WO2011127808A1 - 硬币图像识别方法及装置 - Google Patents

硬币图像识别方法及装置 Download PDF

Info

Publication number
WO2011127808A1
WO2011127808A1 PCT/CN2011/072642 CN2011072642W WO2011127808A1 WO 2011127808 A1 WO2011127808 A1 WO 2011127808A1 CN 2011072642 W CN2011072642 W CN 2011072642W WO 2011127808 A1 WO2011127808 A1 WO 2011127808A1
Authority
WO
WIPO (PCT)
Prior art keywords
coin
image
baffle
template
matching
Prior art date
Application number
PCT/CN2011/072642
Other languages
English (en)
French (fr)
Inventor
唐昌
胡剑峰
李龙
朱东飞
范金富
汤石男
申香梅
廖东玲
彭勤勤
朱柱锦
Original Assignee
高新现代智能系统股份有限公司
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 高新现代智能系统股份有限公司 filed Critical 高新现代智能系统股份有限公司
Publication of WO2011127808A1 publication Critical patent/WO2011127808A1/zh

Links

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D5/00—Testing specially adapted to determine the identity or genuineness of coins, e.g. for segregating coins which are unacceptable or alien to a currency
    • G07D5/005—Testing the surface pattern, e.g. relief
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/0006—Industrial image inspection using a design-rule based approach
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D5/00—Testing specially adapted to determine the identity or genuineness of coins, e.g. for segregating coins which are unacceptable or alien to a currency
    • G07D5/02—Testing the dimensions, e.g. thickness, diameter; Testing the deformation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image

