WO2015032187A1 - 一种纸币处理方法及装置 - Google Patents

一种纸币处理方法及装置 Download PDF

Info

Publication number
WO2015032187A1
WO2015032187A1 PCT/CN2014/072179 CN2014072179W WO2015032187A1 WO 2015032187 A1 WO2015032187 A1 WO 2015032187A1 CN 2014072179 W CN2014072179 W CN 2014072179W WO 2015032187 A1 WO2015032187 A1 WO 2015032187A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
white light
infrared
double
sided
Prior art date
Application number
PCT/CN2014/072179
Other languages
English (en)
French (fr)
Inventor
梁添才
余元超
王锟
王卫锋
Original Assignee
广州广电运通信息科技有限公司
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 广州广电运通信息科技有限公司 filed Critical 广州广电运通信息科技有限公司
Publication of WO2015032187A1 publication Critical patent/WO2015032187A1/zh

Links

Classifications

    • GPHYSICS
    • G07CHECKING-DEVICES
    • G07DHANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00Testing 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
    • G07D7/06Testing 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 wave or particle radiation
    • G07D7/12Visible light, infrared or ultraviolet radiation
    • GPHYSICS
    • G07CHECKING-DEVICES
    • G07DHANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00Testing 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
    • G07D7/20Testing patterns thereon
    • G07D7/2016Testing patterns thereon using feature extraction, e.g. segmentation, edge detection or Hough-transformation

