WO2018214627A1 - Procédé et système de détection de l'authenticité de billets de banque - Google Patents

Procédé et système de détection de l'authenticité de billets de banque Download PDF

Info

Publication number
WO2018214627A1
WO2018214627A1 PCT/CN2018/079783 CN2018079783W WO2018214627A1 WO 2018214627 A1 WO2018214627 A1 WO 2018214627A1 CN 2018079783 W CN2018079783 W CN 2018079783W WO 2018214627 A1 WO2018214627 A1 WO 2018214627A1
Authority
WO
WIPO (PCT)
Prior art keywords
banknote
image
positioning
feature points
key feature
Prior art date
Application number
PCT/CN2018/079783
Other languages
English (en)
Chinese (zh)
Inventor
崔山领
王倩文
徐靖
Original Assignee
广州广电运通金融电子股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州广电运通金融电子股份有限公司 filed Critical 广州广电运通金融电子股份有限公司
Publication of WO2018214627A1 publication Critical patent/WO2018214627A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
    • G07D7/16—Testing the dimensions
    • G07D7/162—Length or width

Definitions

  • the invention relates to a method for detecting the authenticity of banknotes, in particular to a method and a system for realizing the authenticity detection of banknotes by accurately detecting the image size of banknotes.
  • the size of normal banknotes must meet a certain size. If the size of the banknotes is too large or too small, it is an abnormal banknote, and circulation is prohibited. Therefore, if the size of the banknotes circulating in the market is abnormal, it can usually be considered to be caused by artificial cutting or splicing, and this part of the banknote is what we call an abnormal banknote that is not allowed to circulate.
  • the overall width and height of the banknote are calculated by detecting the image of the banknote, and then compared with the geometrical size of the standard banknote to see if the size requirement is met, and if it is satisfied, it is judged as Normal size banknotes, otherwise judged as abnormal banknotes, for normal size banknotes can be combined with other security features to further determine the authenticity of banknotes.
  • the banknote image collected by the scan is indirectly measured and calculated, and this is inevitably subject to errors such as actual physical size of the banknote, mechanical motion error of the image capturing device, and image correction error.
  • the influence of one or several factors causes a large error in the image of the collected banknote (up to 6 mm or more), which eventually causes a large deviation between the size of the banknote to be tested and the size of the standard banknote, resulting in the determination of the size of the real banknote due to the size problem.
  • the banknote processing apparatus makes an erroneous operation of rejecting the real banknotes. For cropped or spliced non-normal banknotes of the same size as standard banknotes, which are mistaken for genuine banknotes due to their physical size, may eventually lead to erroneous recognition results that cause abnormal banknotes to enter the market.
  • the banknote authenticity detecting method comprises the following steps:
  • Step 1 image acquisition and correction, collecting banknote images, rotating and cropping the image according to the measured vertices and angle information of the banknotes, and finding a valid banknote image area;
  • Step 2 positioning the key feature points, performing key feature point positioning on the banknote image to be tested, and calculating the size feature value of the banknote image to be tested according to the relative coordinates of the key feature point positioning;
  • Step 3 authenticity judgment, comparing the proximity of the size feature value of the banknote image to be tested and the standard size feature value corresponding to the preset standard banknote, and if the error exceeds the threshold value, it is determined as an abnormal banknote, and if the error does not exceed the threshold value, For normal banknotes.
  • the performing key feature point positioning on the banknote image to be tested specifically includes:
  • Step 2.1 performing coarse positioning of key feature points on the image of the banknote to be tested, finding the approximate position of the key feature points according to the information and setting information of the face value of the banknote, and calculating the coarse positioning coordinates of the key feature points by means of projection positioning or integral map positioning;
  • the key feature points include a lateral feature point of the banknote image and a longitudinal feature point, wherein the horizontal feature point is used to calculate a lateral size feature value of the banknote image, and the longitudinal feature point is used to calculate a longitudinal size feature of the banknote image. value.
  • the number of the lateral feature points and the longitudinal feature points are respectively greater than two.
  • At least one of the lateral feature points and the longitudinal feature points is a reuse feature point.
  • the collecting banknote images comprises collecting a banknote front image and a reverse image.
  • the invention also provides a banknote authenticity detecting system, comprising:
  • a banknote image acquisition module configured to collect an image of the banknote, and rotate and crop the image according to the measured vertices and angle information of the banknote to find an effective banknote image area
  • a feature point positioning module is configured to perform key feature point positioning on the banknote image to be tested, and calculate a size feature value of the banknote image to be tested according to the relative coordinates of the key feature point positioning;
  • a banknote authenticity judging module is configured to compare the feature value of the banknote image to be tested with the approximation degree of the feature value corresponding to the preset standard banknote, and if the error exceeds the threshold value, it is determined as an abnormal banknote, and if the error does not exceed the threshold value, it is determined as Normal banknotes.
  • the feature point location module specifically includes:
  • a coarse positioning module performs rough positioning of key feature points on the image of the banknote to be tested, finds the approximate position of the key feature points according to the information and setting information of the face value of the banknote, and calculates the coarse positioning of the key feature points by means of projection positioning or integral map positioning. coordinate;
  • a fine positioning module performs precise positioning of key feature points on the banknote image to be measured, expands N pixels on the basis of the above coarse positioning coordinates, and calculates accurate coordinates by template matching, wherein N is a natural number.
  • the scheme performs key feature point positioning by measuring the banknote image, and calculates the size feature value of the banknote image to be tested according to the relative coordinates of the key feature point positioning; in particular, the banknote image is measured by the relative coordinates between the plurality of key feature points.
  • the overall error localization avoids the influence of the large size error of the whole banknote, refines the overall error to the local, and through the relativeization of multiple absolute errors, the weakening error is achieved to improve the recognition accuracy, and the solution is effectively solved.
  • the overall size meets the requirements of the wrong decision of cutting or splicing banknotes.
  • FIG. 1 is a schematic view showing a processing flow of a banknote authenticity detecting method provided by the present invention
  • FIG. 2 is a schematic diagram showing main processes of performing key feature point positioning on a banknote image in FIG. 1;
  • FIG. 3 is a schematic diagram showing the selection of key feature points on one side of a banknote image of one currency
  • FIG. 4 is a schematic diagram of a system composition for realizing the method for detecting the authenticity of the banknote of FIG. 1.
  • the banknote authenticity detecting method provided by the present invention mainly comprises the following steps:
  • the key feature point positioning of the banknote image to be tested specifically includes:
  • the key feature points include a lateral feature point and a longitudinal feature point of the banknote image, wherein the horizontal feature point is used to calculate a lateral size feature value of the banknote image,
  • the longitudinal feature points are used to calculate the longitudinal dimension feature values of the banknote image.
  • the number of the horizontal feature points and the vertical feature points are respectively greater than two, and considering the contradiction between the processing efficiency and the accuracy, the number of the horizontal feature points and the vertical feature points is preferably selected from 3 to 8 Of course, in the case of extremely high processing capacity of the device, the number of feature points can be appropriately increased to improve the accuracy while ensuring processing efficiency.
  • At least one of the horizontal feature points and the vertical feature points is a reuse feature point, so as to detect and determine the constraint feature of the horizontal and vertical dimensions of the banknote image.
  • the collecting the banknote image comprises collecting the front side image and the reverse side image of the banknote, and performing the above-mentioned authenticity detecting step on the front and back sides of the banknote image respectively to obtain a more accurate detection result.
  • the key feature points can cover the horizontal direction and the longitudinal direction of the banknote respectively, as shown in Fig. 3, which are respectively the selection of the horizontal and vertical key feature points of the front and back of a banknote, and the selected key features.
  • the points are marked accordingly, for example, the front lateral key feature points are recorded as (FH1, FH2, FH3, FHV4), and the longitudinal key feature points are marked as (FHV4, FV5, FV6); wherein FHV4 is the horizontal and vertical reuse feature points;