Definitions

  • the present invention relates to the technical field of coin image recognition, and more particularly to a method and apparatus for recognizing a dynamic coin image using image pattern matching techniques. Background technique
  • the core unit of the coin processing module of the automatic coin vending machine in the domestic coin-operated vending machine and the rail-track automatic ticket-checking system is a coin discriminator, and the main identification indicators of the coin discriminator are: diameter, material and thickness.
  • the coin discriminator In recent years, there has been a market currency, counterfeit currency, etc., which are very similar or identical in material, size and weight to RMB coins, but which are obviously inconsistent in appearance.
  • the current coin discriminator it is possible to accurately identify fake coins with different materials but similar appearances, but it is difficult to identify materials, coins, counterfeit coins, etc. that are close in appearance. In response to this situation, the coin discriminator is improved, and the identification parameter is adjusted.
  • the recognition rate of the fake coin can reach about 90%, but at the same time, because the coin is in motion during the recognition process, the difficulty of recognition is increased, and the rejection rate of the real coin is increased. It has also risen by about 2%, which still does not completely solve the problem of mis-collection of fake coins.
  • Image matching technology is a very important technology in the field of modern information processing, especially in the field of image information processing.
  • Image matching is to select certain features, similarity criteria, and search strategies based on reference images and real-time images to perform related operations to determine the best spatial correspondence point for matching.
  • feature space, similarity measure and search strategy are three main research problems.
  • the key to image matching is to determine an effective matching method, which requires high matching probability, small error, fast speed and good timeliness.
  • Image matching methods are generally divided into two categories: gray-based matching methods and feature-based matching methods.
  • gray-scale matching The basic idea of gray-scale matching is to treat the image as a two-dimensional signal from a statistical point of view, and use statistical correlation methods to find correlations between signals. Using the correlation functions of the two signals, their similarities are evaluated to determine the point of the same name.
  • the most classic gray matching method is the normalized gray matching method.
  • the basic principle is to pixel-by-pixel a gray matrix of a real-time image window with a certain size, and all possible window grayscale arrays of the reference image.
  • the matching method of searching and comparing according to a similarity measure method theoretically adopts image correlation technology.
  • Feature matching refers to an algorithm that performs parameter matching by extracting features (points, lines, faces, etc.) of two or more images, and then using the described parameters to perform matching.
  • Common features include point features, edge features, and region features.
  • the technical problem to be solved by the present invention is to avoid the above-mentioned deficiencies of the prior art, and to provide an automatic detection of fake coins, which uses image pattern matching technology to identify dynamic target images, and rejects the same material, weight and size of RMB coins.
  • a coin image recognition method and apparatus for fake coins having different appearances.
  • the technical solution adopted by the present invention to solve the above technical problem is a coin image recognition method, which comprises the following steps:
  • step (8) Determine whether the geometric feature matching is successful. If the matching is successful, go directly to step (11). If the matching is not successful, enter step (9);
  • the template library is extracted as follows:
  • Extracting and calculating the diameter of the outer contour circle of the coin in the image is extracted as follows:
  • the geometric feature matching is performed by the outer contour circle geometric feature template, and the diameter of the circle is calculated. Extracting the region of interest of the target image by extracting as follows - determining the center of the circle by the extracted outer contour circle of the coin;
  • the center of the circle is a square heart
  • the target image is obtained by cropping the image with the actual size of the coin in the image plus a 10-pixel square length.
  • An apparatus for implementing the coin image recognition method of claim 1, the apparatus comprising:
  • a coin channel module a channel and an image imaging area for connecting the coin slot to the coin discriminator
  • An image acquisition module configured to trigger an image captured by the coin currency module
  • the image processing module is configured to process the image captured by the image acquisition module in real time.
  • the coin channel module is composed of a first baffle, a second baffle, a third baffle, a fourth baffle, a first baffle and a second baffle, wherein the first baffle and the second baffle are arranged up and down Forming a coin channel, the first baffle and the second baffle are wedge-shaped near the coin channel, and the third baffle and the fourth baffle are respectively located on the left and right sides of the coin channel, and the first baffle and the second baffle are respectively disposed on On the front and rear sides of the coin channel, the first baffle is provided with a rectangular coin slot, and the second baffle is provided with a rectangular coin outlet, and the two ends of the coin channel are respectively connected with the coin slot and the coin outlet.
  • the third baffle and the fourth baffle are glass partitions.
  • the image collection module is composed of a sensor, a camera, a lens and a light source, wherein the sensor is disposed on the coin path near the coin slot, the light source is located between the coin channel module and the camera, and the center distance of the light source coincides with the lens center distance.
  • the sensor is an on-beam sensor.
  • the image processing module is composed of a PC with image processing software.
  • the coin image recognition method and apparatus of the present invention provides the possibility of accurately distinguishing fake coins having different diameters, thicknesses, and materials while having different patterns.
  • FIG. 1 is a schematic flow chart of a coin image recognition method according to the present invention.
  • FIG. 2 is a schematic block diagram of an embodiment of a coin image recognition apparatus according to the present invention.
  • FIG. 3 is a schematic structural view of an embodiment of a coin image recognition apparatus according to the present invention.
  • FIG. 4 is a schematic view of a coin slot device of an embodiment of the coin image recognition device of the present invention.
  • Figure 5 is a cross-sectional view showing the structure of an embodiment of the coin image discriminating device of the present invention, specifically, a cross-sectional view taken along line A-A of Figure 4; detailed description
  • the coin image recognition method of the present invention comprises the following steps - step S101, extracting coin image features, and acquiring a template library.
  • the rational selection of templates is the key to template matching.
  • the present invention is described by taking an example for identifying RMB one yuan and five corner coins.
  • the RMB issuance version has four versions of a new one-yuan coin, an old one-yuan coin, a new five-corner coin, and an old five-corner coin.
  • At least 8 geometric feature templates are constructed and masked as MODEL 1 by masking techniques, and at least 2 grayscale feature templates are constructed according to different materials of 1 and 5 angle coins. Marked as MODEL 2; Next, create the outer contour geometric template of the one-dollar coin and the five-corner coin, respectively, and mark it as M0DEL-3.
  • the above template library can be increased with the increase of the coin new coin, or with the increase of the number of different face values of the coin.
  • Step S102 importing a template into an image recognition system and performing preprocessing.
  • the geometric feature template and the gray feature template are added to the coin image recognition system, the recognition parameters of the template are adjusted, the template has a rotation size invariance, and has a scale conversion characteristic, and the scale size can be selected according to requirements, and the preferred ratio in this embodiment 1 ⁇
