WO2018130119A1 - 一种基于双冠字号的纸币鉴伪方法和系统 - Google Patents

一种基于双冠字号的纸币鉴伪方法和系统 Download PDF

Info

Publication number
WO2018130119A1
WO2018130119A1 PCT/CN2018/071569 CN2018071569W WO2018130119A1 WO 2018130119 A1 WO2018130119 A1 WO 2018130119A1 CN 2018071569 W CN2018071569 W CN 2018071569W WO 2018130119 A1 WO2018130119 A1 WO 2018130119A1
Authority
WO
WIPO (PCT)
Prior art keywords
crown
size
feature
projection
area
Prior art date
Application number
PCT/CN2018/071569
Other languages
English (en)
French (fr)
Inventor
陈盛福
杨伟明
陈健
张恩富
Original Assignee
广州广电运通金融电子股份有限公司
广州中智融通金融科技有限公司
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, 广州中智融通金融科技有限公司 filed Critical 广州广电运通金融电子股份有限公司
Publication of WO2018130119A1 publication Critical patent/WO2018130119A1/zh

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
    • G07D7/004—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 using digital security elements, e.g. information coded on a magnetic thread or strip
    • G07D7/0047—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 using digital security elements, e.g. information coded on a magnetic thread or strip using checkcodes, e.g. coded numbers derived from serial number and denomination

