JP2008243098A

JP2008243098A - Paper sheet discriminating device and paper sheet discriminating method

Info

Publication number: JP2008243098A
Application number: JP2007086138A
Authority: JP
Inventors: Hiroshi Fukuda; 浩福田
Original assignee: Toshiba Corp
Current assignee: Toshiba Corp
Priority date: 2007-03-29
Filing date: 2007-03-29
Publication date: 2008-10-09
Anticipated expiration: 2027-03-29
Also published as: JP5044255B2

Abstract

<P>PROBLEM TO BE SOLVED: To provide a paper sheet discriminating device capable of always performing discrimination of a stable type or the like by effectively using color information even in the case of having to cope with paper sheets of various colors. <P>SOLUTION: A color image on a paper sheet is acquired by a color camera, R, G and B patterns constituting the color image are extracted as input patterns from the acquired color image, evaluation values representing a matching rate between both patterns are respectively calculated by collating the extracted R, G and B input patterns with R, G and B dictionary patterns prepared on the basis of a color image previously acquired from a plurality of types of paper sheets of a discrimination target in each corresponding color, each of the calculated evaluation values is subjected to prescribed weighting and added in each color, and the type of the paper sheet is discriminated on the basis of the addition result. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

本発明は、たとえば、紙幣やそれ以外の有価証券などの紙葉類を区分処理する紙葉類処理装置において、紙葉類の種類や真偽等を判別する紙葉類判別装置および紙葉類判別方法に関する。 The present invention relates to a paper sheet discriminating apparatus and a paper sheet for discriminating the type and authenticity of paper sheets in a paper sheet processing apparatus that sorts paper sheets such as banknotes and other securities. It relates to a determination method.

一般に、たとえば、紙幣やそれ以外の有価証券などの紙葉類を区分処理する紙葉類処理装置において、紙葉類の種類等を判別する紙葉類判別装置にあっては、あらかじめ判別対象とする複数種の紙葉類から取得した辞書パターンを辞書パターン記憶部に記憶しておき、入力された紙葉類の画像から取得した入力パターンと辞書パターン記憶部内の辞書パターンとを照合することにより両パターンの一致度を表す評価値を計算し、求めた評価値に基づき当該紙葉類の種類等を判別する方法が用いられている（たとえば、特許文献１、特許文献２参照）。
特開２００３−２８８６２７号公報特開２００５−２９３２８５号公報 In general, for example, in a paper sheet processing apparatus that sorts paper sheets such as banknotes and other securities, a paper sheet discrimination apparatus that discriminates the type of the paper sheet, By storing a dictionary pattern acquired from a plurality of types of paper sheets in the dictionary pattern storage unit, and collating the input pattern acquired from the input paper sheet image with the dictionary pattern in the dictionary pattern storage unit A method is used in which an evaluation value representing the degree of coincidence of both patterns is calculated, and the type of the paper sheet is discriminated based on the obtained evaluation value (see, for example, Patent Document 1 and Patent Document 2).
JP 2003-288627 A JP 2005-293285 A

しかし、従来の種類（以降、券種ともいう）判別は、全ての紙葉類に対する種類判別を同一色のデータあるいは濃淡データで実施している。この方法では、対応すべき紙葉類の種類が少なく、色味が似ている場合は問題ないが、対応種類が増え、様々な色の紙葉類に対応しなければいけない場合には、どうしてもパターンの出にくい種類ができてしまい、特定種類で種類判別の性能が悪くなることがあるという問題がある。 However, in the conventional type discrimination (hereinafter also referred to as ticket type), type discrimination for all paper sheets is performed using data of the same color or grayscale data. In this method, there is no problem when the types of paper sheets to be handled are small and the colors are similar, but there is no problem if the number of supported types increases and it is necessary to handle paper sheets of various colors. There is a problem that types that are difficult to produce patterns are produced, and that the type discrimination performance may be deteriorated for specific types.

そこで、本発明は、様々な色の紙葉類に対応しなければいけない場合でも、色情報を有効に用いることにより常に安定した種類等の判別が行なえる紙葉類判別装置および紙葉類判別方法を提供することを目的とする。 Therefore, the present invention provides a paper sheet discriminating apparatus and a paper sheet discriminating apparatus that can always perform stable type discrimination by effectively using color information even when paper sheets of various colors must be supported. It aims to provide a method.

本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色ごとの辞書パターンをあらかじめ記憶している辞書パターン記憶手段と、前記パターン抽出手段により抽出された複数の色ごとの入力パターンと前記辞書パターン記憶手段に記憶された複数の色ごとの辞書パターンとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段と、この評価値計算手段により求められた各評価値を色ごとに所定の重みを付けて加算し、その加算結果に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 The paper sheet discriminating apparatus according to the present invention includes a transport unit that transports a paper sheet to be discriminated, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color image acquisition unit. Pattern extracting means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image obtained by the means, and the color image obtained from the color image respectively corresponding to the type of paper sheet Dictionary pattern storage means for storing in advance a dictionary pattern for each of a plurality of colors, an input pattern for each of a plurality of colors extracted by the pattern extraction means, and a plurality of colors stored in the dictionary pattern storage means An evaluation value calculator that calculates an evaluation value representing the degree of coincidence between the two patterns by matching each dictionary pattern with the corresponding color And a discriminating unit that adds each evaluation value obtained by the evaluation value calculating unit with a predetermined weight for each color, and discriminates the type of the sheet based on the addition result. .

また、本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、このパターン抽出手段により抽出された複数の色ごとの入力パターンをそれぞれ別の表色系のパターンに変換するパターン変換手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色ごとのパターンがそれぞれ別の表色系に変換されてなる辞書パターンをあらかじめ記憶している辞書パターン記憶手段と、前記パターン変換手段により変換された表色系の各入力パターンと前記辞書パターン記憶手段に記憶された表色系の各辞書パターンとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段と、この評価値計算手段により求められた各評価値を色ごとに所定の重みを付けて加算し、その加算結果に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 Further, the paper sheet discriminating apparatus of the present invention includes a transport unit that transports a paper sheet to be discriminated, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color A pattern extraction unit that extracts a pattern for each of a plurality of colors constituting the color image from the color image acquired by the image acquisition unit as an input pattern, and an input pattern for each of the plurality of colors that is extracted by the pattern extraction unit. Pattern conversion means for converting to a different color system pattern, and a pattern for each of a plurality of colors constituting the color image acquired from a color image corresponding to each type of paper sheet is in a different color system Dictionary pattern storage means for storing the converted dictionary pattern in advance, and each input system of the color system converted by the pattern conversion means. Evaluation value calculating means for calculating an evaluation value representing the degree of coincidence between the two patterns by matching each color of the color pattern system dictionary pattern stored in the dictionary pattern storage means for each corresponding color, and Each evaluation value obtained by the evaluation value calculation means is added with a predetermined weight for each color, and a discrimination means for discriminating the type or the like of the paper sheet based on the addition result is provided.

また、本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色のうち特定の１つの色を指定する指定色情報と当該色の辞書パターンをあらかじめ記憶している辞書パターン記憶手段と、前記パターン抽出手段により抽出された複数の色ごとの入力パターンのうち前記辞書パターン記憶手段に記憶された指定色情報で指定される色の入力パターンを選択し、この選択した入力パターンと前記辞書パターン記憶手段に記憶された辞書パターンとを照合することにより両パターンの一致度を表す評価値を計算する評価値計算手段と、この評価値計算手段により求められた評価値に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 Further, the paper sheet discriminating apparatus of the present invention includes a transport unit that transports a paper sheet to be discriminated, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color A pattern extracting means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the image acquiring means, and the color acquired from the color image corresponding to each type of paper sheet Designated color information for designating a specific color among a plurality of colors constituting the image, dictionary pattern storage means for storing a dictionary pattern of the color in advance, and a plurality of colors extracted by the pattern extraction means The input pattern of the color designated by the designated color information stored in the dictionary pattern storage means is selected from among the input patterns of the selected input pattern, and the selected input pattern And an evaluation value calculation means for calculating an evaluation value representing the degree of coincidence of both patterns by comparing the dictionary pattern stored in the dictionary pattern storage means, and based on the evaluation value obtained by the evaluation value calculation means And discriminating means for discriminating the type of the paper sheet.