  • the scale is between 0. 9 ⁇ 1.
  • Step S103 collecting an image and transmitting the image.
  • the camera accepts a sensor trigger signal to collect an image of the coin moving in the coin track and transmits it to the image recognition system via the USB data line.
  • Step S104 extracting and calculating the diameter of the outer contour circle of the coin in the image.
  • the outer contour of the coin is highlighted in white against the background.
  • the geometric feature matching is performed by the above MODEL 3 template, and the outer contour circle of the coin is found in a short time, and the diameter of the outer contour circle is calculated.
  • Step S105 it is judged whether the diameter of the circle extracted by the step S104 is within the allowable diameter range.
  • the conditional judgment formula is: 0. 9D ⁇ d ⁇ l . lD (where D is an empirical value of the outer diameter of the 5-angle or 1-yuan coin, d is the diameter of the outer contour circle extracted in step S104), if d is in a given range If it is considered that the basic characteristics of the coin are met, the process proceeds to step S106; otherwise, it is considered to be a fake coin and is rejected, and the process proceeds directly to step S1012.
  • Step S106 extracting a region of interest of the target image.
  • the center of the coin is determined by the outer contour circle of the coin extracted in step S104, and the center of the circle is centered from the original image.
  • the actual size of the coin in the image is added by 10 pixels square length. Crop the image to get the target image. If the crop area exceeds the original image boundary, the crop area is translated according to the opposite direction of the border beyond the side of the crop area on the border of the image.
  • Step S107 the geometric feature template is matched.
  • the geometric template matching is performed on the clipped target image in step S106 by the template library M0DEL-1, and the rotation angle, the scale factor and the matching score of the geometric template matching are constrained by the geometric feature matching, and the matching accuracy is ensured. Minimize the matching time.
  • Step S108 Determine whether the geometric feature matching is successful.
  • the score of the geometric feature template matching is obtained by step S107, and the judgment is made by the formula Geometry_M Match_SC0>60 (Geometry_Jatch strigSC0RE indicates the geometric feature template matching score, the percentage system is used, 60 is the empirical value, and can be adjusted according to needs), if the score exceeds 60 points, the matching is successful, the system considers it to be a renminbi coin, and proceeds to step S1011. If the score is lower than 60 points, the matching fails, and the system considers that it is necessary to continue the authentication, and proceeds to step S109.
  • Step S109 the grayscale feature template is matched.
  • the grayscale feature matching is performed on the clipped target image in step S106 by the template library M0DEL-2, and the rotation angle of the grayscale feature matching and the matching score of the grayscale template matching are constrained, and the matching accuracy is ensured. It is possible to shorten the matching time.
  • step S1010 it is determined whether the gradation feature matching is successful.
  • the score of the grayscale feature template matching is obtained by step S109, and the grayscale feature template matching score is one hundred percent by the Gray Gray Jfatch_SC0RE>45 (Gray_Match_SCORE) System, 45 is the experience value, can be adjusted according to needs), if the score exceeds 45 points, the match is successful, the system considers it to be a RMB coin, and proceeds to step S1011. If the score is lower than 45 points, the match fails, and the system considers it to be a fake coin such as a game currency, and proceeds to step S1012.
  • Step S101 1 judge the RMB currency. It is judged whether it belongs to a 5-point coin or a 1-yuan coin by the matching of steps S108 and S1010, and counts the number of coins of the 5-corner coin and the 1-yuan coin, respectively.
  • Step S1012 saving the recognition result. Save the recognition result to the identification library, where the RMB coin is marked as 1 and the non-RMB coin is marked as 0 for other modules to call in real time.
  • the present invention further provides a coin image recognition device, as shown in FIG. 2 to FIG. 5, the device includes: a channel for connecting a coin slot and a coin discriminator, and an image imaging region.
  • the currency channel module 20 includes: an image acquisition module 30 for triggering an image acquired through the coin bank module; and an image processing module 40 for real-time processing of the image captured by the image collection module.
  • the currency channel module 20 is composed of a first baffle 201, a second baffle 202, a third baffle 203, a fourth baffle 204, a first baffle 205 and a second baffle 206, wherein the first baffle 205
  • the second partition plate 206 is disposed above and below the constituent coin channel 209.
  • the first partition plate 205 and the second partition plate 206 are wedge-shaped near the coin channel 209, and the third baffle plate 203 and the fourth baffle plate 204 are respectively located at the left and right sides of the coin channel 209.
  • the first baffle 201 and the second baffle 202 are respectively disposed on the front and rear sides of the coin 209.
  • the first baffle 201 is provided with a rectangular coin slot 207
  • the second baffle 202 is provided with a rectangular coin.
  • both ends of the coin path 209 are connected to the coin slot 207 and the coin slot 208, respectively.
  • the third baffle 203 and the fourth baffle 204 are preferably thin glass sheets as baffles.
  • the first partition 205 and the second partition 206 constituting the coin passage 209 are thinned toward the position of the coin passage, which is equivalent to the thickness of the coin.
  • the coin width is slightly larger than the thickness of all RMB coins, D « 1. 3d, where D is the width of the coin and d is the thickness of the coin.
  • the coin channel is too wide, the angle of the coin is too large in the coin channel, and the reflection phenomenon is easy to occur locally, and the recognition reliability is reduced.
  • the coin channel is too narrow, the coin inclination is large, and the coin currency phenomenon is easy to occur. Therefore, the appropriate coin width is reasonably selected. It is also an important factor in image imaging.
  • the first partition 205 and the second partition 206 which form the upper and lower sides of the coin channel are wedge-shaped or thinned near the coin channel, which can increase the uniform illumination of the light in the coin channel, and increase The image contrast and the features on the image are more pronounced, which improves the recognition rate of coin image recognition.
  • the image acquisition module is composed of a sensor 301, a camera 302, a lens 303, and a light source 304.
  • the sensor 301 is located on the coin 209 near the coin slot 207, and the light source 304 is located between the currency module 20 and the camera 302.
  • the center distance of the light source coincides with the center distance of the lens.
  • the sensor 301 uses an on-beam type sensor. When the coin passes the sensor, the sensor generates a trigger signal and sends a signal to the camera to trigger the camera to capture the image.
  • Camera 302 is an industrial camera.
  • the light source 304 can be selected from high-brightness red-emitting LEDs.
  • the main applications of the ring light are edge detection, wafer scratches and stains on the surface of the wafer, metal, CD, and reading of the imprinted text.
  • the inner diameter of the light source satisfies the field of view of the coin image.
  • the coin inserts the coin from the coin slot 207 into the coin lane, the coin generates a trigger signal through the sensor 301, and the camera 302 accepts the trigger signal.
  • the present invention considers that the collected image is mostly located in the middle of the field of view formed by the light source. The position is guaranteed by setting the camera trigger delay time, and the captured image is transmitted to the image processing module 40 for processing through the USB interface data line.
  • Image Processing Module 40 I I I consists of a PC with image processing software. It accepts image data transmitted through the USB interface, and processes each frame of the image in real time, and stores the processing result in a dynamic library for the next link.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Geometry (AREA)
  • Testing Of Coins (AREA)

