JP7688984B2 - Paper sheet identification device and paper sheet identification method - Google Patents

Description

This disclosure relates to a paper sheet recognition device, a paper sheet recognition method, and a paper sheet recognition program.

In a paper sheet identification device that identifies paper sheets such as banknotes, various sensors are used to acquire the characteristics of the paper sheet. Then, based on the acquired characteristics of the paper sheet, it is common to identify (determine) the type (denomination), authenticity, and fitness of the paper sheet.

For example, Patent Documents 1 and 2 disclose an optical sensor that irradiates a banknote with infrared light of multiple wavelengths from a light source and receives the light reflected or transmitted by the banknote with a light receiving unit.

International Publication No. 2019/082251 International Publication No. 2020/208806

However, the output data of the light receiving unit may vary due to environmental factors such as characteristic variations between light emitting elements (e.g., LEDs) of the light source, shifts in peak wavelength due to temperature, and fluctuations in the transmittance of the light guide of the light source (e.g., a light guide made of acrylic resin) according to wavelength. In such cases, the paper sheet identification process is performed based on output data with variations, which reduces the accuracy of the paper sheet identification process. Furthermore, this variation in output data due to environmental factors tends to be particularly noticeable in the infrared range, so accuracy is particularly likely to deteriorate when the paper sheet identification process is performed based on output data in the infrared range.

In addition, when irradiating a banknote with light of multiple wavelengths and identifying paper sheets based on output data relating to the light of multiple wavelengths obtained from the light receiving unit, it is common to perform the identification process using only the output data relating to the light of each wavelength. However, this may result in insufficient performance in distinguishing between genuine and counterfeit identification objects (e.g., special ink) on the paper sheets. Insufficient performance in distinguishing between genuine and counterfeit identification objects leads to a decrease in the identification accuracy of the paper sheets.

This disclosure has been made in consideration of the current situation, and aims to provide a paper sheet recognition device, a paper sheet recognition method, and a paper sheet recognition program that can improve the accuracy of paper sheet recognition.

In order to solve the above-mentioned problems and achieve the objective, (1) a paper sheet recognition device according to a first aspect of the present disclosure is a paper sheet recognition device that recognizes paper sheets having an identification object provided thereon, and includes a light source that irradiates light onto the paper sheet, a light receiving unit that receives light coming from the paper sheet, and a control unit that acquires output data from the light receiving unit, and the control unit acquires first data that is output data from the light receiving unit corresponding to an area including the identification object and second data that is output data from the light receiving unit corresponding to an area not including the identification object, generates third data by correcting the first data using the second data, and recognizes the paper sheet based on the third data.

(2) In the paper sheet identification device described in (1) above, the light source may irradiate light of multiple wavelengths onto the paper sheet, the light receiving unit may receive light of multiple wavelengths arriving from the paper sheet, the first data and the second data may each include data relating to the light of the multiple wavelengths, the control unit may generate third data relating to the multiple wavelengths by correcting the first data relating to the light of the multiple wavelengths with second data relating to the corresponding light of the same wavelength, and the paper sheet may be identified based on the third data relating to the light of the multiple wavelengths.

(3) In the paper sheet identification device described in (2) above, the control unit may calculate first multiplication data, which is data obtained by multiplying together data related to the first wavelength and the second wavelength of the first data related to the light of the multiple wavelengths, and second multiplication data, which is data obtained by multiplying together data related to the first wavelength and the second wavelength of the second data related to the light of the multiple wavelengths, and may generate fourth data by correcting the first multiplication data with the second multiplication data, and may identify the paper sheet based on the third data related to the light of the multiple wavelengths and the fourth data.

(4) In the paper sheet recognition device described in any one of (1) to (3) above, the control unit may use a representative value of output data of an area that does not include the recognition object as the second data.

(5) In the paper sheet identification device described in any one of (1) to (4) above, the light source may include a rod-shaped light guide made of acrylic resin and a light-emitting element facing at least one of the two end faces of the light guide, and the light source may irradiate light onto the paper sheet through the light guide.

(6) In the paper sheet recognition device described in any one of (1) to (5) above, the light source may irradiate the paper sheet with linear light in the main scanning direction, the light receiving unit may receive linear light in the main scanning direction coming from the paper sheet, and the control unit may use output data according to the position in the main scanning direction of the area including the recognition object as the second data.

(7) A paper sheet identification device according to a second aspect of the present disclosure is a paper sheet identification device that identifies a paper sheet having an identification target provided thereon, and includes a light source that irradiates the paper sheet with light of multiple wavelengths, a light receiving unit that receives the light of multiple wavelengths arriving from the paper sheet, and a control unit that acquires output data relating to the light of the multiple wavelengths from the light receiving unit, and the control unit calculates multiplication data that is data obtained by multiplying data relating to light of a first wavelength and data relating to light of a second wavelength among the acquired output data relating to the light of the multiple wavelengths, and identifies the paper sheet based on the output data relating to the light of the multiple wavelengths and the multiplication data.

(8) In the paper sheet identification device described in any one of (1) to (7) above, the light source may irradiate the paper sheet with infrared light, and the infrared light may include light with a wavelength of 850 nm or more and 950 nm or less.

(9) In the paper sheet identification device described in any one of (1) to (8) above, the light receiving unit may receive light irradiated from the light source and reflected by the paper sheet, and the output data may include data related to the light reflected by the paper sheet.

(10) In the paper sheet recognition device described in any one of (1) to (9) above, the paper sheet to be recognized may include, as an object to be recognized, at least one of an ink whose reflectance decreases as the wavelength becomes longer in the infrared region and an ink whose reflectance increases as the wavelength becomes longer in the infrared region.

(11) A paper sheet identification method according to a third aspect of the present disclosure is a paper sheet identification method for identifying a paper sheet having an identification object provided thereon, and includes the steps of: (A) acquiring output data from a light receiving unit that receives light irradiated from a light source and transmitted from the paper sheet; (B) acquiring first data, which is output data of the light receiving unit corresponding to an area including the identification object, and second data, which is output data of the light receiving unit corresponding to an area not including the identification object; (C) generating third data by correcting the first data using the second data; and (D) identifying the paper sheet based on the third data.

(12) In the paper sheet identification method described in (11) above, the light source may irradiate light of multiple wavelengths onto the paper sheet, the light receiving unit may receive light of multiple wavelengths arriving from the paper sheet, the first data and the second data may each include data relating to the light of the multiple wavelengths, and in the step (C), third data relating to the multiple wavelengths may be generated by correcting the first data relating to the light of the multiple wavelengths with second data relating to the corresponding light of the same wavelength, and in the step (D), the paper sheet may be identified based on the third data relating to the light of the multiple wavelengths.

(13) The paper sheet identification method described in (12) above may further include a step (E) of calculating first multiplication data, which is data obtained by multiplying together data related to the first wavelength and the second wavelength of the first data related to the light of the multiple wavelengths, and second multiplication data, which is data obtained by multiplying together data related to the first wavelength and the second wavelength of the second data related to the light of the multiple wavelengths. In the step (C), fourth data may be generated by correcting the first multiplication data with the second multiplication data, and in the step (D), the paper sheet may be identified based on the third data related to the light of the multiple wavelengths and the fourth data.

(14) In the paper sheet recognition method described in any one of (11) to (13) above, in step (C), a representative value of output data in an area that does not include the recognition object may be used as the second data.

(15) In the paper sheet identification method described in any one of (11) to (14) above, the light source may include a rod-shaped light guide made of acrylic resin and a light-emitting element facing at least one of the two end faces of the light guide, and the light source may irradiate light onto the paper sheet through the light guide.

(16) In the paper sheet identification method described in any of (11) to (15) above, the light source may irradiate the paper sheet with linear light in the main scanning direction, and the light receiving unit may receive linear light in the main scanning direction coming from the paper sheet, and in the above step (C), output data corresponding to the position in the main scanning direction of the area including the object to be identified may be used as the second data.

(17) A paper sheet identification method according to a fourth aspect of the present disclosure is a paper sheet identification method for identifying a paper sheet having an identification target provided thereon, and includes the steps of: acquiring output data relating to light of multiple wavelengths from a light receiving unit that receives light of multiple wavelengths irradiated from a light source and arriving from the paper sheet; calculating multiplication data that is data obtained by multiplying data relating to light of a first wavelength and data relating to light of a second wavelength among the acquired output data relating to light of multiple wavelengths; and identifying the paper sheet based on the output data relating to light of multiple wavelengths and the multiplication data.

(18) In the paper sheet identification method described in any of (11) to (17) above, the light source may irradiate the paper sheet with infrared light, and the infrared light may include light with a wavelength of 850 nm or more and 950 nm or less.

(19) In the paper sheet identification method described in any one of (11) to (18) above, the light receiving unit may receive light irradiated from the light source and reflected by the paper sheet, and the output data may include data related to the light reflected by the paper sheet.

(20) In the paper sheet identification method described in any of (11) to (19) above, the paper sheet to be identified may include, as an identification object, at least one of an ink whose reflectance decreases as the wavelength becomes longer in the infrared region and an ink whose reflectance increases as the wavelength becomes longer in the infrared region.