Definitions

  • the invention relates to a banknote authentication technology, in particular to a banknote authentication method for self-checking by a double crown type unique to a banknote, and a system for implementing the method.
  • money plays an irreplaceable role in the commodity economy.
  • Money plays a role in regulating, promoting and stabilizing the national economy.
  • counterfeit banknotes There are two types of counterfeit banknotes: counterfeit coins and altered coins.
  • counterfeit coins because of their material, ink, and printing process, the characteristics of the real coin are quite different. Therefore, the current financial instruments can be effectively identified, and the coin is removed by cutting the front or back of the ticket.
  • Banknotes made by illegal means such as cut-and-spread, altered denominations, etc., commonly seen in the market, "changing coins” mainly have a piece of coin, splitting a real coin into two, and at the same time half-true and half-false by patching.
  • banknotes in many countries have double-crowned font sizes.
  • the 100 yuan of the 2015 edition and the 100 yuan and 50 yuan of the 1999 edition have double crowns, namely the horizontal crown and the vertical crown, and the two crowns.
  • the crown size is the same. It is better to detect the coinage by comparing the counterfeit technology to compare the left and right crown numbers.
  • the existing crown number extraction technology obtains the feature number of the banknote by extracting the image of the crown number of the banknote, and obtains the crown number of the banknote. If the double crown is used for the counterfeiting, the OCR (Optical Character Recognition) extraction of the crown number is required separately. Furthermore, the matching of the characters is the same. Therefore, the prior art has a technical problem that the efficiency is low and the calculation amount is large, and the crown number cannot be recognized after being uniformly changed.
  • the present invention provides a method for confirming whether the horizontal and vertical crown numbers are consistent by extracting and comparing the character image features of the horizontal crown size and the vertical crown size. Whether it is a coin-changing coin, this method does not require the OCR extraction comparison of characters to effectively solve the problem that the horizontal and vertical crown size is unrecognizable and can not be recognized, and at the same time, the algorithm processing efficiency is greatly improved under the premise of ensuring the accuracy of the quasi-authentication.
  • the method for authenticating banknotes based on double-crown fonts specifically includes the following processing steps:
  • Step 1 collecting banknote image information, respectively extracting a first crown size area and a second crown size area;
  • Step 2 Normalize the size of the first crown size area and the second crown size area respectively, and perform binarization processing to form a first crown size binarized image and a second crown size binarized image;
  • Step 3 Perform the same at least 2n equal division on the first crown number binary image and the second crown size binary image respectively, where n is a natural number, and calculate the proportion of black points in each divided area, and then Absolute difference statistics are performed on the ratio of the black points of each corresponding area of the first crown size and the second crown size, and all the differences are summed, and the summation value is recorded as the feature T1;
  • Step 4 Perform black point projection in the X direction and the Y direction on the binarized image of the first crown size and the binarized image of the second crown, respectively, and then calculate the projection energy in the X direction and the Y direction, respectively, and the first crown number
  • the projection energy in the X direction and the projection energy corresponding to the second crown number in the Y direction are absolute differences, which are recorded as the feature T2; the projection energy in the Y direction of the first crown is equal to the projection energy in the X direction corresponding to the second crown number. Value, recorded as feature T3;
  • Step 5 Perform coupling degree statistics on the feature T1, the feature T2, and the feature T3, and determine whether the crown number A and the crown number B are consistent according to the coupling degree value, to obtain whether the corresponding banknote is a counterfeit result of the altered coin.
  • step 1 collecting banknote image information, respectively extracting the first crown size area and the second crown number area, specifically:
  • Step 1.1 collecting banknote image information, respectively extracting a first crown size area and a second crown size area;
  • the first crown size area and the second crown size area are respectively normalized and binarized to form a first crown size binary image and a second crown number binary image.
  • step 3 performing the same at least 2n equal division on the first crown number binarized image and the second crown size binarized image, respectively, n is a natural number, and calculating the black dot occupied by each divided region.
  • the ratio is obtained by performing absolute difference statistics on the ratio of the black points of each corresponding region of the first crown size and the second crown size, and summing all the differences, and the summation value is recorded as the feature T1, specifically:
  • the absolute value difference and CharBlackRatioSameLevel(i) form a feature T1.
  • the black point projection in the X direction and the Y direction is performed on the binarized image of the first crown size and the binarized image of the second crown, respectively, and then the projection energies in the X direction and the Y direction are respectively calculated, and
  • the projection energy of the first crown number in the X direction and the projection energy corresponding to the second crown number in the Y direction are absolute differences, which are recorded as the feature T2; and the projection energy of the first crown size Y direction corresponds to the projection of the second crown size in the X direction
  • the absolute difference of energy is recorded as characteristic T3, specifically:
  • step 5 performing the coupling degree statistics on the feature T1, the feature T2 and the feature T3, and determining whether the crown number A and the crown number B are consistent according to the coupling degree value, to obtain whether the corresponding banknote is a counterfeit result of the altered coin, Specifically:
  • Step 5.1 Perform coupling statistics according to the following formula:
  • Step 5.2 determining the coupling degree value according to the following conditions
  • T is the statistical threshold
  • Step 5.3 finally, whether the banknote is a counterfeit result of the altered coin.
  • Another object of the present invention is to provide a banknote authentication system based on a double-crowned font, which specifically includes:
  • An image acquisition unit is configured to collect banknote image information, and respectively extract a first crown size area and a second crown size area;
  • An image processing unit is configured to normalize the size of the first crown area and the second crown area, and perform binarization processing to form a first crown number binarized image and a second crown number binarized image;
  • An image segmentation and blackpoint statistical unit is configured to perform the same at least 2n equal division on the first crown size binarized image and the second crown size binarized image, where n is a natural number, and each segment is calculated
  • the proportion of the black points, the black point ratio of each corresponding area of the first crown size and the second crown size is sequentially subjected to absolute difference statistics, and all the differences are summed, and the summation value is recorded as the feature T1;
  • An image black point projection and energy difference calculation unit is configured to perform black point projection in the X direction and the Y direction on the first crown type binarized image and the second crown type binarized image, and then calculate the X direction and the Y direction respectively.
  • the projection energy of the direction is the absolute difference between the projection energy of the first crown size X direction and the projection energy of the second crown size corresponding to the Y direction, and is recorded as the feature T2; the projection energy of the first crown size Y direction and the second crown
  • the font size corresponds to the projection energy in the X direction to make an absolute difference, which is recorded as feature T3;
  • a feature coupling degree judging unit is configured to perform the coupling degree statistics on the feature T1, the feature T2 and the feature T3, and determine whether the crown number A and the crown number B are consistent according to the coupling degree value, so as to obtain whether the corresponding banknote is a counterfeit coin result.
