、エッジ画像のエッジ画素のうち、モデル画像のエッジ画素と一致するエッジ画素の数である。NumD1は、エッジ画像のエッジ画素のうちでモデル画像の何れのエッジ画素とも一致しないエッジ画素の数である。そしてNumD2は、モデル画像のエッジ画素のうちでエッジ画像の何れのエッジ画素とも一致しないエッジ画素の数である。すなわち、第１の類似度S1は、識別対象画像のメダル領域内のエッジ画素のうち、モデル画像上のメダル領域内の何れかのエッジ画素と一致するものの数が、モデル画像上のメダル領域内の何れのエッジ画素とも一致しないものの数より多いほど高くなる。 When the adjustment of the number of edge pixels in each partial area is completed, the similarity calculation unit 69 aligns the center of the medal area on the model image and the center of the medal area on the edge image. Then, the similarity calculation unit 69 rotates each edge pixel on the edge image by the correction rotation angle with the center of the medal area as the rotation center. Then, the similarity calculation unit 69 calculates a first similarity S1 between the pattern on the surface of the model of the appropriate medal and the pattern on the surface of the identification target medal according to the following equation.

Here, NumC is the number of edge pixels that match the edge pixels of the model image among the edge pixels of the edge image. NumD1 is the number of edge pixels that do not match any edge pixel of the model image among the edge pixels of the edge image. NumD2 is the number of edge pixels that do not match any edge pixel of the edge image among the edge pixels of the model image. That is, the first similarity S1 is the number of edge pixels in the medal area of the identification target image that match any one of the edge pixels in the medal area on the model image. The higher the number of pixels that do not match any of the edge pixels, the higher.

  Furthermore, the similarity calculation unit 69 divides the medal area of the edge image into a predetermined number of ring-shaped partial areas along the radial direction from the center of the medal area. Also in this case, the similarity calculation unit 69 preferably makes each divided region larger than the ring-shaped partial region in the model edge number adjustment unit 65. Therefore, for example, the similarity calculation unit 69 sets the radial width of each partial region to 10 pixels. Then, the similarity calculation unit 69 performs the same processing as described above for each partial region, so that the ratio of the number of edge pixels to the area of each partial region is less than a predetermined ratio (for example, 0.4). Adjust the number of. However, the adjustment target of the number of edge pixels is an edge image before the number of edge pixels for each partial region in the circumferential direction is adjusted.

When the adjustment of the number of edge pixels in each partial area is completed, the similarity calculation unit 69 aligns the center of the medal area on the model image and the center of the medal area on the edge image. Then, the similarity calculation unit 69 rotates each edge pixel on the edge image by the correction rotation angle with the center of the medal area as the rotation center. Then, the similarity calculation unit 69 calculates a second similarity S2 between the pattern on the surface of the appropriate medal model and the pattern on the surface of the identification target medal according to the following equation. The second similarity S2 is also calculated according to the equation (2).
The similarity calculation unit 69 may calculate the first similarity S1 and the second similarity S2 using the expression (1) instead of the expression (2).

The similarity calculation unit 69 uses the average value S (= (S1 + S2) / 2) of the first similarity S1 and the second similarity S2 as the surface pattern of the model of the appropriate medal finally obtained. The similarity of the pattern on the surface of the identification target medal. Alternatively, the similarity calculation unit 69 uses the larger one of the first similarity S1 and the second similarity S2 as the surface pattern of the model of the appropriate medal finally obtained and the surface of the identification target medal. It is good also as the similarity of a pattern. Although the degree of similarity obtained in this way is affected by the individual difference of the identification medals, even if the surface pattern of the appropriate medals and the identification medals is a rotationally symmetric pattern, it can represent the degree of similarity between them. it can.
The similarity calculation unit 69 stores the calculated similarity in the memory 52.

  The determination unit 70 determines whether or not the identification target medal is a proper medal based on the orientation coincidence and the similarity. In the present embodiment, the determination unit 70 uses a value obtained by multiplying the orientation coincidence by the similarity as a total evaluation score, and when the total evaluation score is equal to or higher than a predetermined determination threshold, the identification target medal is determined as a proper medal. On the other hand, if the overall evaluation score is less than the determination threshold, it is determined that the identification target medal is not a proper medal. The determination threshold is set to 1.0, for example, and is stored in the memory 52 in advance.