  • the coordinates of the four key feature points after positioning are:
  • FX1 ⁇ FX8 are the calculated standard size feature values, where FX6 ⁇ FX8 are the standard size feature values across the feature area, and width is the identified long edge width value of the banknote.
  • the coarse positioning is to find the key feature points in the approximate position of the banknote according to the face value of the banknote, and calculate the key by means of projection positioning or integral map positioning.
  • FX1 - FX8 are the calculated size characteristic values of the banknote sample to be tested, wherein FX6 - FX8 are cross-feature area size feature values of the banknote sample to be tested, and width is the measured long side width of the banknote;
  • the authenticity judgment compares the size feature value of the banknote image to be tested and the preset standard size feature value, and if the error exceeds the threshold T, it is judged to be an abnormal banknote.
  • the resolution of the banknote image is 100dpi.
  • the deviation of the maximum and minimum values of the dimension values of each dimension is calculated.
  • the local feature detection method based on the scheme is proposed.
  • the present invention also provides a preferred banknote authenticity detecting system, which comprises the following modules:
  • a banknote image acquisition module 1 is configured to collect an image of a banknote, and rotate and crop the image according to the measured vertices and angle information of the banknote to find a valid banknote image area;
  • a feature point positioning module 2 is configured to perform key feature point positioning on the banknote image to be tested, and calculate a size feature value of the banknote image to be tested according to the relative coordinates of the key feature point positioning;
  • a banknote authenticity judging module 3 is configured to compare the feature value of the banknote image to be tested with the approximation degree of the feature value corresponding to the preset standard banknote, and if the error exceeds the threshold value, it is determined as an abnormal banknote, and if the error does not exceed the threshold value, then For normal banknotes.
  • the feature point location module specifically includes:
  • a coarse positioning module 21 performs rough positioning of key feature points on the banknote image to be tested, finds the approximate position of the key feature points according to the face value information and setting information of the banknote, and calculates the thickness of the key feature points by means of projection positioning or integral map positioning. Positioning coordinates
  • a fine positioning module 22 performs fine positioning of key feature points on the banknote image to be measured, expands N pixels on the basis of the above coarse positioning coordinates, and calculates accurate coordinates by using template matching, wherein N is a natural number.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Inspection Of Paper Currency And Valuable Securities (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un procédé et un système de détection de l'authenticité d'un billet de banque. Le procédé de détection de l'authenticité d'un billet de banque comprend les étapes suivantes : étape 1, acquisition et correction d'image : acquérir une image d'un billet de banque, et mettre en rotation et recadrer l'image selon des informations de sommet et des informations d'angle mesurées, pour rechercher une région d'image de billet de banque valide; étape 2, positionnement de points caractéristiques clés : positionner des points caractéristiques clés de l'image du billet de banque à détecter, et calculer les valeurs de caractéristiques de taille de l'image du billet de banque à détecter selon les coordonnées relatives des points caractéristiques clés positionnés; 3, détermination d'authenticité : comparer les valeurs de caractéristique de taille de l'image du billet de banque à détecter avec des valeurs de caractéristique de taille standard correspondant à un billet de banque standard prédéfini; si l'erreur entre les valeurs de caractéristique de taille de l'image du billet de banque à détecter et les valeurs de caractéristique de taille standard dépasse un seuil, déterminer que le billet de banque à détecter est un billet de banque anormal; et si l'erreur ne dépasse pas le seuil, déterminer que le billet de banque à détecter est un billet de banque normal. Selon les présents procédé et système de détection de l'authenticité d'un billet de banque, l'erreur globale de la mesure de l'image du billet de banque est localisée au moyen des coordonnées relatives d'une pluralité de points caractéristiques clés, et une pluralité d'erreurs absolues sont relativisées, de façon à atteindre l'objectif de diminution d'erreurs pour améliorer la précision de reconnaissance.
PCT/CN2018/079783 2017-05-24 2018-03-21 Procédé et système de détection de l'authenticité de billets de banque WO2018214627A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710374836.1A CN107123187A (zh) 2017-05-24 2017-05-24 一种钞票真伪检测方法及系统
CN201710374836.1 2017-05-24