Description

硬币图像识别方法及装置 技术领域
本发明涉及硬币图像识别的技术领域，特别涉及采用图像模式匹配技术对处于动态的 硬币图像进行识别的方法和装置。 背景技术
国内现有投币自动售货机和轨道交通自动售检票系统中自动售票机的硬币处理模块 的核心单元是硬币鉴别器， 而硬币鉴别器的主要识别指标是： 直径、 材质和厚度等。 近年 来市场上出现了一种材质、 大小和重量与人民币硬币非常接近或相同， 但是外观明显不一 致的游戏币、假币等异币。 使用目前的硬币鉴别器可准确识别材质不同但外观相似的假硬 币， 但对于材质、 外型接近的游戏币、 假币等识别有困难。 针对此情况， 对硬币鉴别器进 行改进， 调整识别参数， 假硬币的识别率可达约 90%, 但同时因识别的过程中硬币处于运 动状态， 增加了识别的难度， 真币的拒收率也上升了约 2%左右， 这样仍不能完全解决假硬 币的误收问题。
图像匹配技术是近代信息处理， 特别是图像信息处理领域中极为重要的技术。 图像匹 配就是要根据参考图像和实时图像来选定某些特征、 相似性准则及搜索策略进行相关运 算， 以确定匹配的最佳空间对应点。 在图像匹配技术中， 特征空间、 相似性度量和搜索策 略是三个主要研究问题， 图像匹配关键是要确定有效的匹配方法， 要求匹配概率高、 误差 小、速度快且适时性好。 图像匹配的方法一般分为基于灰度的匹配方法和基于特征的匹配 方法两大类。
灰度匹配的基本思想就是以统计的观点将图像看成是二维信号，釆用统计相关的方法 寻找信号间的相关匹配。 利用两个信号的相关函数， 评价它们的相似性以确定同名点。 最 经典的灰度匹配法是归一化的灰度匹配法，其基本原理是逐像素的把一个以一定大小的实 时图像窗口的灰度矩阵， 与参考图像的所有可能的窗口灰度阵列， 按某种相似性度量方法 进行搜索比较的匹配方法， 从理论上说就是采用图像相关技术。
特征匹配是指通过分别提取两个或多个图像的特征（点、 线、 面等特征）， 对特征进行 参数描述， 然后运用所描述的参数来进行匹配的一种算法。 常见的特征包括点特征、 边缘 特征和区域特征等。
比较上述两种图像匹配方法： 灰度匹配的计算量较大； 特征匹配的计算量较小， 对位 置变化敏感， 可大大提高匹配的精确程度， 特征点的提取过程可以减少噪声的影响， 对灰 度变化、 图像形变以及局部遮挡等都有较好的适应能力。 在实际应用中， 应根据具体情况 选择其中一种方法或者两者结合的方法。
鉴于硬币鉴别器不能完全解决假硬币识别问题， 因此， 在原有硬币鉴别器识别的基础 上增加图像识别装置， 专门用于区分材质、 重量和大小相同而外观明显不同的假硬币， 具 有广泛而重要的应用前景。 发明内容
本发明要解决的技术问题在于避免上述现有技术的不足之处，而提出一种对假硬币自 动检测，采用图像模式匹配技术对动态目标图像进行识别，剔除与人民币硬币材质、重量、 大小相同但外观不同的假硬币的硬币图像识别方法及装置。
本发明解决上述技术问题采用的技术方案是， 一种硬币图像识别方法，该方法包 括如下步骤：
(1) 提取硬币图像特征， 获取模板库；
(2) 导入模板到图像识别装置中并进行预处理；
(3) 采集图像并传输图像数据；
(4) 提取并计算图像中的硬币外轮廓圆直径；
(5) 判断提取的硬币外轮廓圆直径是否在允许的直径范围内， 若在给定范围内， 则 认为符合硬币基本特征， 直接进入步骤 （6) 若不在给定范围内， 则认为是假硬币， 进入 步骤 （12) ；
(6) 提取目标图像的感兴趣区域；
(7) 几何特征模板匹配；
(8) 判断几何特征匹配是否成功， 若匹配成功则直接进入步骤 （11) ， 若匹配不成 功则进入歩骤 (9) ；
(9) 灰度特征模板匹配；
(10) 判断灰度特征匹配是否成功；
(11) 判断硬币的面值， 并依面值的不同进行统计；
(12) 保存识别结果。
所述模板库按以下方法提取：
分别提取人民币硬币的正反面的几何特征；
通过掩膜技术忽略次要特征；
构建几何特征模板；
' 根据硬币材质不同构建灰度特征模板；
分别制作硬币的外轮廓圆几何特征模板。
所述的导入模板到图像识别装置中并进行预处理， 其方法为- 将几何特征模板和灰度特征模板加入硬币图像识别装置中；
调整模板的识别参数， 使模板具有旋转尺寸不变和比例变换特性。
所述的提取并计算图像中的硬币外轮廓圆直径， 按如下方法提取：
通过外轮廓圆几何特征模板进行几何特征匹配， 并计算该圆直径。 所述的提取目标图像的感兴趣区域， 按如下方法提取- 通过提取的硬币外轮廓圆确定圆心；
从原图像中以圆心为正方形形心；
以硬币在图像中的实际大小外加 10个像素正方形边长剪裁图像， 获取目标图像。 一种实现权利要求 1所述硬币图像识别方法的装置， 该装置包括：
币道模块， 用于硬币投币口与硬币鉴别器连接的通道和图像成像区域；
图像采集模块， 用于触发采集通过硬币币道模块的图像；