  • the image collecting unit further includes a character area extracting unit, configured to perform character area extraction on the first crown size area and the second crown size area.
  • the binarized image and the second crown image binarized image are respectively projected in the X direction and the Y direction, and the projection energy is statistically calculated.
  • the second judgment dimension and the third are formed by combining the black point proportion similarity and the projection energy similarity. Judging the dimension, and finally determining the final coupling degree according to the first judgment dimension, the second judgment dimension, and the third judgment dimension, thereby determining whether the first crown number and the second crown number are consistent.
  • the method has high accuracy, simple algorithm implementation and less calculation, which is more suitable for embedded financial instruments, and can effectively improve the counterfeiting performance of financial instruments.
  • FIG. 1 is a schematic diagram of the composition of a banknote counterfeiting system based on double crown font features provided by the present invention
  • FIG. 2 is a schematic diagram showing the composition of an optimized functional unit of the banknote counterfeiting system shown in FIG. 1;
  • FIG. 3 is a process flow diagram of a banknote authentication method based on a double crown feature provided by the present invention.
  • the present invention provides a banknote counterfeiting system based on a double crown feature, which specifically includes an image capturing unit 1 for collecting banknote image information and extracting a first crown size area and a second crown number respectively. region;
  • An image processing unit 2 is configured to normalize the size of the first crown area and the second crown area, and perform binarization processing to form a first crown number binary image and a second crown number binary image ;
  • An image segmentation and black point statistic unit 3 is configured to perform the same at least 2n equal division on the first crown number binarized image and the second crown size binarized image, where n is a natural number, ensuring accuracy and efficiency.
  • n is a natural number
  • the black point ratio of each corresponding area of the first crown size and the second crown size is subjected to absolute difference statistics, and all the differences are summed, and the summation value is recorded as feature T1;
  • An image black point projection and energy difference calculating unit 4 is configured to perform black point projection in the X direction and the Y direction on the first crown type binarized image and the second crown type binarized image, and then calculate the X direction and
  • the projection energy in the Y direction is the absolute difference between the projection energy of the first crown size X direction and the projection energy of the second crown size corresponding to the Y direction, and is recorded as the feature T2; the projection energy of the first crown size Y direction and the second The crown size corresponds to the projection energy in the X direction to make an absolute difference, which is recorded as feature T3;
  • a characteristic coupling degree judging unit 5 is configured to perform the coupling degree statistics on the feature T1, the feature T2 and the feature T3, and determine whether the crown number A and the crown number B are consistent according to the coupling degree value, so as to obtain whether the corresponding banknote is a coin for alteration. Pseudo results.
  • the image capturing unit further includes a character region extracting unit 11 configured to perform character region extraction on the first crown region and the second crown region.
  • the image processing unit 2 is configured to normalize the size of the character region of the first crown region and the character region of the second crown region, and perform binarization processing to form a character region of the first crown number. Binarized image of the binarized image and the character region of the second crown size;
  • the ratio of the black points of the area in turn, the absolute difference statistics are performed on the ratio of the black points of each corresponding area of the character area of the first crown size and the character area of the second crown size, and all the differences are summed, And the value is recorded as feature T1;
  • the image black dot projection and energy difference calculation unit 4 is configured to perform black point projection in the X direction and the Y direction on the character area binarized image of the first crown size and the character area binarized image of the second crown size And then calculating the projection energy in the X direction and the Y direction, respectively, and the absolute difference between the projection energy of the character region X direction of the first crown size and the projection energy of the second crown size corresponding to the Y direction is recorded as the feature T2;
  • the projection energy of the character region Y direction of the first crown size is compared with the projection energy of the second crown symbol corresponding to the projection energy in the X direction, and is denoted as feature T3;
  • the characteristic coupling degree judging unit 5 is configured to perform the coupling degree statistics on the feature T1, the feature T2 and the feature T3, and determine whether the crown number A and the crown number B are consistent according to the coupling degree value, so as to obtain whether the corresponding banknote is a coin. Counterfeit results.
  • the above system obtains the effect of detecting the coinage by extracting the features of the transverse crown size and the longitudinal crown size on the banknote, and then comparing the lateral feature with the longitudinal feature to confirm whether the transverse crown size and the longitudinal crown size are consistent.
  • the method has high accuracy and convenient method, and can effectively improve the counterfeiting performance of financial instruments.
  • a method for authenticating a banknote based on a double crown feature is provided by the implementation of the present invention, which specifically includes the following processing steps:
  • S1 collecting banknote image information, respectively extracting the first crown size area and the second crown size area; taking RMB as an example: respectively extracting the horizontal crown size area and the vertical crown size area on the banknote, respectively, as the crown number A and the crown number B;
  • S2 Normalize the size of the first crown size area and the second crown size area respectively, and perform binarization processing to form a first crown size binarized image and a second crown size binarized image; taking RMB as an example : N characters of the crown number A and the crown number B are respectively cut and extracted, where N is equal to 10, and the characters of the crown number A and the corresponding characters of the crown size B are normalized and then binarized;
  • the operation feature T1, the feature T2 and the feature T3 may be processed one by one or in multiple parallel processes. Due to the limitation of the text description, the functions can only be expressed sequentially, but in the technical implementation, the computer can select one by one according to the performance of the computer. It can also be processed in multiple parallels.
  • the method for authenticating the banknote based on the double crown feature can also be implemented by the following steps:
  • S1.1 collecting banknote image information, respectively extracting a first crown size area and a second crown size area;
  • the absolute value difference and CharBlackRatioSameLevel(i) form a feature T1.
  • T is the statistical threshold
  • the banknote image signal is first acquired; then the image preprocessing is performed, including bilinear interpolation and binarization; then the character is equally divided into four regions, and the black dot proportion is separately counted, and the horizontal and vertical black are performed.
  • the point ratio is compared with the similarity degree; then the X and Y directions are respectively projected horizontally and vertically, and the projection energy is calculated; finally, combined with the black point ratio similarity and the projection energy similarity, the final coupling degree is judged, and the horizontal and vertical directions are judged. Whether the characters are consistent.
  • the method has high accuracy, simple algorithm implementation and less calculation, which is more suitable for embedded financial instruments, and can effectively improve the counterfeiting performance of financial instruments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Inspection Of Paper Currency And Valuable Securities (AREA)
  • Character Discrimination (AREA)
  • Image Analysis (AREA)

