JP6214036B2 - Coin identification device and coin identification method - Google Patents

Description

  The present invention relates to a coin identifying device and a coin identifying method for identifying the type of a coin, and more particularly to a coin identifying device and a coin identifying method capable of stably identifying a coin without being affected by a temperature change or the like.

  Conventionally, a coin image is imaged by irradiating light toward a coin conveyed on a conveyance path and receiving the reflected light on an image pickup device such as a CCD, and based on the coin image, A coin identifying device for identifying authenticity is known. In order to guide the reflected light from the coin to the image sensor, a lens unit in which a plurality of lenses are combined is used. In many cases, inexpensive plastic lenses such as acrylic resin (PMMA) lenses and polycarbonate resin (PC) lenses are used for the lens unit. However, these resin lenses have a refractive index change or volume change caused by a temperature change. There is a problem that the lens characteristics are easily changed. For example, if the image magnification of the coin image changes due to a change in lens characteristics, the coin diameter obtained from the coin image changes, and the accuracy of the identification process decreases. For this reason, it is necessary to correct the influence on the image due to the change in the lens characteristics and then perform the identification process by obtaining a feature amount such as a coin diameter.

  For example, Patent Document 1 discloses a medium identification device that corrects an image using a reference medium. In this apparatus, a reference medium image acquired under a predetermined condition is stored before the start of medium processing. Then, when the medium processing is started, the reference medium is imaged again, and the correction contents of the image are set based on the obtained reference medium image and the reference medium image stored in advance.

  Moreover, for example, Patent Literature 2 discloses a coin identifying device including a temperature sensor therein. In this apparatus, the temperature at the time of coin image capturing is measured by a temperature sensor, and the image magnification of the coin image is corrected based on the measured temperature. Thereby, even when the temperature at the time of coin processing changes, a coin image can be acquired with a fixed image magnification, and the influence of the temperature change on the coin identification accuracy can be suppressed.

JP 2009-110325 A JP 2010-26992 A

  However, in the above prior art 1, it is necessary to prepare and manage a reference medium separately from the apparatus, and it is necessary to insert the reference medium into the apparatus and capture an image before starting the process. In addition, the temperature may change during the execution of the process. If it is desired to correct this, the process being executed must be stopped in order to input the reference medium into the apparatus and capture an image.

  Further, in the above-described prior art 2, it is necessary to measure the change in the image magnification due to the temperature change in advance and acquire information indicating the relationship between the temperature and the correction amount. Specifically, in order to correct the image magnification in accordance with the measured temperature at the time of image acquisition, the image magnification when the temperature changes with respect to the image magnification of the image captured at a predetermined temperature set as the reference temperature. It is necessary to investigate changes and save the obtained data in a memory or the like, but these operations are time-consuming.

  In addition, in the coin identification device, there is a factor that lowers the coin identification accuracy in addition to the change in the image magnification caused by the temperature change. For example, in the coin identification device, coins passing through a predetermined position on the conveyance path are imaged, and there are cases where coin powder or the like deposited on the conveyance path is reflected in the captured image. In such a case, an image with no coins is taken as a background image, and by subtracting the background image from the coin image obtained by picking up the coins, an image such as coin dust contained in the background image is used as noise. Can be removed. However, if the lens is displaced due to impact or vibration and the position of the noise is shifted between the background image and the coin image, the noise cannot be removed even if the background image is subtracted from the coin image. And, if noise is included in the coin image, there is a possibility that coin identification cannot be performed accurately.

  The present invention has been made to solve the above-described problems of the prior art, and even when a temperature change, an impact, or the like occurs, a coin identification device and a coin identification that can stably acquire a highly accurate coin image. It aims to provide a method.

In order to solve the above-described problems and achieve the object, the present invention is a coin identifying apparatus that performs identification processing based on an image obtained by irradiating light from a light source toward a coin and imaging the coin, an image sensor for imaging by the imaging an imaging range by the lens unit on the image sensor coin includes a plurality of reference members which are arranged fixedly on the outside of the coin passage region and the coin passage region through the image A storage unit that stores a distance reference value indicating a distance between the reference members above, and an image captured by the image sensor, and the distance between the reference members calculated on the image is the distance reference value And an image correction unit that corrects the images to match.

  Further, the present invention is the above invention, wherein the distance reference value indicates a distance between the reference members under a reference temperature, and the image correction unit detects the distance reference value and the reference temperature by the image sensor. A change in image magnification caused by a temperature change is corrected based on a distance between the reference members on images taken at different temperatures.

  In the invention described above, the light intensity ratio may be a ratio between a luminance reference value indicating a luminance value of the image of the reference member and a luminance value of the image of the reference member included in the image captured by the image sensor. A luminance calculation unit to calculate, and a light source control unit to control the light emission amount of the light source based on the light amount ratio calculated by the luminance calculation unit are further provided.

Further, the present invention provides the image processing apparatus according to the above invention, wherein a deviation amount between a reference position on an image indicating the position of the reference member and a position of the reference member included in an image captured by the image sensor is calculated, and the image And further comprising an image cutout unit that corrects a cutout reference that defines a position for cutting out the partial area image from the image based on the shift amount, and cuts out the partial area image using the corrected cutout reference. .

  Further, the present invention is characterized in that, in the above invention, the distance and the shift amount calculated on the image are calculated in units of one pixel.

  Moreover, the present invention is characterized in that, in the above invention, the reference member is fixed in a sensor case constituting the image sensor.

  Further, the present invention is the above invention, wherein the sensor case and the reference member are made of a material having a smaller linear expansion coefficient than the material forming the lens unit, and the reference member is more reflective than the sensor case. It is made of a high material.

Further, the present invention is a coin identification method for performing identification processing based on an image obtained by irradiating light from a light source toward a coin and capturing the coin, and indicating a distance between reference members on the image A storage step for storing values, a coin passing region through which a coin to be identified passes, and an imaging range including a plurality of reference members fixed outside the coin passing region are imaged on an image sensor by a lens unit. An image capturing step for capturing an image, and an image correcting step for correcting the image captured in the image capturing step so that a distance between the reference members calculated on the image matches the distance reference value. It is characterized by that.