また、本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色ごとの辞書パターンおよび各色ごとの重み情報をあらかじめ記憶している辞書パターン記憶手段と、前記パターン抽出手段により抽出された複数の色ごとの入力パターンと前記辞書パターン記憶手段に記憶された複数の色ごとの辞書パターンとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段と、この評価値計算手段により求められた各評価値を前記辞書パターン記憶手段に記憶されている各色ごとの重み情報による重みを対応する色ごとに付けて加算し、その加算結果に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 Further, the paper sheet discriminating apparatus of the present invention includes a transport unit that transports a paper sheet to be discriminated, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color A pattern extracting means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the image acquiring means, and the color acquired from the color image corresponding to each type of paper sheet Dictionary pattern storage means for storing in advance a dictionary pattern for each of a plurality of colors constituting the image and weight information for each color, an input pattern for each of a plurality of colors extracted by the pattern extraction means, and the dictionary pattern storage means An evaluation value that represents the degree of coincidence of both patterns by matching the dictionary pattern for each of the colors stored in Each of the evaluation value calculation means for calculating, and each evaluation value obtained by the evaluation value calculation means is added with weights corresponding to the weight information for each color stored in the dictionary pattern storage means for each corresponding color, Discriminating means for discriminating the type of the paper sheet based on the addition result.

また、本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、このパターン抽出手段により抽出された複数の色ごとの入力パターンをそれぞれ別の表色系のパターンに変換するパターン変換手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色のうち特定の１つの色を指定する指定色情報と当該色のパターンが別の表色系に変換されてなる辞書パターンをあらかじめ記憶している辞書パターン記憶手段と、前記パターン変換手段により変換された表色系の各入力パターンのうち前記辞書パターン記憶手段に記憶された指定色情報で指定される色の入力パターンを選択し、この選択した入力パターンと前記辞書パターン記憶手段に記憶された辞書パターンとを照合することにより両パターンの一致度を表す評価値を計算する評価値計算手段と、この評価値計算手段により求められた評価値に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 Further, the paper sheet discriminating apparatus of the present invention includes a transport unit that transports a paper sheet to be discriminated, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color A pattern extraction unit that extracts a pattern for each of a plurality of colors constituting the color image from the color image acquired by the image acquisition unit as an input pattern, and an input pattern for each of the plurality of colors that is extracted by the pattern extraction unit. Pattern conversion means for converting to another color system pattern, and designation for designating one specific color among a plurality of colors constituting the color image acquired from the color image corresponding to each type of paper sheet Dictionary pattern storage means for storing in advance a dictionary pattern formed by converting color information and a pattern of the color into another color system, and the pattern conversion The input pattern of the color designated by the designated color information stored in the dictionary pattern storage means is selected from the color system input patterns converted by the stage, and the selected input pattern and the dictionary pattern storage means are selected. Evaluation value calculation means for calculating an evaluation value representing the degree of coincidence of both patterns by collating with the stored dictionary pattern, and the type of the paper sheet based on the evaluation value obtained by this evaluation value calculation means Discriminating means for discriminating.

さらに、本発明の紙葉類判別装置は、判別対象の紙葉類を搬送する搬送手段と、この搬送手段により搬送される紙葉類上のカラー画像を取得するカラー画像取得手段と、このカラー画像取得手段により取得されたカラー画像から当該カラー画像を構成する複数の色ごとのパターンを入力パターンとして抽出するパターン抽出手段と、このパターン抽出手段により抽出された複数の色ごとの入力パターンをそれぞれ別の表色系のパターンに変換するパターン変換手段と、紙葉類の種類にそれぞれ対応するカラー画像から取得された当該カラー画像を構成する複数の色ごとのパターンがそれぞれ別の表色系に変換されてなる辞書パターンおよび各色ごとの重み情報をあらかじめ記憶している辞書パターン記憶手段と、前記パターン変換手段により変換された表色系の各入力パターンと前記辞書パターン記憶手段に記憶された表色系の各辞書パターンとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段と、この評価値計算手段により求められた各評価値を前記辞書パターン記憶手段に記憶されている各色ごとの重み情報による重みを対応する色ごとに付けて加算し、その加算結果に基づき当該紙葉類の種類等を判別する判別手段とを具備している。 Further, the paper sheet discriminating apparatus of the present invention includes a transport unit that transports a paper sheet to be determined, a color image acquisition unit that acquires a color image on the paper sheet transported by the transport unit, and the color A pattern extraction unit that extracts a pattern for each of a plurality of colors constituting the color image from the color image acquired by the image acquisition unit as an input pattern, and an input pattern for each of the plurality of colors that is extracted by the pattern extraction unit. Pattern conversion means for converting to a different color system pattern, and a pattern for each of a plurality of colors constituting the color image acquired from a color image corresponding to each type of paper sheet is in a different color system Dictionary pattern storage means for storing the converted dictionary pattern and weight information for each color in advance, and the pattern conversion means Each of the converted color system input patterns and each color system dictionary pattern stored in the dictionary pattern storage means are collated for each corresponding color to calculate an evaluation value representing the degree of coincidence of both patterns. An evaluation value calculating means for adding the weights obtained by weight information for each color stored in the dictionary pattern storage means for each corresponding color and adding each evaluation value obtained by the evaluation value calculating means. Discriminating means for discriminating the type of the paper sheet based on the result.

本発明によれば、様々な色の紙葉類に対応しなければいけない場合でも、色情報を有効に用いることにより常に安定した種類等の判別が行なえる紙葉類判別装置および紙葉類判別方法を提供できる。 According to the present invention, a paper sheet discriminating apparatus and a paper sheet discriminating apparatus that can always perform stable classification etc. by effectively using color information even when paper sheets of various colors have to be handled. Can provide a method.

以下、本発明の実施の形態について図面を参照して説明する。
まず、第１の実施の形態について説明する。
図１は、本発明の実施の形態に係る紙葉類判別装置の構成を概略的に示すものである。この紙葉類判別装置は、判別対象としての紙幣などの紙葉類Ｐを搬送する搬送手段としての搬送路１１、搬送路１１で搬送される紙葉類Ｐの表面を照明する照明装置１２、照明装置１２で照明された紙葉類Ｐの表面のカラー画像を取得するカラー画像取得手段としてのカラービデオカメラ（以降、単にカラーカメラと称す）１３、カラーカメラ１３で得られたカラー画像を処理することにより、たとえば紙葉類Ｐの種類を判別する処理部１４、および、処理部１４の判別結果に基づき搬送路１１および図示しない紙葉類区分手段などの制御を行なう搬送制御部１５を有して構成されている。 Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, the first embodiment will be described.
FIG. 1 schematically shows the configuration of a paper sheet discriminating apparatus according to an embodiment of the present invention. The paper sheet discriminating apparatus includes a transport path 11 as a transport unit that transports a paper sheet P such as a bill as a discrimination target, an illumination device 12 that illuminates the surface of the paper sheet P transported in the transport path 11, A color video camera (hereinafter simply referred to as a color camera) 13 as a color image acquisition means for acquiring a color image of the surface of the paper sheet P illuminated by the illuminating device 12, and a color image obtained by the color camera 13 are processed. Thus, for example, a processing unit 14 that determines the type of the paper sheet P, and a transport control unit 15 that controls the transport path 11 and a paper sheet sorting unit (not shown) based on the determination result of the processing unit 14 are provided. Configured.