Abstract

一种纸币鉴伪技术，尤其是一种通过纸币特有的双冠字号进行自我校验的纸币鉴伪方法以及实现该方法的系统，首先获取纸币图像信号；然后进行图像预处理，包括双线性插值和二值化；然后将字符等分成四个区域，分别统计黑点占比，并进行横向和纵向黑点占比相似度比较；然后横向和纵向分别进行X方向和Y方向投影，并统计投影能量；最后，结合黑点占比相似度和投影能量相似度，进行最终耦合度判断，判断横向和纵向字符是否一致来高效地进行伪造币的鉴别。该方法准确度较高，算法实施简洁，计算量较少，比较适合嵌入式金融机具，并且可以有效提升金融机具鉴伪性能。

Description

一种基于双冠字号的纸币鉴伪方法和系统
本申请要求于2017年01月13日提交中国专利局、申请号为201710026017.8、发明名称为“一种基于双冠字号的纸币鉴伪方法和系统”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及一种纸币鉴伪技术，尤其是一种通过纸币特有的双冠字号进行自我校验的纸币鉴伪方法以及实现该方法的系统。
背景技术
货币作为价值符号和价值交换媒介，在商品经济中充当着不可替代的重要角色，货币发挥着调控、促进和稳定国民经济的作用。
近年来，中国经济持续高速健康发展，社会财富得到快速的积累，制贩假币的现象也随之迅速蔓延。假币的流通影响了市场规律，扰乱社会、经济秩序，损害一个国家货币的信誉和社会公众的利益。因此，世界各国无不将假币列为收缴和打击的对象，反假货币已成为一个世界性的话题。
假钞的种类分为两种：伪造币、变造币。对于伪造币由于其材质、油墨以及印刷制作工艺与真币的特性具有较大的差别，因此目前的金融机具基本可以有效识别，而变造币则是采用将票券正面或背面揭开、剪割拼凑、涂改面额等非法手段所变造的纸币，市面上常见的“变造币”主要有拼凑币，将一张真币一分为二，同时通过粘补将一半真的和一半假的拼凑在一起；将小面额真币进行裁剪，将真币上的变色荧光数字挖剪，再粘补在大面额假币同一位置上，以便在金融机具上蒙混过关，因此金融机具如何有效对变造币的鉴伪是金融机具行业需要解决的一大技术难题。
目前，很多国家的纸币具有双冠字号，以人民币为例，人民币2015版的100元以及1999版的100元和50元具有双冠字号，即横向冠字号和纵向冠字号，两个冠字号区域的冠字号一致。通过鉴伪技术对比左右冠字号是否一致，可以较好地检测出变造币。
而现有冠字号提取技术通过获取纸币冠字号图像，进行特征提取，获得纸币冠字号，如果用常规的方式进行双冠字号鉴伪，需要分别进行两次冠字 号的OCR(Optical Character Recognition)提取再进行字符是否相同的比对，因此现有技术存在效率低运算量大、冠字号统一串改后无法识别的技术问题。
发明内容
本发明为了解决现有技术运算量大而导致的处理速度慢的技术问题，提供一种通过提取和比对横向冠字号和纵向冠字号的字符图像特征，确认横纵冠字号是否一致，从而确认是否为变造币，该方法不需要字符的OCR提取对比可以有效解决横纵冠字号统一串改而无法识别的问题，同时在保证准鉴伪确性的前提下使算法处理效率大大提高。
这种基于双冠字号的纸币鉴伪方法，具体包括如下处理步骤：
步骤1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
步骤2：分别对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；
步骤3：分别对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
步骤4：分别对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；
步骤5：对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
优选的，上述步骤1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域，具体为：
步骤1.1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
步骤1.2：分别对第一冠字号区域和第二冠字号区域进行字符区域提取，分别标记为HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)，其中N表示冠字号包含字符的个数。
进一步的，上述步骤2：分别对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像，具体为：
步骤2.1：分别对所述字符区域HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)进行大小归一化，形成字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)，其中N表示冠字号包含字符的个数。
步骤2.2：分别对字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)进行二值化处理形成字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)；
进一步的，上述步骤3：分别对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1，具体为：
步骤3.1：分别对字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)分别进行相同的至少2n等分分块，标记为H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)，其中j表示分割后的块数，n为自然数；
步骤3.2：分别统计H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分块中黑点所占比例值，分别标记为H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)，其中j表示分割后的块数，n为自然数；
步骤3.3：对比例值H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)进行耦合度对比，具体为：
Figure PCTCN2018071569-appb-000001
该绝对值差值和CharBlackRatioSameLevel(i)形成特征T1。
进一步的，上述步骤4：分别对第一冠字号二值化图像和第二冠字号二 值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3，具体为：
步骤4.1：对所述分块H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分别进行X方向和Y方向黑点投影，第一冠字号X方向和Y方向投影分别形成投影标记为H3CharShadowX(i)(i＝1...2n)和H3CharShadowY(i)(i＝1...2n)，第二冠字号X方向和Y方向投影分别标记为V3CharShadowX(i)(i＝1...2n)和V3CharShadowY(i)(i＝1...2n)；