  According to the present invention, a distance between reference members is obtained on a reference image obtained by imaging a reference member fixedly arranged within an imaging range by an image sensor under a predetermined condition, and stored as a distance reference value in a storage unit. If the distance between the reference members changes on images taken under different conditions, the image can be corrected so that this distance matches the distance reference value. Even when a change in image magnification or a shift of a region to be cut out in a captured image occurs due to the above, a highly accurate coin image can be acquired. And a coin can be correctly identified using the acquired coin image.

  Further, according to the present invention, for example, even when the refractive index of the lens unit constituting the image sensor changes and the image magnification with respect to the reference image changes as the temperature changes, the distance between the reference members is increased. Since the change in the image magnification can be corrected based on this, a highly accurate coin image can be acquired without investigating the relationship between the temperature change and the image magnification change in advance and setting the correction amount.

  Further, according to the present invention, the coin identification process is performed using the luminance value of the image of the reference member on the reference image taken by adjusting the light emission amount of the light source so that a clear coin image can be acquired as the luminance reference value. Since the light emission amount of the light source can be controlled so that the luminance value of the image of the reference member on the captured image matches the luminance reference value, a clear coin image can be acquired.

  Further, according to the present invention, for example, due to the positional deviation of the lenses constituting the lens unit, the cut-out range of the partial region image set on the reference image and the cut-out on the image captured when performing the coin identification process Even when the position of the extraction range is deviated, the amount of deviation can be calculated from the positional relationship of the reference members in both images, and the position of the extraction reference that defines the extraction range on the reference image can be corrected. . Even if temperature changes, shocks, vibrations, etc. occur, the same partial area can always be cut out based on the cut-out criteria after correction, so the noise is obtained by subtracting the background image that picked up noise from the image picked up from coins. Can be removed.

  Further, according to the present invention, the distance and displacement amount of the reference member can be calculated in units of pixels, and the change in the image magnification of the image and the cut-out position of the partial area image can be accurately corrected in units of pixels.

  Further, according to the present invention, since the reference member is fixedly arranged in the sensor case of the image sensor, it is not necessary to separately prepare a reference medium, a reference chart, etc., and image magnification or positional deviation is required. Can be easily corrected.

  Further, according to the present invention, by fixing and arranging a reference member formed of a material having a low linear expansion coefficient and a high light reflectance on a sensor case formed of a material having a low linear expansion coefficient. Since the position of the reference member does not change even in an environment in which the refractive index of the lens unit changes with a temperature change, image correction using the reference member can be performed.

FIG. 1 is a schematic diagram illustrating a configuration example of a coin identifying device according to the present embodiment. FIG. 2 is a diagram illustrating an image captured by an image sensor included in the coin identification device and a configuration of the image sensor. FIG. 3 is a block diagram showing a schematic configuration of the coin identifying device. FIG. 4 is a diagram for explaining a method of correcting an image using a reference plate. FIG. 5 is a diagram for explaining a method of removing a noise component using a background image. FIG. 6 is a flowchart for describing processing for acquiring information serving as a reference for image correction. FIG. 7 is a flowchart for explaining a processing method for identifying coins while performing image correction. FIG. 8 is a flowchart for explaining another processing method for identifying coins while performing image correction.

  In the coin discriminating apparatus according to the present embodiment, when a change in image magnification or distortion occurs in an image obtained by picking up a coin using the lens due to a change in the refractive index of the lens caused by a temperature change, the correction is made. In addition, a predetermined partial region image including the entire coin image can be cut out from the captured image. Hereinafter, preferred embodiments of a coin identifying device and a coin identifying method according to the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 1 is a schematic diagram illustrating a configuration example of a coin identifying device according to the present embodiment. As shown in FIG. 1, in the coin discriminating apparatus, when the coin 100 transported in the direction indicated by the arrow on the transport path 25 passes through the magnetic sensor 20, a temporary denomination is determined. Then, when passing through the image sensor 10 </ b> A provided on the upper surface side of the transport path 25 and the image sensor 10 </ b> B provided on the lower surface side of the transport path 25, images of the front and back surfaces of the coin 100 are acquired. From the result of the temporary denomination determination by the magnetic sensor 20, after narrowing down the denomination of the coin 100, the denomination of the coin 100 is obtained using the front and back images of the coin 100 acquired by the image sensors 10 </ b> A and 10 </ b> B. Authenticity, damage, etc. are determined. Since the image sensor 10A and the image sensor 10B have the same structure, the image sensor 10 will be described below.

  The image sensor 10 has a measurement window 14 made of a transparent glass plate or the like on the conveyance path 25 side. When the coin 100 reaches the imaging position on the conveyance path 25, light is emitted from the light source 13 toward the coin 100. Irradiate. The light reflected by the coin 100 enters the lens unit 12 through the measurement window 14. Then, the light condensed by the lens unit 12 is imaged on the image sensor 11, and the image of the coin 100 is acquired by the image sensor 11.

  The magnetic sensor 20 is a sensor for detecting the characteristics of the coin such as the material and thickness of the coin 100 and the rough outer diameter, and outputs the detected signal to the denomination provisional determination unit 31 of the control unit 30 described later ( (See FIG. 3).

  Next, a detailed configuration of the image sensor 10 will be described. FIG. 2 is a diagram for explaining an image captured by the image sensor 10 and a configuration of the image sensor 10. A schematic cross-sectional view of the image sensor 10 is shown on the lower side of FIG. 2, and an example of an image captured by the image sensor 10 is shown on the upper side.

  As shown in the lower diagram of FIG. 2, the image sensor 10 includes an imaging element 11 such as a CCD disposed on a substrate 17 inside an upper case 18 and a lower case 19 made of resin that block light, and the imaging element 11. A lens unit 12 for forming an image on the substrate, a light source 13 composed of a plurality of LEDs and the like disposed on the substrate 16, and a mirror 15 that reflects the light from the light source 13 and irradiates it toward the coin 100. Have. Further, the measurement window 14 is attached to the upper case 18 so that the upper surface of the measurement window and the upper surface of the case are on the same plane.