Definitions

  • the invention relates to the field of banknote processing, and in particular to a banknote processing method and device. Background technique
  • the image of the banknote crown area is often saved only when the banknote image is saved, because the crown number is the "DNA" of the banknote.
  • some devices also have the character recognition function of the crown character. By quickly searching for the banknote number, the "suspicious banknotes" are quickly locked, which improves the efficiency of banknote circulation.
  • the financial self-service devices search and cut the image of the banknote crown area on the white light reflection image of the banknote, thereby saving the image, identifying and recording the banknote crown number.
  • the white light image of the banknote is loud and has a lot of interference, which is not conducive to the effective implementation of the crown area.
  • Positioning, therefore, the financial self-service equipment often has the problem of inaccurate interception of the image of the banknote crown image.
  • the inaccurate image capture of the crown size area makes the banknote tracing function of the financial self-service equipment greatly reduced, which seriously affects the "anti-counterfeiting and anti-counterfeiting" of the financial self-service equipment.
  • Capabilities, critical moments, cannot provide judicial evidence with reference value to the judiciary. Summary of the invention
  • the invention provides a banknote processing method and device, which utilizes an infrared image to be affected by new or old banknotes, a white light reflection map, an infrared reflection image and a transmission image of an optical sensor collecting banknotes, and the banknotes are identified by image feature template matching,
  • the infrared reflection image and the infrared transmission image of the banknote are used to accurately position the crown number of the banknote, and the positioning result is mapped to the white light reflection image to realize the precise positioning of the crown number on the white light reflection image, thereby implementing character cutting and
  • the character recognition can effectively eliminate the influence of the dirty area on the positioning of the crown size area, thereby effectively improving the recognition rate of the crown number.
  • the banknote processing method provided by the invention comprises:
  • Determining the white light image according to position information of the crown area on the white light image The gray scale information of the crown area is obtained and stored.
  • the types of the banknotes include: currency of the banknote, denomination.
  • the method further comprises:
  • the double-sided white light reflection image includes a front white light reflection image and a reverse white light reflection image;
  • the double-sided infrared reflection image includes a front side infrared reflection image and a reverse side infrared reflection image.
  • Steps: selecting the double-sided white light reflection image and the double-sided infrared reflection image according to the facing direction, and obtaining the required white light image and the infrared image include:
  • the white light image and the infrared image that meet the requirements are a front white light image and a front infrared image, respectively;
  • the white light image and the infrared image that meet the requirements are the reverse white light image and the reverse side infrared image, respectively.
  • the step of determining the crown size area according to the type includes:
  • a crown area is determined based on the banknote data.
  • the banknote processing device provided by the present invention comprises:
  • control unit controlling the entire processing flow after the banknotes enter the banknote processing device; collecting unit, collecting a double-sided white light reflection image and a double-sided infrared reflection image of the banknote; and a banknote recognition unit, configured to reflect according to the double-sided white light Image recognition type and face of banknotes
  • a crown size area determining unit configured to select the double-sided white light reflection image and the double-sided infrared reflection image according to the facing direction, obtain a white light image and an infrared image that meet the requirements, and determine a crown size area according to the type;
  • the crown size area locating unit 1 is configured to search the area corresponding to the gray level requirement in the preset initial search area of the infrared image by using the crown size area as a search object, and obtain the preset initial search in the crown size area. Location information on the area;
  • the crown location area locating unit 2 is configured to map location information of the crown area in the preset initial search area to the infrared image according to location information of the initial search area on the infrared image, to obtain Position information of the crown area on the infrared image;
  • a crown area positioning unit 3 configured to: in the infrared image, according to a resolution relationship between the white light image and the infrared image And mapping the position information on the white light image to obtain position information of the crown area on the white light image, and completing positioning of the crown area on the white light image;
  • a crown area obtaining unit configured to acquire gray level information of the crown area on the white light image according to position information of the crown area on the white light image
  • a storage unit configured to store the grayscale information and template data required by the control unit to process.
  • the device also includes:
  • An image cutting unit configured to perform image cutting on the double-sided white light reflection image and the double-sided infrared reflection image according to a boundary between the double-sided white light reflection image and the double-sided infrared reflection image to obtain a cut image;
  • a rotation unit configured to perform rotation correction on the cut image by using an image rotation technique.
  • a double-sided white light reflection image and a double-sided infrared reflection image of the banknote are first collected; then the type and orientation of the banknote are identified according to the double-sided white light reflection image; and then the double-sided white light reflection is selected according to the orientation
  • An image and the double-sided infrared reflected image obtaining a desired white light image and an infrared image, and determining a crown size area according to the type;
  • the area of the crown is a search object, and the area that meets the gray level requirement is searched in the preset initial search area of the infrared image, and the position information of the crown area on the preset initial search area is obtained; and then according to the initial search area.
  • the position information on the infrared image maps position information of the crown area in the preset initial search area to the infrared image, and obtains position information of the crown area on the infrared image. And then mapping location information of the crown area on the infrared image to the white light image according to a resolution relationship between the white light image and the infrared image, to obtain the crown area in the white light Positioning information on the image, completing positioning of the crown area on the white light image; finally acquiring and storing the crown on the white light image according to position information of the crown area on the white light image Grayscale information for the font size area.
  • the infrared image is affected by the new or old banknotes
  • the white light reflection map, the infrared reflection image and the transmission image of the banknote are collected by the optical sensor, and the currency type and the face value are identified by the image feature template matching.
  • the infrared reflection image and the infrared transmission image of the banknote are used to accurately position the crown number of the banknote, and the positioning result is mapped to the white light reflection image to realize accurate positioning of the crown number on the white light reflection image, thereby implementing character cutting and character recognition. It can effectively eliminate the influence of the dirty area on the location of the crown size area, and effectively improve the recognition rate of the crown number.
  • FIG. 1 is a flow chart of a first embodiment of a banknote processing method according to an embodiment of the present invention
  • FIG. 2 is a flow chart of a second embodiment of a banknote processing method according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a front white light reflection image in a second embodiment of the present invention.
  • FIG. 4 is a schematic view showing a reverse white light reflection image in a second embodiment of the present invention.
  • FIG. 5 is a schematic diagram of a front side infrared reflection image in a second embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a reverse side infrared reflection image in a second embodiment of the present invention.
  • FIG. 7 is a schematic view of a front white light reflection image in a coordinate system " ⁇ " according to a second embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a front side white light reflection image after rotation correction according to a second embodiment of the present invention
  • 9 is a schematic diagram of a reverse white light reflection image after rotation correction in a coordinate system according to a second embodiment of the present invention
  • FIG. 10 is a schematic diagram of a banknote crown number on an infrared image in a second embodiment of the present invention
  • FIG. 11 is a schematic view of a front side infrared image in a coordinate system according to a second embodiment of the present invention
  • FIG. 12 is a schematic diagram of a banknote crown number and an initial search area on an infrared image in a second embodiment of the present invention.
  • FIG. 13 is a schematic diagram of sliding a search for a font area in an initial setting search area according to a second embodiment of the present invention.
  • FIG. 14 is a schematic view showing the position of an entire crown area in a white light image according to a second embodiment of the present invention.
  • FIG. 15 is a schematic structural view of an embodiment of a banknote processing apparatus according to an embodiment of the present invention.
  • the invention provides a banknote processing method and device, which utilizes an infrared image to be affected by new or old banknotes, a white light reflection map, an infrared reflection image and a transmission image of an optical sensor collecting banknotes, and the banknotes are identified by image feature template matching,
  • the infrared reflection image and the infrared transmission image of the banknote are used to accurately position the crown number of the banknote, and the positioning result is mapped to the white light reflection image to realize the precise positioning of the crown number on the white light reflection image, thereby implementing character cutting and
  • the character recognition can effectively eliminate the influence of the dirty area on the positioning of the crown size area, thereby effectively improving the recognition rate of the crown number.
  • the method of the embodiment of the present invention can be used not only for detecting a banknote, but also for detecting a sheet-like valuable magnetic file such as a check.
  • the device of the embodiment of the present invention is not limited to being applied to an ATM machine, and can also be applied to In the bill processing apparatus such as a sorting machine, the method and apparatus of the embodiment of the present invention are described below by taking the banknote detection in the ATM machine as an example. Although the banknote detection in the ATM machine is taken as an example, it should not be taken as an example. Definition of inventive method and apparatus.
  • a first embodiment of a banknote processing method includes:
  • the double-sided white light reflection image of the banknote and the double-sided infrared reflection can be collected by the collecting unit. Image.
  • the type of the banknote can be identified according to the double-sided white light reflection image and the image feature template. And oriented.
  • the image feature template described above can be built into the database of the control system, identifying the type of banknote and making a call to the database.
  • the selected double-sided white light reflection image and the double-sided infrared reflection image obtain a white light image and an infrared image that meet the requirements, and determine a crown size area according to the type;
  • the white light image and the infrared image which meet the requirements can be obtained according to the selected double-sided white light reflection image and the double-sided infrared reflection image, and the crown number is determined according to the type. region.
  • the crown area can be used as a search object to search for an area that meets the gray level requirement in the preset initial search area of the infrared image.
  • the above process may specifically be: searching according to the method of preceding and following the column, first searching from left to right or from right to left, after completing the search of the line, proceeding to a column, repeatedly searching by row, until the search matches the gray scale.
  • Required area may be: The search result with the lowest gray mean value in the initial search area is the desired one.
  • the position information of the crown area in the preset initial search area may be mapped to the infrared image according to the position information of the initial search area on the infrared image, and then The position information of the crown area on the infrared image is obtained.
  • Information complete the positioning of the crown area on the white light image
  • the position information of the crown area on the infrared image may be mapped to the white light image according to the resolution relationship between the white light image and the infrared image, thereby obtaining the crown image area in the white light image.
  • the position information on the top thereby completing the positioning of the crown area on the white light image.