(21) A paper sheet recognition program according to a fifth aspect of the present disclosure is a paper sheet recognition program for identifying a paper sheet having an identification object provided thereon using a paper sheet recognition device, and causes the paper sheet recognition device to execute a process (A) of acquiring output data from a light receiving unit that receives light irradiated from a light source and transmitted from the paper sheet, a process (B) of acquiring first data that is output data of the light receiving unit corresponding to an area including the identification object and second data that is output data of the light receiving unit corresponding to an area not including the identification object, a process (C) of generating third data by correcting the first data using the second data, and a process (D) of identifying the paper sheet based on the third data.

(22) In the paper sheet identification program described in (21) above, the light source may irradiate light of multiple wavelengths onto the paper sheet, the light receiving unit may receive light of multiple wavelengths arriving from the paper sheet, the first data and the second data may each include data relating to the light of the multiple wavelengths, and in the process (C), third data relating to the multiple wavelengths may be generated by correcting the first data relating to the light of the multiple wavelengths with second data relating to the corresponding light of the same wavelength, and in the process (D), the paper sheet may be identified based on the third data relating to the light of the multiple wavelengths.

(23) The paper sheet identification program described in (22) above may further include a process (D) for calculating first multiplication data, which is data obtained by multiplying together data related to the first and second wavelengths of the first data related to the light of the multiple wavelengths, and second multiplication data, which is data obtained by multiplying together data related to the first and second wavelengths of the second data related to the light of the multiple wavelengths. In the process (C), fourth data may be generated by correcting the first multiplication data with the second multiplication data, and in the process (D), the paper sheet may be identified based on the third data related to the light of the multiple wavelengths and the fourth data.

(24) In the paper sheet recognition method described in any one of (21) to (23) above, in the process (C), a representative value of output data in an area that does not include the recognition object may be used as the second data.

(25) In the paper sheet identification program described in any one of (21) to (24) above, the light source may include a rod-shaped light guide made of acrylic resin and a light-emitting element facing at least one of the two end faces of the light guide, and the light source may irradiate light onto the paper sheet through the light guide.

(26) In the paper sheet identification program described in any one of (21) to (25) above, the light source may irradiate the paper sheet with linear light in the main scanning direction, and the light receiving unit may receive linear light in the main scanning direction coming from the paper sheet, and in the process (C) above, output data corresponding to the position in the main scanning direction of the area including the object to be identified may be used as the second data.

(27) A paper sheet identification program according to a sixth aspect of the present disclosure is a paper sheet identification program for identifying a paper sheet having an identification target provided thereon using a paper sheet identification device, and causes the paper sheet identification device to execute the following processes: acquiring output data relating to light of multiple wavelengths from a light receiving unit that receives light of multiple wavelengths irradiated from a light source and arriving from the paper sheet; calculating multiplication data that is data obtained by multiplying data relating to light of a first wavelength and data relating to light of a second wavelength among the acquired output data relating to light of multiple wavelengths; and identifying the paper sheet based on the output data relating to light of the multiple wavelengths and the multiplication data.

(28) In the paper sheet identification program described in any one of (21) to (27) above, the light source may irradiate the paper sheet with infrared light, and the infrared light may include light with a wavelength of 850 nm or more and 950 nm or less.

(29) In the paper sheet identification program described in any one of (21) to (28) above, the light receiving unit may receive light irradiated from the light source and reflected by the paper sheet, and the output data may include data related to the light reflected by the paper sheet.

(30) In the paper sheet identification method described in any one of (21) to (29) above, the paper sheet to be identified may include, as an identification object, at least one of an ink whose reflectance decreases as the wavelength becomes longer in the infrared region and an ink whose reflectance increases as the wavelength becomes longer in the infrared region.

According to the present disclosure, it is possible to provide a paper sheet recognition device, a paper sheet recognition method, and a paper sheet recognition program that can improve the accuracy of paper sheet recognition.

FIG. 2 is a schematic diagram illustrating an example of a configuration of the paper sheet recognition device according to the first embodiment, showing a banknote transport path as viewed from the side. FIG. 2 is a schematic plan view showing an example of a banknote to be recognized in the first embodiment. 4 is an example of a graph showing the reflectance in the infrared region of ink provided on a banknote to be recognized in the first embodiment. FIG. 1 is a schematic diagram illustrating an example of a configuration of a paper sheet recognition device according to a first embodiment, as viewed from an oblique direction. FIG. 2 is a schematic diagram illustrating an example of the configuration of the paper sheet recognition device according to the first embodiment, showing a banknote transport path viewed from above. FIG. 2 is a schematic diagram showing image data of an entire banknote. 5 is a flowchart illustrating an example of an operation of the paper sheet recognition device according to the first embodiment. FIG. 11 is a schematic diagram illustrating an example of a configuration of a paper sheet recognition device according to a second embodiment, in which a banknote transport path is viewed from the side. FIG. 11 is a schematic plan view showing an example of a banknote to be recognized in the second embodiment. FIG. 11 is a schematic diagram illustrating an example of a configuration of a paper sheet recognition device according to a second embodiment, as viewed from an oblique direction. FIG. 11 is a schematic diagram illustrating an example of the configuration of a paper sheet recognition device according to a second embodiment, in which a conveyance path for banknotes is viewed from above. 10 is a flowchart illustrating an example of an operation of the paper sheet recognition device according to the second embodiment. FIG. 11 is a schematic perspective view showing the appearance of an example of a paper sheet processing apparatus in which a paper sheet recognition apparatus according to a third embodiment can be mounted. FIG. 11 is a block diagram illustrating an example of a configuration of a paper sheet recognition device according to a third embodiment. 13 is a schematic cross-sectional view illustrating an example of a configuration of an optical line sensor included in a paper sheet recognition device according to a third embodiment. FIG.

Hereinafter, an embodiment of a paper sheet recognition device, a paper sheet recognition method, and a paper sheet recognition program according to the present disclosure will be described with reference to the drawings. Although various paper sheets such as banknotes, checks, gift certificates, bills, documents, securities, and card-like media can be applied as the paper sheets to which the present disclosure is applied, the present disclosure will be described below using an apparatus for banknotes as an example. The paper sheet recognition program may be pre-installed in the paper sheet recognition device, or may be recorded on a computer-readable recording medium or provided to an operator via a network. In addition, in the following description, the same reference numerals are appropriately used in common between different drawings for the same parts or parts having similar functions, and repeated description thereof will be omitted as appropriate. In addition, mutually orthogonal XYZ coordinate systems are appropriately shown in the drawings explaining the structure.

(Embodiment 1)
The configuration of the paper sheet recognition device according to the present embodiment will be described with reference to Fig. 1. Fig. 1 is a schematic diagram for explaining an example of the configuration of the paper sheet recognition device according to the present embodiment, and is a diagram showing a banknote transport path from the side. As shown in Fig. 1, the paper sheet recognition device 1a according to the present embodiment includes a light source 11a that irradiates light onto a banknote BN, a light receiving unit 13a that receives light coming from the banknote BN, and a control unit 20a that acquires output data of the light receiving unit 13a, and can be used by being mounted on, for example, a paper sheet processing device that processes banknotes. The banknote BN to be recognized may be transported in the X direction within an XY plane.

Fig. 2 is an example of a schematic plan view showing an example of a banknote to be recognized in this embodiment. Fig. 3 is an example of a graph showing the reflectance in the infrared region of ink provided on a banknote to be recognized in this embodiment. As shown in Fig. 2, an identification object S is provided (e.g., printed) on at least one main surface of a banknote BN to be recognized in this embodiment. On the main surface of the banknote BN, an area ROI _A , for example, rectangular, that includes the identification object S, and an area ROI _B , for example, rectangular, that does not include the identification object S are set.

The banknote BN to be recognized may contain, as the recognition target S, at least one of ink whose reflectance decreases as the wavelength becomes longer in the infrared region (hereinafter, negative ink) and ink whose reflectance increases as the wavelength becomes longer in the infrared region (hereinafter, positive ink), as shown in FIG. 3. The paper sheet recognition device 1a is capable of identifying banknotes BN containing such inks with high accuracy. Hereinafter, negative ink and positive ink may be collectively referred to as special ink. In FIG. 3, IR1 is, for example, 800 nm, IR2 is, for example, 870 nm, and IR3 is 940 nm.

Light source 11a irradiates light onto banknote BN. The light irradiated by light source 11a may be light in a wavelength band including a peak wavelength and wavelengths in the vicinity of the peak wavelength. The type (wavelength) of light irradiated from light source 11a is not particularly limited, and may include visible light such as white light, red light, green light, and blue light, as well as infrared light.