图像处理模块， 用于实时处理通过图像采集模块触发采集的图像。
所述的币道模块由第一挡板、 第二挡板、 第三挡板、 第四挡板、 第一隔板和第二隔板 组成， 其中第一隔板和第二隔板上下设置组成币道， 第一隔板和第二隔板靠近币道处为楔 形， 第三挡板和第四挡板分别位于币道的左右两侧， 第一挡板和第二挡板分别设于币道的 前后两侧， 第一挡板上设有一长方形投币口， 第二挡板上设有一长方形出币口， 币道的两 端分别与投币口和出币口相连接。
所述的第三挡板和第四挡板为玻璃隔板。
所述的图像釆集模块由传感器、 相机、 镜头和光源组成， 其中传感器设于币道上接近 投币口处， 光源位于币道模块与相机之间， 光源的中心距与镜头中心距重合。
所述的传感器为对射传感器。
所述的图像处理模块由带有图像处理软件的 PC组成。
同现有技术相比较， 本发明的硬币图像识别方法及装置为实现准确区分直径、 厚度和 材质相同而图案不同的假硬币提供了可能。 附图说明
图 1为本发明硬币图像识别方法的流程示意图；
图 2为本发明硬币图像识别装置实施例的原理框图；
图 3为本发明硬币图像识别装置实施例的结构示意图；
图 4为本发明硬币图像识别装置实施例的投币道装置的示意图；
图 5为本发明硬币图像识别装置实施例的结构剖视图，具体为图 4沿 A-A向的剖视图。 具体实施方式
下面结合说明书附图和具体实施方式对本发明进一步解释和说明。
参见图 1所示， 本发明的硬币图像识别方法包括如下步骤- 步骤 S101 , 提取硬币图像特征， 获取模板库。考虑人民币硬币币种的多样性和人民币 硬币的磨损及新旧程度， 合理选择模板是进行模板匹配的关键。 在本实施例中以用于识别 人民币一元和 5角硬币为例对本发明予以说明， 目前人民币发行版本有新版一元硬币、 老 版一元硬币、新版 5角硬币和老版 5角硬币四个版本， 分别提取 4个版本硬币正反面的几 何特征， 根据统计学特点， 通过掩膜技术忽略次要特征， 构建至少 8个几何特征模板， 并 标记为 MODEL 1 ; 同时， 根据 1元和 5角硬币材质不同构建至少 2个灰度特征模板， 标记 为 MODEL 2 ; 接着， 分别制作一元硬币和 5 角硬币的外轮廓圆几何特征模板， 并标记为 M0DEL-3。 以上模板库可随着硬币新版币的增加而增加 , 也可随硬币不同面值数量的增加 而增加。
歩骤 S102 , 导入模板到图像识别系统中并进行预处理。将几何特征模板和灰度特征模 板加入硬币图像识别系统中， 调整模板的识别参数， 使模板具有旋转尺寸不变性， 并具有 比例变换特性， 比例尺寸可根据需要选择， 本实施例中优选的比例尺度在 0. 9〜1. 1之间。
歩骤 S103 ,采集图像并传输图像。相机接受传感器触发信号釆集在币道内运动的硬币 的图像， 并通过 USB数据线传送至图像识别系统中。
步骤 S104, 提取并计算图像中的硬币外轮廓圆直径。在环形光照射下， 硬币外轮廓与 背景对比成白色高亮显示， 通过上述 MODEL 3模板进行几何特征匹配， 在较短的时间内寻 找硬币的外轮廓圆， 并计算该外轮廓圆直径。
步骤 S105, 判断歩骤 S104提取的圆直径是否在允许的直径范围内。 通过条件判断公 式： 0. 9D<d<l . lD (其中， D为 5角或 1元硬币的外径的经验值， d为步骤 S104提取的外 轮廓圆直径） ， 如果 d在给定范围内， 认为符合硬币基本特征， 进入步骤 S106; 否则， 则 认为是假硬币并剔除， 直接进入步骤 S1012。
步骤 S106， 提取目标图像的感兴趣区域。 通过步骤 S104所提取的硬币外轮廓圆确定 圆心， 从原图像中以圆心为正方形形心， 为了减少匹配时产生的边缘效应的影响， 以硬币 在图像中的实际大小外加 10个像素正方形边长剪裁图像， 获得目标图像。 如果剪裁区域 超过原图像边界，则根据超越边界的反方向平移剪裁区域直至剪裁区域的一边在图像的边 界上。
步骤 S107 , 几何特征模板匹配。 通过上述模板库 M0DEL-1对步骤 S106的剪裁后的目 标图像进行几何特征模板匹配， 约束几何特征匹配的旋转角度、 比例系数和几何模板匹配 的匹配得分数， 在保证匹配准确率的情况下， 尽可能缩短匹配时间。
步骤 S108 , 判断几何特征匹配是否成功。 通过步骤 S107获得几何特征模板匹配的得 分， 通过公式 Geometry— Match— SC0RE>60 (Geometry_Jatch„SC0RE表示几何特征模板匹配 得分， 采用百分制， 60为经验值， 可根据需要调整)进行判断， 如果得分超过 60分， 匹配 成功， 系统认为是人民币硬币， 进入步骤 S1011。 如果得分低于 60分， 匹配失败， 系统认 为还需继续鉴别， 进入步骤 S109。