  The control circuit 51 outputs a determination result as to whether or not the medal is an appropriate medal to a higher-level control device (not shown).

  17 and 18 are operation flowcharts of the medal identification process executed by the control circuit 51. FIG. The control circuit 51 executes this medal identification process every time an image is received from the imaging unit 4.

  The medal area detection unit 61 detects a medal area from the identification target image (step S301). The shape correcting unit 66 corrects the identification target image so that the shape of the medal area is substantially circular and the size of the medal area substantially matches the size of the medal area on the model image (step S302). Then, the edge detection unit 62 detects edge pixels in the medal area of the identification target image (step S303).

  The edge position correction unit 67 aligns the center of the medal area of the identification target image and the center of the medal area of the model image with a predetermined angle (for example, with the center of the medal area as the rotation center). , By 1 degree), the parallel movement amount of the identification target image in which the edge distribution in the medal area of the identification target image and the edge distribution in the medal area of the model image most closely match is calculated (step S304). Then, the edge position correcting unit 67 rotates the identification target image by the rotation angle when the edge distribution in the medal area of the identification target image and the edge distribution in the medal area of the model image most closely match, and then rotates the rotation angle. Each edge pixel in the medal area on the identification target image is translated by the translation amount at (step S305). Then, the edge position correction unit 67 stores edge distribution information that represents the corrected position of the edge pixel in polar coordinates in the memory 52.

  The orientation coincidence calculation unit 68 calculates a corrected rotation angle from the rotation angle of the identification target image when the edge distribution in the medal region on the model image and the edge distribution in the medal region of the identification target image most closely match ( Step S306). Further, the orientation coincidence calculation unit 68 determines the identification target medal based on the coincidence when the pattern orientation in the appropriate medal model matches the pattern orientation of the identification medal and the coincidence when the orientation does not match. Is calculated (step S307).

  As shown in FIG. 18, the similarity calculation unit 69 aligns the center of the medal area on the model image with the center of the medal area on the identification target image, and then identifies the identification target image by the obtained rotation angle. Is rotated to match the center and orientation of the model of the appropriate medal with the center and orientation of the identification target medal (step S308). Then, the similarity calculation unit 69 divides the medal area on the identification target image into a plurality of partial areas along the circumferential direction, so that the ratio of the edge pixels included in each partial area is less than a predetermined ratio. For each region, the number of edge pixels is adjusted by selecting edge pixels in order from the pixel with the highest edge strength (step S309). Thereafter, the similarity calculation unit 69 calculates the first similarity S1 by calculating the number of edge pixels that do not match the number of edge pixels that match between the edge pixels on the model image and the edge pixels on the identification target image. (Step S310).

  The similarity calculation unit 69 divides the medal area on the identification target image into a plurality of partial areas along the radial direction from the center of the medal area, and the ratio of the edge pixels included in each partial area is less than a predetermined ratio. For each partial region, the number of edge pixels is adjusted by selecting edge pixels in order from the pixel with the highest edge strength (step S311). Thereafter, the similarity calculation unit 69 calculates the second similarity S2 by obtaining the number of edge pixels that do not match the number of edge pixels that match between the edge pixels on the model image and the edge pixels on the identification target image. (Step S312). The similarity calculation unit 69 calculates the average value of the first similarity S1 and the second similarity S2 as the similarity S (step S313).

The determination unit 70 calculates a value obtained by multiplying the degree of coincidence DM by the similarity S as the overall evaluation score T (step S314). Then, the determination unit 70 determines whether or not the overall evaluation score T is equal to or greater than a determination threshold ThD (step S315). When the total evaluation score T is equal to or greater than the determination threshold ThD (step S315—Yes), the determination unit 70 determines that the identification target medal is an appropriate medal and outputs a notification signal indicating that to the upper control device (step S316). ). On the other hand, when the comprehensive evaluation score T is less than the determination threshold ThD (step S315-No), the determination unit 70 determines that the identification target medal is not an appropriate medal, and outputs a notification signal indicating that fact to the host controller. (Step S317).
After step S316 or S317, the control circuit 51 ends the medal identification process.