処理部１４は、たとえば、図２に示すように、カラーカメラ１３で撮像されたカラー画像から当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ（＝Ｓ）として抽出するパターン抽出手段としての画像処理部２１、画像処理部２１から得られる入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂを一時記憶する入力パターン記憶部２２、あらかじめ判別対象とする複数種の紙葉類から取得したカラー画像に基づき作成された当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の辞書パターンＤ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢ（＝Ｄ_ｉ）を記憶している辞書パターン記憶手段としての辞書パターン記憶部２３、入力パターン記憶部２２内の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂと辞書パターン記憶部２３内の辞書パターンＤ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段、および、求められた各評価値を色ごとに所定の重みを付けて加算し、その加算結果に基づき当該紙葉類Ｐの種類を判別する判別手段として機能する判別処理部２４を有して構成されている。 For example, as illustrated in FIG. 2, the processing unit 14 includes a plurality of colors constituting the color image from the color image captured by the color camera 13, in this example, R (red), G (green), and B (blue). ) As input patterns S _R , S _G , S _B (= S), and temporarily stores the input patterns S _R , S _G , S _B obtained from the image processing unit 21. The input pattern storage unit 22, a plurality of colors constituting the color image created based on color images acquired in advance from a plurality of types of paper sheets to be discriminated, in this example R (red), G (green), Dictionary pattern storage unit 23 as dictionary pattern storage means for storing B (blue) dictionary patterns D _iR , D _iG , D _iB (= D _i ), and input patterns S _R , S in input pattern storage unit 22 _G S _B and dictionary pattern dictionary pattern D _iR in the storage unit _{23, D} _iG, evaluation value calculating means for calculating an evaluation value representing the degree of coincidence of both patterns respectively by matching each color corresponding to the D _iB, and Each of the obtained evaluation values is added with a predetermined weight for each color, and is configured to include a determination processing unit 24 that functions as a determination unit that determines the type of the paper sheet P based on the addition result. ing.

判別処理部２４は、カラーカメラ１３から前述したようにＲ，Ｇ，Ｂの３色のデータが得られる場合、入力パターンＳ＝（Ｓ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ）：Ｓ_Ｒ＝（Ｓ_Ｒ１，Ｓ_Ｒ２，…，Ｓ_ＲＭ）、Ｓ_Ｇ＝（Ｓ_Ｇ１，Ｓ_Ｇ２，…，Ｓ_ＧＭ）、Ｓ_Ｂ＝（Ｓ_Ｂ１，Ｓ_Ｂ２，…，Ｓ_ＢＭ）を入力パターン記憶部２２から読出すとともに、各券種の辞書パターンＤ_ｉ＝（Ｄ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢ）：Ｄ_ｉＲ＝（Ｄ_ｉＲ１，Ｄ_ｉＲ２，…，Ｄ_ｉＲＭ）、Ｄ_ｉＧ＝（Ｄ_ｉＧ１，Ｄ_ｉＧ２，…，Ｄ_ｉＧＭ）、Ｄ_ｉＢ＝（Ｄ_ｉＢ１，Ｄ_ｉＢ２，…，Ｄ_ｉＢＭ）を辞書パターン記憶部２３から読出すことにより、券種判別の処理を行なう。
ここに、図３は辞書パターン記憶部２３の記憶内容の一例を示している。 As described above, when the color camera 13 obtains data of three colors R, G, and B from the color camera 13, the discrimination processing unit 24 has an input pattern S = (S _R , S _G , S _B ): S _R = (S _R1 , S _R2 ,..., S _RM ), S _G = (S _G1 , S _G2 ,..., S _GM ), S _B = (S _B1 , S _B2 ,..., S _BM ) are read from the input pattern storage unit 22. with, each denomination of the dictionary pattern _{_{_{_{D i = (D iR, D}}}} iG, D iB): D iR = (D iR1, D iR2, ..., D iRM), D iG = (D iG1, D iG2, ..., D _iGM ), D _iB = (D _iB1 , D _iB2 ,..., D _iBM ) are read from the dictionary pattern storage unit 23 to perform the ticket type discrimination process.
FIG. 3 shows an example of the contents stored in the dictionary pattern storage unit 23.

図４は、判別処理部２４における券種判別処理の流れを示しており、以下、それについて説明する。まず、評価関数ｆ（Ｓ，Ｄ_ｉ）により券種ｉの評価値Ｅ（ｉ）を求め、この求めた評価値Ｅ（ｉ）とそれまでの評価値の最大値（Ｅｍａｘ）とを比較し、評価値Ｅ（ｉ）が最大値（Ｅｍａｘ）よりも大きければ、Ｅ（ｉ）をＥｍａｘとし、ｉを券種番号Ｄ_ＮＭとして記憶する。この処理は全ての券種に対して繰り返えされ、繰り返し処理が終わった後にＤ_ＮＭに記憶されている券種番号が当該紙葉類Ｐの券種番号として搬送制御部１５に送られる。 FIG. 4 shows the flow of the ticket type discrimination process in the discrimination processing unit 24, which will be described below. First, the evaluation value E ( _i ) of the ticket type i is obtained by the evaluation function f (S, D _i ), and the obtained evaluation value E (i) is compared with the maximum value (Emax) of the evaluation values so far. If the evaluation value E (i) is larger than the maximum value (Emax), E (i) is stored as Emax and i is stored as the ticket type number _DNM . This process is repeated for all the ticket types, and the ticket type number stored in the _DNM is sent to the transport control unit 15 as the ticket type number of the paper sheet P after the repetition process is completed.

評価値Ｅ（ｉ）を求める評価関数ｆ（Ｓ，Ｄ_ｉ）は、対応する色ごとに評価関数ｆ１（Ｓ_Ｃ，Ｄ_ｉＣ）にて評価した各色の評価値に重みを付けて加算する（下記式参照）。
Ｅ（ｉ）＝ｆ（Ｓ，Ｄ_ｉ）＝Ｋｒ×ｆ１（Ｓ_Ｒ，Ｄ_ｉＲ）
＋Ｋｇ×ｆ１（Ｓ_Ｇ，Ｄ_ｉＧ）＋Ｋｂ×ｆ１（Ｓ_Ｂ，Ｄ_ｉＢ）
Ｋｒ，Ｋｇ，Ｋｂ：各色に対する重み情報（Ｋｒ＋Ｋｇ＋Ｋｂ＝１）
評価関数ｆ１（Ｓ_Ｃ，Ｄ_ｉＣ）の一例として、下記数１に示すように、Ｓ_Ｃ＝（Ｓ_Ｃ１，Ｓ_Ｃ２，…，Ｓ_ＣＭ）とＤ_ｉＣ＝（Ｄ_ｉＣ１，Ｄ_ｉＣ２，…，Ｄ_ｉＣＭ）との内積が考えられる。

Obtaining an evaluation value E (i) the evaluation function f (S, _{D i),} the corresponding evaluation for each color function _f1 (S _{C, D iC)} and adding the weighted evaluation values of each color was evaluated by ( Refer to the following formula).
E (i) = f (S, D _i ) = Kr × f 1 (S _R , D _iR )
+ Kg × f1 (S _G , D _iG ) + Kb × f1 (S _B , D _iB )
Kr, Kg, Kb: Weight information for each color (Kr + Kg + Kb = 1)
As an example of the evaluation function f1 (S _C , D _iC ), S _C = (S _C1 , S _C2 ,..., S _CM ) and D _iC = (D _iC1 , D _iC2,. _Di product with D _iCM ) is conceivable.

このように、第１の実施の形態によれば、紙葉類から取得したカラー画像の色ごとに辞書パターンとの評価値を計算し、この求めた評価値を色ごとに重みを付けて加算し、この加算結果を券種判別するための評価値とすることで、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 As described above, according to the first embodiment, the evaluation value with the dictionary pattern is calculated for each color of the color image acquired from the paper sheet, and the obtained evaluation value is added with weighting for each color. By using this addition result as an evaluation value for discriminating the ticket type, stable ticket type discrimination can always be performed even when paper sheets of various colors have to be handled.