  Although not shown in the lower diagram of FIG. 2, on the substrate 17 on which the image sensor 11 is arranged, a drive circuit that drives the image sensor 11, detects that the coin 100 has reached the imaging position, and acquires an image. A timing sensor for determining timing, a control circuit for the light source 13, an interface for communicating with a control unit 30 (see FIG. 3), which will be described later, and the like are mounted. The detection result by the timing sensor is used for controlling the imaging timing by the imaging device 11 and the like.

  2, the lens unit 12 is configured by combining a plurality of lenses. By using the lens unit 12, the imaging element 11 captures an image of a range including the measurement window 14 and a partial region of the upper case 18 (“image imaging range” in the figure). The coin 100 is imaged in an area (“coin passage area” in the figure) corresponding to the measurement window 14 in the image capturing range. In the upper case 18, two reference plates 51 and 52 are fixed outside the coin passage area and inside the image capturing range as reference members used for correcting image magnification and brightness.

  As a result, as shown in the upper diagram of FIG. 2, the image 60 acquired by capturing the image capturing range by the image sensor 11 includes the coin passing region 70 corresponding to the measurement window 14 and the reference plates 51 and 52. The reference plate images 151 and 152 corresponding to are included. The two reference plates 51 and 52 fixed to the upper case 18 are fixed to the upper case 18 by adjusting their positions on the image 60 so as to be diagonal and equidistant from the center of the image. The adjustment of the arrangement positions of the reference plates 51 and 52 is for performing a process of removing dust or the like on the measurement window 14 as noise on the image, and not performing the noise removal process. It is not necessary to adjust the arrangement position that is symmetric with respect to the center of the image. Details of the noise removal processing will be described later.

  The upper case 18 is formed using a material such as a resin having a small linear expansion coefficient and easy to process, and the reference plates 51 and 52 are formed using a material having a small linear expansion coefficient and high light reflectance. . For example, the reference plates 51 and 52 are gray sheet-like members having a predetermined reflectance that is higher than that of the upper case 18 and are formed using metal or the like. Even when the refractive index or volume of the resin lens forming the lens unit 12 changes in an environment where the coin identification device 1 is used, the positions, sizes, and shapes of the upper case 18 and the reference plates 51 and 52 change. Therefore, based on the images of the reference plates 51 and 52 included in the image 60, a change in image magnification and image distortion can be detected and corrected. Further, for example, even when the lens in the lens unit 12 is displaced due to impact or vibration and the position of the coin passage region 70 is displaced in the image 60, the positions of the reference plate images 151 and 152 on the image 60 Thus, the amount of deviation can be calculated and corrected, and the image of the coin passage area 70 can be cut out correctly. Moreover, if the image 60 is acquired after adjusting the light source 13 so that a clear coin image can be imaged, and the brightness values (pixel values) of the reference plate images 151 and 152 are recorded as the brightness reference values, coin processing is performed. When performing the above, it is possible to correct the luminance value of the reference plate images 151 and 152 so as to match the luminance reference value by controlling the light emission amount of the light source 13. Details of each process will be described later.

  FIG. 3 is a block diagram showing a schematic configuration of the coin identifying device 1. As illustrated in FIG. 3, the coin identifying device 1 includes a control unit 30 and a storage unit 40 in addition to the image sensor 10 and the magnetic sensor 20 illustrated in FIGS. 1 and 2. The control unit 30 includes a denomination temporary determination unit 31, an image acquisition unit 32, a luminance calculation unit 33, a light source control unit 34, an image cutout unit 35, an image correction unit 36, and a coin identification unit 37. The storage unit 40 also stores correction reference information 41 that is information for correcting image magnification, brightness, and the like, and a template 42 for identifying the denomination and authenticity of the coin 100.

  The denomination temporary determination unit 31 receives a detection signal when the coin 100 is detected by the magnetic sensor 20, performs a temporary determination of the denomination of the coin 100, and outputs the obtained temporary determination result to the coin identification unit 37. Have

  The image acquisition unit 32 has a function of controlling the image sensor 10 and acquiring the front and back images of the coin 100. Specifically, light is emitted from the light source 13 toward the coin 100 passing through the transport path 25, and the timing at which the coin 100 reaches the imaging position is detected, and the coin 100 is imaged by the imaging element 11.

  The luminance calculation unit 33 corrects the luminance value so that the luminance value of the image of the reference plates 51 and 52 included in the image acquired by the image acquisition unit 32 matches a predetermined luminance reference value. It has a function of calculating the correction amount. Specifically, the ratio between the luminance reference value stored in advance in the storage unit 40 as the correction reference information 41 and the luminance value of the reference plates 51 and 52 on the acquired image is calculated as the light amount ratio. Based on the calculated light amount ratio, the light source controller 34 controls the light source 13 to adjust the luminance value of the image. That is, in the present embodiment, the luminance value (pixel value) is not changed by image processing, but by adjusting the light emission amount of the light source 13 by the functions of the luminance calculation unit 33 and the light source control unit 34, Is corrected.

  The light source control unit 34 has a function of adjusting the light emission amount of the light source 13 based on the light amount ratio calculated by the luminance calculation unit 33. Specifically, when the light source control unit 34 causes the light source 13 to emit light with a predetermined light emission amount set as an initial value, an image is acquired by the image acquisition unit 32. When the light quantity ratio calculated by the luminance calculation unit 33 based on the acquired image and the luminance reference value is input to the light source control unit 34, the light source control unit 34 performs a reference on the image acquired by the image acquisition unit 32. The light emission amount of the light source 13 is adjusted so that the luminance values of the images of the plates 51 and 52 coincide with the luminance reference value. Since the brightness reference value is set in advance so that the pattern on the coin surface clearly appears on the coin image, the pattern of the coin 100 can be obtained with stable brightness by controlling the light source 13 based on the brightness reference value. It is possible to acquire a coin image that clearly shows.