The light receiving unit 13a receives light coming from the banknote BN. That is, it can function as an optical sensor. The light receiving unit 13a may receive light irradiated from the light source 11a and reflected by the banknote BN. That is, while the light source 11a irradiates light, the light receiving unit 13a may receive light reflected by the banknote BN. At this time, the light receiving unit 13a can function as a sensor having sensitivity at least to the wavelength band of the light irradiated from the light source 11a. The light receiving unit 13a may output an electrical signal corresponding to the amount of light received as output data. More specifically, the light receiving unit 13a may be equipped with a light receiving element, and the light receiving element may receive light and convert it into an electrical signal corresponding to the amount of incident light, and the light receiving unit 13a may output the electrical signal. Note that, here, "amount of light" means a physical quantity proportional to the radiation intensity and incident time of the incident light.

FIG. 4 is a schematic diagram illustrating an example of the configuration of a paper sheet recognition device according to this embodiment, viewed from an oblique direction. As shown in FIG. 4, the light source 11a and the light receiving unit 13a may form an optical line sensor 14a extending in the Y direction. In this case, the Y direction corresponds to the main scanning direction of the optical line sensor 14a, and the X direction corresponds to the sub-scanning direction of the optical line sensor 14a. The light source 11a may irradiate light in a straight line extending in the Y direction. The light receiving unit 13a may include a plurality of light receiving elements (light receiving pixels) arranged in a line in the Y direction, and may form a linear image sensor.

The length of the light source 11a and the light receiving unit 13a in the Y direction (main scanning direction) may be longer than the length of the banknote BN in the Y direction. The light source 11a may irradiate light linearly over the entire banknote BN in the Y direction, and the light receiving unit 13a may receive the light reflected over the entire banknote BN in the Y direction. That is, the light receiving unit 13a may output an electrical signal according to the amount of incident light in multiple channels corresponding to multiple light receiving elements (positions in the Y direction). The channels are numbers assigned to the light receiving elements in order in the Y direction. At this time, the light receiving unit 13a may output line data, which is data related to the light received simultaneously in each channel, as output data. While transporting the banknote BN in the X direction (sub-scanning direction), data related to the reflected light of the entire banknote BN may be obtained by repeating the irradiation of light by the light source 11a and the reception of light by the light receiving unit 13a.

Figure 5 is a schematic diagram illustrating an example of the configuration of a paper sheet recognition device according to this embodiment, and is a view of the banknote transport path from above. As shown in Figure 5, the light source 11a may include a light guide 15a and light emitting elements 17a facing two end faces 15aa of the light guide 15a, and may irradiate light onto the banknote BN via the light guide 15a.

The light guide 15a is a transparent rod-shaped optical component that guides the light from the light-emitting element 17a and irradiates linear light toward the banknote BN that is the target of irradiation, linearizing the light emitted from the light-emitting element 17a.

The light guide 15a may be made of acrylic resin. Since an acrylic resin light guide 15a has a large effect on the output data of the light receiving unit 13a, in this case, it is particularly effective in reducing the variability of the first data described below. Therefore, it is possible to particularly effectively improve the recognition accuracy of banknotes BN.

The light-emitting element 17a is an element that emits light toward the opposing end surface 15aa, and may be, for example, an LED (Light Emitting Diode). Note that multiple light-emitting elements 17a may be provided for the corresponding end surface. Furthermore, the light-emitting element 17a may be disposed facing only one of the two end surfaces 15aa.

The light source 11a may irradiate the banknote BN with infrared light including light with a wavelength of 850 nm or more and 950 nm or less. Since the transmittance of the light guide 15a (e.g., a light guide made of acrylic resin) of the light source 11a varies particularly greatly in the infrared region around 900 nm, the above configuration can particularly effectively reduce the variation in the output data (first data described later). Therefore, it is possible to particularly effectively improve the identification accuracy of the banknote BN. In addition, it is possible to identify with high accuracy banknote BN that contains, as the identification object S, ink (e.g., special ink) whose reflectance varies in the infrared region around 900 nm.

Each light receiving pixel of the light receiving unit 13a may be equipped with a light receiving element having sensitivity across multiple mutually different wavelength bands, or may be equipped with multiple types of light receiving elements that selectively receive light of mutually different wavelength bands. In the former case, the light source 11a may irradiate the banknote BN with light of multiple wavelengths in sequence, and the light receiving unit 13a may receive light of each wavelength in accordance with the irradiation timing of the light of each wavelength. In the latter case, the light source 11a may irradiate the banknote BN with light of multiple wavelengths simultaneously, and the light receiving unit 13a may receive the light of the multiple wavelengths with multiple types of light receiving elements, respectively.

The control unit 20a performs a process of acquiring output data from the light receiving unit 13a. That is, it acquires data according to the amount of light received by the light receiving unit 13a. The output data from the light receiving unit 13a acquired by the control unit 20a may be digitalized. The control unit 20a may acquire image data of the entire banknote BN as the output data from the light receiving unit 13a.

Figure 6 is a schematic diagram showing image data of the entire banknote. Image data of the entire banknote BN is data (two-dimensional data) obtained by capturing an image of the entire banknote BN, and may be composed of a number of pixels Pix arranged in a matrix in the Y direction (main scanning direction) and X direction (sub-scanning direction) as shown in Figure 6. The address of each pixel Pix may be specified by the channel of the light receiving unit 13a corresponding to the position in the Y direction and the line corresponding to the position in the X direction. The line is a number assigned in sequence to the line data output sequentially by the light receiving unit 13a.

The output data of the light receiving unit 13a acquired by the control unit 20a may include data related to light reflected by the banknote BN. This makes it possible to identify with high accuracy banknotes BN that contain ink with a characteristic reflectance (e.g., special ink) as the identification target S.

The control unit 20a may also acquire image data relating to the reflected light of the entire banknote BN as output data from the light receiving unit 13a. Hereinafter, the image data relating to the reflected light is referred to as reflected image data.

The resolution of the output data acquired by the control unit 20a may be the same as or different from the resolution of the output data of the light receiving unit 13a, and may be lower, for example, in the Y direction (main scanning direction) and the X direction (sub-scanning direction).

The control unit 20a may control each part of the paper sheet recognition device 1a, and may be configured, for example, with programs for implementing various processes, including a paper sheet recognition program, a CPU (Central Processing Unit) that executes the programs, and various hardware (e.g., FPGA (Field Programmable Gate Array)) controlled by the CPU.

The control unit 20a performs a process of acquiring first data, which is output data of the light receiving unit 13a corresponding to an area ROI _A (see Figure 2) that includes the object to be identified S, and second data, which is output data of the light receiving unit 13a corresponding to an area ROI _B (see Figure 2) that does not include the object to be identified S.

The first data and the second data may both be partial data of the reflection image data of the entire banknote BN, i.e., reflection image data of a portion of the banknote BN. The regions ROI _A and ROI _B may be set in advance according to the denomination of the banknote BN, and the control unit 20a may extract the first data and the second data from the output data of the light receiving unit 13a (e.g., reflection image data of the entire banknote BN) based on the position information of the regions ROI _A and ROI _B set for each denomination.

The control unit 20a then performs a process (hereinafter, sometimes referred to as a correction process) of generating third data by correcting the first data with the second data, and performs a process (hereinafter, sometimes referred to as a recognition process) of identifying the banknote BN based on the third data. The output data of the light receiving unit 13a may vary due to environmental factors such as characteristic variations of each element of the light emitting element 17a (e.g., LED) of the light source 11a, a shift in peak wavelength due to temperature, a change in the transmittance of the light guide 15a (e.g., a light guide made of acrylic resin) of the light source 11a according to the wavelength, and characteristic variations of the light receiving element itself of the light receiving unit 13a. However, according to this embodiment, the first data, which is the output data of the light receiving unit 13a corresponding to the region ROI _A including the identification target S, is corrected with the second data, which is the output data of the light receiving unit 13a corresponding to the region ROI _B not including the identification target S, to generate the third data, so that the variation of the output data (particularly the first data) due to the above environmental factors can be reduced. Since the banknote BN is identified based on the corrected third data, the banknote BN can be identified based on data with reduced variation, thereby improving the accuracy of identifying the banknote BN.

In the correction process, the control unit 20a may generate the third data by normalizing the first data with the second data. For example, the control unit 20a may perform a process of dividing each output value (pixel value) of the first data by an output value related to the second data.

Furthermore, the control unit 20a may use, as the second data, a representative value (e.g., average value, median value, etc.) of the output data of the light receiving unit 13a corresponding to the region ROI _B not including the identification target S. For example, the control unit 20a may standardize (divide) each output value (pixel value) of the first data by using an average value of the output values (pixel values) included in the reflection image data of the region ROI _B not including the identification target S.

The light source 11a may irradiate the banknote BN with light of multiple wavelengths, the light receiving unit 13a may receive the light of multiple wavelengths arriving from the banknote BN, the first data and the second data may each include data relating to the light of multiple wavelengths, the control unit 20a may generate third data relating to the light of multiple wavelengths by correcting the first data relating to the light of multiple wavelengths with the corresponding second data relating to the same wavelength in the correction process, and may identify the banknote BN based on the third data relating to the light of multiple wavelengths in the identification process. This makes it possible to identify with high accuracy the banknote BN provided with an identification object S (e.g., special ink) whose characteristics (e.g., reflectance) change depending on the wavelength.