步骤4.2：根据ShadowPower＝∑k*f(k)，其中k表示信号点位置，f(k)表示信号点幅值，获得第一冠字号X方向投影H3CharShadowX(i)(i＝1...2n)和Y方向投影H3CharShadowY(i)(i＝1...2n)的投影能量H3CharShadowXpower(i)(i＝1...2n)，H3CharShadowYpower(i)(i＝1...2n)；第二冠字号X方向投影V3CharShadowX(i)(i＝1...2n)和Y方向投影V3CharShadowY(i)(i＝1...2n)的投影能量分别标记为V3CharShadowXpower(i)(i＝1...2n)，V3CharShadowYpower(i)(i＝1...2n)；
步骤4.3：将第一冠字号X方向的投影能量H3CharShadowXpower(i)(i＝1...2n)与第二冠字号对应Y方向的投影能量V3CharShadowYpower(i)(i＝1...2n)进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量H3CharShadowYpower(i)(i＝1...2n)与第二冠字号对应X方向的投影能量V3CharShadowXpower(i)(i＝1...2n)进行绝对差值，记为特征T3。
进一步的，上述步骤5：对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果，具体为：
步骤5.1：根据下式进行耦合度统计：
FinalSameLevel(i)＝coef1*T1+coef1*T2+coef1*T3
其中，coef1+coef2+coef3＝1；
步骤5.2，根据如下条件判断上述耦合度值，
Figure PCTCN2018071569-appb-000002
其中，T是统计阈值；
当FinalJudge(i)为1表示第一冠字号与第二冠字号不同，为0表示两个冠字号相同；
步骤5.3，最后给出纸币是否为变造币的鉴伪结果。
本发明的另一目的是提供一种基于双冠字号的纸币鉴伪系统，其具体包括：
一图像采集单元，用于采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
一图像处理单元，用于对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；
一图像分块及黑点统计单元，用于对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
一图像黑点投影及能量差值计算单元，用于对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；
一特征耦合度判断单元，用于对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
优选的，所述图像采集单元还包括一字符区域提取单元，用于对第一冠字号区域和第二冠字号区域进行字符区域提取。
该技术方案与现有技术相比具有如下有益效果：
通过对第一冠字号二值化图像和第二冠字号二值化图像分别统计黑点占比，并进行横向和纵向黑点占比相似度比较，形成第一判断维度；然后第一冠字号二值化图像和第二冠字号二值化图像分别进行X方向和Y方向投影，并统计投影能量；最后，结合黑点占比相似度和投影能量相似度，形成第二 判断维度和第三判断维度，最后根据第一判断维度、第二判断维度和第三判断维度进行最终耦合度判断，从而判断第一冠字号和第二冠字号是否一致。该方法准确度较高，算法实施简单，计算量较少，比较适合嵌入式金融机具，并且可以有效提升金融机具鉴伪性能。
附图说明
图1是本发明提供的一种基于双冠字号特征的纸币鉴伪系统的组成示意图；
图2是图1所示纸币鉴伪系统的优化功能单元组成示意图；
图3是本发明提供的一种基于双冠字号特征的纸币鉴伪方法的处理流程图。
具体实施方式
下面所描述的实施例仅仅是本发明所提供的基于双冠字号的纸币鉴伪方法和系统的一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
参阅图1，本发明提供的一种基于双冠字号特征的纸币鉴伪系统，其具体包括：一图像采集单元1，用于采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
一图像处理单元2，用于对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；
一图像分块及黑点统计单元3，用于对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，在保证精度和效率的前提下，一般n＝2即可以满足要求，当然n取值越大，鉴伪的准确性就越高，但是运算量增加，降低了处理效率，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
一图像黑点投影及能量差值计算单元4，用于对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X 方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；
一特征耦合度判断单元5，用于对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
参阅图2所示，优选的，所述图像采集单元还包括一字符区域提取单元11，用于对第一冠字号区域和第二冠字号区域进行字符区域提取。相应的，所述图像处理单元2，用于对第一冠字号区域的字符区域和第二冠字号区域的字符区域进行大小归一化，并进行二值化处理形成第一冠字号的字符区域二值化图像和第二冠字号的字符区域二值化图像；
所述图像分块及黑点统计单元3，用于对第一冠字号的字符区域二值化图像和第二冠字号的字符区域二值化图像进行相同的至少2n等分分区，n为自然数，在保证精度和效率的前提下，一般n＝2即可以满足要求，当然n取值越大，鉴伪的准确性就越高，但是运算量增加，降低了处理效率，并计算每个分割区域的黑点所占比例，依次将第一冠字号的字符区域和第二冠字号的字符区域每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
所述图像黑点投影及能量差值计算单元4，用于对第一冠字号的字符区域二值化图像和第二冠字号的字符区域二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号的字符区域X方向的投影能量与第二冠字号的字符区域对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号的字符区域Y方向的投影能量与第二冠字号的字符区域对应X方向的投影能量进行绝对差值，记为特征T3；
所述特征耦合度判断单元5，用于对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
上述系统通过提取纸币上横向冠字号和纵向冠字号的特征，然后对比横向特征和纵向特征，从而确认横向冠字号和纵向冠字号是否一致，从而达到 检测变造币的效果。该方法准确性较高，方法便捷，能有效提高金融机具鉴伪性能。
参阅图3所示，本发明实施提供的一种基于双冠字号特征的纸币鉴伪方法，具体包括如下处理步骤：
S1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；以人民币为例：分别提取出纸币上横向冠字号区域和纵向冠字号区域，分别记为冠字号A和冠字号B；
S2：分别对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；以人民币为例：将冠字号A和冠字号B的N个字符分别切割提取出来，这里的N等于10，并将冠字号A的字符与冠字号B的对应字符进行大小归一化，然后进行二值化；