  As described above, this medal identification device has a pattern of the surface of the medal detected from an image obtained by photographing an identification target medal inserted in the medal identification device and a model of a proper medal created in advance. The similarity with the surface pattern is calculated, and it is determined whether or not the identification target medal is an appropriate medal based on the similarity. At this time, this medal identification device equalizes the number of edge pixels in the medal area by adjusting the number of edge pixels corresponding to the edge of the pattern for each of the plurality of partial areas obtained by dividing the medal area of the identification target image. After that, the similarity is calculated based on the coincidence / non-coincidence of the edge pixels. Thus, even if a part of the surface of the identification target medal is detected as an edge pixel, it does not affect the part without the surface, so that the influence on the similarity due to the surface of the medal is suppressed. Therefore, this medal identification device can accurately determine whether or not the identification target medal is an appropriate medal even if there is an individual difference of medals due to a spear or the like. Furthermore, since this medal identification device sets partial areas according to each of a plurality of different division methods and adjusts the number of edge pixels and calculates the similarity for each division method, the partial areas affected by wrinkles and the like Since the medal can be identified by using the similarity obtained based on the division method having a relatively small number, the degradation of the identification accuracy due to individual differences of medals due to a bag or the like can be suppressed.

  In addition, this invention is not limited to said embodiment. For example, the pattern on one side of the appropriate medal may be different from the pattern on the other side. In such a case, the medal identification device may generate the model image for each face of the medal and store it in the memory by executing the above model registration process for each face of the medal. Then, the medal identification device executes the above-described medal identification processing for each model image, and if the determination result that the identification target medal is a proper medal is obtained for any one of the model images, the identification target medal is a proper medal. May be determined.

  Similarly, a plurality of medals of different types may be used. In such a case, the medal identification device may generate a model image for each medal type and store the model image in the memory by executing the model registration process for each medal type. Then, the medal identification device performs the above-described medal identification process for each model image, and if the identification target medal is determined to be an appropriate medal for any one model image, the identification target medal is the model image. It may be determined that the medal is of the type shown in the above, and a notification signal indicating the type of the medal may be output to the host control device. On the other hand, if it is determined that the identification target medal is not a proper medal for any model image, this medal identification device outputs a notification signal indicating the determination result that the identification target medal is not a proper medal to the host controller. That's fine.

  According to another modification, the processing of the shape correction unit 66 may be performed on each sample image. Alternatively, the processing of the shape correction unit 66 may be omitted as long as the inclination of the medal in the passage and the variation in the distance from the imaging unit 4 to the medal are negligible.

  According to still another modified example, when the pattern on the surface of the appropriate medal is rotationally symmetric, the determination unit may use the similarity itself as the comprehensive evaluation value. In this case, it is not necessary to calculate the orientation coincidence, and it is only necessary to align the center of the medal area of the identification target image and the center of the medal area of the model image when calculating the similarity, Since there is no need to match the orientation of the pattern, the processing of the orientation coincidence calculation unit 68 may be omitted.

  According to another modification, the similarity calculation unit 69 divides the medal area of the edge image corresponding to the identification target image into a plurality of partial areas according to any one of the division methods, and the edge of each partial area. The similarity may be calculated after selecting the pixels as in the above example. The similarity calculation unit 69 may divide the medal area of the edge image into a plurality of partial areas according to a division method different from the above division method. For example, the similarity calculation unit 69 may divide the medal area of the edge image into a plurality of partial areas having a grid shape or a strip shape.

  According to still another modification, the model registration process may be executed in a device different from the medal identification device, and the finally obtained model image may be stored in the memory of the medal identification device. In this case, for example, every time a sample image is generated by the imaging unit 4, the control circuit 51 outputs the sample image to a device that executes a model registration process via the communication interface 53.

A device that executes the model registration process includes a processor, a memory circuit, and a communication interface for connecting to a medal identification device, for example, like a computer. And the processor of the apparatus performs the process of each part utilized by the model registration process among each part which the control circuit 51 has. When the device executes model registration processing to generate a model image, the control circuit 51 of the medal identification device receives the model image from the device via the communication interface 53 and stores the model image in the memory 52. .
In this modification, since it is not necessary to store many sample images in the memory 52, the memory capacity of the memory 52 can be reduced.