In this case, the control unit 20a may perform a process (hereinafter, sometimes referred to as a multiplication process) to calculate first multiplication data, which is data obtained by multiplying data related to the first and second wavelengths of the first data related to the light of multiple wavelengths, and second multiplication data, which is data obtained by multiplying data related to the first and second wavelengths of the second data related to the light of multiple wavelengths. In the correction process, a process may be further performed to generate fourth data by correcting the first multiplication data with the second multiplication data. In the identification process, the banknote BN may be identified based on the third data related to the light of multiple wavelengths and the fourth data. This can improve the performance of separating the authenticity of the identification target S (especially the special ink). Therefore, it is possible to further improve the identification accuracy of the banknote BN.

In general inks, such as infrared absorbing inks and infrared non-absorbing inks, there is almost no difference in reflectance in the infrared region, and in the space of output data related to infrared light, each output is distributed in the direction of vector (1,1,1) ^T passing through the origin. On the other hand, this is not the case for special inks. Therefore, the product of output data related to light of multiple wavelengths is more susceptible to influence in general inks than in special inks, which contributes to improving the degree of separation between special inks and general inks.

Note that "multiple wavelengths of light" refers to light with different wavelength bands, and includes at least a first wavelength and a second wavelength. For example, it may be two wavelengths of light, i.e., a first wavelength and a second wavelength, or it may be three wavelengths of light, i.e., a first wavelength, a second wavelength, and a third wavelength. For example, the multiple wavelengths of light may be light with different colors for visible light, and for infrared light and ultraviolet light, it may be light whose wavelength bands only partially overlap each other, or light whose wavelength bands do not overlap each other.

In addition, the "first data (second data) relating to light of multiple wavelengths" refers to the first data (second data) based on the data output by the light receiving unit 13a when it receives light of each wavelength, and includes the first data (second data) relating to light of the first wavelength through the first data (second data) relating to light of the Nth wavelength (N is an integer equal to or greater than 2).

When two wavelengths of light are used, the control unit 20a may, in the correction process, generate third data (hereinafter, λ1 third data) related to the light of the first wavelength by correcting the first data (hereinafter, λ1 first data) related to the light of the first wavelength with the second data (hereinafter, λ1 second data) related to the light of the first wavelength, and may generate third data (hereinafter, λ2 third data) related to the light of the second wavelength by correcting the first data (hereinafter, λ2 first data) related to the light of the second wavelength with the second data (hereinafter, λ2 second data). When three wavelengths of light are used, the control unit 20a may, in the correction process, further generate third data (hereinafter, λ3 third data) related to the light of the third wavelength by correcting the first data (hereinafter, λ3 first data) related to the light of the third wavelength with the second data (hereinafter, λ1 second data).

Furthermore, when light of two wavelengths is used, in the multiplication process, the control unit 20a may calculate, as the first multiplication data, data obtained by multiplying the first data of λ1 by the first data of λ2 (hereinafter, λ1×λ2 first data), and may calculate, as the second multiplication data, data obtained by multiplying the second data of λ1 by the second data of λ2 (hereinafter, λ1×λ2 second data), and in the correction process, may generate, as the fourth data, data obtained by correcting the first data of λ1×λ2 by the second data of λ1×λ2 (hereinafter, λ1×λ2 fourth data).

When light of three wavelengths is used, the control unit 20a may further calculate, as the first multiplication data in the multiplication process, data obtained by multiplying the first data of λ1 by the first data of λ3 (hereinafter, λ1×λ3 first data) and data obtained by multiplying the first data of λ2 by the first data of λ3 (hereinafter, λ2×λ3 first data), and may further calculate, as the second multiplication data, data obtained by multiplying the second data of λ1 by the second data of λ3 (hereinafter, λ1×λ3 second data) and data obtained by multiplying the second data of λ2 by the second data of λ3 (hereinafter, λ2×λ3 second data), and may further generate, as the fourth data in the correction process, data obtained by correcting the first data of λ1×λ3 by the second data of λ1×λ3 (hereinafter, λ1×λ3 fourth data) and data obtained by correcting the first data of λ2×λ3 by the second data of λ2×λ3 (hereinafter, λ2×λ3 fourth data).

When light of two wavelengths is used, the first multiplication data may be data obtained by multiplying each output value (pixel value) of the first data related to light of the first wavelength by each output value (pixel value) of the second data related to light of the second wavelength (however, the output values related to the same point in the XY plane are multiplied together). When light of three wavelengths is used, output values (pixel values) may be multiplied together in a similar manner (however, the output values related to the same point in the XY plane are multiplied together).

When light of two wavelengths is used, the second multiplied data may be data obtained by multiplying a representative value of the second data relating to light of the first wavelength by a representative value of the second data relating to light of the second wavelength (however, representative values of the same type (e.g., average values) are multiplied together). When light of three wavelengths is used, representative values may be multiplied together (however, representative values of the same type) in a similar manner.

When using light of two wavelengths, the control unit 20a may identify banknote BN in the identification process based on the third data of λ1 and the third data of λ2 (two-dimensional features), or may identify banknote BN based on the third data of λ1, the third data of λ2, and the fourth data of λ1 x λ2 (three-dimensional features).

When light of three wavelengths is used, the control unit 20a may identify banknote BN in the identification process based on the third data of λ1, the third data of λ2, and the third data of λ3 (three-dimensional features), or may identify banknote BN based on the third data of λ1, the third data of λ2, the third data of λ3, the fourth data of λ1×λ2, the fourth data of λ1×λ3, and the fourth data of λ2×λ3 (six-dimensional features).

When light of multiple wavelengths is used as described above, the light receiving unit 13a may receive the light of multiple wavelengths reflected by the banknote BN while the light source 11a irradiates the light of multiple wavelengths. In this case, the light receiving unit 13a may function as a sensor sensitive to at least the wavelength band of the light irradiated from the light source 11a. The light receiving unit 13a may output an electrical signal corresponding to the amount of light received as output data for each wavelength. More specifically, the light receiving unit 13a may be equipped with a light receiving element, which may receive light and convert it into an electrical signal corresponding to the amount of incident light, and the light receiving unit 13a may output the electrical signal for each wavelength.

When light of multiple wavelengths is used, the light of each wavelength irradiated by the light source 11a may be light of a wavelength band including a peak wavelength and a wavelength in the vicinity thereof, and the light of multiple wavelengths irradiated by the light source 11a may have different peak wavelengths. The light source 11a may also include multiple light-emitting elements 17a having different peak wavelengths. For example, the light source 11a may irradiate infrared light of multiple wavelengths having different peak wavelengths, or may include multiple light-emitting elements 17a each irradiating infrared light of multiple wavelengths having different peak wavelengths. The light-receiving unit 13a may receive infrared light of multiple wavelengths irradiated from the light source 11a and reflected by the banknote BN. Similarly, the first data and the second data may each be data related to infrared light of multiple wavelengths. Furthermore, various data related to light of multiple wavelengths processed by the control unit 20a may be data related to infrared light of multiple wavelengths.

For example, the light source 11a may irradiate infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light) having different peak wavelengths, the light receiving unit 13a may receive the infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light) irradiated from the light source 11a and reflected by the banknote BN, and the first data and the second data may each include data related to the infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light). The various data related to the light of multiple wavelengths processed by the control unit 20a may be data related to the infrared light of multiple wavelengths (for example, the first and second infrared light, or the first to third infrared light).

The first and second infrared lights may be a combination of any two of the following: infrared light having a peak wavelength in the wavelength band of 770 to 830 nm, infrared light having a peak wavelength in the wavelength band of 840 to 900 nm, and infrared light having a peak wavelength in the wavelength band of 910 to 970 nm. Note that either of the wavelengths of the first and second infrared lights may be larger.

Similarly, the first to third infrared lights may be a combination of infrared light having a peak wavelength in the 770 to 830 nm wavelength band, infrared light having a peak wavelength in the 840 to 900 nm wavelength band, and infrared light having a peak wavelength in the 910 to 970 nm wavelength band. Note that there is no particular limitation on the magnitude relationship of the wavelengths of the first to third infrared lights.

As shown in Fig. 4, the light source 11a may irradiate the banknote BN with linear light in the main scanning direction (Y direction), the light receiving unit 13a may receive linear light in the main scanning direction (Y direction) coming from the banknote BN, and the control unit 20a may use output data corresponding to the position in the main scanning direction (Y direction) of an area ROI _A (see Fig. 2) including the object to be identified S as the second data. When the position in the main scanning direction changes, the length of the light guide 15a (e.g., a light guide made of acrylic resin) of the light source 11a through which the light passes also changes, so the output data also changes in the main scanning direction, and the change is particularly noticeable at the center in the main scanning direction (the center of the light guide 15a). However, according to the above configuration, output data corresponding to the position in the main scanning direction of the area ROI _A including the object to be identified S is used as the second data for correction, so that the variation in the output data caused by the position in the main scanning direction can be reduced. Furthermore, although the characteristics of the light receiving elements may vary depending on the position in the main scanning direction, the above-described configuration can reduce such variations in the characteristics of the light receiving elements, thereby further improving the recognition accuracy of the banknotes BN.