  In addition, as a method for adjusting the light amount, for example, there are a method for adjusting a light emission current for causing the light source 13 to emit light, and a method for adjusting a light emission time. In addition, the light source 13 is configured by a plurality of LEDs that emit light at different wavelengths, and the light frequency is changed by changing the color frequency of the light source 13 by changing the LED that emits light according to the color, reflection characteristics, etc. of the coin. There is also a way to adjust. There is also a method of adjusting the gain of the image sensor 11 on the light receiving side instead of or in addition to changing the light amount of the light source 13. Further, a correction amount (for example, a multiplication value necessary for matching the acquired pixel value with the luminance reference value) is calculated from the pixel values of the captured reference plate images 151 and 152, and an image in which the coin 100 is captured is formed. The coin identifying process may be performed after correcting the pixel value of each pixel. Further, a mode in which the light amount is adjusted by combining a plurality of these techniques may be used. The light emission current adjustment and the lighting time adjustment of the light source 13 are performed before the coin identification process is started. For example, after calculating the light amount ratio between the pixel value of the reference plate images 151 and 152 on the image captured in the state where no coin 100 has arrived and the luminance reference value, the relationship between the light amount ratio, the light emission current, and the lighting time. Referring to a table that prescribes in advance or obtaining the correction amount of the light emission current and the light emission time by a calculation formula, perform multiple imaging and adjust to meet the desired condition, and then The coin identification process is started. However, since the pixel values of the reference plate images 151 and 152 can be obtained from the image obtained by imaging the coin 100 even after the coin identification process is started, each time the coin 100 flowing through the transport path 25 is imaged, It is also possible to adjust the light quantity of the light source 13. For example, when changing the gain on the light receiving side, the pixel values of the reference plate images 151 and 152 are acquired and the correction amount is calculated when each coin 100 is imaged, and the coin 100 is imaged based on the correction amount. An image for identification may be created by correcting the image.

  As illustrated in FIG. 2, the image cutout unit 35 has a function of cutting out a partial region image of the coin passage region 70 from the image 60 obtained by picking up the image pickup range by the image sensor 10. For example, even when the position of the coin passage region 70 is shifted on the image 60 because the lens in the lens unit 12 has been displaced, the image cutout unit 35 includes the reference plate image included in the image 60. The shift amount is calculated based on the reference positions 151 and 152, and the partial shift image of the coin passage region 70 is cut out after correcting the shift based on the shift amount.

  The image correction unit 36 has a function of calculating a change or distortion of the image magnification from the positional relationship between the two reference plate images 151 and 152 included in the image 60 and correcting this by image processing. For example, the distance between the two reference plate images 151 and 152 on the image 60 acquired in an environment of a predetermined temperature is set as the distance reference value. Then, when the distance between the two reference plate images 151 and 152 on the image 60 acquired when performing the coin identification process is shorter or longer than the distance reference value, image correction is performed. The unit 36 corrects the image so that the changed distance matches the distance reference value. At this time, the image correction unit 36 can also correct image distortion based on the change rate in the X-axis direction and the change rate in the Y-axis direction between the two reference plate images 151 and 152. In the present embodiment, the two reference plates 51 and 52 are shown. For example, three or more reference plates are arranged in the image capturing range, and the moment feature on the image obtained by capturing these is used. Thus, the inclination of the image can be corrected.

  As described above, the image correction unit 36 has a function of correcting a change, distortion, inclination, and the like of the image magnification of the image. Hereinafter, a case where the image magnification is corrected will be described as an example, and details will be described later. To do.

  After the image obtained by capturing the coin 100 by the image acquisition unit 32 is processed by the image correction unit 36 and the image cutout unit 35, the coin identification unit 37 uses the obtained coin image to determine the denomination of the coin 100. It has a function of identifying authenticity.

  Specifically, after the light emission amount of the light source 13 is adjusted by the luminance calculation unit 33 and the light source control unit 34, the coin 100 conveyed through the conveyance path 25 is imaged by the image acquisition unit 32. The image correction unit 36 corrects the image magnification of the image of the coin passage area 70 cut out by the image cutout unit 35 from the captured image. The coin identifying unit 37 calculates the coin diameter of the coin 100 using the image thus obtained. Further, the coin identifying unit 37 refers to the corresponding template 42 from the storage unit 40 based on the temporary determination result by the denomination temporary determining unit 31, and the coin diameter calculated from the coin diameter and the image included in the template 42. And compare. When the diameter of the coin 100 matches the coin diameter included in the template 42 within a predetermined error range, that is, when it is determined that the coin 100 is a denomination coin according to the provisional determination result, a coin image is further added. Is used to determine the denomination and the like of the coin 100 in detail. For example, the coin image and the coin image included in the template 42 are compared, the degree of coincidence between the two images is evaluated by an image matching technique, and the denomination or the like of the coin 100 is determined. And based on the determination result based on a coin diameter, and the determination result by image comparison, the identification results, such as a money type of the coin 100, are finally determined.

  Regarding the processing for determining the coin diameter by converting the coin image into polar coordinates and the processing for comparing the coin image and the template image, for example, the conventional technique disclosed in Japanese Patent Application Laid-Open No. 2010-26992 can be used, and thus details are described. The detailed explanation is omitted. In the following, description of the processing contents by the luminance calculation unit 33, the image cutout unit 35, and the image correction unit 36, which is one of the features according to the present embodiment, will be continued.

  FIG. 4 is a diagram for explaining processing contents of luminance correction by the luminance calculation unit 33 and the light source control unit 34, image extraction by the image extraction unit 35, and image correction by the image correction unit 36. 4A and 4B are reference images 61 and 62 acquired by the image acquisition unit 32 in order to acquire the correction reference information 41 stored in the storage unit 40 under the environment of the reference temperature. FIG. 4C shows an image acquisition unit 32 for calculating the correction amount of the brightness, the cutout position, and the image magnification when the coin identification process is started in the environment where the coin identification device 1 is installed. It is the acquired image 63 for correction. FIG. 4D shows, from the left side, an image 64 acquired at the time of the coin identification process, a cutout image 73 obtained by cutting out the coin passage area 72 by correcting the positional deviation from the image 64, and an image of the cutout image 73. An identification image 74 with the magnification corrected is shown.