FIG. 19 is a schematic perspective view of a spinning game machine 100 including a medal identification device according to an embodiment or a modification of the present invention. FIG. 20 is a schematic perspective view of the spinning machine 100 with the front panel opened. FIG. 21 is a circuit block diagram of the spinning cylinder game machine 100. As shown in FIGS. 19 and 20, the spinning machine 100 includes a main body housing 101 that is a gaming machine body, drums 102 a to 102 c, a start lever 103, stop buttons 104 a to 104 c, and a front panel 105. Have
The spinning machine 100 includes a medal identification device 106 attached to the back side of the front panel 105, a control box 107 in which various control circuits are accommodated, and a main control circuit 110. In response to the control signal, a medal storage / discharge mechanism 108 is provided for temporarily storing medals and discharging medals.
As shown in FIG. 21, in the control box 107, for example, a main control circuit 110 that controls the entire spinning game machine 100, an effect control circuit 111 that controls each part related to the effect of the game, A power supply circuit 112 that supplies power to each part of the spinning machine 100 is accommodated.

  An opening 105a is formed at the upper center of the front surface of the front panel 105, and the drums 102a to 102c are visible through the opening 105a. An opening 105c for inserting medals is formed on the upper surface of the frame 105b on the lower side of the opening 105a. The opening 105c and the medal insertion slot provided on the upper surface of the medal identification device 106 coincide with each other. In addition, a medal identification device 106 is arranged. The medal discharge port of the medal identification device 106 is connected to the medal storage / discharge mechanism 108, and medals discharged from the medal identification device 106 are collected by the medal storage / discharge mechanism 108.

  A decoration device including a plurality of light emitting diodes may be provided as a lower panel below the front panel 105. Further, a medal discharge port 105d for discharging medals is formed in the front panel 105 below the decoration device. A medal tray 105e for preventing the discharged medal from falling is attached below the medal discharge port 105d. A speaker (not shown) may be attached near the upper left end and near the upper right end of the front panel 105.

The drums 102a to 102c are individually separated around a rotation axis (not shown) that is substantially parallel and substantially horizontal to the front surface of the main body housing 101 in accordance with a control signal from the effect control circuit 111. It can be rotated. The surfaces of the drums 102a to 102c are each divided into a plurality of regions having substantially the same width along the rotation direction, and various designs are drawn for each region. Instead of the drums 102a to 102c, a display device such as a liquid crystal display may be provided so that the display screen can be viewed through the opening 105a. In this case, the display device displays an image showing a plurality of drums on the display screen in response to the control signal from the effect control circuit 111.
The start lever 103 is provided on the left side toward the front surface of the frame 105 b of the front panel 105. Further, stop buttons 104a to 104c are provided substantially at the front center of the frame 105b. The stop buttons 104a to 104c correspond to the drums 102a to 102c, respectively.

  The medal identification device 106 is a medal identification device according to the above-described embodiment or modification. When the medal is inserted into the medal insertion slot of the medal identification device 106 through the opening 105c, the medal identification device 106 determines whether the inserted medal is an appropriate medal and sends a notification signal indicating the determination result to the main control circuit. To 110. Then, if the inserted medal is an appropriate medal, the main control circuit 110 determines the number of games according to the medal and permits the spinning machine 100 to start the game. Thereafter, when the start lever 103 is operated, a signal indicating that the start lever 103 has been operated is transmitted to the main control circuit 110. Then, the main control circuit 110 causes the effect control circuit 111 to start rotating the drums 102a to 102c. Further, the production control circuit 111 outputs a sound effect different from that when the drums 102a to 102c are stopped through the speaker in order to enhance the player's interest during the rotation of the drums 102a to 102c. The light emitting diode may be blinked.