In this case, the region ROI _A including the identification target S and the region ROI _B not including the identification target S may be set in the same location (same range) of the banknote BN in the main scanning direction (Y direction), and the first data and the second data may both be data in the same channel range included in the reflection image data of the entire banknote BN. For example, the first data and the second data may be data from channel n to channel m (where n and m are natural numbers satisfying n<m) of the reflection image data of the entire banknote BN.

The range of the first data and the second data in the sub-scanning direction (X direction), i.e., the number of lines, can be set appropriately and may be the same or different from each other.

In the recognition process, the control unit 20a may apply a discrimination function to each pixel and determine for each pixel whether or not the pixel has special ink. That is, the above-mentioned various data used in the recognition process may be input to the discrimination function as features (e.g., two-dimensional, three-dimensional, or six-dimensional features), and based on the output of the discrimination function, it may be determined whether or not an object to be recognized S is provided on the banknote BN. This allows the control unit 20a to speed up the recognition process.

Here, the discriminant function can be generated by supervised machine learning. Specifically, for example, discriminant analysis, support vector machines, neural networks, etc. can be used. For example, image data including the ink portion of a banknote on which special ink is printed and its background portion (peripheral portion) can be used as the teacher data. Here, the special ink portion and the background portion, i.e., the non-special ink portion, can be used as class labels. For example, in the case of negative ink, image data related to infrared light having a peak wavelength in the wavelength band of 910 to 970 nm can be binarized (separated into the ink portion and the background portion), and the result can be used as the class label for each pixel as it is. In the case of positive ink, image data related to infrared light having a peak wavelength in the wavelength band of 770 to 830 nm can be binarized (separated into the ink portion and the background portion), and the result can be used as the class label for each pixel as it is. In either case, the binarization process may be Otsu binarization.

More specifically, for example, when light of three wavelengths is used, the control unit 20a may determine whether or not an object to be recognized S is provided on the banknote BN in the recognition process based on the following formulas (1) to (6). Here, formula (3') may be used instead of formula (3).

In formulas (1) and (2), L is a threshold value, and in formula (1), if λ is equal to or greater than L, it is determined that the banknote BN has an object to be recognized S, and if λ is less than L, it is determined that the banknote BN does not have an object to be recognized S.

In formula (2), λ is an evaluation value, and the sum of _Pj within the region ROI _A including the recognition target S is calculated.

In each equation, _Pj is the result of the discrimination function f(x), and can be the sum of pixels determined to contain special ink (equation (3)), or the sum of classification scores that are the output of the discrimination function f(x) (equation (3')). In equations (3) and (3'), c is the threshold value of the classification score.

The discrimination function f(x) is expressed by formula (4), and three-dimensional features are input as shown in formula (6). In formula (5), ω(w ₀ to w ₃ ) is a coefficient vector (weight) obtained as a result of machine learning, and φ(x _j ) expressed by formula (6) is a feature vector of the target pixel j (j is a number indicating the target pixel). In formula (6), x _1j , x _2j and x _3j are pixel values related to the first, second and third lights (which may be the first, second and third infrared lights) in the target pixel _j , respectively, and μ ₁ , μ ₂ and μ ₃ are representative values of the output data related to the first, second and third lights (which may be the first, second and third infrared lights) in the region ROI _B that does not include the identification target S, respectively. Here, the discrimination function f(x) is expressed by a linear discrimination formula.

In addition, in the formula (6), x _1j /μ ₁ , x _2j /μ ₂ and x _3j /μ ₃ correspond to the third data of λ1, the third data of λ2 and the third data of λ3, respectively.

Furthermore, when light of three wavelengths is used, the control unit 20a may determine whether or not an object to be recognized S is provided on the banknote BN in the recognition process based on the following formulas (11) to (16). Here, formula (13') may be used instead of formula (13).

In equations (11) and (12), L is a threshold value. In equation (11), if λ is equal to or greater than L, it is determined that an object to be recognized S is provided on the banknote BN, and if λ is less than L, it is determined that an object to be recognized S is not provided on the banknote BN.

In formula (12), λ is an evaluation value, and the sum of _Pj within the region ROI _A including the recognition target S is calculated.

In each equation, _Pj is the result of the discrimination function f(x), and can be the sum of pixels determined to contain special ink (equation (13)), or the sum of classification scores that are the output of the discrimination function f(x) (equation (13')). In equations (13) and (13'), c is the threshold value of the classification score.

The discrimination function f(x) is expressed by equation (14), and six-dimensional features are input as shown in equation (16). In equation (15), ω(w ₀ to w ₆ ) is a coefficient vector (weight) obtained as a result of machine learning, and φ(x _j ) expressed by equation (16) is a feature vector of the target pixel j (j is a number indicating the target pixel). In equation (16), x _1j , x _2j and x _3j are pixel values related to the first, second and third lights (which may be the first, second and third infrared lights) in the target pixel _j , respectively, and μ ₁ , μ ₂ and μ ₃ are representative values of the output data related to the first, second and third lights (which may be the first, second and third infrared lights) in the region ROI _B that does not include the identification target object S, respectively. Here, the discrimination function f(x) is expressed by an equation obtained by extending a linear discrimination equation to a nonlinear model.

In addition, _in equation (16), _x1j / _μ1 , _x2j / _μ2 , x3j/ _μ3 , _x1jx2j / _μ1μ2 , _x1jx3j / _{μ1μ3 , and x2jx3j} / _μ2μ3 correspond _to the _third data of _λ1 , the third data of λ2, the _third data of λ3, the fourth data of λ1×λ2, the fourth data of λ1×λ3, and the fourth data of λ2×λ3, respectively.

The control unit 20a may perform the multiplication process simultaneously with the correction process within the same calculation process, for example as described above in equations (6) and (16).

The control unit 20a may identify the authenticity of the banknote BN in the identification process. For example, if it is determined that the banknote BN has an identification object S, it may determine that the banknote BN is a genuine note, and if it is determined that the banknote BN does not have an identification object S, it may determine that the banknote BN is a counterfeit note.

Next, the operation of the paper sheet recognition device 1a according to this embodiment will be described with reference to FIG. 7. FIG. 7 is a flowchart illustrating an example of the operation of the paper sheet recognition device according to this embodiment.

As shown in FIG. 7, first, the control unit 20a acquires output data from the light receiving unit 13a, which receives light irradiated from the light source 11a and arriving (e.g., reflected) from the banknote BN (step S11).

Next, the control unit 20a acquires first data, which is output data of the light receiving unit 13a corresponding to the region ROI _A that includes the object to be identified S, and second data, which is output data of the light receiving unit 13a corresponding to the region ROI _B that does not include the object to be identified S (step S12).

Next, the control unit 20a performs a correction process to generate third data by correcting the first data with the second data (step S13). In step S13, the control unit 20a may perform a multiplication process in addition to the correction process.

Then, the control unit 20a performs a recognition process to identify the banknote BN based on the third data (step S14), and the operation of the paper sheet recognition device 1a ends.

(Embodiment 2)
The configuration of the paper sheet recognition device according to the present embodiment will be described with reference to Fig. 8. Fig. 8 is a schematic diagram for explaining an example of the configuration of the paper sheet recognition device according to the present embodiment, and is a diagram showing a banknote transport path from the side. As shown in Fig. 8, the paper sheet recognition device 1b according to the present embodiment includes a light source 11b for irradiating a banknote BN with light of multiple wavelengths, a light receiving unit 13b for receiving the light of multiple wavelengths coming from the banknote BN, and a control unit 20b for acquiring output data related to the light of multiple wavelengths of the light receiving unit 13b, and can be used, for example, by being mounted on a paper sheet processing device that processes banknotes. The banknote BN to be recognized may be transported in the X direction in the XY plane.

Fig. 9 is a schematic plan view showing an example of a banknote to be recognized in this embodiment. As shown in Fig. 9, a recognition target S is provided (e.g., printed) on at least one main surface of the banknote BN to be recognized in this embodiment. On the main surface of the banknote BN, for example, a rectangular area ROI _A including the recognition target S is set.

The banknote BN to be recognized may contain, as the recognition object S, at least one of ink (negative ink) whose reflectance decreases as the wavelength becomes longer in the infrared region, and ink (positive ink) whose reflectance increases as the wavelength becomes longer in the infrared region, as shown in FIG. 3. The paper sheet recognition device 1b is capable of identifying banknotes BN containing such ink with high accuracy.

Light source 11b irradiates the banknote BN with light of multiple wavelengths. The type (wavelength) of light irradiated from light source 11b is not particularly limited, and examples include visible light such as white light, red light, green light, and blue light, as well as infrared light.