S3：分别对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，在保证精度和效率的前提下，一般n＝2即可以满足要求，当然n取值越大，鉴伪的准确性就越高，但是运算量增加，降低了处理效率，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；以人民币为例：将冠字号A的每个字符图像等分成4个区域，并计算出每个分割区域的黑点占整个字符的比例，将冠字号B按照同样方法统计黑点占比；然后依次将冠字号A和冠字号B各对应字符4个区域的黑点比例进行绝对差值统计，并将4个差值求和，求和值记为特征T1；
S4：分别对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；以人民币为例：将冠字号A的每个字符进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，冠字号B按同样方法计算出X方向和Y方向的投影能量；将冠字号A 每个字符X方向的投影能量与冠字号B对应字符Y方向的投影能量进行绝对差值，记为特征T2；将冠字号A每个字符Y方向的投影能量与冠字号B对应字符X方向的投影能量进行绝对差值，记为特征T3；
S5：对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果，以人民币为例：对特征1、特征2和特征3进行耦合度判断，判断冠字号A和冠字号B的各对应字符是否相同，从而判断冠字号A与冠字号B是否一致。
对于上述方法的处理步骤运算特征T1、特征T2和特征T3可以逐一进行也可以多个并行处理，由于文字描述的局限性，只能顺序表达，但是在技术实现上计算机根据运算性能可以选择逐一进行也可以多个并行处理。
优选的所述该基于双冠字号特征的纸币鉴伪方法，还可以通过如下步骤实现：
S1.1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
S1.2：分别对第一冠字号区域和第二冠字号区域进行字符区域提取，分别标记为HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)，其中N表示冠字号包含字符的个数；以人民币为例：N为10。
S2.1：分别对所述字符区域HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)进行大小归一化，形成字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)，其中N表示冠字号包含字符的个数。
S2.2：分别对字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)进行二值化处理形成字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)；
S3.1：分别对字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)分别进行相同的至少2n等分分块，n为自然数，本实施例中n＝2即可用满足精度与效率的要求，即每个字符二值化图像进行4等分分块，假定H3Char(i)和V3Char(i)的图像尺寸为R*S，其中i表示第i个字符，R和S分别表示字符高度和字符宽度，将H3Char(i)和V3Char(i)分别等份分割成R/2*S/2的4个小块，标记为H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)，其中j表示分割后的块数，n＝2；
S3.2：分别统计H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分块中黑点所占比例值，分别标记为H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)，其中j表示分割后的块数，n为自然数；
S3.3：对比例值H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)进行耦合度对比，具体为：
Figure PCTCN2018071569-appb-000003
该绝对值差值和CharBlackRatioSameLevel(i)形成特征T1。
S4.1：对所述每个字符分块H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分别进行X方向和Y方向黑点投影，第一冠字号X方向和Y方向投影分别形成投影标记为H3CharShadowX(i)(i＝1...2n)和H3CharShadowY(i)(i＝1...2n)，第二冠字号X方向和Y方向投影分别标记为V3CharShadowX(i)(i＝1...2n)和V3CharShadowY(i)(i＝1...2n)；
S4.2：根据ShadowPower＝∑k*f(k)，其中k表示信号点位置，f(k)表示信号点幅值，获得第一冠字号X方向投影H3CharShadowX(i)(i＝1...2n)和Y方向投影H3CharShadowY(i)(i＝1...2n)的投影能量H3CharShadowXpower(i)(i＝1...2n)，H3CharShadowYpower(i)(i＝1...2n)；第二冠字号X方向投影V3CharShadowX(i)(i＝1...2n)和Y方向投影V3CharShadowY(i)(i＝1...2n)的投影能量分别标记为V3CharShadowXpower(i)(i＝1...2n)，V3CharShadowYpower(i)(i＝1...2n)；
S4.3：将第一冠字号X方向的投影能量H3CharShadowXpower(i)(i＝1...2n)与第二冠字号对应Y方向的投影能量V3CharShadowYpower(i)(i＝1...2n)进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量H3CharShadowYpower(i)(i＝1...2n)与第二冠字号对应X方向的投影能量V3CharShadowXpower(i)(i＝1...2n)进行绝对差值，记为特征T3。
S5.1：根据下式进行耦合度统计：
FinalSameLevel(i)＝coef1*T1+coef1*T2+coef1*T3
其中，coef1+coef2+coef3＝1；
S5.2，根据如下条件判断上述耦合度值，
Figure PCTCN2018071569-appb-000004
其中，T是统计阈值；
当FinalJudge(i)为1表示第一冠字号与第二冠字号不同，为0表示两个冠字号相同；
S5.3，最后给出纸币是否为变造币的鉴伪结果。
本发明实施例中，首先获取纸币图像信号；然后进行图像预处理，包括双线性插值和二值化；然后将字符等分成四个区域，分别统计黑点占比，并进行横向和纵向黑点占比相似度比较；然后横向和纵向分别进行X方向和Y方向投影，并统计投影能量；最后，结合黑点占比相似度和投影能量相似度，进行最终耦合度判断，判断横向和纵向字符是否一致。该方法准确度较高，算法实施简洁，计算量较少，比较适合嵌入式金融机具，并且可以有效提升金融机具鉴伪性能。
以上所揭露的仅为本发明一种较佳实施例而已，当然不能以此来限定本发明之权利范围，因此依本发明申请专利范围所作的等同变化，仍属本发明所涵盖的范围。