  • the gray scale information of the crown area can be acquired and stored on the white light image according to the position information of the crown area on the white light image.
  • the method in the embodiment of the invention utilizes the infrared image to be affected by the new or old banknotes, the white light reflection map, the infrared reflection image and the transmission image of the banknote collected by the optical sensor, and the banknote type and the face value are identified by the image feature template matching.
  • the accurate positioning of the banknote crown number is achieved, and the positioning result is mapped to the white light reflection image to realize the precise positioning of the crown number on the white light reflection image, thereby implementing character cutting and character recognition. Effectively eliminate the influence of the dirty area on the location of the crown size area, and effectively improve the recognition rate of the crown number.
  • the second embodiment of the banknote processing method of the present invention includes :
  • the double-sided white light reflection image includes a front white light reflection image and a reverse white light reflection image;
  • the double-sided infrared reflection image includes a front side infrared reflection image and a reverse side infrared reflection image.
  • the boundary between the double-sided white light reflection image and the double-sided infrared reflection image may be used as a cutting line to respectively perform image on the double-sided white light reflection image and the double-sided infrared reflection image. Cutting, you can get the cut that only contains the part of the banknote Cut the image.
  • the cut image After the cut image is obtained, the cut image can be rotated and corrected by the image rotation technique to obtain a center image.
  • the type and orientation of the banknote may be identified according to the double-sided white light reflection image and the image feature template.
  • the image feature template described above can be built into the database of the control system, and the type of banknote can be identified and the database can be called.
  • the type of the above banknotes may include the currency type and denomination of the banknotes.
  • the selected double-sided white light reflection image and the double-sided infrared reflection image obtain a white light image and an infrared image that meet the requirements, and determine a crown size area according to the type;
  • the white light image and the infrared image which meet the requirements can be obtained according to the selected double-sided white light reflection image and the double-sided infrared reflection image, and the crown number is determined according to the type. region.
  • the white light image and the infrared image which are required to be the front side are the front white light image and the front side infrared image respectively; when the surface is facing the reverse side,
  • the white light image and the infrared image that meet the requirements are the reverse white light image and the reverse side infrared image, respectively.
  • the specific process of determining the crown size area according to the type may be: searching for the corresponding banknote data from the preset database according to the type; determining the crown size area based on the banknote data.
  • the crown area can be used as a search object to search for an area that meets the gray level requirement in the preset initial search area of the infrared image.
  • the above process may specifically be: searching according to the method of preceding and following the column, first searching from left to right or from right to left, after completing the search of the line, proceeding to a column, repeatedly searching by row, until the search matches the gray scale.
  • Required area may be: The search result with the lowest gray mean value in the initial search area is the desired one.
  • the crown size area is pre- Positioning the position information in the initial search area onto the infrared image to obtain position information of the crown area on the infrared image;
  • the position information of the crown area in the preset initial search area may be mapped to the infrared image according to the position information of the initial search area on the infrared image, and then The position information of the crown area on the infrared image is obtained.
  • the position information of the crown area on the infrared image may be mapped to the white light image according to the resolution relationship between the white light image and the infrared image, thereby obtaining the crown image area in the white light image.
  • the position information on the top thereby completing the positioning of the crown area on the white light image.
  • the gray scale information of the crown area can be acquired and stored on the white light image according to the position information of the crown area on the white light image.
  • the coin size of the ATM machine has 10 characters, of which the first 4 characters will disappear on the infrared image, the last 6 characters will not disappear on the infrared image, and the physical length of the 10 characters of the banknote in the face is A. Inches, the height of a single character is ⁇ inches.
  • the function of the optical sensor is to collect the front white light reflection image WHR_Front.bmp, the reverse white light reflection image WHR_Black.bmp, the front infrared reflection image IRR_Front.bmp, and the reverse infrared reflection image IRR_Back.bmp, as shown in Figure 3-6. Show. Let the white light image resolution be M * w d pi and the infrared image resolution be p * 2dpi;
  • the front side of the white light reflection image WHR_Front.bmp collected by the image sensor is in the y-axis, the front white light reflection image WHR_Front.bmp, the lower boundary is the J-axis, and the front white light reflection image WHR_Front.bmp is the lower left vertex.
  • the coordinate system x . ⁇ As shown in Figure 7.
  • the four vertices of the H ⁇ foreground region are A, B, C, and D, respectively.
  • the image is cut, and the cut foreground area is rotated and corrected by the image rotation technique.
  • the straight line of the left boundary of the foreground image of the banknote image is taken as the ordinate direction
  • the straight line of the lower boundary of the foreground image of the banknote image is the horizontal coordinate direction
  • the lower left vertex of the foreground area of the banknote is the coordinate origin.
  • the coordinates are established and the rotated white light image is placed in the coordinate system. , as shown in Figure 8, Figure 9.
  • the banknote crown number is located on the front side of the banknote, there are 10 characters, of which the first 4 characters will disappear on the infrared image, and the last 6 characters will not disappear on the infrared image, and the banknote
  • the 10 characters of the crown size have a physical length of ⁇ ⁇ inches and a single character height of ⁇ inches, as shown in Figure 10.
  • the initial search area set is the rectangular area ⁇ ⁇ / ⁇
  • the lower left vertex r of the area is in the coordinate system ⁇ 3 ⁇ 4.
  • the middle coordinate is 7 ⁇ . ,>
  • the initial search area is high H (pixels), width w (pixels), as shown in Figure 12. Taking r in the region as the coordinate origin, and the edge ⁇ direction is the ordinate direction, edge?
  • the direction of the ⁇ is the direction of the abscissa, and the coordinate system is established, as shown in Figure 13.
  • slide the search height in the first and last columns A rectangular area with pixels and width ⁇ pixels, satisfying the condition
  • the average value of the gray area of the pixel and the rectangular area of the pixel is ⁇ , [the corresponding height is the pixel, and the rectangular area with the width of ⁇ pixel is the 6 crown character area that has not disappeared, and the lower left vertex W is at the coordinates.
  • the corresponding coordinates in 2 y 2 are
  • R Xl R Xl +T Xi w 7
  • the coordinates of the lower left vertex of the crown area in the white light image coordinate system, and the upper right vertex of the upper left vertex of the upper left vertex are respectively: 10
  • the coordinates of the four vertices of the crown region in the white light image coordinate system can be determined.
  • the method of the embodiment of the invention utilizes the infrared image to be affected by the new or old banknotes, the white light reflection map, the infrared reflection image and the transmission image of the banknote collected by the optical sensor, and the banknote type and the face value are identified by the image feature template matching.
  • the infrared reflection image and the infrared transmission image of the banknote are used to accurately position the crown number of the banknote, and the positioning result is mapped to the white light reflection image to realize accurate positioning of the crown number on the white light reflection image, thereby implementing character cutting and character recognition. It can effectively eliminate the influence of the dirty area on the location of the crown size area, and effectively improve the recognition rate of the crown number.
  • an embodiment of a banknote processing apparatus in an embodiment of the present invention includes:
  • the control unit 1501 controls the entire processing flow after the banknotes enter the banknote processing device; the collecting unit 1502 collects the double-sided white light reflection image and the double-sided infrared reflection image of the banknote; the banknote identification unit 1503 is configured to identify the banknote according to the double-sided white light reflection image Type and face To
  • a crown size determining unit 1504 configured to select a white light image and an infrared image according to the selection of the double-sided white light reflection image and the double-sided infrared reflection image, and determine the crown size area according to the type;
  • the crown size area locating unit 1505 is configured to search for the area corresponding to the gray level requirement in the preset initial search area of the infrared image by using the crown size area as the search object, and obtain the position information of the crown size area on the preset initial search area. ;
  • the crown size area locating unit 2506 is configured to map the position information of the crown area in the preset initial search area to the infrared image according to the position information of the initial search area on the infrared image, and obtain the crown area on the infrared image. Location information;
  • the crown size area locating unit 3 1507 is configured to map the position information of the crown size area on the infrared image to the white light image according to the resolution relationship between the white light image and the infrared image, and obtain the position information of the crown size area on the white light image.
  • the positioning of the crown area on the white light image is completed;
  • the crown area obtaining unit 1508 is configured to obtain gray information of the crown area on the white light image according to the position information of the crown area on the white light image;
  • the storage unit 1509 is configured to store gray scale information and template data required for the control unit to process.
  • the image cutting unit 1510 is configured to perform image cutting on the double-sided white light reflection image and the double-sided infrared reflection image according to the boundary between the double-sided white light reflection image and the double-sided infrared reflection image to obtain a cut image;
  • the rotating unit 1511 is configured to perform rotation correction on the cut image by using an image rotation technique.
  • the collecting unit 1502 of the embodiment of the present invention first collects the double-sided white light reflection image and the double-sided infrared reflection image of the banknote; then the banknote recognition unit 1503 identifies the paper image and the double-sided infrared reflection image according to the double-sided white light reflection image, and obtains the required requirements.
  • the white light image and the infrared image, and the crown size area is determined according to the type; then the crown size area locating unit 1505 searches for the crown size area, and searches for the area that meets the gray level requirement in the preset initial search area of the infrared image.
  • the crown area positioning unit 2506 maps the position information of the crown area in the preset initial search area according to the position information of the initial search area on the infrared image.
  • the crown area positioning unit 3 1507 maps the position information of the crown area on the infrared image to white light according to the resolution relationship between the white light image and the infrared image.
  • the position information of the crown area on the white light image is obtained, the position information of the crown area on the white image is completed, and the gray scale information of the crown area is obtained on the white light image.
  • the storage unit 1509 stores the gradation information and the template data required for the control unit processing.
  • the device of the embodiment of the invention utilizes the infrared image to be affected by the new or old banknotes, the white light reflection map, the infrared reflection image and the transmission image of the banknote collected by the optical sensor, and the banknote type and the face value are identified by the image feature template matching.
  • the infrared reflection image and the infrared transmission image of the banknote are used to accurately position the crown number of the banknote, and the positioning result is mapped to the white light reflection image to realize accurate positioning of the crown number on the white light reflection image, thereby implementing character cutting and character recognition. It can effectively eliminate the influence of the dirty area on the location of the crown size area, and effectively improve the recognition rate of the crown number.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Toxicology (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Inspection Of Paper Currency And Valuable Securities (AREA)
  • Character Input (AREA)