次に、第２の実施の形態について説明する。
図５は、第２の実施の形態に係る処理部１４の構成を示している。第２の実施の形態に係る処理部１４は、たとえば、カラーカメラ１３で撮像されたカラー画像から当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ（＝Ｓ）として抽出するパターン抽出手段としての画像処理部３１、画像処理部２１から得られる入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂを一時記憶する入力パターン記憶部３２、入力パターン記憶部２２内の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_ＢをＨ（Ｈｕｅ：色相）、Ｓ（Ｓａｔｕｒａｔｉｏｎ：彩度）、Ｖ（Ｖａｌｕｅ：バリュー）の表色系のパターンＳ_Ｈ，Ｓ_Ｓ，Ｓ_Ｖに変換するパターン変換手段としてのパターン変換部３３、あらかじめ判別対象とする複数種の紙葉類から取得したカラー画像に基づき作成された当該カラー画像を構成する複数の色、この例ではＲ、Ｇ、ＢのパターンがそれぞれＨ、Ｓ、Ｖの表色系に変換されてなる辞書パターンＤ_ｉＨ，Ｄ_ｉＳ，Ｄ_ｉＶ（＝Ｄ_ｉ）を記憶している辞書パターン記憶手段としての辞書パターン記憶部３４、パターン変換部３３で変換された入力パターンＳ_Ｈ，Ｓ_Ｓ，Ｓ_Ｖと辞書パターン記憶部２３内の辞書パターンＤ_ｉＨ，Ｄ_ｉＳ，Ｄ_ｉＶとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段、および、求められた各評価値を色ごとに所定の重みを付けて加算し、その加算結果に基づき当該紙葉類Ｐの種類を判別する判別手段として機能する判別処理部３５を有して構成されている。 Next, a second embodiment will be described.
FIG. 5 shows a configuration of the processing unit 14 according to the second embodiment. The processing unit 14 according to the second embodiment, for example, from a color image captured by the color camera 13, a plurality of colors constituting the color image, in this example R (red), G (green), B ( Blue) input patterns S _R , S _G , S _B (= S) are extracted as an image processing unit 31 as pattern extracting means, and input patterns S _R , S _G , S _B obtained from the image processing unit 21 are temporarily stored. input pattern storage unit 32, the input pattern _S R in the input pattern storage unit _22, S G, the _{S B} H (hue: hue), S (saturation: saturation), V (Value: Value) color system pattern S _{_H,} S _S, the conversion part 33 as a pattern conversion means for converting the S _V, the mosquito that is created based on the color image obtained from a plurality of kinds of sheets to be pre-determined target A plurality of colors constituting the over images, in this example R, G, H pattern B, respectively, S, color system conversion has been made dictionary pattern _D iH of _{_{V, D iS, D iV (}} = D i) The dictionary pattern storage unit 34 as a dictionary pattern storage means for storing the input patterns S _H , S _S , S _V converted by the pattern conversion unit 33 and the dictionary patterns D _iH , D _iS in the dictionary pattern storage unit 23 , _DiV for each corresponding color, and an evaluation value calculation means for calculating an evaluation value representing the degree of coincidence between both patterns, and a predetermined weight for each of the obtained evaluation values for each color It is configured to include a determination processing unit 35 that functions as a determination unit that adds and determines the type of the paper sheet P based on the addition result.

パターン変換部３３は、たとえば、図６に示すフローチャートにしたがって処理することで、Ｒ，Ｇ，Ｂの入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_ＢをＨ、Ｓ、Ｖの表色系のパターンＳ_Ｈ，Ｓ_Ｓ，Ｓ_Ｖに変換する。図６のフローチャートは、文献「Ｓｍｉｔｈ，ＡｌｖｅｙＲａｙ，”ＣｏｌｏｒＧａｍｕｔＴｒａｎｓｆｏｒｍａｔｉｏｎＰａｉｒｓ”，ＣｏｍｐｕｔｅｒＧｒａｐｈｉｃｓ，Ｖｏｌ．１２，ｐｐ．１２−１９，１９７８（ＰｒｏｃＳＩＧＧＲＡＰＨ７８）」を参照した表色系変換アルゴリズムである。 Pattern conversion section 33, for example, by processing according to the flowchart shown in FIG. 6, R, G, the input pattern _S R of _B, S G, _{S B} and H, S, pattern _{S H} of the color system of V , to convert _{S S,} to _{S V.} The flowchart of FIG. 6 is a color system conversion algorithm with reference to the document “Smith, Alley Ray,“ Color Gamut Transformation Pairs ”, Computer Graphics, Vol. 12, pp. 12-19, 1978 (Proc SIGGRAPH 78). .

判別処理部３５は、カラーカメラ１３から前述したようにＲ，Ｇ，Ｂの３色のデータが得られる場合、入力パターンＳ＝（Ｓ_Ｈ，Ｓ_Ｓ，Ｓ_Ｖ）：Ｓ_Ｈ＝（Ｓ_Ｈ１，Ｓ_Ｈ２，…，Ｓ_ＨＭ）、Ｓ_Ｓ＝（Ｓ_Ｓ１，Ｓ_Ｓ２，…，Ｓ_ＳＭ）、Ｓ_Ｖ＝（Ｓ_Ｖ１，Ｓ_Ｖ２，…，Ｓ_ＶＭ）をパターン変換部３３から取得すとともに、各券種の辞書パターンＤ_ｉ＝（Ｄ_ｉＨ，Ｄ_ｉＳ，Ｄ_ｉＶ）：Ｄ_ｉＨ＝（Ｄ_ｉＨ１，Ｄ_ｉＨ２，…，Ｄ_ｉＨＭ）、Ｄ_ｉＳ＝（Ｄ_ｉＳ１，Ｄ_ｉＳ２，…，Ｄ_ｉＳＭ）、Ｄ_ｉＶ＝（Ｄ_ｉＶ１，Ｄ_ｉＶ２，…，Ｄ_ｉＶＭ）を辞書パターン記憶部３４から読出すことにより、券種判別の処理を行なう。
ここに、図７は辞書パターン記憶部３４の記憶内容の一例を示している。 As described above, when the three color data of R, G, and B are obtained from the color camera 13, the discrimination processing unit 35 is input pattern S = (S _H , S _S , S _V ): S _H = (S _H1 , S _H2 ,..., S _HM ), S _S = (S _S1 , S _S2 ,..., S _SM ), S _V = (S _V1 , S _V2 ,..., S _VM ) are acquired from the pattern conversion unit 33. , each denomination of the dictionary pattern _{_{_{_{D i = (D iH, D}}}} iS, D iV): D iH = (D iH1, D iH2, ..., D iHM), D iS = (D iS1, D iS2, ..., D _iSM ), D _iV = (D _iV1 , D _iV2 ,..., D _iVM ) is read from the dictionary pattern storage unit 34 to perform the bill type discrimination process.
FIG. 7 shows an example of the contents stored in the dictionary pattern storage unit 34.

なお、判別処理部３５における券種判別処理の流れは、前述した第１の実施の形態と基本的には同様であるので説明は省略する。 Note that the flow of the ticket type discrimination process in the discrimination processing unit 35 is basically the same as that of the first embodiment described above, and thus the description thereof is omitted.

評価値Ｅ（ｉ）を求める評価関数ｆ（Ｓ，Ｄ_ｉ）は、対応する色ごとに評価関数ｆ３（Ｓ_Ｃ，Ｄ_ｉＣ）にて評価した各色の評価値に重みを付けて加算する（下記式参照）。
Ｅ（ｉ）＝ｆ（Ｓ，Ｄ_ｉ）＝ｆ３（Ｓ_Ｈ，Ｓ_Ｓ，Ｓ_Ｖ，Ｄ_ｉＨ，Ｄ_ｉＳ，Ｄ_ｉＶ）
＝Ｋｈ×ｆ３（Ｓ_Ｈ，Ｄ_ｉＨ）＋Ｋｓ×ｆ３（Ｓ_Ｓ，Ｄ_ｉＳ）
＋Ｋｖ×ｆ３（Ｓ_Ｖ，Ｄ_ｉＶ）
Ｋｈ，Ｋｓ，Ｋｖ：各色に対する重み情報（Ｋｈ＋Ｋｓ＋Ｋｖ＝１）
このように、第２の実施の形態によれば、紙葉類から取得したカラー画像を表色系に変換した後、辞書パターンとの評価値を計算し、この求めた評価値を重み付けして加算し、この加算結果を券種判別するための評価値とすることで、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 The evaluation function f (S, D _i ) for obtaining the evaluation value E (i) is weighted and added to the evaluation value of each color evaluated by the evaluation function f 3 (S _C , D _iC ) for each corresponding color ( Refer to the following formula).
E (i) = f (S, D _i ) = f 3 (S _H , S _S , S _V , D _iH , D _iS , D _iV )
= Kh × f3 (S _H , D _iH ) + Ks × f3 (S _S , D _iS )
+ Kv × f3 (S _V , D _iV )
Kh, Ks, Kv: Weight information for each color (Kh + Ks + Kv = 1)
Thus, according to the second embodiment, after converting a color image acquired from a paper sheet into a color system, an evaluation value with the dictionary pattern is calculated, and the obtained evaluation value is weighted. By adding the values and using the addition result as an evaluation value for discriminating the ticket type, stable ticket type discrimination can always be performed even when paper sheets of various colors have to be handled.