Thereafter, when any one of the stop buttons 104a to 104c is pressed, the main control circuit 110 receives a signal indicating that the button is pressed from the pressed switch, and rotates the drum corresponding to the pressed switch. Is stopped via the production control circuit 111. Alternatively, the main control circuit 110 controls the drum 102a to 102c for which the corresponding stop button has not been pressed until the predetermined period elapses after the rotation starts, after the predetermined period elapses. Stop via the circuit 111.
When all the drums are stopped and the same symbols are arranged in a line across all the drums, the main control circuit 110 discharges a predetermined number of medals corresponding to the symbols through the medal discharge port. Further, in this case, the production control circuit 111 outputs a sound effect different from that when the drum is rotated and when the same design is not aligned in a line across all the drums, and the light emitting diode of the decoration device is output from the speaker. It may be lit at maximum brightness.

  On the other hand, if the determination result notified from the medal identification device 106 indicates that the inserted medal is an improper medal, the main control circuit 110, for example, in the hall where the spinning machine 100 is installed. This is notified to the management system together with the identification number of the spinning machine 100. Further, the main control circuit 110 may discharge medals determined to be inappropriate medals via the medal storage / discharge mechanism 108.

  The medal identification device according to the present invention can be applied to various devices that are required to identify inserted medals. For example, the medal identification device according to the present invention may be implemented as a medal counting device used for counting the number of acquired medals.

  22 is an external perspective view of the medal counting device 200, and FIG. 23 is a perspective view of the medal counting device 200 with the top cover removed. FIG. 24 is a three-dimensional view of the transport unit 201, FIG. 25 is a plan view of the transport unit 201, FIG. 26 is a diagram illustrating an operation when a medal is transported in the vicinity of the guide, and FIG. FIG. 4 is an exploded view of a transport unit 201.

  The medal counting device 200 is provided with an opening 202 on the left side of the front, and a medal can be inserted through the opening 202. A transport unit 201 is provided on the bottom surface of the opening 202. The transport unit 201 is an example of a passage forming unit, and in order to physically count inserted medals, the inserted medals are arranged along a medal path including a guide unit 224, a discharge unit 222, and the like. It flows down through the imaging range of the imaging unit 226.

  The display unit 203 includes, for example, a liquid crystal display and displays a medal counting result. The printing unit 204 prints the medal counting result on a paper such as a receipt, and outputs the paper on which the coefficient result is printed. The operation unit 205 has buttons and the like, and is operated when counting of medals is started or when counting is ended. The operation unit 205 generates an operation signal corresponding to the operation content and supplies the operation signal to the control unit.

  The separation unit 211 of the transport unit 201 is configured such that the bottom of the disk provided in the cylindrical basket 211a rotates in the direction of the arrow shown in FIG. To be moved so as to be attracted to the basket 211a. By this operation, even if the medals 231 are solidified to some extent, that is, in a collective state, the medals 231 behave individually so as to be attracted to the basket 211a, and thus are separated individually. In addition, when the inserted medal 231 is attracted to and brought into contact with the basket 211a, it moves on the basket 211a in the arrow direction shown in FIG. Further, the medal 231 moving on the basket 211a is conveyed to the guide part 224 through the guide part 212 from the opening provided in the tangential direction of the disk of the separation part 211 as shown in the lower part of FIG.

  As shown in FIG. 26, the guide portion 224 provided on the transport plate 221 is formed in a U-shaped plate shape, and forms a medal passage from the U-shaped opening to the hole 223. . The width of the opening of the guide part 224 is substantially equal to the diameter of the medal 231, and the guide part 224 guides the medal 231 to the hole part 223. As shown in FIG. 26, the medal 231 guided through the guiding portion 212 is supplied to the storage disk 251 (FIG. 27) through the guide portion 224 and the hole portion 223 having a diameter substantially equal to the diameter of the medal 231. The

  As shown in FIG. 27, the storage disk 251 has a plurality of holes 251 e to 251 h that are substantially the same shape as the medal 231. The storage disk 251 stores the four medals 231 in the respective holes, and then loads the medals 231 into the recesses 253e to 253h of the discharge disk 253. The discharge disk 253 discharges the loaded four medals 231 from the discharge unit 222 at a predetermined interval. The discharge unit 222 is provided with a sensor 225 that is a proximity sensor, an imaging unit 226, and a light source 227. When the sensor 225 detects the passage of the medal 231, a corresponding detection signal is generated and output to the control unit 261 (FIG. 29) described later. The control unit 261 causes the image pickup unit 226 whose shooting range is the downstream side of the medal path from the sensor 225 to photograph the medal 231 illuminated by the light source 227. Then, the imaging unit 226 outputs an image showing the medal 231 to the control unit 261.