Abstract

一种纸币处理装置及方法,利用红外图像受纸币新旧或脏污影响小、光学传感器采集纸币的白光反射图、红外反射图像和透射图像,并通过图像特征模板匹配识别纸币币种、面值,在此基础上利用纸币的红外反射图像和红外透射图像实现对纸币冠字号精确定位,将定位结果映射到白光反射图像后实现冠字号在白光反射图像上的精确定位,进而实施字符切割和字符识别,能够有效消除冠字号区域存在脏污时对冠字号区域定位的影响,进而有效提高冠字号的识别率。

Description

一种纸币处理方法及装置 本申请要求于 2013 年 09 月 06 日提交中国专利局、 申请号为 201310405433.0、 发明名称为"一种纸币处理方法及装置"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
本发明涉及纸币处理领域, 具体涉及一种纸币处理方法及装置。 背景技术
近年来, 随着世界经济和科技的迅速发展, 各式各样的金融自助设备 在人们日常生活的各个领域得到应用, 以 ATM 为例, 在城市、 城镇以及 甚至乡村都能看到银行铺设的 ATM网点, 这些自助网点能 24小时为人们 提供转账、 存款、 取款等现金业务, 在为银行省去了大量的人力成本的同 时实现了银行营业时间和空间的延拓, 大大提高了银行的营业效率。
但与之相伴的也带来了许多问题, 比如金融自助设备误收假币等。 随 着不法分子造假技术的日益先进, 许多假币甚至达到了以假乱真的地步, 因此经常能看到媒体报道客户在某某银行网点取出 "假钞" 而产生纠纷或 者某某银行在进行行内现金清点时出现因误收 "假币" 而 "短款" 的现象, 无疑银行和客户都是假币的受害者。 为了有效跟踪假钞的真正来源, 为司 法机关在处理类似经济纠纷时提供相应的司法依据, 保护客户和金融机构 的权益, 大多数的金融设备都通过保留交易纸币图像等手段建立了钞票流 通追溯系统, 为了节约数据传输时间和存储空间, 在保存钞票图像时往往 只保存钞票冠字号区域图像, 因为冠字号码是钞票的 "DNA"。 同时, 一 些设备还具备了冠字号字符识别功能,依靠检索钞票冠字号码迅速锁定"嫌 疑钞票", 提高了钞票流通溯源的效率。
目前金融自助设备都是在钞票白光反射图像上搜索并切割钞票冠字号 区域图像, 进而保存图像、 识别并记录钞票冠字号。 然而当钞票比较残旧、 时, 由于钞票白光图像噪声大, 干扰多, 不利于实现对冠字号区域的有效 定位, 因此金融自助设备经常出现钞票冠字号图像截取不准确的问题, 冠 字号区域图像截取的不准确使得金融自助设备的钞票追溯功能大大下降, 严重影响了金融自助设备的 "反假防假" 能力, 关键时刻也不能为司法机 关提供有参考价值的司法证据。 发明内容
本发明提出一种纸币处理方法及装置, 利用红外图像受纸币新旧或脏 污影响小、光学传感器采集纸币的白光反射图、红外反射图像和透射图像, 并通过图像特征模板匹配识别纸币币种、 面值, 在此基础上利用纸币的红 外反射图像和红外透射图像实现对纸币冠字号精确定位, 将定位结果映射 到白光反射图像后实现冠字号在白光反射图像上的精确定位, 进而实施字 符切割和字符识别, 能够有效消除冠字号区域存在脏污时对冠字号区域定 位的影响, 进而有效提高冠字号的识别率。
本发明提供的纸币处理方法, 包括:
采集所述纸币的双面白光反射图像和双面红外反射图像;
根据所述双面白光反射图像及图像特征模板识别纸币的类型和面向; 根据所述面向选择所述双面白光反射图像和所述双面红外反射图像, 得到符合要求的白光图像和红外图像, 并根据所述类型确定冠字号区域; 以所述冠字号区域为搜索对象, 在红外图像的预设初始搜索区域内搜 索符合灰度要求的区域, 得到所述冠字号区域在所述预设初始搜索区域上 的位置信息;
根据初始搜索区域在所述红外图像上的位置信息, 将所述冠字号区域 在所述预设初始搜索区域内的位置信息映射至所述红外图像上, 得到所述 冠字号区域在所述红外图像上的位置信息;
根据所述白光图像和所述红外图像的分辨率关系, 将所述冠字号区域 在所述红外图像上的位置信息映射至所述白光图像上, 得到所述冠字号区 域在所述白光图像上的位置信息, 完成所述冠字号区域在所述白光图像上 的定位;
根据所述冠字号区域在所述白光图像上的位置信息, 在所述白光图像 上获取并存储所述冠字号区域的灰度信息。
可选地 ,
所述纸币的类型包括: 纸币的币种、 面额。
可选地 ,
在步骤所述采集所述纸币的双面白光反射图像和双面红外反射图像之 后及步骤所述根据所述双面白光反射图像识别纸币的类型和面向之前还包 括:
根据所述双面白光反射图像和所述双面红外反射图像的边界对所述双 面白光反射图像和所述双面红外反射图像进行图像切割, 得到切割图像; 利用图像旋转技术对所述切割图像进行旋转纠偏。
可选地 ,
所述双面白光反射图像包括正面白光反射图像和反面白光反射图像; 所述双面红外反射图像包括正面红外反射图像和反面红外反射图像。 可选地 ,
步骤所述根据所述面向选择所述双面白光反射图像和所述双面红外反 射图像, 得到符合要求的白光图像和红外图像包括:
当面向为正面时, 符合要求的白光图像和红外图像分别为正面白光图 像和正面红外图像;
当面向为反面时, 符合要求的白光图像和红外图像分别为反面白光图 像和反面红外图像。
可选地 ,
步骤所述根据所述类型确定冠字号区域包括:
根据所述类型从预置数据库中查找相应的纸币数据;
根据所述纸币数据确定冠字号区域。
本发明提供的纸币处理装置, 包括:
控制单元, 控制纸币进入所述纸币处理装置后的整个处理流程; 采集单元, 采集所述纸币的双面白光反射图像和双面红外反射图像; 钞票识别单元, 用于根据所述双面白光反射图像识别纸币的类型和面 向;
冠字号区域确定单元, 用于根据所述面向选择所述双面白光反射图像 和所述双面红外反射图像, 得到符合要求的白光图像和红外图像, 并根据 所述类型确定冠字号区域;
冠字号区域定位单元一, 用于以所述冠字号区域为搜索对象, 在红外 图像的预设初始搜索区域内搜索符合灰度要求的区域, 得到所述冠字号区 域在所述预设初始搜索区域上的位置信息;
冠字号区域定位单元二, 用于根据初始搜索区域在所述红外图像上的 位置信息, 将所述冠字号区域在所述预设初始搜索区域内的位置信息映射 至所述红外图像上, 得到所述冠字号区域在所述红外图像上的位置信息; 冠字号区域定位单元三, 用于根据所述白光图像和所述红外图像的分 辨率关系, 将所述冠字号区域在所述红外图像上的位置信息映射至所述白 光图像上, 得到所述冠字号区域在所述白光图像上的位置信息, 完成所述 冠字号区域在所述白光图像上的定位;
冠字号区域获取单元, 用于根据所述冠字号区域在所述白光图像上的 位置信息, 在所述白光图像上获取所述冠字号区域的灰度信息;
存储单元, 用于存储所述灰度信息及所述控制单元处理所需要的模板 数据。
可选地 ,
所述装置还包括:
图像切割单元, 用于根据所述双面白光反射图像和所述双面红外反射 图像的边界对所述双面白光反射图像和所述双面红外反射图像进行图像切 割, 得到切割图像;
旋转单元, 用于利用图像旋转技术对所述切割图像进行旋转纠偏。 本发明中, 首先采集所述纸币的双面白光反射图像和双面红外反射图 像; 然后根据所述双面白光反射图像识别纸币的类型和面向; 接着根据所 述面向选择所述双面白光反射图像和所述双面红外反射图像, 得到符合要 求的白光图像和红外图像, 并根据所述类型确定冠字号区域; 然后以所述 冠字号区域为搜索对象, 在红外图像的预设初始搜索区域内搜索符合灰度 要求的区域,得到所述冠字号区域在所述预设初始搜索区域上的位置信息; 接着根据初始搜索区域在所述红外图像上的位置信息, 将所述冠字号区域 在所述预设初始搜索区域内的位置信息映射至所述红外图像上, 得到所述 冠字号区域在所述红外图像上的位置信息; 然后根据所述白光图像和所述 红外图像的分辨率关系, 将所述冠字号区域在所述红外图像上的位置信息 映射至所述白光图像上, 得到所述冠字号区域在所述白光图像上的位置信 息, 完成所述冠字号区域在所述白光图像上的定位; 最后根据所述冠字号 区域在所述白光图像上的位置信息, 在所述白光图像上获取并存储所述冠 字号区域的灰度信息。 由于本发明的方法及装置, 利用红外图像受纸币新 旧或脏污影响小、 光学传感器采集纸币的白光反射图、 红外反射图像和透 射图像, 并通过图像特征模板匹配识别纸币币种、 面值, 在此基础上利用 纸币的红外反射图像和红外透射图像实现对纸币冠字号精确定位, 将定位 结果映射到白光反射图像后实现冠字号在白光反射图像上的精确定位, 进 而实施字符切割和字符识别, 能够有效消除冠字号区域存在脏污时对冠字 号区域定位的影响, 进而有效提高冠字号的识别率。
附图说明
图 1为本发明实施例中纸币处理方法第一实施例流程图;
图 2为本发明实施例中纸币处理方法第二实施例流程图;
图 3为本发明第二实施例中正面白光反射图像示意图;
图 4为本发明第二实施例中反面白光反射图像示意图;
图 5为本发明第二实施例中正面红外反射图像示意图;
图 6为本发明第二实施例中反面红外反射图像示意图;
图 7 为本发明第二实施例中正面白光反射图像在坐标系 "^^中的示意 图;