  First, the installation positions of the lens unit 12 and the image sensor 11 of the image sensor 10 shown in FIG. 2 are adjusted under the environment of the reference temperature. As a result, the image pickup range including the reference plates 51 and 52 is picked up by the image pickup device 11 in a focused state. 4A, on the rectangular image 61, the central pixel of the image 61 and the central pixel of the coin passage area 71 coincide, and the four sides of the coin passage area 71 corresponding to the four sides of the image 61 are matched. And become parallel. In addition, the two reference plate images 151 and 152 included in the image 61 have a diagonal and equidistant positional relationship with respect to the central pixel of the image 61. However, when it is not necessary to correct the positional deviation of the coin passage area 70 in the captured image, the reference plate images 151 and 152 do not have to be diagonal and equidistant, and the reference plate images 151 and 152 are in the imaging range. As long as it is included.

  An image obtained by placing a jig such as a gray card having a predetermined reflectance on the measurement window 14 of the image sensor 10 shown in FIG. 2 is a reference image 61 shown in FIG. The reference image 61 is picked up after adjusting the light emission amount of the light source 13 using a gray card. The left and right X coordinates (X1 and X2) that define the position of the coin passage area 71 are specified with the upper left corner of the reference image 61 as the origin and the number of pixels as the coordinate value. These pieces of information obtained from the reference image 61 are stored in the storage unit 40 as correction reference information 41.

  Subsequently, an image captured with the jig removed from the upper surface of the measurement window 14 is a reference image 62 shown in FIG. In this reference image 62, with the upper left corner as the origin, the upper left corner 151a (a1, b1) and lower right edge 151b (a2, b2) of the reference plate image 151, the upper left edge 152a (a3, b3) of the reference plate image 152, and The right lower end 152b (a4, b4) is specified. Then, a distance reference value L0 shown in FIG. 4B is calculated using the identified coordinate values. In addition, the luminance values at the center positions of the two reference plate images 151 and 152 are acquired, and these are set as the luminance reference values C1 and C2. These pieces of information obtained from the reference image 62 are stored in the storage unit 40 as correction reference information 41.

  The storage of the correction reference information 41 in the storage unit 40 using the reference images 61 and 62 shown in FIGS. 4A and 4B is performed, for example, when the coin identifying device 1 is manufactured. Then, when the user of the coin identifying device 1 performs the coin identifying process, the correction of the luminance, the cutout position, and the image magnification using the correction reference information 41 stored in the storage unit 40 is performed.

  FIG. 4C is a correction image 63 that is captured in the previous stage after the coin 100 is conveyed on the conveyance path 25 after the coin identification process is started. Since the correction image 63 is an image captured at a temperature different from the reference temperature when the reference images 61 and 62 illustrated in FIGS. 4A and 4B are captured, a change in the refractive index of the lens unit 12 is performed. Thus, the image has an image magnification different from that of the reference images 61 and 62. Further, if there is a displacement of the lens constituting the lens unit 12, the position of the coin passage area 72 is displaced on the correction image 63, and the center pixel and the center pixel of the coin passage area 72 on the correction image 63 are shifted. Will not match.

  2 has a configuration in which the center of the lens unit 12 coincides with the center of the image capturing range of the upper case 18 and the center of the image sensor 11. For this reason, the correction image 63 in FIG. 4C obtained in a state where the refractive index of the lens constituting the lens unit 12 is changed due to the temperature change is the image center of the reference image 61 shown in FIG. As a reference, the entire image is enlarged or reduced.

  In FIG. 4C, for the sake of simplicity of explanation, a description will be given of a method for correcting misalignment when the coin passage area 72 is deviated in the X-axis direction in the correction image 63. That is, FIG. 4C shows a state where the cutting references X1 and X2 shown in FIG. 4A are shifted to X3 and X4 shown in FIG.

  First, image brightness correction performed by adjusting the light emission amount of the light source 13 will be described. The luminance calculation unit 33 acquires the luminance values C11 and C12 of the center pixel of the reference plate images 251 and 252 on the correction image 63 shown in FIG. Then, the luminance calculation unit 33 calculates a light amount ratio K = {(C1 + C2) / (C11 + C12)} with the luminance reference values C1 and C2 included in the correction reference information 41. When imaging the coin 100 to be processed in the coin identification process, the light source control unit 34 controls the light emission amount of the light source 13 based on the light amount ratio K input from the luminance calculation unit 33. In addition, about the luminance values C11 and C12, the aspect which calculates | requires and utilizes the average value of the luminance value of the center pixel of the reference | standard plate image 251,252 and / or the several pixel periphery of a center pixel may be sufficient.

  Next, correction of the cutting position of the coin passage area 72 will be described. Based on the distance reference value L0 on the reference image 62 shown in FIG. 4B and the distance L1 on the correction image 63 shown in FIG. = L1 / L0 is calculated. Subsequently, the image cutout unit 35 uses the reference positions which are the coordinates 151a and 152b of the reference plate images 151 and 152 on the reference image 62 shown in FIG. 4B, and the correction image shown in FIG. 63. Based on the corresponding coordinates 252a and 252b on 63, a deviation amount Dx between the cutting references X1 and X2 shown in FIG. 5A and the cutting references X3 and X4 shown in FIG. To do. The deviation amount Dx in the X-axis direction is calculated as Dx = {(a11 + a14) − (a1 + a4)} / 2. Further, the image cutout unit 35 includes a distance Lx between the cutout references X1 and X2 shown in FIG. 4A and the X coordinate (Xc) of the center line of the correction image 63 shown in FIG. To get. Then, based on these values, the image cutout unit 35 calculates the coordinate values of the cutout references X3 and X4 on the correction image 63. Specifically, as shown in FIG. 4C, the coordinate values of the cutting references X3 and X4 after the positional shift are X3 = Xc− (Lx / 2) S + Dx and X4 = Xc + (Lx / 2) Calculated as S + Dx.