Note that "multiple wavelengths of light" refers to light with different wavelength bands, and includes at least a first wavelength and a second wavelength. For example, it may be two wavelengths of light, i.e., a first wavelength and a second wavelength, or it may be three wavelengths of light, i.e., a first wavelength, a second wavelength, and a third wavelength. For example, the multiple wavelengths of light may be light with different colors for visible light, and for infrared light and ultraviolet light, it may be light whose wavelength bands only partially overlap each other, or light whose wavelength bands do not overlap each other.

The light receiving unit 13b receives light of multiple wavelengths coming from the banknote BN. That is, it can function as an optical sensor. The light receiving unit 13b may receive light of multiple wavelengths irradiated from the light source 11b and reflected by the banknote BN. That is, while the light source 11b irradiates light of multiple wavelengths, the light receiving unit 13b may receive light of multiple wavelengths reflected by the banknote BN. At this time, the light receiving unit 13b may function as a sensor having sensitivity to at least the wavelength band of the light irradiated from the light source 11b. The light receiving unit 13b may output an electrical signal corresponding to the amount of light received as output data for each wavelength. More specifically, the light receiving unit 13b may be equipped with a light receiving element, and the light receiving element may receive light and convert it into an electrical signal corresponding to the amount of incident light, and the light receiving unit 13b may output the electrical signal for each wavelength.

FIG. 10 is a schematic diagram illustrating an example of the configuration of a paper sheet recognition device according to this embodiment, viewed from an oblique direction. As shown in FIG. 10, the light source 11b and the light receiving unit 13b may form an optical line sensor 14b extending in the Y direction. In this case, the Y direction corresponds to the main scanning direction of the optical line sensor 14b, and the X direction corresponds to the sub-scanning direction of the optical line sensor 14b. The light source 11b may irradiate light in a straight line extending in the Y direction. The light receiving unit 13b may include a plurality of light receiving elements (light receiving pixels) arranged in a line in the Y direction, and may form a linear image sensor.

The length of the light source 11b and the light receiving unit 13b in the Y direction (main scanning direction) may be longer than the length of the banknote BN in the Y direction. The light source 11b may irradiate light linearly over the entire banknote BN in the Y direction, and the light receiving unit 13b may receive the light reflected over the entire banknote BN in the Y direction. That is, the light receiving unit 13b may output an electrical signal according to the amount of incident light in multiple channels corresponding to multiple light receiving elements (positions in the Y direction). The channels are numbers assigned to the light receiving elements in order in the Y direction. At this time, the light receiving unit 13b outputs line data, which is data related to the light received simultaneously in each channel, as output data. While transporting the banknote BN in the X direction (sub-scanning direction), data related to the reflected light of the entire banknote BN may be obtained by repeating the irradiation of light by the light source 11b and the reception of light by the light receiving unit 13b.

Figure 11 is a schematic diagram illustrating an example of the configuration of a paper sheet recognition device according to this embodiment, and is a view of the banknote transport path from above. As shown in Figure 11, the light source 11b may include a light guide 15b and light emitting elements 17b facing two end faces 15ba of the light guide 15b, and may irradiate light onto the banknote BN via the light guide 15b.

The light guide 15b is a transparent rod-shaped optical component that guides the light from the light-emitting element 17b and irradiates linear light toward the banknote BN that is the target of irradiation, linearizing the light emitted from the light-emitting element 17b.

The light guide 15b may be made of acrylic resin.

The light-emitting element 17b is an element that emits light toward the opposing end surface 15ba, and may be, for example, an LED. Note that multiple light-emitting elements 17b may be provided for the corresponding end surface. Furthermore, the light-emitting element 17b may be disposed facing only one of the two end surfaces 15ba.

Light source 11b may irradiate banknote BN with infrared light including light with a wavelength of 850 nm or more and 950 nm or less. This makes it possible to identify with high accuracy banknote BN that contains, as the identification object S, ink (e.g., special ink) whose reflectance varies in the infrared region around 900 nm.

Each light receiving pixel of the light receiving unit 13b may be equipped with a light receiving element having sensitivity across multiple mutually different wavelength bands, or may be equipped with multiple types of light receiving elements that selectively receive light of mutually different wavelength bands. In the former case, the light source 11b may irradiate the banknote BN with light of multiple wavelengths in sequence, and the light receiving unit 13b may receive light of each wavelength in accordance with the irradiation timing of the light of each wavelength. In the latter case, the light source 11b may irradiate the banknote BN with light of multiple wavelengths simultaneously, and the light receiving unit 13b may receive the light of the multiple wavelengths with multiple types of light receiving elements, respectively.

The control unit 20b performs a process of acquiring output data relating to light of multiple wavelengths from the light receiving unit 13b. That is, it acquires data for each wavelength according to the amount of light received by the light receiving unit 13b. Note that "output data relating to light of multiple wavelengths" refers to data output by the light receiving unit 13b by receiving light of each wavelength, and includes output data relating to light of the first wavelength to output data relating to light of the Nth wavelength (N is an integer equal to or greater than 2). The output data of the light receiving unit 13b acquired by the control unit 20b may be digitized. The control unit 20b may acquire image data of the entire banknote BN as the output data of the light receiving unit 13b.

The image data of the entire banknote BN is data (two-dimensional data) obtained by capturing an image of the entire banknote BN, and may be composed of a plurality of pixels Pix arranged in a matrix in the Y direction (main scanning direction) and the X direction (sub-scanning direction) as shown in FIG. 6. The address of each pixel Pix may be specified by the channel of the light receiving unit 13b corresponding to the position in the Y direction and the line corresponding to the position in the X direction. The line is a number assigned in sequence to the line data output sequentially by the light receiving unit 13b.

The output data of the light receiving unit 13b acquired by the control unit 20b may include data related to light reflected by the banknote BN. This makes it possible to identify with high accuracy banknotes BN that contain ink with a characteristic reflectance (e.g., special ink) as the identification target S.

The control unit 20b may also acquire image data relating to the reflected light of the entire banknote BN, i.e., reflected image data of the entire banknote BN, as output data of the light receiving unit 13b.

The resolution of the output data acquired by the control unit 20b may be the same as or different from the resolution of the output data of the light receiving unit 13b, and may be lower, for example, in the Y direction (main scanning direction) and the X direction (sub-scanning direction).

The control unit 20b may control each part of the paper sheet recognition device 1b, and may be configured, for example, with programs for implementing various processes, including a paper sheet recognition program, a CPU that executes the programs, and various hardware (e.g., FPGA) controlled by the CPU.

The control unit 20b performs a process (hereinafter, sometimes referred to as a multiplication process) to calculate multiplication data, which is data obtained by multiplying data relating to light of the first wavelength and data relating to light of the second wavelength (output data) among the output data relating to light of the multiple wavelengths acquired, and performs a process (hereinafter, sometimes referred to as an identification process) to identify the banknote BN based on the output data relating to light of the multiple wavelengths and the multiplication data. This makes it possible to improve the performance of distinguishing between genuine and counterfeit objects S to be identified (especially special ink). Therefore, it is possible to improve the identification accuracy of the banknote BN.

In general inks, such as infrared absorbing inks and infrared non-absorbing inks, there is almost no difference in reflectance in the infrared region, and in the space of output data related to infrared light, each output is distributed in the direction of vector (1,1,1) ^T passing through the origin. On the other hand, this is not the case for special inks. Therefore, the product of output data related to light of multiple wavelengths is more susceptible to influence in general inks than in special inks, which contributes to improving the degree of separation between special inks and general inks.

The output data relating to the light of the multiple wavelengths acquired by the control unit 20b from the light receiving unit 13b may be data corresponding to the region ROI _A including the recognition target S. In other words, the output data relating to the light of the multiple wavelengths may be a portion of the reflected image data of the entire banknote BN, i.e., reflected image data of a portion of the banknote BN. The region ROI _A may be set in advance according to the denomination of the banknote BN, and the control unit 20b may extract data corresponding to the region ROI _A including the recognition target S from the output data of the light receiving unit 13b (e.g., the reflected image data of the entire banknote BN) based on position information of the region ROI _A set for each denomination.

When two wavelengths of light are used, in the multiplication process, the control unit 20b may calculate multiplication data by multiplying the output data related to the light of the first wavelength (hereinafter, output data of λ1) by the output data related to the light of the second wavelength (hereinafter, output data of λ2) (hereinafter, output data of λ1×λ2), and in the recognition process, the control unit 20b may recognize the banknote BN based on the output data of λ1, the output data of λ2, and the output data of λ1×λ2 (three-dimensional features).

When light of three wavelengths is used, the control unit 20b may further calculate, as multiplication data, data obtained by multiplying the output data of λ1 by the output data relating to the light of the third wavelength (hereinafter, output data of λ3) in the multiplication process (hereinafter, output data of λ1×λ3) and data obtained by multiplying the output data of λ2 by the output data of λ3 (hereinafter, output data of λ2×λ3). In the identification process, the control unit 20b may identify the banknote BN based on the output data of λ1, the output data of λ2, the output data of λ3, the output data of λ1×λ2, the output data of λ1×λ3, and the output data of λ2×λ3 (six-dimensional features).