图 8为本发明第二实施例中旋转纠偏后的正面白光反射图像在坐标系 中的示意图; 图 9为本发明第二实施例中旋转纠偏后的反面白光反射图像在坐标系 中的示意图;
图 10为本发明第二实施例中红外图像上的钞票冠字号示意图; 图 11 为本发明第二实施例中正面红外图像在坐标系 中的示意 图;
图 12 为本发明第二实施例中红外图像上的钞票冠字号及初始搜索区 域示意图;
图 13 为本发明第二实施例在初始设定搜索区域中滑动搜素冠字号区 域的示意图;
图 14为本发明第二实施例整个冠字号区域在白光图像中位置示意图; 图 15为本发明实施例中纸币处理装置实施例结构示意图。 具体实施方式
本发明提出一种纸币处理方法及装置, 利用红外图像受纸币新旧或脏 污影响小、光学传感器采集纸币的白光反射图、红外反射图像和透射图像, 并通过图像特征模板匹配识别纸币币种、 面值, 在此基础上利用纸币的红 外反射图像和红外透射图像实现对纸币冠字号精确定位, 将定位结果映射 到白光反射图像后实现冠字号在白光反射图像上的精确定位, 进而实施字 符切割和字符识别, 能够有效消除冠字号区域存在脏污时对冠字号区域定 位的影响, 进而有效提高冠字号的识别率。
需要说明的是, 本发明实施例的方法不但可以用于检测钞票, 还可以 用于检测支票等薄片类有价磁性文件, 本发明实施例的装置不限于应用于 ATM机中, 还可以应用于清分机等票据处理设备中, 下面以 ATM机中的 钞票检测为例对本发明实施例的方法和装置进行说明, 虽然仅以 ATM机 中的钞票检测为例进行说明,但是不应将此作为本发明方法和装置的限定。
请参阅图 1 , 本发明实施例中纸币处理方法的第一实施例包括:
101、 采集纸币的双面白光反射图像和双面红外反射图像;
当纸币在 ATM机进行存款, 通过传送通道进入磁信号传感器的扫描 区域时, 可以通过采集单元采集纸币的双面白光反射图像和双面红外反射 图像。
102、 根据双面白光反射图像及图像特征模板识别纸币的类型和面向; 采集纸币的双面白光反射图像和双面红外反射图像之后, 可以根据双 面白光反射图像及图像特征模板识别纸币的类型和面向。 上述的图像特征 模板可以内置在控制系统的数据库中, 当识别纸币的类型和面向对数据库 进行调用即可。
103、根据面向选择双面白光反射图像和双面红外反射图像,得到符合 要求的白光图像和红外图像, 并根据类型确定冠字号区域;
根据双面白光反射图像及图像特征模板识别纸币的类型和面向之后, 可以根据面向选择双面白光反射图像和双面红外反射图像, 得到符合要求 的白光图像和红外图像, 并根据类型确定冠字号区域。
104、 以冠字号区域为搜索对象,在红外图像的预设初始搜索区域内搜 索符合灰度要求的区域, 得到冠字号区域在预设初始搜索区域上的位置信 息;
完成步骤 103之后, 可以将冠字号区域作为搜索对象, 在红外图像的 预设初始搜索区域内搜索符合灰度要求的区域。 上述的过程具体可以是: 按照先行后列的方式进行搜索, 先从左往右或从右往左地进行搜索, 完成 本行搜索之后, 前进一列, 重复按行搜索, 直至搜索到符合灰度要求的区 域。 上述灰度要求可以为: 初始搜索区域内灰度均值最低的搜索结果即为 所求。
105、根据初始搜索区域在红外图像上的位置信息,将冠字号区域在预 设初始搜索区域内的位置信息映射至红外图像上, 得到冠字号区域在红外 图像上的位置信息;
得到冠字号区域在预设初始搜索区域上的位置信息之后, 可以根据初 始搜索区域在红外图像上的位置信息, 将冠字号区域在预设初始搜索区域 内的位置信息映射至红外图像上, 进而得到冠字号区域在红外图像上的位 置信息。
106、根据白光图像和红外图像的分辨率关系,将冠字号区域在红外图 像上的位置信息映射至白光图像上, 得到冠字号区域在白光图像上的位置 信息, 完成冠字号区域在白光图像上的定位;
得到冠字号区域在红外图像上的位置信息之后, 可以根据白光图像和 红外图像的分辨率关系, 将冠字号区域在红外图像上的位置信息映射至白 光图像上, 进而得到冠字号区域在白光图像上的位置信息, 从而完成冠字 号区域在白光图像上的定位。
107、根据冠字号区域在白光图像上的位置信息,在白光图像上获取并 存储冠字号区域的灰度信息。
完成冠字号区域在白光图像上的定位之后, 可以根据冠字号区域在白 光图像上的位置信息, 在白光图像上获取并存储冠字号区域的灰度信息。
本发明实施例中的方法利用红外图像受纸币新旧或脏污影响小、 光学 传感器采集纸币的白光反射图、 红外反射图像和透射图像, 并通过图像特 征模板匹配识别纸币币种、 面值, 在此基础上利用纸币的红外反射图像和 红外透射图像实现对纸币冠字号精确定位, 将定位结果映射到白光反射图 像后实现冠字号在白光反射图像上的精确定位, 进而实施字符切割和字符 识别, 能够有效消除冠字号区域存在脏污时对冠字号区域定位的影响, 进 而有效提高冠字号的识别率。
上面筒单介绍了本发明纸币处理方法的第一实施例, 下面对本发明纸 币处理方法的第二实施例进行详细的描述, 请参阅图 2, 本发明实施例中 纸币处理方法第二实施例包括:
201、 采集纸币的双面白光反射图像和双面红外反射图像;
当纸币在 ATM机进行存款, 通过传送通道进入磁信号传感器的扫描 区域时, 可以通过采集单元采集纸币的双面白光反射图像和双面红外反射 图像。上述双面白光反射图像包括正面白光反射图像和反面白光反射图像; 上述双面红外反射图像包括正面红外反射图像和反面红外反射图像。
202、根据双面白光反射图像和双面红外反射图像的边界对双面白光反 射图像和双面红外反射图像进行图像切割, 得到切割图像;
采集纸币的双面白光反射图像和双面红外反射图像之后, 可以将双面 白光反射图像和双面红外反射图像的边界作为切割线, 分别对双面白光反 射图像和双面红外反射图像进行图像切割, 可以得到仅包含纸币部分的切 割图像。
203、 利用图像旋转技术对切割图像进行旋转纠偏;
得到切割图像之后,可以利用图像旋转技术对切割图像进行旋转纠偏, 得到正中图像。
204、 根据双面白光反射图像及图像特征模板识别纸币的类型和面向; 利用图像旋转技术对切割图像进行旋转纠偏之后, 可以根据双面白光 反射图像及图像特征模板识别纸币的类型和面向。 上述的图像特征模板可 以内置在控制系统的数据库中, 当识别纸币的类型和面向对数据库进行调 用即可。 上述纸币的类型可以包括纸币的币种、 面额。
205、根据面向选择双面白光反射图像和双面红外反射图像,得到符合 要求的白光图像和红外图像, 并根据类型确定冠字号区域;
根据双面白光反射图像及图像特征模板识别纸币的类型和面向之后, 可以根据面向选择双面白光反射图像和双面红外反射图像, 得到符合要求 的白光图像和红外图像, 并根据类型确定冠字号区域。 根据面向选择双面 白光反射图像和双面红外反射图像的具体过程可以是: 当面向为正面时, 符合要求的白光图像和红外图像分别为正面白光图像和正面红外图像; 当 面向为反面时, 符合要求的白光图像和红外图像分别为反面白光图像和反 面红外图像。 根据类型确定冠字号区域的具体过程可以是: 根据类型从预 置数据库中查找相应的纸币数据; 根据纸币数据确定冠字号区域。
206、 以冠字号区域为搜索对象,在红外图像的预设初始搜索区域内搜 索符合灰度要求的区域, 得到冠字号区域在预设初始搜索区域上的位置信 息;
完成步骤 205之后, 可以将冠字号区域作为搜索对象, 在红外图像的 预设初始搜索区域内搜索符合灰度要求的区域。 上述的过程具体可以是: 按照先行后列的方式进行搜索, 先从左往右或从右往左地进行搜索, 完成 本行搜索之后, 前进一列, 重复按行搜索, 直至搜索到符合灰度要求的区 域。 上述灰度要求可以为: 初始搜索区域内灰度均值最低的搜索结果即为 所求。
207、根据初始搜索区域在红外图像上的位置信息,将冠字号区域在预 设初始搜索区域内的位置信息映射至红外图像上, 得到冠字号区域在红外 图像上的位置信息;
得到冠字号区域在预设初始搜索区域上的位置信息之后, 可以根据初 始搜索区域在红外图像上的位置信息, 将冠字号区域在预设初始搜索区域 内的位置信息映射至红外图像上, 进而得到冠字号区域在红外图像上的位 置信息。
208、根据白光图像和红外图像的分辨率关系,将冠字号区域在红外图 像上的位置信息映射至白光图像上, 得到冠字号区域在白光图像上的位置 信息, 完成冠字号区域在白光图像上的定位;
得到冠字号区域在红外图像上的位置信息之后, 可以根据白光图像和 红外图像的分辨率关系, 将冠字号区域在红外图像上的位置信息映射至白 光图像上, 进而得到冠字号区域在白光图像上的位置信息, 从而完成冠字 号区域在白光图像上的定位。
209、根据冠字号区域在白光图像上的位置信息,在白光图像上获取并 存储冠字号区域的灰度信息。
完成冠字号区域在白光图像上的定位之后, 可以根据冠字号区域在白 光图像上的位置信息, 在白光图像上获取并存储冠字号区域的灰度信息。 假设投入 ATM机中钞票冠字号有 10个字符, 其中前 4个字符在红外 图像上会消失,后 6个字符在红外图像上不消失,且钞票冠字号 10个字符 在票面的物理长度为 A英寸, 单个字符高度为 ^英寸。
1、 投入钞票, 流程开始;