Claims (8)

  1. 一种基于双冠字号的纸币鉴伪方法，其包括如下步骤：
    步骤1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
    步骤2：分别对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；
    步骤3：分别对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
    步骤4：分别对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；
    步骤5：对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
  2. 如权利要求1所述的基于双冠字号的纸币鉴伪方法，其特征在于，步骤1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域，具体为：
    步骤1.1：采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
    步骤1.2：分别对第一冠字号区域和第二冠字号区域进行字符区域提取，分别标记为HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)，其中N表示冠字号包含字符的个数。
  3. 如权利要求2所述的基于双冠字号的纸币鉴伪方法，其特征在于，所述步骤2：分别对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像，具体为：
    步骤2.1：分别对所述字符区域HorizontalChar(i)(i＝1...N)和VerticalChar(i)(i＝1...N)进行大小归一化，形成字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)，其中N表示冠字号包含字符的个数。
    步骤2.2：分别对字符归一化图像H2Char(i)(i＝1...N)和V2Char(i)(i＝1...N)进行二值化处理形成字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)；
  4. 如权利要求3所述的基于双冠字号的纸币鉴伪方法，其特征在于，所述步骤3：分别对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1，具体为：
    步骤3.1：分别对字符二值化图像H3Char(i)(i＝1...N)和V3Char(i)(i＝1...N)分别进行相同的至少2n等分分块，标记为H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)，其中j表示分割后的块数，n为自然数；
    步骤3.2：分别统计H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分块中黑点所占比例值，分别标记为H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)，其中j表示分割后的块数，n为自然数；
    步骤3.3：对比例值H3CharBlackPointRatio(i,j)(j＝1...2n)和V3CharBlackPointRatio(i,j)(j＝1...2n)进行耦合度对比，具体为：
    Figure PCTCN2018071569-appb-100001
    该绝对值差值和CharBlackRatioSameLevel(i)形成特征T1。
  5. 如权利要求4所述的基于双冠字号的纸币鉴伪方法，其特征在于，所述步骤4：分别对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3，具体为：
    步骤4.1：对所述每个字符二值化图像的分块H3CharPart(i,j)(j＝1...2n)和V3CharPart(i,j)(j＝1...2n)分别进行X方向和Y方向黑点投影，第一冠字号的每个字符二值化图像X方向和Y方向投影分别形成投影标记为H3CharShadowX(i)(i＝1...2n)和H3CharShadowY(i)(i＝1...2n)，第二冠字号的每个字符二值化图像X方向和Y方向投影分别标记为V3CharShadowX(i)(i＝1...2n)和V3CharShadowY(i)(i＝1...2n)；
    步骤4.2：根据ShadowPower＝∑k*f(k)，其中k表示信号点位置，f(k)表示信号点幅值，获得第一冠字号的每个字符二值化图像X方向投影H3CharShadowX(i)(i＝1...2n)和Y方向投影H3CharShadowY(i)(i＝1...2n)的投影能量H3CharShadowXpower(i)(i＝1...2n)，H3CharShadowYpower(i)(i＝1...2n)；第二冠字号的每个字符二值化图像X方向投影V3CharShadowX(i)(i＝1...2n)和Y方向投影V3CharShadowY(i)(i＝1...2n)的投影能量分别标记为V3CharShadowXpower(i)(i＝1...2n)，V3CharShadowYpower(i)(i＝1...2n)；
    步骤4.3：将第一冠字号的每个字符二值化图像X方向的投影能量H3CharShadowXpower(i)(i＝1...2n)与第二冠字号的每个字符二值化图像对应Y方向的投影能量V3CharShadowYpower(i)(i＝1...2n)进行绝对差值，记为特征T2；将第一冠字号的每个字符二值化图像Y方向的投影能量H3CharShadowYpower(i)(i＝1...2n)与第二冠字号的每个字符二值化图像对应X方向的投影能量V3CharShadowXpower(i)(i＝1...2n)进行绝对差值，记为特征T3。
  6. 如权利要求5所述的基于双冠字号的纸币鉴伪方法，其特征在于，所述步骤5：对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果，具体为：
    步骤5.1：根据下式进行耦合度统计：
    FinalSameLevel(i)＝coef1*T1+coef1*T2+coef1*T3
    其中，coef1+coef2+coef3＝1；
    步骤5.2，根据如下条件判断上述耦合度值，
    Figure PCTCN2018071569-appb-100002
    其中，T是统计阈值；
    当FinalJudge(i)为1表示第一冠字号与第二冠字号不同，为0表示两个冠 字号相同；
    步骤5.3，最后给出纸币是否为变造币的鉴伪结果。
  7. 一种基于双冠字号的纸币鉴伪系统，其包括：
    一图像采集单元，用于采集纸币图像信息，分别提取第一冠字号区域和第二冠字号区域；
    一图像处理单元，用于对第一冠字号区域和第二冠字号区域进行大小归一化，并进行二值化处理形成第一冠字号二值化图像和第二冠字号二值化图像；
    一图像分块及黑点统计单元，用于对第一冠字号二值化图像和第二冠字号二值化图像进行相同的至少2n等分分区，n为自然数，并计算每个分割区域的黑点所占比例，依次将第一冠字号和第二冠字号每个对应区域的黑点比例进行绝对差值统计，并将所有差值进行求和，其求和值记为特征T1；
    一图像黑点投影及能量差值计算单元，用于对第一冠字号二值化图像和第二冠字号二值化图像进行X方向和Y方向的黑点投影，然后分别计算X方向和Y方向的投影能量，将第一冠字号X方向的投影能量与第二冠字号对应Y方向的投影能量进行绝对差值，记为特征T2；将第一冠字号Y方向的投影能量与第二冠字号对应X方向的投影能量进行绝对差值，记为特征T3；
    一特征耦合度判断单元，用于对特征T1、特征T2和特征T3进行耦合度统计，根据耦合度值判断冠字号A与冠字号B是否一致，以获得对应纸币是否为变造币的鉴伪结果。
  8. 如权利要求7所述的基于双冠字号的纸币鉴伪系统，其特征在于，所述图像采集单元还包括一字符区域提取单元，用于对第一冠字号区域和第二冠字号区域进行字符区域提取。
PCT/CN2018/071569 2017-01-13 2018-01-05 一种基于双冠字号的纸币鉴伪方法和系统 WO2018130119A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710026017.8A CN106780953B (zh) 2017-01-13 2017-01-13 一种基于双冠字号的纸币鉴伪方法和系统
CN201710026017.8 2017-01-13

Publications (1)