  In this way, the light quantity ratio K, the image magnification S, and the deviation amount Dx in the X-axis direction are calculated, and when the coin 100 to be processed in the coin identification process begins to be transported through the transport path 25, the brightness calculation unit 33 inputs it. Based on the light quantity ratio K, the light source controller 34 controls the light emission amount of the light source 13. Then, at the timing when the coin 100 conveyed through the conveyance path 25 reaches the imaging position by the image sensor 10, the image acquisition unit 32 acquires an image 64 including the coin image 101 as shown in FIG. Is done. Subsequently, the image cutout unit 35 obtains a partial region image of the coin passage region including the coin image 101 from the acquired image 64 based on the cutout references X3 and X4 in which the positional deviation is corrected, that is, FIG. ) Of the partial region image corresponding to the coin passage region 72. As shown in FIG. 4D, the cut-out image 73 cut out in this way is in a state before image magnification correction is performed, and therefore a different width from the width Lx of the coin passage region 71 shown in FIG. (Lx × S). Subsequently, the image correcting unit 36 generates an identification image 74 enlarged or reduced based on the image magnification S from the cut image 73 cut out by the image cutout unit 35. As a result, the identification image 74 is an image whose horizontal width matches the horizontal width of the coin passage region 71 of the reference image 61, and the coin image 102 is also an accurate image with the image magnification corrected.

  The identification image 74 corrected by the image correcting unit 36 is input to the coin identifying unit 37. And in the coin identification part 37, the coin identification process compared with the process which determines a coin diameter, and a template image is performed using the coin image 102 contained in this image 74 for identification.

  The correction of the image magnification is performed in order to obtain a coin image 102 that can be compared with the data of the template 42 stored in the storage unit 40 in advance. On the other hand, the correction of the cut-out position of the coin passage area is performed for the purpose of noise removal processing for removing noise components from the identification image. The fact that the cutout position needs to be corrected in order to remove the noise component will be described with reference to FIG.

  FIG. 5 is a diagram for explaining the removal of noise components performed in the identification image. For ease of explanation, FIG. 5A shows an image 65 in which the image magnification and the positional deviation have been corrected, and FIGS. 5B and 5C show the correction of the image magnification and the X-axis direction. An image 66 before the position shift is corrected is shown.

  FIG. 5A shows an image 65 acquired without the coin 100 during the coin identification process, and a background image 75 obtained by cutting out the coin passage area 75 from the image 65. For example, if dust such as coin powder adheres on the measurement window 14 of the image sensor 10 shown in FIG. 2, the acquired image 65 corresponds to the measurement window 14 as shown in FIG. The dust 80 is reflected in the coin passage area 75.

  Note that the image 65 shown in FIG. 5A has been corrected for image magnification and positional deviation, and thus can be regarded as a diagram showing noise 80 on the reference image. Therefore, the value obtained from the coordinates 251a (5, 5), 252b (250, 250) of the reference plate images 251 and 252 with the upper left corner of the image 65 as the origin is the distance reference value L0 = 346.5 pixels. In addition, the X coordinates of the cutting reference X10 (20, 0) and X11 (220, 0) indicate that the coin passing area 75 has a width of 100 pixels on the left and right with the center line as a boundary.

  In the image 66 shown in FIG. 5B in which the image 100 is picked up by the image acquisition unit 32 during the coin identification process, a change in image magnification due to a temperature change and the lens unit 12 are configured. Misalignment of the coin passage area 76 caused by the misalignment of the lens is generated. For this reason, the image cutout unit 35 and the image correction unit 36 perform a process of correcting them.

  Specifically, as shown in FIG. 5B, the distance (355 pixels) obtained from the coordinates 251a (4, 2) and 252b (255, 253) of the reference plate images 251 and 252 and the distance reference value ( The image magnification S = 1.024 (= 355 / 346.5) is calculated based on (346.5 pixels). This indicates that in the reference image, the image is cut out in the range of 100 pixels left and right from the center line of the image, and in the image 66, this distance is 102.4 (= 100 × S) pixels. Although it is necessary to move the extraction reference to the outside by 2.4 pixels, the image extraction unit 35 enlarges the extraction reference to the outside by 3 pixels because the image extraction is performed in units of one pixel.

  At this time, when the image is cut out by correcting only the image magnification without taking the positional deviation into consideration, as shown in FIG. 5B, the cutting references X10 (20, 0), X11 ( The partial region image of the coin passage region 76 is cut out between the cutting references X20 (17, 0) and X21 (223, 0) obtained by enlarging 220,0) three pixels to the left and right based on the image magnification S. Then, the image correction unit 36 performs correction based on the image magnification S, and a corrected image 77 shown in FIG. 5B is obtained. However, in this corrected image 77, since the positional deviation of the coin passage area is not taken into consideration, the position of the noise 81 included in this corrected image 77 is the position of the noise 80 included in the background image 75 shown in FIG. It will be out of position. As a result, when the difference image obtained by subtracting the background image 75 from the corrected image 77 is used as the identification image 78, as shown in FIG. 5B, a part of the noise 81 remains in the identification image 78. It becomes an image. Thus, even if an image including the noise 81 is input to the coin identifying unit 37 as the identification image 78, the coin 100 may not be correctly identified.

  In the coin discriminating apparatus 1, the image cutout unit 35 uses the coordinates 251a (5, 5) and 252b (250, 250) of the reference plate images 251 and 252 of the reference image as reference positions, and uses the coordinates and coins of the reference positions. From the coordinates 251a (4, 2) and 252b (255, 253) of the reference plate images 251 and 252 of the captured image 66, the displacement amount Dx of the positional displacement is calculated. In this case, the shift amount Dx is 2 pixels (= {(4 + 255) − (5 + 250)} / 2). For this reason, the image cutout unit 35 further calculates the shift amount Dx = 2 pixels from the cutout references X20 (17,0) and X21 (223,0) in FIG. 5B corrected based on the image magnification S. As shown in FIG. 6C, X30 (19, 0) and X31 (225, 0) shifted by two pixels in the positive direction of the X axis are used as the extraction reference. That is, the cutting reference is not set to 103 pixels equally from the center line Xc of the image 66 as shown in FIG. 5B, but left from the center line Xc as shown in FIG. 5C. 101 pixels and 105 pixels to the right are set.

  Thus, as shown in FIG. 5C, the image magnification S is corrected by cutting out the image 176 between the extraction references X30 and X31 in which the image magnification S and the positional deviation amount Dx are corrected. A corrected image 177 is obtained. In this corrected image 177, since the positional deviation is taken into consideration in addition to the image magnification, the position of the noise 80 included in the background image 75 shown in FIG. 5A and the corrected image 177 shown in FIG. The position of the noise 81 matches. When the image correction unit 36 subtracts the background image 75 from the correction image 177, the noise 81 is removed from the correction image 177, and an identification image 178 shown in FIG. 5C can be obtained. In this way, by inputting the image from which the noise 81 is removed to the coin identifying unit 37 as the identification image 178, it is possible to correctly identify the coin.