步骤 S109, 灰度特征模板匹配。 通过上述模板库 M0DEL-2对步骤 S106的剪裁后的目 标图像进行灰度特征匹配， 约束灰度特征匹配的旋转角度和灰度模板匹配的匹配得分数， 在保证匹配准确率的情况下， 尽可能缩短匹配时间。
步骤 S1010，判断灰度特征匹配是否成功。通过步骤 S109获得灰度特征模板匹配的得 分， 通过公 Gray Jfatch—SC0RE〉45 (Gray_Match— SCORE表示灰度特征模板匹配得分一百分 制， 45为经验值， 可根据需要调整)判断， 如果得分超过 45分， 匹配成功， 系统认为是人 民币硬币，进入步骤 S1011。如果得分低于 45分， 匹配失败，系统认为是游戏币等假硬币， 进入歩骤 S1012。
歩骤 S101 1 ,判断人民币币种。通过步骤 S108和 S1010的匹配判断是属于 5角硬币还 是 1元硬币， 并分别统计 5角硬币和 1元硬币的投币数。
歩骤 S1012, 保存识别结果。 保存识别结果至识别库中， 其中人民币硬币标记为 1 , 非人民币硬币标记为 0， 供其他模块实时调用。
相应于上面的方法实施例， 本发明还提供一种硬币图像识别装置， 参见图 2〜图 5 所示， 该装置包括： 用于硬币投币口与硬币鉴别器连接的通道和图像成像区域的币道模块 20；用于触发采集通过硬币币道模块的图像的图像采集模块 30;用于实时处理通过图像釆 集模块触发采集的图像的图像处理模块 40。
所述币道模块 20由第一挡板 201、 第二挡板 202、 第三挡板 203、 第四挡板 204、 第 一隔板 205和第二隔板 206组成，其中第一隔板 205和第二隔板 206上下设置组成币道 209， 第一隔板 205和第二隔板 206靠近币道 209处为楔形，第三挡板 203和第四挡板 204分别 位于币道 209的左右两侧， 第一挡板 201和第二挡板 202分别设于币道 209的前后两侧， 第一挡板 201上设有一长方形投币口 207 , 第二挡板 202上设有一长方形出币口 208， 币 道 209的两端分别与投币口 207和出币口 208相连接。为了保证硬币在币道内成像清晰并 且为了获得较高的对比度， 且考虑挡板的耐磨性和透光性， 第三挡板 203和第四挡板 204 优选薄玻璃片作为挡板。 币道 209两侧的玻璃挡板越薄， 环形光离图像的距离越近， 采集 的图像的对比度越好， 图像的表面特征越清晰， 识别准确率和速度越高， 因此合理选择合 适的透光材料， 是图像成像好坏的一个重要因素。考虑币道的透光性， 如图 4、 图 5所示， 组成币道 209的第一隔板 205和第二隔板 206在接近币道位置厚度变薄，与硬币厚度相当。 币道宽度稍大于所有人民币硬币的厚度， D« 1. 3d， 其中， D为币道宽度， d为人民币硬币 的厚度。 币道太宽， 硬币在币道内倾斜角度过大， 局部容易发生反光现象， 识别可靠性减 低； 币道过窄， 硬币倾斜度大， 容易发生卡币现象， 因此， 合理选择合适的币道宽度， 也 是图像成像好坏的一个重要因素。
如图 4、 图 5所示， 组成币道的上下的第一隔板 205和第二隔板 206在靠近币道处做 成楔形或者切薄， 可增加光在币道内均匀照亮， 增大了图像对比度和使图像上的特征更加 明显， 能提高硬币图像识别的识别率。
所述的图像采集模块由传感器 301、 相机 302、 镜头 303和光源 304组成， 其中传感 器 301位于币道 209上靠近投币口 207处， 光源 304位于币道模块 20与相机 302之间， 其中， 光源的中心距与镜头中心距重合。 传感器 301釆用对射型传感器， 硬币通过传感器 时， 传感器产生触发信号， 同时将信号发送至相机， 触发相机采集图像。 相机 302为工业 相机。光源 304可选用高亮度发红光 LED, 考虑环形光主要用途为边缘检测、 晶片、 金属、 CD的表面划痕和污点、 读取刻印文字， 光源的内径满足硬币成像的视野范围。 参见图 2，投币者将硬币从投币口 207投入币道，硬币通过传感器 301产生触发信号， 相机 302 接受触发信号， 本发明考虑釆集的图像绝大部分位于光源形成的视场的中间位 置， 通过设置相机触发延迟时间来保障， 采集的图像通过 USB接口数据线传送到图像处理 模块 40进行处理。
图像处理模块 40 I I I带有图像处理软件的 PC组成，接受通过 USB接口传输过来的图像 数据， 并对采集的每帧图像进行实时处理， 将处理结果存入动态库中供下一环节调用。
以上所述仅为本发明的优选实施例而己， 并不用于限制本发明， 对于本领域的技术人 员来说， 本发明可以有各种更改和变化。 如可根据需识别的币种不一样而做适应性修改， 比如说也可用于识别其他国家或地区的硬币币种或是用于游戏机中用于识别游戏币与非 游戏币。 凡在本发明的精神和原则之内， 所作的任何修改、 等同替换、 改进等， 均应包含 在本发明的权利要求范围之内。