  The hole 221 a of the transport plate 221, the hole 251 i formed at the center of the storage disk 251, and the hole 253 i formed at the center of the discharge disk 253 are coaxial, and the shaft 259 passes therethrough. The shaft 259 transmits the torque of the motor 257 to the discharge disk 253 via the belt 260. Therefore, the discharge disk 253 rotates at a predetermined rotational speed in the direction of the arrow in FIG. 27 by the torque transmitted from the shaft 259.

  At this time, if all four medals 231 are not accumulated on the accumulation disk 251, the accumulation disk 251 and the ejection disk 253 are engaged with each other by the ejection stopper pin 258 in a state of being overlapped with each other. With the rotation of 253, the storage disk 251 also rotates.

  As described above, the storage disk 251 has the holes 251e to 251h that are substantially the same shape as the medal 231 at equal intervals. In addition, the discharge disc 253 is formed with indentations 253e to 253h at equal intervals in which medals 231 can be loaded. When the storage disk 251 and the discharge disk 253 are engaged by the stopper pins 252-1 to 252-4, the holes 251e to 251h and the recesses 253e to 253h are engaged with each other in different phases. It is in the state. Accordingly, when the medal 231 is inserted from the hole 223, the medal 231 is stopped by the discharge disk 253, fixed by the holes 251e to 251h of the storage disk 251, and accumulated in the holes 251e to 251h. become.

  Further, the stopper pins 252-1 to 252-4 can be divided into the storage disk 251 side and the discharge disk 253 side, respectively. More specifically, as shown in FIG. 28, the stopper pin 252 includes an upper stopper pin 252a and a lower stopper pin 252b. The bottom portions 253a ′ to 253d ′ of the holes 253a to 253d in which the upper stopper pin 252a and the lower stopper pin 252b are accommodated are provided with an elastic body (not shown) such as a spring or a spring so as to repel upward. It comes to give power. The bottom portion 253a ′ (shown for the bottom portion 253a ′ in FIG. 28 is the same for the bottom portions 253b ′ to 253d ′). When the medal 231 falls from the top, the weight of the medal 231 causes the elastic body When the repulsive force is defeated, the upper stopper pin 252a and the lower stopper pin 252b slide downward as shown by the right bottom 253a '' in the drawing by pressing the bottom 253a ′, The holes 251a to 251d and the holes 253a to 253d are inserted.

  At this time, the lengths of the upper stopper pin 252a and the lower stopper pin 252b are substantially the same as the thicknesses of the storage disk 251 and the discharge disk 253, respectively, so that the transport plate 221, the storage disk 251 and the discharge disk 253 are sequentially stacked. In the state, the storage disk 251 and the discharge disk 253 are respectively housed in the holes 251a to 251d and the holes 253a to 253d. As a result, the engagement state of the storage disk 251 and the discharge disk 253 is released.

  That is, when the storage disk 251 and the discharge disk 253 rotate and move into the range of the hole 223, the stopper pin 252 is moved by the transport plate 221 and the discharge disk 253 as shown in the left part of FIG. By releasing from the state of being biased from above and below, the convex portion 252a ′ protrudes out of the hole portion 223, so that the transport plate 221 and the storage disk 251 are locked. Further, at this time, the lower stopper pin 252b locks the storage disk 251 and the discharge disk 253, and the rotation of the storage disk 251 and the discharge disk 253 is stopped.

  Further, in this state, when the medal 231 is inserted into the hole 223, the stopper pin 252 is pressed from above by the medal 231 as shown in the right part of FIG. The lower stopper pin 252b presses the bottom 253a ′ and is slid downward as indicated by the bottom 253a ″, so that it is accommodated in the holes 251a to 251d and the holes 253a to 253d, respectively.

  Therefore, when the medal 231 is guided by the guide portion 224 and sequentially inserted into the hole portion 223, for example, it is accumulated in the hole portion 251e of the accumulation disk 251. At this time, the stopper pin 252 is released from the state where it is locked to the transport plate 221, so that the storage disk 251 starts rotating together with the discharge plate 253. Then, when rotated at an angle of approximately 45 °, the rotation is stopped again by the stopper pin 252 at the position where the hole 251f arranged next to the hole 251e where the medal 231 is not accumulated overlaps the hole 223. To do.