  Next, the processing procedure of each process described above will be described with reference to the flowcharts of FIGS. FIG. 6 is a flowchart illustrating a process for acquiring information to be stored as the correction reference information 41 in the storage unit 40 in the coin identification device 1. FIG. 6A shows a process of acquiring information related to the cutting references X1 and X2 shown in FIG. 4A, and FIG. 6B shows the luminance reference values C1 and C2 and the distance reference shown in FIG. The process which acquires the information regarding the value L0 is shown.

  First, under a reference temperature environment, a jig made of a gray chart is placed in the measurement window 14 of the image sensor 10, and the reference image 61 shown in FIG. S1). Then, using the reference image 61, as described with reference to FIG. 4A, the positions of the cutting references X1 and X2 of the image of the coin passage area 71 are determined (step S2). Information relating to the cutting standards X1 and X2 is stored in the storage unit 40 as correction standard information 41.

  Moreover, the reference image 62 shown in FIG.4 (b) is acquired by the image acquisition part 32 in the state of a reference temperature, without the jig | tool in the measurement window 14 (step S11). Then, using the reference image 62, as described with reference to FIG. 4B, the luminance value at the center of the reference plate images 151 and 152 is acquired, and this is used as the luminance reference values C1 and C2. (Step S12). Further, the distance between the reference plate images 151 and 152 is acquired, and this is set as the distance reference value L0 (step S13). Information regarding the luminance reference values C1 and C2 and the distance reference value L0 is stored in the storage unit 40 as correction reference information 41.

  Thus, the coin identifying process is started with the correction reference information 41 stored in the storage unit 40. FIG. 7 is a flowchart for explaining a coin identifying process performed by correcting misalignment, image magnification, and the like.

  First, in the image sensor 10, the light source 13 is turned on by the light source control unit 34 with the light emission amount set as the initial value. And the image 63 for correction shown in FIG.4 (c) is acquired by the image acquisition part 32 in the state without the coin 100 (step S21).

  Using the acquired correction image 63, as described with reference to FIG. 4C, the luminance calculation unit 33 acquires the luminance values C11 and C12 of the central pixels of the reference plate images 251 and 252. (Step S22), the light amount ratio K is calculated from the luminance values C11 and C12 and the luminance reference values C1 and C2 (Step S23).

  Next, the image cutout unit 35 calculates the distance L1 between the reference plate images 251 and 252 (step S24), and calculates the image magnification S from the distance L1 and the distance reference value L0 (step S25). The image cutout unit 35 calculates a deviation amount Dx of the reference image 61 from the cutout references X1 and X2 (step S26), and cuts out the partial area image of the coin passage area 72 from the correction image 63. The coordinates of the cutting references X3 and X4 are determined (step S27).

  Next, the light source control unit 34 controls the light emission amount of the light source 13 based on the light amount ratio K input from the luminance calculation unit 33 (step S28). And the image acquisition part 32 images the coin 100 and acquires an image according to the timing at which the coin 100 to be processed in the coin identification process has been transported through the transport path 25 (step S29). Then, as described with reference to FIG. 4D, the image cutout unit 35 determines from the acquired image 64 between the cutout references X3 and X4 that have corrected the position shift based on the shift amount Dx. A partial area image 73 corresponding to the coin passage area is cut out (step S30). Then, the image correction unit 36 corrects the image magnification of the cut-out partial area image 73 based on the calculated image magnification S (step S31). In this way, an identification image 74 shown in FIG. 4D can be obtained.

  If the correction image previously acquired in step S21 includes noise in which dust or the like due to coin dust is reflected, processing for removing this noise is performed (step S32). Specifically, as described with reference to FIG. 5, the image correction unit 36 captures the coin 100 with the background image 75 showing the noise 80 on the reference image as shown in FIG. Then, the image obtained by subtracting from the corrected image 177 shown in FIG.

  The identification image thus obtained is input from the image correction unit 36 to the coin identification unit 37. And the coin identification part 37 performs the identification process which identifies the money type etc. of the coin 100 using the input image for identification (step S33).

  While there is a coin 100 to be processed in the coin identification process (step S34; No), an image of each coin 100 from which noise is removed by correcting the image magnification and displacement is acquired, and the coin is based on this image. The processes of steps S29 to S33 for identifying 100 denominations and the like are repeated. And if the process of all the coins 100 made into a process target is completed (step S34; Yes), a coin identification process will be completed.

  In FIG. 7, the correction amounts for the positional deviation and the image magnification are obtained first (steps S24 to S27), and thereafter, using these correction amounts, a plurality of coins 100 to be processed are identified. Although an embodiment has been shown, the processing order is not limited to this. For example, it may be a mode in which the correction amount of the positional deviation and the image magnification is obtained for each coin 100 to be processed. FIG. 8 is a flowchart for explaining a coin identifying process performed by obtaining a correction amount of the positional deviation and the image magnification in each coin 100 to be processed.

  In the flowchart shown in FIG. 8, the same step numbers as in FIG. 7 are assigned to the corresponding steps in order to show the difference from FIG. 7, but each step is performed in the order indicated by the arrows in the flowchart. Further, the processing contents of each step shown in FIG. 8 are the same as the contents shown in FIG.

  In the coin identification process shown in FIG. 8, after the process is started, an image is acquired without the coin 100 (step S21), and brightness values C11 and C12 are acquired from the image (step S22). The light source 13 is calculated (step S23), and the light source 13 is controlled based on the calculated light quantity ratio K (step S28). And the image of each coin 100 is acquired (step S29), the process which calculates | requires the position shift Dx and the image magnification S is performed (steps S24-27), and based on the calculated | required correction amount, the image of which the position shift was corrected is performed. Cutout processing (step S30) and image magnification correction processing (step S31) are performed.