2、 光学传感器作用 , 采集投入纸币的正面白光反射图像 WHR_Front.bmp,反面白光反射图像 WHR_Black.bmp , 正面红外反射图像 IRR_Front.bmp ,反面红外反射图像 IRR_Back.bmp , 分别如图 3-图 6所示。 设白光图像分辨率为 M * w dpi , 红外图像分辨率为 p*2dpi;
3、 以图像传感器采集到的正面白光反射图像 WHR_Front.bmp的左边 界所在直线方向为 y轴、正面白光反射图像 WHR_Front.bmp下边界所在直 线方向为 J轴、正面白光反射图像 WHR_Front.bmp左下顶点为坐标原点建 立坐标系 x。\ 如图 7所示。 H殳前景区域的四个顶点分别为 A, B, C和 D。 通过探测前景区域的边界点, 再利用边界点进行线性拟合, 可得到: 左边界直线方程(即边 AB所在直线方程): y = x
右边界直线方程(即边 CD所在直线方程): y = k2X
上边界直线方程(即边 AD所在直线方程): y ^ k'x + b'
下边界直线方程(即边 BC所在直线方程): y = x
4、根据前景区域四个边界直线, 完成图像切割, 并将切割出来的前景 区域利用图像旋转技术进行旋转纠偏。 以钞票图像前景区域左边界所在直 线为纵坐标方向, 钞票图像前景区域下边界所在直线为横坐标方向, 钞票 前景区域左下顶点为坐标原点, 建立坐标 并将旋转后的白光图像置 于坐标系 中, 如图 8、 图 9所示。
5、根据双面白光反射图像及图像特征模板识别纸币的类型和面向。在 白光图像上选择指定区域,与存储单元存储的图像特征模板进行匹配识别, 确定投入纸币的币种、 面值和面向。 殳识别结果为某币种 XXX100元, 入钞方向为正面正立向上。 假设根据存储单元中存储的 XXX100元纸币信 息, 投入钞票冠字号位于钞票正面, 有 10个字符, 其中前 4个字符在红外 图像上会消失,后 6个字符在红外图像上不消失,且钞票冠字号 10个字符 在票面的物理长度为 Ζι英寸, 单个字符高度为^英寸, 如图 10所示。
6、 将钞票正面红外反射图像置于坐标系 Α , 如图 11所示。根据钞 票冠字号的物理长度、 单个字符的物理高度、 冠字号在红外图像上的成像 特点以及红外图像的分辨率, 可推算红外图像上, 未消失的 6个字符区域 高度: pa = l2 * Q (个像素),
宽度: ^ =[^ ^*(10-4)]/10 (个像素), 因此, 红外图像上一个高度为 像素、 宽度为 W 像素的矩形框恰好 能框住冠字号在红外图像上的成像区域。 假定设定的初始搜索区域为矩形 区域^ ι/ν , 区域左下顶点 r在坐标系 ·¾。 中坐标为7^。,> , 初始搜索区 域高 H (个像素), 宽 w (个像素), 如图 12所示。 以区域 中 r为坐 标原点, 边 ^方向为纵坐标方向, 边?^方向为横坐标方向, 建立坐标系 , 如图 13 所示。 在区域 中, 按先行后列的方式滑动搜索高度 为 像素、 宽度为^ ^像素的矩形区域, 满足条件
Greyij] = min Grey[i],U = 0,1, ···,(〃 - hpix + 1) * ( — w pix + I) , J <= i) 其中 为区 中所有满足高度为 h 象素、 宽度为 像素的矩形 区域的灰度均值, 则 ^ [ 对应的高度为 像素、 宽度为^ ^像素的矩形 区域即是未消失的 6个冠字号字符区域, 其左下顶点 W在坐标 。2y2中对 应的坐标为
Figure imgf000014_0001
R ― W . +
Υv2, = J pix
因: 顶点 R在坐标系 "V^i中对应的坐标为
RXl =RXl +TXi w 7
T yi = yo ,
因此得:
R =j%(W-wpix +l) + x0
Figure imgf000014_0002
即完成了 (未消失 6个的)冠字号字符区域的左下顶点 R在红外图像 上的定位。
7、根据红外反射图像和白光反射图像分辨率关系,则如图 14中所示, R在坐标系 中坐标为
M、 ― ,M_,
= * (-) = υ% ( ― wpix +ΐ)+¾]* (-)
P
8、 进而可得如图 14中所示, 白光图像坐标系¾ ^中冠字号区域左下 顶点 /, 左上顶点 右下顶点 右上顶点 的坐标分别为: 10
XR - * x*(
10
Y,
下顶点
右下顶点
Figure imgf000015_0001
左上顶点 Υ =ΥΏ+Ι^Ν
L
右上顶点
Figure imgf000015_0002
由此, 可以确定冠字号区域四个顶点在白光图像坐标系 中的坐 标'
9、 将白光图像中四个顶点 , J , L, 围成的区域内的灰度数据存 储到指定空间, 完成在白光图像中对冠字号图像区域的截取, 实现冠字号 区域图像存储, 以备后用。
本发明实施例的方法利用红外图像受纸币新旧或脏污影响小、 光学传 感器采集纸币的白光反射图、 红外反射图像和透射图像, 并通过图像特征 模板匹配识别纸币币种、 面值, 在此基础上利用纸币的红外反射图像和红 外透射图像实现对纸币冠字号精确定位, 将定位结果映射到白光反射图像 后实现冠字号在白光反射图像上的精确定位, 进而实施字符切割和字符识 另' h 能够有效消除冠字号区域存在脏污时对冠字号区域定位的影响, 进而 有效提高冠字号的识别率。
上面对本发明纸币处理方法的第二实施例作了详细描述, 特别是检测 通信链路的通信质量和根据排序数据和分配规则向接入设备分配通信链路 的过程, 下面介绍本发明纸币处理装置实施例, 请参阅图 15, 本发明实施 例中纸币处理装置实施例包括:
控制单元 1501, 控制纸币进入纸币处理装置后的整个处理流程; 采集单元 1502, 采集纸币的双面白光反射图像和双面红外反射图像; 钞票识别单元 1503, 用于根据双面白光反射图像识别纸币的类型和面 向;
冠字号区域确定单元 1504, 用于根据面向选择双面白光反射图像和双 面红外反射图像, 得到符合要求的白光图像和红外图像, 并根据类型确定 冠字号区域;
冠字号区域定位单元一 1505 , 用于以冠字号区域为搜索对象, 在红外 图像的预设初始搜索区域内搜索符合灰度要求的区域, 得到冠字号区域在 预设初始搜索区域上的位置信息;
冠字号区域定位单元二 1506, 用于根据初始搜索区域在红外图像上的 位置信息, 将冠字号区域在预设初始搜索区域内的位置信息映射至红外图 像上, 得到冠字号区域在红外图像上的位置信息;
冠字号区域定位单元三 1507, 用于根据白光图像和红外图像的分辨率 关系, 将冠字号区域在红外图像上的位置信息映射至白光图像上, 得到冠 字号区域在白光图像上的位置信息,完成冠字号区域在白光图像上的定位; 冠字号区域获取单元 1508, 用于根据冠字号区域在白光图像上的位置 信息, 在白光图像上获取冠字号区域的灰度信息;
存储单元 1509, 用于存储灰度信息及控制单元处理所需要的模板数 据。
本发明实施例的装置进一步还包括:
图像切割单元 1510, 用于根据双面白光反射图像和双面红外反射图像 的边界对双面白光反射图像和双面红外反射图像进行图像切割, 得到切割 图像;
旋转单元 1511 , 用于利用图像旋转技术对所述切割图像进行旋转纠 偏。
本发明实施例的采集单元 1502 首先采集纸币的双面白光反射图像和 双面红外反射图像;然后钞票识别单元 1503根据双面白光反射图像识别纸 射图像和双面红外反射图像, 得到符合要求的白光图像和红外图像, 并根 据类型确定冠字号区域;然后冠字号区域定位单元一 1505以冠字号区域为 搜索对象, 在红外图像的预设初始搜索区域内搜索符合灰度要求的区域, 得到冠字号区域在预设初始搜索区域上的位置信息; 接着冠字号区域定位 单元二 1506根据初始搜索区域在红外图像上的位置信息,将冠字号区域在 预设初始搜索区域内的位置信息映射至红外图像上, 得到冠字号区域在红 外图像上的位置信息;然后冠字号区域定位单元三 1507根据白光图像和红 外图像的分辨率关系, 将冠字号区域在红外图像上的位置信息映射至白光 图像上, 得到冠字号区域在白光图像上的位置信息, 完成冠字号区域在白 像上的位置信息, 在白光图像上获取冠字号区域的灰度信息。 存储单元 1509中存储了灰度信息及控制单元处理所需要的模板数据。
本发明实施例的装置利用红外图像受纸币新旧或脏污影响小、 光学传 感器采集纸币的白光反射图、 红外反射图像和透射图像, 并通过图像特征 模板匹配识别纸币币种、 面值, 在此基础上利用纸币的红外反射图像和红 外透射图像实现对纸币冠字号精确定位, 将定位结果映射到白光反射图像 后实现冠字号在白光反射图像上的精确定位, 进而实施字符切割和字符识 另' h 能够有效消除冠字号区域存在脏污时对冠字号区域定位的影响, 进而 有效提高冠字号的识别率。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分步 骤是可以通过程序来指令相关的硬件完成, 其中的程序可以存储于一种计 算机可读存储介质中, 上述提到的存储介质可以是只读存储器, 磁盘或光 盘等。
以上对本发明所提供的一种纸币处理方法及装置进行了详细介绍, 对 于本领域的一般技术人员, 依据本发明实施例的思想, 在具体实施方式及 应用范围上均会有改变之处, 综上所述, 本说明书内容不应理解为对本发 明的限制。