Publication Number Publication Date
WO2018130119A1 true WO2018130119A1 (zh) 2018-07-19

Family

ID=58945406

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/071569 WO2018130119A1 (zh) 2017-01-13 2018-01-05 一种基于双冠字号的纸币鉴伪方法和系统

Country Status (2)

Country Link
CN (1) CN106780953B (zh)
WO (1) WO2018130119A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106780953B (zh) * 2017-01-13 2020-06-02 广州广电运通金融电子股份有限公司 一种基于双冠字号的纸币鉴伪方法和系统
CN111046866B (zh) * 2019-12-13 2023-04-18 哈尔滨工程大学 一种结合ctpn和svm的人民币冠字号区域检测方法
CN111445620B (zh) * 2020-03-09 2022-05-13 广州中智融通金融科技有限公司 一种库房现金存储方法以及装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010117803A (ja) * 2008-11-12 2010-05-27 Hitachi Omron Terminal Solutions Corp 紙葉類処理装置及び紙葉類処理プログラム
CN101809631A (zh) * 2007-09-27 2010-08-18 光荣株式会社 纸张处理装置
CN103136845A (zh) * 2013-01-23 2013-06-05 浙江大学 一种基于冠字号图像特征的人民币鉴伪方法
CN104916029A (zh) * 2014-03-13 2015-09-16 广州南沙资讯科技园有限公司博士后科研工作站 一种纸币验证系统及基于该系统的纸币验证方法
CN204856686U (zh) * 2015-08-16 2015-12-09 浙江维融电子科技股份有限公司 一种用于2015版人民币的新型鉴伪系统
CN105335741A (zh) * 2015-10-28 2016-02-17 深圳怡化电脑股份有限公司 一种脏污冠字号分类的方法和系统
CN106780953A (zh) * 2017-01-13 2017-05-31 广州广电运通金融电子股份有限公司 一种基于双冠字号的纸币鉴伪方法和系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102890840B (zh) * 2012-08-22 2016-03-23 山东新北洋信息技术股份有限公司 纸币鉴别方法和装置
CN205670338U (zh) * 2016-06-17 2016-11-02 光荣电子工业（苏州）有限公司 具有拼接币检伪功能的清分机

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101809631A (zh) * 2007-09-27 2010-08-18 光荣株式会社 纸张处理装置
JP2010117803A (ja) * 2008-11-12 2010-05-27 Hitachi Omron Terminal Solutions Corp 紙葉類処理装置及び紙葉類処理プログラム
CN103136845A (zh) * 2013-01-23 2013-06-05 浙江大学 一种基于冠字号图像特征的人民币鉴伪方法
CN104916029A (zh) * 2014-03-13 2015-09-16 广州南沙资讯科技园有限公司博士后科研工作站 一种纸币验证系统及基于该系统的纸币验证方法
CN204856686U (zh) * 2015-08-16 2015-12-09 浙江维融电子科技股份有限公司 一种用于2015版人民币的新型鉴伪系统
CN105335741A (zh) * 2015-10-28 2016-02-17 深圳怡化电脑股份有限公司 一种脏污冠字号分类的方法和系统
CN106780953A (zh) * 2017-01-13 2017-05-31 广州广电运通金融电子股份有限公司 一种基于双冠字号的纸币鉴伪方法和系统

Also Published As

Publication number Publication date
CN106780953A (zh) 2017-05-31
CN106780953B (zh) 2020-06-02

Similar Documents

Publication Publication Date Title
Abburu et al. Currency recognition system using image processing
Mirza et al. Paper currency verification system based on characteristic extraction using image processing
CN103886309A (zh) 人像识别鉴别美元面额的方法
Zeggeye et al. Automatic recognition and counterfeit detection of Ethiopian paper currency
WO2018130119A1 (zh) 一种基于双冠字号的纸币鉴伪方法和系统
EP3417429A1 (en) Method and apparatus for detection of document tampering
Murthy et al. Design and implementation of paper currency recognition with counterfeit detection
Sharan et al. Detection of counterfeit Indian currency note using image processing
Sawant et al. Currency recognition using image processing and minimum distance classifier technique
Viraktamath et al. Review on detection of fake currency using image processing techniques
KR102094234B1 (ko) 지폐 일련번호 인식 장치 및 방법
CN107527423B (zh) 纸币防伪识别方法及纸币识别装置
CN106447904B (zh) 一种纸币的鉴别方法和装置
Atchaya et al. Fake currency detection using Image processing
Lamsal et al. Counterfeit paper banknote identification based on color and texture
Rajan et al. An extensive study on currency recognition system using image processing
CN106296975B (zh) 一种美元纸币面值的识别方法及装置
Ponishjino et al. Bogus currency authorization using HSV techniques
KR101232684B1 (ko) 베이시안 접근법을 이용한 지폐 진위 감별 방법
CN108154596B (zh) 一种基于图像匹配的双冠号纸币鉴伪方法
Sooruth et al. Automatic South African Coin Recognition Through Visual Template Matching
Sarkar et al. Detecting counterfeit currency and identifying its source
CN109410420B (zh) 一种图像检测方法、装置和票据鉴别仪
Shoeb et al. Software system to detect counterfeit Egyptian currency
Ghosh et al. A study on diverse recognition techniques for Indian currency note

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: 18739429

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: 18739429

Country of ref document: EP

Kind code of ref document: A1