When light of two wavelengths is used, the multiplication data may be data obtained by multiplying each output value (pixel value) of the output data related to light of the first wavelength by each output value (pixel value) of the output data related to light of the second wavelength (however, output values related to the same point in the XY plane are multiplied together). When light of three wavelengths is used, output values (pixel values) may be multiplied together in a similar manner (however, output values related to the same point in the XY plane are multiplied together).

In this embodiment, the light of each wavelength irradiated by the light source 11b may be light of a wavelength band including a peak wavelength and a wavelength in the vicinity thereof, and the light of the multiple wavelengths irradiated by the light source 11b may have different peak wavelengths. The light source 11b may also include multiple light-emitting elements 17b having different peak wavelengths. For example, the light source 11b may irradiate infrared light of multiple wavelengths having different peak wavelengths, or may include multiple light-emitting elements 17b that each irradiate infrared light of multiple wavelengths having different peak wavelengths. The light-receiving unit 13b may also receive infrared light of multiple wavelengths irradiated from the light source 11b and reflected by the banknote BN. Similarly, the output data related to the light of multiple wavelengths acquired by the control unit 20b from the light-receiving unit 13b may each be data related to infrared light of multiple wavelengths. Furthermore, various data related to light of multiple wavelengths processed by the control unit 20b may be data related to infrared light of multiple wavelengths.

For example, the light source 11b may irradiate infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light) having different peak wavelengths, the light receiving unit 13b may receive the infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light) irradiated from the light source 11b and reflected by the banknote BN, and the control unit 20b may acquire data related to the infrared light of multiple wavelengths (first and second infrared light, or first to third infrared light) from the light receiving unit 13b. The various data related to the light of multiple wavelengths processed by the control unit 20b may be data related to the infrared light of multiple wavelengths (for example, the first and second infrared light, or the first to third infrared light).

The first and second infrared lights may be a combination of any two of the following: infrared light having a peak wavelength in the wavelength band of 770 to 830 nm, infrared light having a peak wavelength in the wavelength band of 840 to 900 nm, and infrared light having a peak wavelength in the wavelength band of 910 to 970 nm. Note that either of the wavelengths of the first and second infrared lights may be larger.

Similarly, the first to third infrared lights may be a combination of infrared light having a peak wavelength in the 770 to 830 nm wavelength band, infrared light having a peak wavelength in the 840 to 900 nm wavelength band, and infrared light having a peak wavelength in the 910 to 970 nm wavelength band. Note that there is no particular limitation on the magnitude relationship of the wavelengths of the first to third infrared lights.

In the recognition process, the control unit 20b may apply a discrimination function to each pixel and determine for each pixel whether or not the pixel has special ink. That is, the above-mentioned various data used in the recognition process may be input to the discrimination function as features (e.g., three-dimensional or six-dimensional features), and based on the output of the discrimination function, it may be determined whether or not an object to be recognized S is provided on the banknote BN. This allows the control unit 20b to speed up the recognition process.

Here, the discriminant function can be generated by supervised machine learning. Specifically, for example, discriminant analysis, support vector machines, neural networks, etc. can be used. For example, image data including the ink portion of a banknote on which special ink is printed and its background portion (peripheral portion) can be used as the teacher data. Here, the feature quantities (for example, three-dimensional or six-dimensional feature quantities) related to the above-mentioned various data are input, and the special ink portion and the background portion, i.e., the non-special ink portion, can be used as class labels. For example, in the case of negative ink, image data related to infrared light having a peak wavelength in the wavelength band of 910 to 970 nm can be binarized (separated into the ink portion and the background portion), and the result can be used as the class label for each pixel as it is. In the case of positive ink, image data related to infrared light having a peak wavelength in the wavelength band of 770 to 830 nm can be binarized (separated into the ink portion and the background portion), and the result can be used as the class label for each pixel as it is. In either case, the binarization process may be Otsu binarization.

More specifically, for example, when light of three wavelengths is used, the control unit 20b may determine in the recognition process whether or not an object to be recognized S is provided on the banknote BN based on the following formulas (21) to (26). Here, formula (23') may be used instead of formula (23).

In equations (21) and (22), L is a threshold value. In equation (21), if λ is equal to or greater than L, it is determined that an object to be recognized S is provided on the banknote BN, and if λ is less than L, it is determined that an object to be recognized S is not provided on the banknote BN.

In formula (22), λ is an evaluation value, and the sum of _Pj within the region ROI _A including the recognition target S is calculated.

In each equation, _Pj is the result of the discrimination function f(x), and can be the sum of pixels determined to contain special ink (equation (23)), or the sum of classification scores that are the output of the discrimination function f(x) (equation (23')). In equations (23) and (23'), c is the threshold value of the classification score.

The discrimination function f(x) is expressed by equation (24), and six-dimensional features are input as shown in equation (26). In equation (25), ω(w ₀ to w ₆ ) is a coefficient vector obtained as a result of machine learning, and φ(x _j ) expressed by equation (26) is a feature vector of the target pixel j (j is a number indicating the target pixel). In equation (26), x _1j , x _2j and x _3j are pixel values related to the first, second and third light (which may be the first, second and third infrared light) at the target pixel _j , respectively. Here, the discrimination function f(x) is expressed by an equation obtained by extending a linear discrimination equation to a nonlinear model.

In addition, in equation (26), _x1j , _x2j , _x3j , _x1jx2j , _{x1jx3j ,} and _x2jx3j correspond to the output data of _λ1 , the output data of λ2, the output data of λ3, the output data of λ1×λ2, _the output data of λ1×λ3, and the output data of λ2×λ3, respectively.

The control unit 20b may identify the authenticity of the banknote BN in the identification process. For example, if it is determined that the identification object S is provided on the banknote BN, it may determine that the banknote BN is a genuine note, and if it is determined that the identification object S is not provided on the banknote BN, it may determine that the banknote BN is a counterfeit note.

Next, the operation of the paper sheet recognition device 1b according to this embodiment will be described with reference to FIG. 12. FIG. 12 is a flowchart illustrating an example of the operation of the paper sheet recognition device according to this embodiment.

As shown in FIG. 12, first, the control unit 20b acquires output data from the light receiving unit 13b, which receives light of multiple wavelengths irradiated from the light source 11b and arriving (e.g., reflected) from the banknote BN (step S21).

Next, the control unit 20b performs a multiplication to calculate multiplication data, which is data obtained by multiplying together data relating to the first wavelength and data relating to the second wavelength of light among the acquired output data relating to the light of multiple wavelengths (step S22).

If the data relating to the light of the first wavelength and the second wavelength corresponds to the region ROI _A including the object to be identified S, before step S22, a process may be performed for the output data relating to the light of each wavelength to extract data corresponding to the region ROI _A including the object to be identified S (e.g., reflection image data of the region ROI A including the object to be identified S) from data corresponding to the entire banknote BN (e.g., reflection image data of the entire banknote _BN ).

Thereafter, the control unit 20b performs a recognition process to identify the banknote BN based on the output data relating to the multiple wavelengths of light acquired from the light receiving unit 13b (which may be data corresponding to the area ROI _A including the recognition target S) and the calculated multiplication data (step S23), and the operation of the paper sheet recognition device 1b is terminated.

(Embodiment 3)
The configuration of a paper sheet processing apparatus in which the paper sheet recognition device according to the present embodiment can be installed will be described with reference to Fig. 13. Fig. 13 is a schematic perspective view showing the exterior of an example of a paper sheet processing apparatus in which the paper sheet recognition device according to the present embodiment can be installed. The paper sheet processing apparatus in which the paper sheet recognition device according to the present embodiment is installed may have the configuration shown in Fig. 13, for example. The paper sheet processing apparatus 300 shown in Fig. 13 includes a banknote recognition device according to the present embodiment (not shown in Fig. 13) that performs a banknote recognition process, a hopper 301 on which multiple banknotes to be processed are placed in a stacked state, two reject units 302 from which rejected banknotes are discharged, an operation unit 303 for inputting instructions from an operator, four stacking units 306a to 306d for sorting and stacking banknotes whose denomination, authenticity, and fitness have been identified in a housing 310, and a display unit 305 for displaying information such as the results of recognition and counting of banknotes and the stacking status of each stacking unit 306a to 306d.

Next, the configuration of the paper sheet recognition device according to this embodiment will be described with reference to FIG. 14. FIG. 14 is a block diagram illustrating an example of the configuration of the paper sheet recognition device according to this embodiment. As shown in FIG. 14, the paper sheet recognition device 100 according to this embodiment includes an optical line sensor 110, a control unit 120, a memory unit 130, and a conveying unit 140.

The optical line sensor 110 detects various optical characteristics of the transported banknotes, and may include a light source 111 and a light receiving unit 113 along the transport path of the banknotes. The light source 111 irradiates the banknotes with light of multiple wavelengths, and the light receiving unit 113 receives the light of multiple wavelengths irradiated from the light source 111 and reflected by the banknotes, and outputs data related to the light of the multiple wavelengths for each wavelength.