Claims

权利要求
1、 一种硬币图像识别方法， 其特征在于， 该方法包括如下步骤：
(1) 提取硬币图像特征， 获取模板库；
( 2 ) 导入模板到图像识别装置中并进行预处理；
(3) 采集图像并传输图像数据；
(4) 提取并计算图像中的硬币外轮廓圆直径；
(5) 判断提取的硬币外轮廓圆直径是否在允许的直径范围内， 若在给定范围内， 则 认为符合硬币基本特征， 直接进入步骤 （6) 若不在给定范围内， 则认为是假硬币， 进入 步骤 （12) ；
(6) 提取目标图像的感兴趣区域；
(7) 几何特征模板匹配；
(8) 判断几何特征匹配是否成功， 若匹配成功则直接进入步骤 （11) ， 若匹配不成 功则进入歩骤 （9) ；
(9) 灰度特征模板匹配；
(10) 判断灰度特征匹配是否成功；
(11) 判断硬币的面值， 并依面值的不同进行统计；
(12) 保存识别结果。
2、 根据权利要求 1所述的硬币图像识别方法， 其特征在于， 所述模板库按以下方法 提取：
分别提取人民币硬币的正反面的几何特征；
通过掩膜技术忽略次要特征 ·，
构建几何特征模板；
根据硬币材质不同构建灰度特征模板；
分别制作硬币的外轮廓圆几何特征模板。
3、 根据权利要求 1所述的硬币图像识别方法， 其特征在于， 所述的导入模板到图像 识别装置中并进行预处理， 其方法为：
将几何特征模板和灰度特征模板加入硬币图像识别装置中；
调整模板的识别参数， 使模板具有旋转尺寸不变和比例变换特性。
4、 根据权利要求 1所述的硬币图像识别方法， 其特征在于， 所述的提取并计算图像 中的硬币外轮廓圆直径， 按如下方法提取：
通过外轮廓圆几何特征模板进行几何特征匹配， 并计算该圆直径。
5、 根据权利要求 1所述的硬币图像识别方法， 其特征在于， 所述的提取目标图像的 感兴趣区域， 按如下方法提取：
通过提取的硬币外轮廓圆确定圆心； 从原图像中以圆心为正方形形心；
以硬币在图像中的实际大小外加 10个像素正方形边长剪裁图像， 获取目标图像。
6、 一种实现权利要求 1所述硬币图像识别方法的装置， 其特征在于， 该装置包括： 币道模块， 用于硬币投币口与硬币鉴别器连接的通道和图像成像区域；
图像采集模块， 用于触发采集通过硬币币道模块的图像；
图像处理模块， 用于实时处理通过图像采集模块触发釆集的图像。
7、 根据权利要求 6所述的硬币图像识别装置， 其特征在于， 所述的币道模块由第一 挡板、 第二挡板、 第三挡板、 第四挡板、 第一隔板和第二隔板组成， 其中第一隔板和第二 隔板上下设置组成币道，第一隔板和第二隔板靠近币道处为楔形，第三挡板和第四挡板分 别位于币道的左右两侧，第一挡板和第二挡板分别设于币道的前后两侧，第一挡板上设有 一长方形投币口，第二挡板上设有一长方形出币口， 币道的两端分别与投币口和出币口相 连接。
8、 根据权利要求 7所述的硬币图像识别装置， 其特征在于， 所述的第三挡板和第四 挡板为玻璃隔板。
9、 根据权利要求 6所述的硬币图像识别装置， 其特征在于， 所述的图像采集模块由 传感器、 相机、 镜头和光源组成， 其中传感器设于币道上接近投币口处， 光源位于币道模 块与相机之间， 光源的中心距与镜头中心距重合。
10、 根据权利要求 9所述的硬币图像识别装置， 其特征在于，所述的传感器为对射传 感器。
11、根据权利要求 6所述的硬币图像识别装置， 其特征在于， 所述的图像处理模块由 带有图像处理软件的 PC组成。
PCT/CN2011/072642 2010-04-12 2011-04-12 硬币图像识别方法及装置 WO2011127808A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201010148110.4 2010-04-12
CN2010101481104A CN101819692B (zh) 2010-04-12 2010-04-12 硬币图像识别方法及装置

Publications (1)

Publication Number Publication Date
WO2011127808A1 true WO2011127808A1 (zh) 2011-10-20

Family

ID=42654779

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/072642 WO2011127808A1 (zh) 2010-04-12 2011-04-12 硬币图像识别方法及装置

Country Status (2)

Country Link
CN (1) CN101819692B (zh)
WO (1) WO2011127808A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104050745A (zh) * 2013-03-13 2014-09-17 苏州日宝科技有限责任公司 一种基于图像识别的高速硬币清分技术