Publications (1)

Publication Number Publication Date
WO2018214627A1 true WO2018214627A1 (fr) 2018-11-29

Family

ID=59729533

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/079783 WO2018214627A1 (fr) 2017-05-24 2018-03-21 Procédé et système de détection de l'authenticité de billets de banque

Country Status (2)

Country Link
CN (1) CN107123187A (fr)
WO (1) WO2018214627A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107123187A (zh) * 2017-05-24 2017-09-01 广州广电运通金融电子股份有限公司 一种钞票真伪检测方法及系统
CN113255641A (zh) * 2020-12-31 2021-08-13 深圳怡化电脑股份有限公司 一种图像识别的方法及装置、电子设备、存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101681532A (zh) * 2007-03-30 2010-03-24 德国捷德有限公司 检查有价文件的方法和设备
WO2016037895A1 (fr) * 2014-09-09 2016-03-17 Sicpa Holding Sa Billets de banque présentant des éléments interdépendants
CN105551133A (zh) * 2015-11-16 2016-05-04 新达通科技股份有限公司 一种纸币拼接缝或折痕的识别方法及系统
CN107123187A (zh) * 2017-05-24 2017-09-01 广州广电运通金融电子股份有限公司 一种钞票真伪检测方法及系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100879611B1 (ko) * 2007-05-15 2009-01-21 엘지엔시스(주) 지폐판독장치 및 이의 지폐판독방법
EP2246825B1 (fr) * 2009-04-28 2014-10-08 Banqit AB Procédé pour dispositif de détection de billets de banque et dispositif de détection de billets de banque
CN102722708B (zh) * 2012-05-16 2015-04-15 广州广电运通金融电子股份有限公司 一种薄片类介质分类方法及其装置
CN105989659B (zh) * 2016-04-15 2019-03-22 新达通科技股份有限公司 一种相似字符识别方法及纸币冠字码识别方法
CN106204889B (zh) * 2016-07-27 2019-03-12 深圳怡化电脑股份有限公司 一种纸币荧光区域的定位方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101681532A (zh) * 2007-03-30 2010-03-24 德国捷德有限公司 检查有价文件的方法和设备
WO2016037895A1 (fr) * 2014-09-09 2016-03-17 Sicpa Holding Sa Billets de banque présentant des éléments interdépendants
CN105551133A (zh) * 2015-11-16 2016-05-04 新达通科技股份有限公司 一种纸币拼接缝或折痕的识别方法及系统
CN107123187A (zh) * 2017-05-24 2017-09-01 广州广电运通金融电子股份有限公司 一种钞票真伪检测方法及系统

Also Published As

Publication number Publication date
CN107123187A (zh) 2017-09-01

Similar Documents

Publication Publication Date Title
CN104101600B (zh) 连铸坯断面微小裂纹检测方法
JP2012525618A (ja) 紙幣検出器デバイスのための方法、および紙幣検出器デバイス
CN110443278B (zh) 一种太阳能电池片栅线粗细异常的检测方法、装置及设备
CN109462999B (zh) 通过数据平衡基于学习的视觉检查方法以及利用其的视觉检查装置
TW201237363A (en) Image processing apparatus and computer program
CN113513991B (zh) 一种电池极片毛刺高度检测方法及装置
WO2018214627A1 (fr) Procédé et système de détection de l'authenticité de billets de banque
CN106875540B (zh) 一种纸币厚度异常检测方法及装置
WO2019001191A1 (fr) Procédé et dispositif de détection de billet de banque collé
CN110084807B (zh) 一种焊缝探伤底片图像造假的检测方法
CN106204616B (zh) 一种伊朗纸币币值的识别方法及装置
US11913912B2 (en) Hole expansion ratio testing device, hole expansion ratio testing method, and operating program
CN106296975B (zh) 一种美元纸币面值的识别方法及装置
CA2902244C (fr) Dispositif et procede de controle de documents de valeur, en particulier de billets de banque, ainsi que systeme de traitement de documents de valeur
CN112837285B (zh) 一种板面图像的边缘检测方法及装置
CN111046876B (zh) 基于纹理检测技术的车牌字符快速识别方法及其系统
CN107991322A (zh) 智能识别电力设备内部部件缺陷的方法及装置
CN108573563A (zh) 一种纸币磁信号检测校准方法
CN114240984A (zh) 一种圆形标志点边缘提取方法及其应用
CN108961530B (zh) 一种纸币缺陷识别方法以及系统
CN108492446B (zh) 一种纸币边缘查找方法以及系统
CN106651960B (zh) 一种图像传感器的故障检测方法及装置
Lam et al. Reference point detection for arch type fingerprints
CN108956617A (zh) 一种基于微形变智能分类器的电子器件焊点热循环失效的检测方法
JP2001175911A (ja) 画像イメージによる硬貨真偽判定方法及び装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18805359

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18805359

Country of ref document: EP

Kind code of ref document: A1