次に、第３の実施の形態について説明する。
図８は、第３の実施の形態に係る処理部１４の構成を示している。第３の実施の形態に係る処理部１４は、たとえば、カラーカメラ１３で撮像されたカラー画像から当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ（＝Ｓ）として抽出するパターン抽出手段としての画像処理部４１、画像処理部４１から得られる入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂを一時記憶する入力パターン記憶部４２、あらかじめ判別対象とする複数種の紙葉類から取得したカラー画像に基づき作成された当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）のうち特定の１つの色を指定する指定色情報（Ｄ_ｉＣ）と当該色の辞書パターンＤ_ｉ＝（Ｄ_ｉ１，Ｄ_ｉ２，…，Ｄ_ｉＭ）を記憶している辞書パターン記憶手段としての辞書パターン記憶部４３、入力パターン記憶部４２内の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂのうち辞書パターン記憶部４３に記憶された指定色情報（Ｄ_ｉＣ）で指定される色の入力パターンを選択し、この選択した入力パターンと辞書パターン記憶部４３内の辞書パターンＤ_ｉ＝（Ｄ_ｉ１，Ｄ_ｉ２，…，Ｄ_ｉＭ）とを照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段、および、求められた各評価値に基づき当該紙葉類Ｐの種類を判別する判別手段として機能する判別処理部４４を有して構成されている。
ここに、上記指定色情報（Ｄ_ｉＣ）は、この実施の形態では他の紙葉類と最もパターンに差が表われやすい１つの色を指定するものとする。 Next, a third embodiment will be described.
FIG. 8 shows the configuration of the processing unit 14 according to the third embodiment. The processing unit 14 according to the third embodiment, for example, from a color image captured by the color camera 13, a plurality of colors constituting the color image, in this example, R (red), G (green), B ( Blue) input patterns S _R , S _G , S _B (= S) are extracted as image extraction unit 41, and input patterns S _R , S _G , S _B obtained from image processing unit 41 are temporarily stored. Input pattern storage unit 42, a plurality of colors constituting the color image created based on color images acquired in advance from a plurality of types of paper sheets to be discriminated, in this example R (red), G (green) , B (blue), a designated color information (D _iC ) for designating a specific color, and a dictionary pattern D _i = (D _i1 , D _i2 ,..., D _iM ) for that color. Dictionary pattern as storage means Down storage unit 43, selects the input pattern S _R, S _G, the color of the input pattern specified by the designated color information stored in the dictionary pattern storage unit 43 of the S _B (D _iC) in the input pattern memory 42 Then, by comparing the selected input pattern with the dictionary pattern D _i = (D _i1 , D _i2 ,..., D _iM ) in the dictionary pattern storage unit 43, evaluation values representing the degree of coincidence of both patterns are calculated. And a discrimination processing unit 44 that functions as discrimination means for discriminating the type of the paper sheet P based on the obtained evaluation values.
Here, in the present embodiment, the designated color information (D _iC ) designates one color that most easily shows a difference in pattern from other paper sheets.

判別処理部４４は、カラーカメラ１３から前述したようにＲ，Ｇ，Ｂの３色のデータが得られる場合、入力パターンＳ＝（Ｓ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ）：Ｓ_Ｒ＝（Ｓ_Ｒ１，Ｓ_Ｒ２，…，Ｓ_ＲＭ）、Ｓ_Ｇ＝（Ｓ_Ｇ１，Ｓ_Ｇ２，…，Ｓ_ＧＭ）、Ｓ_Ｂ＝（Ｓ_Ｂ１，Ｓ_Ｂ２，…，Ｓ_ＢＭ）のうち、指定色情報（Ｄ_ｉＣ）で指定された色の入力パターンＳ_ＤｉＣ＝（Ｓ_ＤｉＣ１，Ｓ_ＤｉＣ２，…，Ｓ_ＤｉＣＭ）を入力パターン記憶部４２から読出すとともに、各券種の辞書パターンＤ_ｉ＝（Ｄ_ｉ１，Ｄ_ｉ２，…，Ｄ_ｉＭ）を辞書パターン記憶部４３から読出すことにより、券種判別の処理を行なう。
ここに、図９は辞書パターン記憶部４３の記憶内容の一例を示している。 As described above, when the color camera 13 obtains data of three colors R, G, and B from the color camera 13, the discrimination processing unit 44 has an input pattern S = (S _R , S _G , S _B ): S _R = (S _R1 , S _R2 ,..., S _RM ), S _G = (S _G1 , S _G2 ,..., S _GM ), S _B = (S _B1 , S _B2 ,..., S _BM ), the designated color information (D _iC ) input pattern of the specified color _{_{_{S DiC = (S DiC1, S}}} DiC2, ..., with reading from the input pattern storage unit 42 _{S DICM),} the dictionary pattern _D of each note denomination i ₌ _{(D i1,} D _i2 ,..., D _iM ) are read from the dictionary pattern storage unit 43 to perform ticket type discrimination processing.
FIG. 9 shows an example of the contents stored in the dictionary pattern storage unit 43.

図１０は、判別処理部４４における券種判別処理の流れを示しており、以下、それについて説明する。まず、評価関数ｆ（Ｓ_ＤｉＣ，Ｄ_ｉ）により券種ｉの評価値Ｅ（ｉ）を求め、この求めた評価値Ｅ（ｉ）とそれまでの評価値の最大値（Ｅｍａｘ）とを比較し、評価値Ｅ（ｉ）が最大値（Ｅｍａｘ）よりも大きければ、Ｅ（ｉ）をＥｍａｘとし、ｉを券種番号Ｄ_ＮＭとして記憶する。この処理は全ての券種に対して繰り返えされ、繰り返し処理が終わった後にＤ_ＮＭに記憶されている券種番号が当該紙葉類Ｐの券種番号として搬送制御部１５に送られる。 FIG. 10 shows the flow of the ticket type discrimination process in the discrimination processing unit 44, which will be described below. First, the evaluation value E ( _i ) of the ticket type i is obtained by the evaluation function f (S _DiC , D _i ), and the obtained evaluation value E (i) is compared with the maximum value (Emax) of the evaluation values so far. If the evaluation value E (i) is larger than the maximum value (Emax), E (i) is stored as Emax and i is stored as the ticket type number _DNM . This process is repeated for all the ticket types, and the ticket type number stored in the _DNM is sent to the transport control unit 15 as the ticket type number of the paper sheet P after the repetition process is completed.

評価値Ｅ（ｉ）を求める評価関数ｆ（Ｓ_ＤｉＣ，Ｄ_ｉ）の一例として、下記数２に示すように、Ｓ_ＤｉＣ＝（Ｓ_ＤｉＣ１，Ｓ_ＤｉＣ２，…，Ｓ_ＤｉＣＭ）とＤ_ｉ＝（Ｄ_ｉ１，Ｄ_ｉ２，…，Ｄ_ｉＭ）との内積が考えられる。

The evaluation function _{f (S} DiC, _{D i)} for obtaining the evaluation value E (i) as an example of, as shown in following equation _{_{_{2, S DiC = (S DiC1}}} , S DiC2, ..., S DiCM) and _D i = ( _Di products with D _i1 , D _i2 ,..., D _iM ) are conceivable.

このように、第３の実施の形態によれば、紙葉類から取得したカラー画像のうち、辞書パターン記憶部にあらかじめ記憶された指定色情報で指定された色を用いて辞書パターンとの評価値を計算し、この求めた評価値を券種判別するための評価値とすることで、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 As described above, according to the third embodiment, the color image acquired from the paper sheet is evaluated with the dictionary pattern using the color specified by the specified color information stored in advance in the dictionary pattern storage unit. By calculating the value and using the obtained evaluation value as an evaluation value for discriminating the ticket type, stable ticket type discrimination can always be performed even when paper sheets of various colors have to be dealt with.