The control unit 120 is composed of programs (including a paper sheet recognition program) for implementing various processes stored in the storage unit 130, a CPU that executes the programs, and various hardware (e.g., FPGA) controlled by the CPU. The control unit 120 controls each unit of the paper sheet recognition device 100 according to the programs stored in the storage unit 130. The control unit 120 also has the function of performing processes such as acquisition of output data from the light receiving unit 113, correction of the acquired output data, multiplication of the acquired output data, and recognition using various corrected and/or multiplied data, according to the programs stored in the storage unit 130. These processes performed by the control unit 120 are similar to the processes performed by the control unit 10a or 10b described in the first or second embodiment, and therefore will not be described in detail.

The control unit 120 performs a process of identifying at least the denomination and authenticity of the banknote as the identification process. According to the identification process using various corrected and/or multiplied data described in the first or second embodiment, it is possible to identify the authenticity of the banknote. The control unit 120 may have a function of determining whether the banknote is fit or unfit. In that case, the control unit 120 has a function of detecting dirt, folds, tears, etc. on the banknote to determine whether the banknote should be treated as a fit note that can be reused in the market or an unfit note that is not suitable for market circulation.

The memory unit 130 is composed of a non-volatile and/or volatile storage device such as a semiconductor memory or a hard disk, and stores various programs and various data for controlling the paper sheet recognition device 100.

The conveying unit 140 rotates and drives multiple rollers, belts, etc. to convey banknotes one by one along a conveying path provided within the paper sheet recognition device 100.

Next, the configuration of the optical line sensor 110 will be described with reference to FIG. 15. FIG. 15 is a schematic cross-sectional view illustrating an example of the configuration of an optical line sensor provided in a paper sheet recognition device according to this embodiment. As shown in FIG. 15, the optical line sensor 110 is composed of a contact image sensor facing the transport path 311 of the paper sheet processing device, and forms part of the transport path 311. Banknotes BN are transported in the X direction within the transport path 311 (XY plane). The Y direction corresponds to the main scanning direction of the optical line sensor 110, and the X direction corresponds to the sub-scanning direction of the optical line sensor 110.

15, the optical line sensor 110 includes two reflection light sources 111, a condensing lens 112, a light receiving unit 113, and a substrate 114. The reflection light source 111 includes, for example, a light guide extending in the main scanning direction and multiple types of light emitting elements that face at least one end face of the light guide and irradiate multiple wavelengths of light, and sequentially irradiates multiple wavelengths of light on the main surface (hereinafter, surface A) of the banknote BN on the light receiving unit 113 side. The condensing lens 112 is, for example, composed of a rod lens array in which multiple rod lenses are arranged in the main scanning direction, and condenses light emitted from the reflection light source 111 and reflected by surface A of the banknote BN. The light receiving unit 113 includes, for example, a linear image sensor in which multiple light receiving elements (light receiving pixels) are arranged in the main scanning direction, and each light receiving element has sensitivity to the wavelength band of multiple wavelengths of light irradiated by the light source 111. Each light receiving element may be, for example, a silicon (Si) photodiode that is sensitive from at least the visible region to the infrared region of wavelength 1100 nm. Each light receiving element is mounted on a substrate 114, receives light focused by a focusing lens 112, converts it into an electrical signal according to the amount of incident light, and outputs it to the substrate 114. Each light receiving element receives light of the wavelength in accordance with the timing of irradiation of the light of each wavelength by the light source 111. The substrate 114 includes, for example, a drive circuit for driving the light receiving element and a signal processing circuit for processing and outputting the signal from the light receiving element. The substrate 114 amplifies the output signal of the light receiving unit 113 (each light receiving element), A/D converts it to digital data, and outputs it.

Light source 111 emits at least infrared light of multiple wavelengths as light of multiple wavelengths, for example, first to third infrared light having different peak wavelengths. Light source 111 may also emit visible light. Examples of visible light that can be used include red light (R), green light (G), blue light (B), and white light (W) that includes these three colors of light.

In this embodiment, the control unit 120 performs the same processing as the control unit 10a or 10b described in the first or second embodiment, so it is possible to improve the accuracy of identifying banknotes, as in the first or second embodiment.

In the above embodiment, the output data relating to the light irradiated from the light source and reflected by the banknote is used for correction processing and multiplication processing by the control unit. However, the output data relating to the light irradiated from the light source and transmitted through the banknote may also be used for correction processing and multiplication processing by the control unit.

Although the embodiments have been described above with reference to the drawings, the present disclosure is not limited to the above-described embodiments. Furthermore, the configurations of the respective embodiments may be appropriately combined or modified without departing from the spirit and scope of the present disclosure.

As described above, the present disclosure is a useful technology for improving the accuracy of identifying paper sheets.

1a, 1b, 100: Paper sheet recognition device 11a, 11b, 111: Light source 13a, 13b, 113: Light receiving unit 14a, 14b, 110: Optical line sensor 15a, 15b: Light guide 15aa, 15ba: End surface 17a, 17b of light guide: Light emitting element 20a, 20b, 120: Control unit 112: Condenser lens 114: Substrate 130: Memory unit 140: Transport unit 300: Banknote processing device 301: Hopper 302: Reject unit 303: Operation unit 305: Display unit 306a to 306d: Stacking unit 311: Transport path BN: Banknote S: Recognition target ROI _A : Region including recognition target ROI _B : Region not including recognition target Pix: Pixel

Claims (10)

A paper sheet recognition device that recognizes a paper sheet having a recognition object provided thereon,
A light source that irradiates light onto the paper sheet;
A light receiving unit that receives light coming from the paper sheet;
A control unit that acquires output data of the light receiving unit,
The control unit is
Obtaining first data, which is output data of the light receiving unit corresponding to an area of the paper sheet that includes an identification object, and second data, which is output data of the light receiving unit corresponding to an area of the paper sheet that does not include the identification object;
generating third data by correcting the first data using the second data;
The third data is input as a feature to a discrimination function;
A paper sheet recognition device that recognizes paper sheets based on an output of the discrimination function . The light source irradiates light of multiple wavelengths onto the paper sheet,
The light receiving unit receives light of a plurality of wavelengths coming from a paper sheet,
the first data and the second data each include data related to the light of the multiple wavelengths,
the control unit generates third data relating to the light of the plurality of wavelengths by correcting each of the first data relating to the light of the plurality of wavelengths with a corresponding second data relating to the light of the same wavelength;
2. The paper sheet recognition device according to claim 1, wherein the paper sheet is recognized based on the third data related to the light of the plurality of wavelengths. the control unit calculates first multiplication data, which is data obtained by multiplying together data related to the light of a first wavelength and data related to the light of a second wavelength among the first data related to the light of the multiple wavelengths, and second multiplication data, which is data obtained by multiplying together data related to the light of the first wavelength and data related to the light of the second wavelength among the second data related to the light of the multiple wavelengths,
generating fourth data by correcting the first multiplied data with the second multiplied data;
3. The paper sheet recognition device according to claim 2, wherein the paper sheet is recognized based on the third data related to the light of the plurality of wavelengths and the fourth data. The paper sheet recognition device according to any one of claims 1 to 3, characterized in that the control unit uses, as the second data, a representative value of output data of an area of the paper sheet that does not include the recognition object. The light source is
A rod-shaped light guide made of acrylic resin;
a light emitting element facing at least one of the two end faces of the light guide;
Equipped with
The light source is
5. The paper sheet identifying device according to claim 1, wherein light is irradiated onto the paper sheet through the light guide. The light source irradiates a linear light beam in a main scanning direction onto the paper sheet,
the light receiving unit receives linear light in a main scanning direction coming from a paper sheet,
The paper sheet recognition device according to any one of claims 1 to 5, characterized in that the control unit uses, as the second data, output data of the light receiving unit corresponding to an area of the paper sheet that does not include the recognition object, which is set in the same range as the position in the main scanning direction of the area including the recognition object of the paper sheet. The light source irradiates infrared light onto the paper sheet,
7. The paper sheet identifying device according to claim 1, wherein the infrared light includes light having a wavelength of 850 nm or more and 950 nm or less. The light receiving unit receives light emitted from the light source and reflected by the paper sheet,
8. The paper sheet recognition device according to claim 1, wherein the output data includes data relating to light reflected by the paper sheet. A paper sheet recognition device according to any one of claims 1 to 8, characterized in that the paper sheet to be recognized includes, as an object to be recognized, at least one of an ink whose reflectance decreases as the wavelength becomes longer in the infrared region and an ink whose reflectance increases as the wavelength becomes longer in the infrared region. A paper sheet recognition method for identifying a paper sheet having a recognition object provided thereon, comprising:
acquiring output data from a light receiving unit that receives light irradiated from a light source and transmitted from the paper sheet;
acquiring first data, which is output data of the light receiving unit corresponding to an area of the paper sheet that includes an identification object, and second data, which is output data of the light receiving unit corresponding to an area of the paper sheet that does not include the identification object;
generating third data by correcting the first data using the second data;
inputting the third data into a discriminant function as a feature;
A step of classifying paper sheets based on an output of the discrimination function ;
A paper sheet identification method comprising:

Images