  Further, when the next medal 231 is inserted from the hole 223, it is accumulated in the hole 251f, and the accumulation disk 251 and the discharge disk 253 start rotating. The above operation is repeated four times continuously every time the medal 231 is inserted into the hole 223.

  A discharge stopper pin 258 is provided on the discharge disk 253. The discharge stopper pin 258 includes a convex portion 258b for locking the hole portion 251h of the storage disk 251 in which the fourth medal 231 is stored, and a convex portion 258a for locking the recess 253h of the discharge disk 253. It is a letter-shaped pin, and the storage disk 251 and the discharge disk 253 are engaged by locking each. When the fourth medal 231 is stored in the hole 251h of the storage disk 251, the discharge stopper pin 258 slides downward due to the weight of the medal 231, and the storage disk 251 engages with the discharge disk 253. To be able to rotate independently.

  As a result, when the ejection disk 253 is in phase with the storage disk 241, that is, when the holes 251 e to 251 h of the storage disk 251 and the recesses 253 e to 253 h of the discharge disk 253 move to the same position, The four medals 231 stored in the holes 251e to 251h of the storage disk 251 are loaded in the corresponding recesses 253e to 253h at almost the same timing.

  Further, in this state, the discharge disk 253 continues to rotate at a predetermined rotation speed, so that the medals 231 loaded in the recesses 253e to 253h move along the disk guide 256 and are discharged from the discharge guides 254 and 255. It is sent to 222 and discharged.

  In addition, four medals 231 can be stored in the storage disk 251, but when the four medals 231 cannot be stored due to a fraction or the like, the forcible discharge unit 241 directly presses the stopper pin 2521 so that the medals 231 can be stored. The accumulation disk 251 and the discharge disk 253 are rotated even when the holes 223 are not accumulated, and finally the discharge stopper pin 258 is pressed to discharge the fractional medal 231 from the discharge disk 253.

  FIG. 29 is a circuit block diagram of the medal counting device 200. The medal counting device 200 includes a control unit 261, a storage unit 262, and a communication interface 263. Each of these units is provided as a separate circuit on a circuit board provided in the medal counting device 200.

  The control unit 261 includes, for example, one or a plurality of processors, and controls each unit of the medal counting device 200. In addition, the control unit 261 executes the function of each unit included in the control circuit of the medal identification device according to the above-described embodiment or modification on the image obtained from the imaging unit 226 so that the conveyed medal 231 is appropriate. It is determined whether or not it is a medal. Then, the control unit 261 increases the medal count by 1 when the transported medal 231 is an appropriate medal. When the counting of inserted medals is completed, the control unit 261 causes the display unit 203 to display the counting result. Further, the control unit 261 transmits the counting result to the printing unit 204 and prints the counting result on paper or the like.

  The storage unit 262 corresponds to the memory of the medal identification device according to the above-described embodiment or modification, has a volatile or nonvolatile readable / writable semiconductor memory, and stores various data necessary for the operation of the control unit 261. To do. For example, the storage unit 262 stores various intermediate data such as edge images, model images, images received from the imaging unit 226, and the like generated in the model registration process or the medal identification process.

  The communication interface 263 corresponds to the communication interface of the medal identification device according to the above-described embodiment or modification, and is connected to a higher-level control device (not shown) such as the sensor 225, the imaging unit 226, the light source 227, and the hall computer. For receiving a control signal to the sensor 225, the imaging unit 226, or the light source 227 from the control unit 261 and outputting the control signal to the sensor 225, the imaging unit 226, or the light source 227. The communication interface 263 receives various control signals from the host control device, signals from the sensor 225, or images from the imaging unit 226 via a signal line, and passes them to the control unit 261.