次に、第４の実施の形態について説明する。
図１１は、第４の実施の形態に係る処理部１４の構成を示している。第４の実施の形態に係る処理部１４は、たとえば、カラーカメラ１３で撮像されたカラー画像から当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ（＝Ｓ）として抽出するパターン抽出手段としての画像処理部５１、画像処理部５１から得られる入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂを一時記憶する入力パターン記憶部５２、あらかじめ判別対象とする複数種の紙葉類から取得したカラー画像に基づき作成された当該カラー画像を構成する複数の色、この例ではＲ（赤）、Ｇ（緑）、Ｂ（青）の辞書パターンＤ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢ（＝Ｄ_ｉ）および各色ごとの重み情報（Ｗ_ｉＲ，Ｗ_ｉＧ，Ｗ_ｉＢ）を記憶している辞書パターン記憶手段としての辞書パターン記憶部５３、入力パターン記憶部５２内の入力パターンＳ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂと辞書パターン記憶部５３内の辞書パターンＤ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢとを対応する色ごとに照合することにより両パターンの一致度を表す評価値をそれぞれ計算する評価値計算手段、および、求められた各評価値を辞書パターン記憶部５３に記憶されている重み情報（Ｗ_ｉＲ，Ｗ_ｉＧ，Ｗ_ｉＢ）による重みを対応する色ごとに付けて加算し、その加算結果に基づき当該紙葉類Ｐの種類を判別する判別手段として機能する判別処理部５４を有して構成されている。
ここに、上記重み情報（Ｗ_ｉＲ，Ｗ_ｉＧ，Ｗ_ｉＢ）は、この実施の形態では他の紙葉類と最もパターンに差が表われやすい色間演算の重み情報とする。 Next, a fourth embodiment will be described.
FIG. 11 shows the configuration of the processing unit 14 according to the fourth embodiment. The processing unit 14 according to the fourth exemplary embodiment includes, for example, a plurality of colors constituting the color image from the color image captured by the color camera 13, in this example R (red), G (green), B ( Blue) input patterns S _R , S _G , S _B (= S) are extracted as image extraction unit 51 as pattern extraction means, and input patterns S _R , S _G , S _B obtained from image processing unit 51 are temporarily stored. Input pattern storage unit 52, a plurality of colors constituting the color image created based on color images acquired in advance from a plurality of types of paper sheets to be discriminated, in this example R (red), G (green) , B (blue) dictionary pattern D _iR , D _iG , D _iB (= D _i ) and weight information ( _WiR , _WiG , _WiB ) for each color are stored as dictionary pattern storage means. Memory 53, by comparing the input patterns S _R , S _G , and S _B in the input pattern storage unit 52 with the dictionary patterns D _iR , D _iG , and D _iB in the dictionary pattern storage unit 53 for each corresponding color. An evaluation value calculating means for calculating an evaluation value representing the degree of coincidence of each, and a weight based on weight information (W _iR , _WiG , _WiB ) stored in the dictionary pattern storage unit 53 for each of the obtained evaluation values A discrimination processing unit 54 that functions as a discriminator for discriminating the type of the paper sheet P based on the addition result for each corresponding color is added.
Here, the weight information (W _iR , W _iG , W _iB ) is weight information for inter-color calculation that is most likely to show a difference in pattern with other paper sheets in this embodiment.

判別処理部５４は、カラーカメラ１３から前述したようにＲ，Ｇ，Ｂの３色のデータが得られる場合、入力パターンＳ＝（Ｓ_Ｒ，Ｓ_Ｇ，Ｓ_Ｂ）：Ｓ_Ｒ＝（Ｓ_Ｒ１，Ｓ_Ｒ２，…，Ｓ_ＲＭ）、Ｓ_Ｇ＝（Ｓ_Ｇ１，Ｓ_Ｇ２，…，Ｓ_ＧＭ）、Ｓ_Ｂ＝（Ｓ_Ｂ１，Ｓ_Ｂ２，…，Ｓ_ＢＭ）を入力パターン記憶部５２から読出すとともに、各券種の辞書パターンＤ_ｉ＝（Ｄ_ｉＲ，Ｄ_ｉＧ，Ｄ_ｉＢ）：Ｄ_ｉＲ＝（Ｄ_ｉＲ１，Ｄ_ｉＲ２，…，Ｄ_ｉＲＭ）、Ｄ_ｉＧ＝（Ｄ_ｉＧ１，Ｄ_ｉＧ２，…，Ｄ_ｉＧＭ）、Ｄ_ｉＢ＝（Ｄ_ｉＢ１，Ｄ_ｉＢ２，…，Ｄ_ｉＢＭ）を辞書パターン記憶部５３から読出すことにより、券種判別の処理を行なう。
ここに、図１２は辞書パターン記憶部５３の記憶内容の一例を示している。 The discrimination processing unit 54, when the data of three colors R, G, and B is obtained from the color camera 13 as described above, the input pattern S = (S _R , S _G , S _B ): S _R = (S _R1 , S _R2 ,..., S _RM ), S _G = (S _G1 , S _G2 ,..., S _GM ), S _B = (S _B1 , S _B2 ,..., S _BM ) are read from the input pattern storage unit 52. with, each denomination of the dictionary pattern _{_{_{_{D i = (D iR, D}}}} iG, D iB): D iR = (D iR1, D iR2, ..., D iRM), D iG = (D iG1, D iG2, ..., D _iGM ), D _iB = (D _iB1 , D _iB2 ,..., D _iBM ) are read from the dictionary pattern storage unit 53 to perform the bill type discrimination process.
FIG. 12 shows an example of the contents stored in the dictionary pattern storage unit 53.

なお、判別処理部５４における券種判別処理の流れは、前述した第１の実施の形態（図４）と同様であるので説明は省略する。 Note that the flow of the ticket type discrimination process in the discrimination processing unit 54 is the same as that in the first embodiment (FIG. 4) described above, and thus the description thereof is omitted.

評価値Ｅ（ｉ）を求める評価関数ｆ（Ｓ，Ｄ_ｉ）は、対応する色ごとに評価関数ｆ１（Ｓ_Ｃ，Ｄ_ｉＣ）にて評価した各色の評価値に辞書パターン記憶部５３に記憶されている重み情報（Ｗ_ｉＲ，Ｗ_ｉＧ，Ｗ_ｉＢによる重みを付けて加算する（下記式参照）。
Ｅ（ｉ）＝ｆ（Ｓ，Ｄ_ｉ）＝Ｗ_ｉＲ×ｆ１（Ｓ_Ｒ，Ｄ_ｉＲ）
＋Ｗ_ｉＧ×ｆ１（Ｓ_Ｇ，Ｄ_ｉＧ）＋Ｗ_ｉＢ×ｆ１（Ｓ_Ｂ，Ｄ_ｉＢ）
Ｗ_ｉＲ，Ｗ_ｉＧ，Ｗ_ｉＢ：各色に対する重み情報（Ｗ_ｉＲ＋Ｗ_ｉＧ＋Ｗ_ｉＢ＝１）
評価関数ｆ１（Ｓ_Ｃ，Ｄ_ｉＣ）の一例として、前記数１に示したように、Ｓ_Ｃ＝（Ｓ_Ｃ１，Ｓ_Ｃ２，…，Ｓ_ＣＭ）とＤ_ｉＣ＝（Ｄ_ｉＣ１，Ｄ_ｉＣ２，…，Ｄ_ｉＣＭ）との内積が考えられる。
このように、第４の実施の形態によれば、紙葉類から取得したカラー画像の色ごとに辞書パターンとの評価値を計算し、この求めた評価値を辞書パターン記憶部にあらかじめ記憶されている重みを用いて加算し、この加算結果を券種判別するための評価値とすることで、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 Obtaining an evaluation value E (i) the evaluation function f (S, _{D i),} the corresponding evaluation for each color function _f1 (S _{C, D iC)} stored in the dictionary pattern storage unit 53 in the evaluation value of each color was evaluated by has been that weight information _(W _{_iR,} W iG, and adding the weighted by _{W iB} (see the following formula).
E (i) = f (S, D _i ) = W _iR × f1 (S _R , D _iR )
+ W _iG × f1 (S _G , D _iG ) + W _iB × f1 (S _B , D _iB )
_WiR , _WiG , _WiB : Weight information for each color ( _WiR + _WiG + _WiB = 1)
As an example of the evaluation function f1 (S _C , D _iC ), as shown in _Equation ₁ , S _C = (S _C1 , S _C2 ,..., S _CM ) and D _iC = (D _iC1 , D _iC2,. , D _iCM ).
As described above, according to the fourth embodiment, the evaluation value with the dictionary pattern is calculated for each color of the color image acquired from the paper sheet, and the obtained evaluation value is stored in the dictionary pattern storage unit in advance. By using the weights that are added, and using the result of the addition as an evaluation value for discriminating the ticket type, it is possible to always perform stable ticket type discrimination even when paper sheets of various colors must be supported. The