Claims

权 利 要 求
1、 一种纸币处理方法, 其特征在于, 包括:
采集所述纸币的双面白光反射图像和双面红外反射图像;
根据所述双面白光反射图像及图像特征模板识别纸币的类型和面向; 根据所述面向选择所述双面白光反射图像和所述双面红外反射图像, 得到符合要求的白光图像和红外图像, 并根据所述类型确定冠字号区域; 以所述冠字号区域为搜索对象, 在红外图像的预设初始搜索区域内搜 索符合灰度要求的区域, 得到所述冠字号区域在所述预设初始搜索区域上 的位置信息;
根据初始搜索区域在所述红外图像上的位置信息, 将所述冠字号区域 在所述预设初始搜索区域内的位置信息映射至所述红外图像上, 得到所述 冠字号区域在所述红外图像上的位置信息;
根据所述白光图像和所述红外图像的分辨率关系, 将所述冠字号区域 在所述红外图像上的位置信息映射至所述白光图像上, 得到所述冠字号区 域在所述白光图像上的位置信息, 完成所述冠字号区域在所述白光图像上 的定位;
根据所述冠字号区域在所述白光图像上的位置信息, 在所述白光图像 上获取并存储所述冠字号区域的灰度信息。
2、 根据权利要求 1所述的纸币处理方法, 其特征在于, 所述纸币的 类型包括: 纸币的币种、 面额。
3、 根据权利要求 1所述的纸币处理方法, 其特征在于, 在步骤所述 采集所述纸币的双面白光反射图像和双面红外反射图像之后及步骤所述根 据所述双面白光反射图像识别纸币的类型和面向之前还包括:
根据所述双面白光反射图像和所述双面红外反射图像的边界对所述 双面白光反射图像和所述双面红外反射图像进行图像切割,得到切割图像; 利用图像旋转技术对所述切割图像进行旋转纠偏。
4、 根据权利要求 1至 3中任一项所述的纸币处理方法, 其特征在于, 所述双面白光反射图像包括正面白光反射图像和反面白光反射图像; 所述双面红外反射图像包括正面红外反射图像和反面红外反射图像。
5、 根据权利要求 1至 3中任一项所述的纸币处理方法, 其特征在于, 步骤所述根据所述面向选择所述双面白光反射图像和所述双面红外反射图 像, 得到符合要求的白光图像和红外图像包括:
当面向为正面时,符合要求的白光图像和红外图像分别为正面白光图 像和正面红外图像;
当面向为反面时,符合要求的白光图像和红外图像分别为反面白光图 像和反面红外图像。
6、 根据权利要求 1所述的纸币处理方法, 其特征在于, 步骤所述根 据所述类型确定冠字号区域包括:
根据所述类型从预置数据库中查找相应的纸币数据;
根据所述纸币数据确定冠字号区域。
7、 一种纸币处理装置, 其特征在于, 包括:
控制单元, 控制纸币进入所述纸币处理装置后的整个处理流程; 采集单元, 采集所述纸币的双面白光反射图像和双面红外反射图像; 钞票识别单元, 用于根据所述双面白光反射图像识别纸币的类型和面 向;
冠字号区域确定单元, 用于根据所述面向选择所述双面白光反射图像 和所述双面红外反射图像, 得到符合要求的白光图像和红外图像, 并根据 所述类型确定冠字号区域;
冠字号区域定位单元一, 用于以所述冠字号区域为搜索对象, 在红外 图像的预设初始搜索区域内搜索符合灰度要求的区域, 得到所述冠字号区 域在所述预设初始搜索区域上的位置信息;
冠字号区域定位单元二, 用于根据初始搜索区域在所述红外图像上的 位置信息, 将所述冠字号区域在所述预设初始搜索区域内的位置信息映射 至所述红外图像上, 得到所述冠字号区域在所述红外图像上的位置信息; 冠字号区域定位单元三, 用于根据所述白光图像和所述红外图像的分 辨率关系, 将所述冠字号区域在所述红外图像上的位置信息映射至所述白 光图像上, 得到所述冠字号区域在所述白光图像上的位置信息, 完成所述 冠字号区域在所述白光图像上的定位;
冠字号区域获取单元, 用于根据所述冠字号区域在所述白光图像上的 位置信息, 在所述白光图像上获取所述冠字号区域的灰度信息;
存储单元, 用于存储所述灰度信息及所述控制单元处理所需要的模板 数据。
8、 根据权利要求 7所述的纸币处理装置, 其特征在于, 所述装置还 包括:
图像切割单元, 用于根据所述双面白光反射图像和所述双面红外反射 图像的边界对所述双面白光反射图像和所述双面红外反射图像进行图像切 割, 得到切割图像;
旋转单元, 用于利用图像旋转技术对所述切割图像进行旋转纠偏。
PCT/CN2014/072179 2013-09-06 2014-02-18 一种纸币处理方法及装置 WO2015032187A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310405433.0 2013-09-06
CN201310405433.0A CN103456075B (zh) 2013-09-06 2013-09-06 一种纸币处理方法及装置