  If the image contains noise, a process for removing the noise is performed (step S32), and a process for identifying the denomination of the coin 100 using the obtained image is performed (step S32). S33). And while the coin 100 used as a process target exists (step S34; No), these processes (step S29, S24-S27, S30-33) are repeated. And if the process of all the coins 100 made into a process target is completed (step S34; Yes), a coin identification process will be completed.

  In the process shown in FIG. 7, the identification processing of each coin 100 is performed after the correction amount of the positional deviation and the image magnification is obtained first, whereas in the processing shown in FIG. Since the magnification correction amount is obtained, it can be corrected even when a positional deviation or a change in image magnification occurs during processing.

  In addition, the aspect which the coin identification device 1 selects and performs not only the aspect which performs a process only in any one processing order of the process shown in FIG. 7, and the process shown in FIG. It does not matter. 7 and 8, the process of correcting the luminance on the image by adjusting the light emission amount of the light source 13, the process of correcting the positional deviation of the partial area image on the captured image, and the change of the image magnification are corrected. Although all the processes of the process to remove and the process to remove noise are shown, it goes without saying that each process is executed only when necessary.

  Moreover, although the example which correct | amends the position shift of a X-axis direction was shown in this embodiment, in the coin identification device 1, it can correct also about the position shift of other directions, such as a Y-axis direction, with the same method. it can. In the present embodiment, an example in which the image magnification is corrected using an image cut out by correcting the positional deviation has been described. However, the positional deviation may be corrected after the image magnification is corrected. The shape of the reference plates 51 and 52 may be a polygon such as a triangle or a pentagon or a circle, and the number of the reference plates 51 and 52 may be three or more. If the above-described calculation of the distance reference value and the luminance reference value and the correction of the image magnification and the positional deviation using these values can be executed with the above-described contents, the reference plates 51 and 52 are examples shown in the present embodiment. It is not limited to.

  Further, in the present embodiment, the method of obtaining the distance reference value, the luminance reference value, etc. on the reference image has been described, but for these values, for example, at the time of manufacture, all or some of the coin identification devices 1 use the reference temperature. A value obtained by actually capturing a reference image under a predetermined condition such as the above may be used, or a design value, a theoretical value, or the like may be used.

  As described above, according to the present embodiment, when the image magnification changes due to a change in the refractive index of the lens unit 12 caused by a temperature change, or when the lens unit 12 or individual lenses constituting the lens unit are shocked. Even when a position shift occurs due to vibration or vibration, the image obtained by capturing the coin 100 can be corrected using the correction reference information 41 to obtain an image that can be compared with the template 42 prepared in advance. For this reason, the denomination etc. of the coin 100 can be accurately identified without being affected by temperature change, impact, vibration, or the like.

  Further, since the image magnification and the positional deviation are corrected using images obtained by imaging the reference plates 51 and 52, it is necessary to prepare a reference medium for image correction or to provide a temperature sensor in the image sensor 10. There is no. In addition, since the correction amount is determined based on the measured temperature, it is not necessary to prepare data indicating the relationship between the temperature and the correction amount in advance.

  As described above, the coin discriminating apparatus and the coin discriminating method according to the present invention acquire a coin image in which the influence of a change in image magnification caused by a temperature change or the like and a lens position shift is corrected, Useful for accurately identifying authenticity.

DESCRIPTION OF SYMBOLS 1 Coin identification apparatus 10, 10A, 10B Image sensor 11 Image pick-up element 12 Lens unit 13 Light source 14 Measurement window 15 Mirror 16, 17 Board | substrate 18 Upper case 19 Lower case 20 Magnetic sensor 25 Transport path 30 Control part 31 Denomination provisional determination part 32 Image acquisition unit 33 Luminance calculation unit 34 Light source control unit 35 Image cutout unit 36 Image correction unit 37 Coin identification unit 40 Storage unit

Claims (8)

A coin identification device that performs identification processing based on an image obtained by irradiating light from a light source toward a coin and imaging the coin,
An image sensor for imaging an imaging range including a coin passing area through which a coin to be identified passes and a plurality of reference members fixedly arranged outside the coin passing area on an imaging element by a lens unit; ,
A storage unit for storing a distance reference value indicating a distance between the reference members on the image;
An coin corrector comprising: an image correcting unit that corrects an image captured by the image sensor so that a distance between the reference members calculated on the image matches the distance reference value. The distance reference value indicates a distance between the reference members under a reference temperature,
The image correction unit
Correcting a change in image magnification caused by a temperature change based on the distance reference value and a distance between the reference members on an image captured by the image sensor at a temperature different from the reference temperature. The coin identification device according to claim 1, wherein: A luminance calculating unit that calculates a ratio between a luminance reference value indicating a luminance value of the image of the reference member and a luminance value of the image of the reference member included in the image captured by the image sensor as a light amount ratio;
The coin identifying apparatus according to claim 1, further comprising a light source control unit that controls a light emission amount of the light source based on the light amount ratio calculated by the luminance calculation unit. The amount of deviation between the reference position on the image indicating the position of the reference member and the position of the reference member included in the image captured by the image sensor is calculated, and the position where the partial region image is cut out from the image is defined. 4. The image cutting unit according to claim 1, further comprising an image cutout unit that corrects a cutout reference based on the shift amount, and cuts out the partial region image using the cutout reference after correction. Coin identification device. 5. The coin identifying apparatus according to claim 4 , wherein the distance and the shift amount calculated on the image are calculated in units of one pixel.   The coin identifying apparatus according to any one of claims 1 to 5, wherein the reference member is fixed in a sensor case constituting the image sensor. The sensor case and the reference member are made of a material having a smaller linear expansion coefficient than the material forming the lens unit,
The coin identifying device according to claim 6, wherein the reference member is made of a material having a higher reflectance than the sensor case. A coin identification method for performing identification processing based on an image obtained by irradiating light from a light source toward a coin and imaging the coin,
Storing a distance reference value indicating a distance between reference members on the image;
An image imaging step of imaging an imaging range including a coin passing area through which a coin to be identified passes and a plurality of reference members fixed outside the coin passing area on the imaging element by a lens unit; and
An image correction step of correcting the image captured in the image capturing step so that a distance between the reference members calculated on the image coincides with the distance reference value. Method.

Images