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819692B (zh) * 2010-04-12 2012-08-15 高新现代智能系统股份有限公司 硬币图像识别方法及装置
SG194708A1 (en) * 2011-04-29 2013-12-30 Monnaie Royale Canadienne Royal Canadian Mint Method and apparatus for authentication of a coin or other manufactured item
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
CN102866160B (zh) * 2012-08-29 2014-10-08 上海造币有限公司 硬币模具表面质量检测的系统及方法
CN105205840B (zh) * 2015-08-07 2018-06-01 中国科学院计算技术研究所 一种电子断层图像中胶体金的直径估计及自动识别方法
CN105118136B (zh) * 2015-09-01 2018-10-16 昆山古鳌电子机械有限公司 一种货币处理装置及方法
CN105225331A (zh) * 2015-10-28 2016-01-06 威海华菱光电股份有限公司 硬币检测设备和硬币检测方法
PT3475926T (pt) 2016-06-23 2022-08-01 Ultra Electronics Forensic Tech Inc Deteção de irregularidades de superfície em moedas
CN110610575B (zh) 2019-09-20 2021-09-07 北京百度网讯科技有限公司 硬币识别方法及装置、收银机

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0981822A (ja) * 1995-09-20 1997-03-28 Oki Electric Ind Co Ltd 硬貨識別装置
JPH09161077A (ja) * 1995-12-11 1997-06-20 Oki Electric Ind Co Ltd 硬貨識別装置
CN1274904A (zh) * 1999-05-24 2000-11-29 罗烈尔银行机器股份有限公司 硬币鉴别设备
JP2006164192A (ja) * 2004-12-10 2006-06-22 Univ Waseda 硬貨識別装置および硬貨識別方法
CN101667311A (zh) * 2008-09-03 2010-03-10 吉鸿电子股份有限公司 影像分币装置
CN101819692A (zh) * 2010-04-12 2010-09-01 高新现代智能系统股份有限公司 硬币图像识别方法及装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3609285B2 (ja) * 1999-05-19 2005-01-12 ローレルバンクマシン株式会社 硬貨判別装置
JP3945379B2 (ja) * 2002-11-06 2007-07-18 松下電器産業株式会社 模様識別装置
JP5140510B2 (ja) * 2008-07-11 2013-02-06 ローレル精機株式会社 円板状物体外観検査装置
JP5264343B2 (ja) * 2008-07-24 2013-08-14 グローリー株式会社 硬貨識別方法および硬貨識別装置
CN101339144B (zh) * 2008-08-14 2010-12-29 中国印钞造币总公司 硬币双面自动检测装置
CN201804365U (zh) * 2010-04-12 2011-04-20 高新现代智能系统股份有限公司 硬币图像识别装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0981822A (ja) * 1995-09-20 1997-03-28 Oki Electric Ind Co Ltd 硬貨識別装置
JPH09161077A (ja) * 1995-12-11 1997-06-20 Oki Electric Ind Co Ltd 硬貨識別装置
CN1274904A (zh) * 1999-05-24 2000-11-29 罗烈尔银行机器股份有限公司 硬币鉴别设备
JP2006164192A (ja) * 2004-12-10 2006-06-22 Univ Waseda 硬貨識別装置および硬貨識別方法
CN101667311A (zh) * 2008-09-03 2010-03-10 吉鸿电子股份有限公司 影像分币装置
CN101819692A (zh) * 2010-04-12 2010-09-01 高新现代智能系统股份有限公司 硬币图像识别方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104050745A (zh) * 2013-03-13 2014-09-17 苏州日宝科技有限责任公司 一种基于图像识别的高速硬币清分技术

Also Published As

Publication number Publication date
CN101819692A (zh) 2010-09-01
CN101819692B (zh) 2012-08-15

Similar Documents

Publication Publication Date Title
WO2011127808A1 (zh) 硬币图像识别方法及装置
CN105957238B (zh) 一种纸币管理方法及其系统
CN109684987B (zh) 一种基于证件的身份验证系统及方法
WO2012083711A1 (zh) 一种纸币检测方法及装置
US20060160608A1 (en) Card game system with automatic bet recognition
US20060160600A1 (en) Card game system with automatic bet recognition
CN102096961A (zh) 一种纸币鉴别方法及装置
WO2013131407A1 (zh) 双验证人脸防伪方法及装置
CN107958217A (zh) 一种基于深度学习的指纹分类识别系统及方法
WO2012083713A1 (zh) 一种纸币、票券识别方法和装置
CN104464079A (zh) 基于模板特征点及其拓扑结构的多币种面值识别方法
JP5372183B2 (ja) 硬貨分類装置および硬貨分類方法
CN106898078B (zh) 一种港币版本的识别方法及装置
JP5174917B2 (ja) 硬貨の正損判別装置及び正損判別方法
CN107305716A (zh) 一种纸币鉴别方法和装置
JP2002109596A (ja) 貨幣識別方法及び装置
CN106803307B (zh) 一种基于模板匹配的纸币面值面向识别方法
CN107767534A (zh) 一种全方位检测纸币防伪特征的方法
CN201804365U (zh) 硬币图像识别装置
Ogbuju et al. Deep learning for genuine Naira banknotes
JP2003196699A (ja) 硬貨識別方法および硬貨識別装置
JP5953553B2 (ja) ディスク判別方法、ディスク判別装置およびディスク選別装置
JP5953614B2 (ja) ディスク判別方法、ディスク判別装置およびディスク選別装置
TWI559261B (zh) Identify methods of forging coins
CN118212666B (zh) 基于多角度的活体掌静脉识别系统及其方法

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: 11768420

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 12012502036

Country of ref document: PH

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11768420

Country of ref document: EP

Kind code of ref document: A1