  As described above, those skilled in the art can make various modifications in accordance with the embodiment to be implemented within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Medal identification device 2 Case 2a Medal insertion slot 2b Medal ejection slot 3 Light source 4 Imaging unit 5 Control board 51 Control circuit 52 Memory 53 Communication interface 61 Medal region detection unit 62 Edge detection unit 63 Rotation angle calculation unit 64 Model generation unit 65 Model edge number adjustment unit 66 Shape correction unit 67 Edge position correction unit 68 Orientation coincidence calculation unit 69 Similarity calculation unit 70 Judgment unit 6 Passage 7a, 7b Guide member 100 Cylinder machine 101 Main body case 102a-102c Drum 103 Start Lever 104a-104c Stop button 105 Front panel 106 Medal identification device 107 Control box 108 Medal storage / discharge mechanism 110 Main control circuit 111 Production control circuit 112 Power supply circuit 200 Medal counting device 201 Conveying unit 202 Opening unit 203 Display unit 2 04 Printing section 205 Operation section 211 Separating section 212 Guide section 221 Conveying plate 222 Discharge section 224 Guide section 226 Imaging section 227 Light source 241 Forced discharge section 251 Storage disk 252-1 to 252-4 Stopper pin 253 Discharge disk 254, 255 Discharge guide 256 Disc guide 257 Motor 258 Discharge stopper pin 259 Shaft 260 Belt 261 Control unit 262 Storage unit 263 Communication interface

Claims (4)

A passage forming portion that forms a passage through which medals pass;
An imaging unit that is arranged so that a part of the passage is an imaging range, and generates an image of the imaging range;
A storage unit for storing a model image including an edge pixel corresponding to an edge of a pattern on the surface of the model in a model region representing a model of a predetermined medal and in which the model is captured;
From the image, a medal area that is an area in which an identification target medal that flows down in the passage is shown, and a medal area detection unit that detects the center of the medal area;
An edge detection unit for detecting an edge pixel in which an edge of a pattern of the surface of the identification target medal is reflected in the medal area;
The medal area is divided into a plurality of partial areas, and for each of the plurality of partial areas, the edge so that the ratio of the number of edge pixels included in the partial area to the area of the partial area is less than a predetermined ratio. Select a pixel, align the center of the model area and the center of the medal area, and the number of edge pixels that match one of the edge pixels in the model area among the selected edge pixels is Similarity in which the similarity is calculated so that the similarity between the pattern on the surface of the model and the pattern on the surface of the identification medal increases as the number of edge pixels that do not match any edge pixel in the model region increases. A degree calculator,
A determination unit that determines whether the identification target medal is the predetermined medal according to the similarity;
A medal identification device. The similarity calculation unit divides the medal area into the plurality of partial areas for each of a plurality of different division methods, and the number of edge pixels occupying the area of each partial area for each of the division methods is the predetermined number. Select the edge pixel to be less than a ratio, calculate the similarity, calculate an average value or a maximum value of the similarity for each of the division methods,
The medal identification device according to claim 1, wherein the determination unit determines whether the identification target medal is the predetermined medal according to an average value or a maximum value of the similarity.   The similarity calculation unit, as one of the plurality of division methods, divides the medal area into the plurality of partial areas along a radial direction from the center of the medal area. The medal identification device according to claim 2, wherein the medal area is divided into the plurality of partial areas along a circumferential direction with respect to a center of the medal area as another one. A medal identification method in a medal identification device, comprising: a passage forming unit that forms a passage through which a medal passes; and an imaging unit that is arranged so that a part of the passage is set as a shooting range and generates an image of the shooting range. There,
Detecting, from the image, a medal area that is an area where an identification target medal flowing down in the passage is shown, and a center of the medal area;
Detecting an edge pixel in which an edge of a pattern of the surface of the identification target medal is shown in the medal area;
The medal area is divided into a plurality of partial areas, and for each of the plurality of partial areas, the edge so that the ratio of the number of edge pixels included in the partial area to the area of the partial area is less than a predetermined ratio. A pixel is selected, and the center of the model area that is an area in which the model is captured on the model image representing the model of the predetermined medal stored in the storage unit is aligned with the center of the medal area, Of the selected edge pixels, the number of edge pixels that match any one of the edge pixels in the model area is larger than the number of edge pixels that do not match any edge pixel in the model area. Calculating the similarity so that the similarity between the pattern and the pattern on the surface of the identification medal is high;
Determining whether the identification target medal is the predetermined medal according to the similarity;
Medal identification method including

Images