Publications (1)

Publication Number Publication Date
WO2015032187A1 true WO2015032187A1 (zh) 2015-03-12

Family

ID=49738399

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/072179 WO2015032187A1 (zh) 2013-09-06 2014-02-18 一种纸币处理方法及装置

Country Status (2)

Country Link
CN (1) CN103456075B (zh)
WO (1) WO2015032187A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106683264A (zh) * 2017-02-20 2017-05-17 深圳怡化电脑股份有限公司 一种污损冠字号的识别方法及装置

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6060859B2 (ja) * 2013-08-29 2017-01-18 沖電気工業株式会社 紙幣処理装置及び紙幣処理方法
CN103456075B (zh) * 2013-09-06 2015-11-25 广州广电运通金融电子股份有限公司 一种纸币处理方法及装置
CN104200566B (zh) 2014-09-11 2018-04-20 广州广电运通金融电子股份有限公司 一种基于清分机积灰条件下的钞票识别方法及清分机
CN104899965B (zh) * 2015-05-15 2017-06-23 浙江大学 一种基于清分机的多国纸币序列号识别方法
CN104916037A (zh) * 2015-06-26 2015-09-16 深圳市中钞信达金融科技有限公司 一种纸币冠字号识别装置及方法
WO2017079913A1 (zh) * 2015-11-11 2017-05-18 深圳怡化电脑股份有限公司 一种钞票版本识别的方法和装置
CN105654609A (zh) * 2015-12-29 2016-06-08 深圳怡化电脑股份有限公司 纸币处理方法及系统
CN108074323B (zh) * 2016-11-15 2021-01-01 深圳怡化电脑股份有限公司 一种纸币面向的识别方法及其装置
CN106875540B (zh) * 2017-02-17 2020-04-07 深圳怡化电脑股份有限公司 一种纸币厚度异常检测方法及装置
CN107016363B (zh) * 2017-03-30 2020-06-05 理光图像技术(上海)有限公司 票据图像管理装置、票据图像管理系统以及方法
CN108665606B (zh) * 2017-03-30 2021-02-02 深圳怡化电脑股份有限公司 一种纸币面向信息的识别方法及装置
CN107025716B (zh) * 2017-06-05 2019-12-10 深圳怡化电脑股份有限公司 检测纸币冠字号污损的方法及装置
CN107610319A (zh) * 2017-08-07 2018-01-19 上海灵岳电子设备有限公司 基于单接触式图像传感器的纸币信息识别装置及方法
CN108062822A (zh) * 2017-12-22 2018-05-22 深圳怡化电脑股份有限公司 存取款交易信息的记录方法、装置、终端设备和存储介质
CN108665603B (zh) * 2018-04-11 2020-06-09 深圳怡化电脑股份有限公司 识别纸币币种的方法、装置及电子设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0629926A2 (en) * 1993-06-15 1994-12-21 Sharp Kabushiki Kaisha Image reading apparatus and image processor incorporating the same
CN102222384A (zh) * 2011-05-27 2011-10-19 尤新革 纸币多光谱图像分析方法
CN103021069A (zh) * 2012-11-21 2013-04-03 深圳市兆图电子有限公司 高速票据图像采集处理系统及其采集处理方法
CN103456075A (zh) * 2013-09-06 2013-12-18 广州广电运通金融电子股份有限公司 一种纸币处理方法及装置

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0881603B1 (en) * 1996-01-25 2005-11-16 SANYO ELECTRIC Co., Ltd. Judging method of sheets, notes, etc. for forgery, and judging method of insertion direction of them
JP4634073B2 (ja) * 2004-06-29 2011-02-16 マミヤ・オーピー株式会社 紙葉類識別装置及び識別方法
CN101968903B (zh) * 2010-09-21 2012-08-29 广州广电运通金融电子股份有限公司 有价文件识别方法及系统
KR101865415B1 (ko) * 2011-05-11 2018-06-07 삼성전자주식회사 모바일 통신기를 이용하는 위폐 감별법
CN102682514B (zh) * 2012-05-17 2014-07-02 广州广电运通金融电子股份有限公司 一种纸类识别方法及相关装置
CN202600828U (zh) * 2012-06-25 2012-12-12 金峰 纸币智能数码影像光变特性黑白数据成像装置
CN102890841B (zh) * 2012-10-08 2013-11-27 广州广电运通金融电子股份有限公司 一种有价文件鉴别方法及装置
CN103049750B (zh) * 2013-01-11 2016-06-15 广州广电运通金融电子股份有限公司 字符识别方法
CN103279735B (zh) * 2013-04-24 2017-07-25 广州广电运通金融电子股份有限公司 一种金融票据识别模块中积灰检测方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0629926A2 (en) * 1993-06-15 1994-12-21 Sharp Kabushiki Kaisha Image reading apparatus and image processor incorporating the same
CN102222384A (zh) * 2011-05-27 2011-10-19 尤新革 纸币多光谱图像分析方法
CN103021069A (zh) * 2012-11-21 2013-04-03 深圳市兆图电子有限公司 高速票据图像采集处理系统及其采集处理方法
CN103456075A (zh) * 2013-09-06 2013-12-18 广州广电运通金融电子股份有限公司 一种纸币处理方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106683264A (zh) * 2017-02-20 2017-05-17 深圳怡化电脑股份有限公司 一种污损冠字号的识别方法及装置

Also Published As

Publication number Publication date
CN103456075B (zh) 2015-11-25
CN103456075A (zh) 2013-12-18

Similar Documents

Publication Publication Date Title
WO2015032187A1 (zh) 一种纸币处理方法及装置
CN105957238B (zh) 一种纸币管理方法及其系统
CN110598699B (zh) 一种基于多光谱图像的防伪票据鉴伪系统和方法
US8625875B2 (en) Document imaging and processing system for performing blind balancing and display conditions
US9014459B2 (en) Identification method for valuable file and identification device thereof
US9652658B1 (en) Fingerprint check to reduce check fraud
CN103279735B (zh) 一种金融票据识别模块中积灰检测方法及系统
CN107610322B (zh) 纸币版本的识别方法、装置、电子设备以及存储介质
WO2016018987A1 (en) Detecting specified image identifiers on objects
WO2016037523A1 (zh) 一种基于清分机积灰条件下的钞票识别方法及清分机
CN106952393B (zh) 纸币识别方法和装置、电子设备和存储介质
CN107180479A (zh) 一种票据鉴别方法、装置、设备和存储介质
CN108320373A (zh) 一种纸币防伪标识的检测的方法及装置
CN106683261B (zh) 一种纸币的鉴伪方法和装置
CN105957237A (zh) 一种纸币的版本识别方法和装置
Amirsab et al. An automated recognition of fake or destroyed Indian currency notes
CN101964123A (zh) 纸币验钞模块可以识别钞号的方法
CN110969757A (zh) 一种多国钞票类型快速识别技术
Chavan et al. Design and implementation of automatic coin dispensing machine
CN107240184A (zh) 一种塑料币版本识别的方法、装置及设备
Shiby et al. Detection of Fake Currency Using Image Processing
CN106408748A (zh) 一种检测新版人民币的方法及装置
CN105118139A (zh) 一种鉴钞方法及带有图像采集及运算装置的鉴别仪
CN106570980A (zh) 一种纸币鉴伪的方法和系统

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: 14842400

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: 14842400

Country of ref document: EP

Kind code of ref document: A1