なお、前述した第１〜第４の実施の形態に限らず、たとえば、第２の実施の形態と第３の実施の形態とを合併することで、紙葉類から取得したカラー画像を表色系に変換した後、辞書パターン記憶部にあらかじめ記憶された指定色情報で指定された色を用いて辞書パターンとの評価値を計算し、この求めた評価値を券種判別するための評価値とするようにしてもよく、このようにしても、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 In addition to the first to fourth embodiments described above, for example, by combining the second embodiment and the third embodiment, a color image acquired from paper sheets is represented by color. After conversion to the system, the evaluation value for the dictionary pattern is calculated using the color specified by the specified color information stored in advance in the dictionary pattern storage unit, and the evaluation value for discriminating the obtained evaluation value for the ticket type Even in this case, even when it is necessary to cope with paper sheets of various colors, it is possible to always perform stable ticket type discrimination.

また、たとえば、第２の実施の形態と第４の実施の形態とを合併することで、紙葉類から取得したカラー画像を表色系に変換した後、辞書パターンとの評価値を計算し、この求めた評価値を辞書パターン記憶部にあらかじめ記憶されている重みを用いて加算し、この加算結果を券種判別するための評価値とするようにしてもよく、このようにしても、様々な色の紙葉類に対応しなければいけない場合でも、常に安定した券種判別が行なえる。 Also, for example, by merging the second embodiment and the fourth embodiment, after converting a color image acquired from paper sheets into a color system, the evaluation value with the dictionary pattern is calculated. The obtained evaluation value may be added using the weight stored in advance in the dictionary pattern storage unit, and the addition result may be used as an evaluation value for discriminating the ticket type. Even when paper sheets of various colors have to be dealt with, stable ticket type discrimination can always be performed.

本発明の実施の形態に係る紙葉類判別装置の構成を概略的に示す模式図。The schematic diagram which shows roughly the structure of the paper sheet discrimination | determination apparatus which concerns on embodiment of this invention. 第１の実施の形態に係る処理部の構成および券種判別の様子を説明する模式図。The schematic diagram explaining the mode of the structure of a process part and ticket type discrimination | determination concerning 1st Embodiment. 第１の実施の形態に係る辞書パターン記憶部の記憶内容の一例を示す模式図。The schematic diagram which shows an example of the memory content of the dictionary pattern memory | storage part which concerns on 1st Embodiment. 第１の実施の形態に係る判別処理部における券種判別処理の流れを示すフローチャート。The flowchart which shows the flow of the ticket type discrimination | determination process in the discrimination | determination processing part which concerns on 1st Embodiment. 第２の実施の形態に係る処理部の構成および券種判別の様子を説明する模式図。The schematic diagram explaining the mode of the structure of a process part and ticket type discrimination | determination concerning 2nd Embodiment. 第２の実施の形態に係るパターン変換部における表色系変換アルゴリズムを示すフローチャート。9 is a flowchart showing a color system conversion algorithm in a pattern conversion unit according to the second embodiment. 第２の実施の形態に係る辞書パターン記憶部の記憶内容の一例を示す模式図。The schematic diagram which shows an example of the memory content of the dictionary pattern memory | storage part which concerns on 2nd Embodiment. 第３の実施の形態に係る処理部の構成および券種判別の様子を説明する模式図。The schematic diagram explaining the mode of the structure of a process part and ticket type discrimination | determination concerning 3rd Embodiment. 第３の実施の形態に係る辞書パターン記憶部の記憶内容の一例を示す模式図。The schematic diagram which shows an example of the memory content of the dictionary pattern memory | storage part which concerns on 3rd Embodiment. 第３の実施の形態に係る判別処理部における券種判別処理の流れを示すフローチャート。The flowchart which shows the flow of the ticket type discrimination | determination process in the discrimination | determination processing part which concerns on 3rd Embodiment. 第４の実施の形態に係る処理部の構成および券種判別の様子を説明する模式図。The schematic diagram explaining the mode of the structure of a process part and ticket kind discrimination | determination concerning 4th Embodiment. 第４の実施の形態に係る辞書パターン記憶部の記憶内容の一例を示す模式図。The schematic diagram which shows an example of the memory content of the dictionary pattern memory | storage part which concerns on 4th Embodiment.

Explanation of symbols

Ｐ…紙葉類（紙幣）、１１…搬送路（搬送手段）、１２…照明装置、１３…カラーカメラ（カラー画像取得手段）、１４…処理部、１５…搬送制御部、２１，３１，４１，５１…画像処理部（パターン抽出手段）、２２，３２，４２，５２…入力パターン記憶部、２３，３４，４３，５３…辞書パターン記憶部（辞書パターン記憶手段）、２４，３５，４４，５４…判別処理部（評価値計算手段、判別手段）、３３…パターン変換部（パターン変換手段）。 P ... paper sheets (banknotes), 11 ... conveying path (conveying means), 12 ... illuminating device, 13 ... color camera (color image obtaining means), 14 ... processing unit, 15 ... conveying control unit, 21, 31, 41 , 51 ... Image processing unit (pattern extraction means), 22, 32, 42, 52 ... Input pattern storage unit, 23, 34, 43, 53 ... Dictionary pattern storage unit (dictionary pattern storage means), 24, 35, 44, 54: Discrimination processing unit (evaluation value calculation means, discrimination means), 33 ... Pattern conversion unit (pattern conversion means).

Claims

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
Dictionary pattern storage means for storing in advance a dictionary pattern for each of a plurality of colors constituting the color image acquired from the color image respectively corresponding to the type of paper sheet;
The degree of coincidence of both patterns is expressed by collating the input pattern for each of the plurality of colors extracted by the pattern extraction unit and the dictionary pattern for each of the plurality of colors stored in the dictionary pattern storage unit for each corresponding color. An evaluation value calculating means for calculating the evaluation value, and
A discriminating unit that adds each evaluation value obtained by the evaluation value calculating unit with a predetermined weight for each color, and discriminates the type of the sheet based on the addition result;
A paper sheet discriminating apparatus comprising:

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
Pattern conversion means for converting the input patterns for each of the plurality of colors extracted by the pattern extraction means into patterns of different color systems;
A dictionary pattern storage that stores in advance a dictionary pattern obtained by converting a pattern for each of a plurality of colors constituting a color image acquired from a color image corresponding to each type of paper sheet into a different color system Means,
The degree of coincidence of both patterns is expressed by collating each input system of the color system converted by the pattern conversion means and each dictionary pattern of the color system stored in the dictionary pattern storage means for each corresponding color. An evaluation value calculating means for calculating the evaluation value, and
A discriminating unit that adds each evaluation value obtained by the evaluation value calculating unit with a predetermined weight for each color, and discriminates the type of the sheet based on the addition result;
A paper sheet discriminating apparatus comprising:

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
A dictionary preliminarily storing designated color information for designating a specific color among a plurality of colors constituting the color image acquired from a color image corresponding to each type of paper sheet and a dictionary pattern of the color Pattern storage means;
The input pattern of the color designated by the designated color information stored in the dictionary pattern storage means is selected from the input patterns for each of the plurality of colors extracted by the pattern extraction means, and the selected input pattern and the dictionary pattern Evaluation value calculation means for calculating an evaluation value representing the degree of coincidence of both patterns by collating with a dictionary pattern stored in the storage means;
Discriminating means for discriminating the type of the paper sheet based on the evaluation value obtained by the evaluation value calculating means;
A paper sheet discriminating apparatus comprising:

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
Dictionary pattern storage means for storing in advance a dictionary pattern for each of a plurality of colors and weight information for each color constituting the color image acquired from a color image corresponding to each type of paper sheet;
The degree of coincidence of both patterns is expressed by collating the input pattern for each of the plurality of colors extracted by the pattern extraction unit and the dictionary pattern for each of the plurality of colors stored in the dictionary pattern storage unit for each corresponding color. An evaluation value calculating means for calculating the evaluation value, and
Each evaluation value obtained by the evaluation value calculating means is added by adding the weight based on the weight information for each color stored in the dictionary pattern storage means for each corresponding color, and based on the addition result, the paper sheet Discriminating means for discriminating the type of the
A paper sheet discriminating apparatus comprising:

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
Pattern conversion means for converting the input patterns for each of the plurality of colors extracted by the pattern extraction means into patterns of different color systems;
Designated color information for designating a specific color out of a plurality of colors constituting the color image acquired from the color image corresponding to each type of paper sheet and the color pattern are converted into different color systems Dictionary pattern storage means for storing a dictionary pattern in advance,
The input pattern of the color designated by the designated color information stored in the dictionary pattern storage means is selected from the color system input patterns converted by the pattern conversion means, and the selected input pattern and the dictionary pattern are selected. Evaluation value calculation means for calculating an evaluation value representing the degree of coincidence of both patterns by collating with a dictionary pattern stored in the storage means;
Discriminating means for discriminating the type of the paper sheet based on the evaluation value obtained by the evaluation value calculating means;
A paper sheet discriminating apparatus comprising:

A conveying means for conveying the paper sheet to be discriminated;
Color image acquisition means for acquiring a color image on the paper sheet conveyed by the conveyance means;
Pattern extraction means for extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition means;
Pattern conversion means for converting the input patterns for each of the plurality of colors extracted by the pattern extraction means into patterns of different color systems;
A dictionary pattern in which a pattern for each of a plurality of colors constituting a color image acquired from a color image corresponding to each type of paper sheet is converted into a different color system and weight information for each color are stored in advance. Dictionary pattern storage means,
The degree of coincidence of both patterns is expressed by collating each input system of the color system converted by the pattern conversion means and each dictionary pattern of the color system stored in the dictionary pattern storage means for each corresponding color. An evaluation value calculating means for calculating the evaluation value, and
Each evaluation value obtained by the evaluation value calculating means is added by adding the weight based on the weight information for each color stored in the dictionary pattern storage means for each corresponding color, and based on the addition result, the paper sheet Discriminating means for discriminating the type of the
A paper sheet discriminating apparatus comprising:

The plurality of colors constituting the color image are R (red), G (green), and B (blue), and the pattern conversion means inputs the R (red), G (green), and B (blue). 6. The paper sheet discriminating apparatus according to claim 2, wherein the pattern is converted into a color system of H (hue), S (saturation), and V (value).

4. The designated color information for designating one specific color is designated color information for designating one color that is most likely to be different in pattern from other paper sheets. 5. The paper sheet discriminating device according to 5.

7. The paper sheet discriminating apparatus according to claim 4 or 6, wherein the weight information for each color is weight information of an inter-color calculation in which a pattern is most likely to differ from other paper sheets.

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
An input pattern for each of a plurality of colors extracted by the pattern extraction unit and a dictionary pattern for each of a plurality of colors constituting the color image acquired from a color image corresponding to each type of paper sheet are stored in advance. An evaluation value calculation step of calculating an evaluation value representing the degree of coincidence between the two patterns by matching the dictionary pattern for each of a plurality of colors stored in the dictionary pattern storage means,
Each evaluation value obtained by this evaluation value calculation step is added with a predetermined weight for each color, and a determination step for determining the type of the paper sheet based on the addition result,
A paper sheet discriminating method comprising:

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
A pattern conversion step for converting an input pattern for each of a plurality of colors extracted by this pattern extraction step into a different color system pattern;
Each color system input pattern converted by this pattern conversion process and a pattern for each of a plurality of colors constituting the color image acquired from a color image respectively corresponding to the type of paper sheet are different color specifications. An evaluation value representing the degree of coincidence between the two patterns by matching each dictionary color of the color system stored in the dictionary pattern storage means that stores the dictionary pattern converted into the system in advance for each corresponding color An evaluation value calculation process for calculating each,
Each evaluation value obtained by this evaluation value calculation step is added with a predetermined weight for each color, and a determination step for determining the type of the paper sheet based on the addition result,
A paper sheet discriminating method comprising:

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
Among the input patterns for each of the plurality of colors extracted by the pattern extraction unit, a specific one color is selected from among the plurality of colors constituting the color image acquired from the color image corresponding to the type of paper sheet. The input pattern of the color specified by the specified color information stored in the dictionary pattern storage means for storing the specified color information to be specified and the dictionary pattern of the color in advance is selected, and the selected input pattern and the dictionary pattern storage are selected. An evaluation value calculation step of calculating an evaluation value representing the degree of coincidence of both patterns by collating with a dictionary pattern stored in the means;
A discriminating step for discriminating the type of the paper sheet based on the evaluation value obtained by the evaluation value calculating step,
A paper sheet discriminating method comprising:

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
An input pattern for each of a plurality of colors extracted by this pattern extraction process, a dictionary pattern for each of a plurality of colors that constitutes the color image acquired from a color image corresponding to each type of paper sheet, and a weight for each color An evaluation value calculation step for calculating an evaluation value representing the degree of coincidence between both patterns by collating the dictionary pattern for each of a plurality of colors stored in the dictionary pattern storage means storing information in advance for each corresponding color When,
Each evaluation value obtained by this evaluation value calculation step is added by adding a weight based on the weight information for each color stored in the dictionary pattern storage means for each corresponding color, and the paper sheet is based on the addition result A discrimination process for discriminating the type of the
A paper sheet discriminating method comprising:

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
A pattern conversion step for converting an input pattern for each of a plurality of colors extracted by this pattern extraction step into a different color system pattern;
Among each input pattern of the color system converted by this pattern conversion step, a specific color among a plurality of colors constituting the color image acquired from the color image respectively corresponding to the type of paper sheet is obtained. Select the input pattern of the color specified by the specified color information stored in the dictionary pattern storage means that stores the specified color information to be specified and the dictionary pattern obtained by converting the color pattern into another color system. And an evaluation value calculating step for calculating an evaluation value representing the degree of coincidence of both patterns by comparing the selected input pattern with the dictionary pattern stored in the dictionary pattern storage means;
A discriminating step for discriminating the type of the paper sheet based on the evaluation value obtained by the evaluation value calculating step,
A paper sheet discriminating method comprising:

A color image acquisition step of acquiring a color image on the paper sheet conveyed by the conveyance means by the color image acquisition means;
A pattern extraction step of extracting, as an input pattern, a pattern for each of a plurality of colors constituting the color image from the color image acquired by the color image acquisition step;
A pattern conversion step for converting an input pattern for each of a plurality of colors extracted by this pattern extraction step into a different color system pattern;
A dictionary pattern in which a pattern for each of a plurality of colors constituting a color image acquired from a color image corresponding to each type of paper sheet is converted into a different color system and weight information for each color are stored in advance. Dictionary pattern storage means,
Each input system of the color system converted by this pattern conversion process and a pattern for each of a plurality of colors constituting the color image acquired from the color image respectively corresponding to the type of the paper sheet are displayed in the color system. The degree of coincidence between the two patterns can be determined by collating the converted dictionary pattern and each color system dictionary pattern stored in the dictionary pattern storage means storing the weight information for each color in advance for each corresponding color. An evaluation value calculation process for calculating each evaluation value to be represented;
Each evaluation value obtained by this evaluation value calculation step is added by adding a weight based on the weight information for each color stored in the dictionary pattern storage means for each corresponding color, and the paper sheet is based on the addition result A discrimination process for discriminating the type of the
A paper sheet discriminating method comprising: