WO2014023109A1 - 一种图像识别***及图像存储控制方法 - Google Patents

一种图像识别***及图像存储控制方法 Download PDF

Info

Publication number
WO2014023109A1
WO2014023109A1 PCT/CN2013/073932 CN2013073932W WO2014023109A1 WO 2014023109 A1 WO2014023109 A1 WO 2014023109A1 CN 2013073932 W CN2013073932 W CN 2013073932W WO 2014023109 A1 WO2014023109 A1 WO 2014023109A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
node
storage area
pointer
storage
Prior art date
Application number
PCT/CN2013/073932
Other languages
English (en)
French (fr)
Inventor
黎明
刘梦涛
Original Assignee
广州广电运通金融电子股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州广电运通金融电子股份有限公司 filed Critical 广州广电运通金融电子股份有限公司
Priority to EP13828011.0A priority Critical patent/EP2884467B1/en
Priority to US14/416,626 priority patent/US9838598B2/en
Priority to AU2013302149A priority patent/AU2013302149B2/en
Publication of WO2014023109A1 publication Critical patent/WO2014023109A1/zh
Priority to ZA2015/00981A priority patent/ZA201500981B/en

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
    • 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
    • G07D7/06—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 wave or particle radiation
    • G07D7/12—Visible light, infrared or ultraviolet radiation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/04—Payment circuits
    • G06Q20/042—Payment circuits characterized in that the payment protocol involves at least one cheque
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0007—Image acquisition
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/94—Hardware or software architectures specially adapted for image or video understanding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition
    • 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
    • G07D7/20—Testing patterns thereon
    • G07D7/202—Testing patterns thereon using pattern matching
    • G07D7/206—Matching template patterns
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/80—Camera processing pipelines; Components thereof
    • 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
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30176—Document

Definitions

  • the present invention relates to an image recognition system, and more particularly to an image recognition system for image recognition of a value document in an automatic rejection machine and a corresponding image storage control method.
  • An authentication device for identifying the characteristics of the processed financial instrument is required to be installed on the financial instrument processing machine.
  • a cash machine such as an automatic teller machine (ATM), a sorting machine, and a money counter has a banknote identification device, which is mainly used to identify the currency, face value, face, authenticity and the like of the banknote. Since the banknotes are added with identifiable specific anti-counterfeiting marks during the design and printing process, different sensors are usually arranged in the above-mentioned banknote identification device to detect and acquire information about the relevant anti-counterfeiting marks, for example, by different wavelengths of light irradiation. To obtain a surface image or a transmission image of the banknote, and then identify the corresponding banknote according to the collected image information.
  • the image information collected by the optical image sensor is first stored in a predetermined identification image memory, and then the image recognition module performs currency, face value, orientation, and true on the image of each value document stored.
  • the identification of pseudo-equal information because the identification module does not successfully identify the image information collected, therefore, in order to understand and study the reasons for the unsuccessful identification, we need to image information that cannot be successfully identified.
  • the conventional banknote rejecting image storing method is as shown in Fig. 1. By setting a banknote rejecting image memory in advance, image information that cannot be authenticated is copied from the recognized image memory to the banknote rejecting image memory.
  • One object of the present invention is to provide an image recognition system that does not require copying of a banknote-rejecting image.
  • the system can greatly reduce the occupation time of the CPU by the non-authentication computing task, and effectively improve the authentication speed of the image recognition system.
  • This image recognition system includes:
  • An image sensor for collecting an image of the recognition object
  • An image memory divides the image storage space into N image storage areas according to the maximum data amount of the captured image, N is a natural number greater than 1, and each storage area stores one image of the collection; an image recognition module, the image memory The stored image is identified;
  • An image storage area controller for controlling a storage area of the captured image and a storage area for controlling the unidentifiable image
  • the image storage area controller includes a storage area index linked list, and the storage area index linked list is divided into N nodes.
  • N is a natural number greater than 1
  • each node corresponds to the image storage area
  • each node stores the storage location of the previous image and the next image linked with the node
  • the node pointers are: a set image write pointer, pointing to a write node position corresponding to the image set storage area; an image read pointer, pointing to a read node position corresponding to the image storage area to be authenticated;
  • the tail pointer points to the location of the reject node corresponding to the last banknote image storage area.
  • the banknote end pointer points to the invalid position.
  • the storage area index linked list includes N nodes, each node includes at least two units, and the first unit stores an address of an image storage area of a previous image linked with the node, and the second unit stores and The address of the image storage area where the next image linked by this node is located.
  • the storage area index linked list is divided into two linked lists, which are respectively used as an image storage linked list and a counterfeit image linked list, and the image storage linked list is composed of a storage area index linked list node corresponding to a storage area that allows writing of the collected image, and the banknote is rejected.
  • the image linked list is composed of a storage area index linked list node corresponding to the storage area occupied by the banknote-rejecting image.
  • the image storage area controller further sets a counterfeit image head pointer for pointing to a node position corresponding to the first banknote rejecting image storage area.
  • the node pointed by the identification image read pointer needs to be added to the end of the counterfeit list, and the node pointed by the counterfeit image head pointer is inserted into the position of the identification image read pointer, and the identification list is added.
  • the discriminating image reading pointer moves to the next node, the counterfeit image head pointer points to the node corresponding to the storage area where the counterfeit image is located, and the previous image and the next link linked to the corresponding node stored by the two nodes The storage location of the image is modified.
  • Another object of the present invention is to provide a corresponding image storage control method comprising the following steps:
  • Step 1 Identifying an image of the object by using an image sensor
  • Step 2 The main control unit writes the captured image into the corresponding image storage area according to the write node pointed by the image storage pointer of the image storage area controller;
  • Step 3 moving the captured image writing pointer to the node corresponding to the storage location of the next image written by the writing node; performing step 1 or step 4 according to the control of the main control unit;
  • Step 4 The image recognition module discriminates the image in the image storage area corresponding to the read node indicated by the authentication image read pointer. If the authentication is successful, step 5 is performed. Otherwise, step 6 is performed; step 5, the identification image is read out. The pointer moves to the node corresponding to the storage location of the next image recorded by the read node, and continues to step 4;
  • Step 6 the tail pointer movement of the counterfeit image and the location address of the relevant node are modified, and the tail pointer of the counterfeit image is pointed to the read node; and the storage location of the previous image stored by the node is read.
  • the storage location of the next image stored in the corresponding node is modified to the next image storage location stored by the read node, and the previous node stored in the storage location of the next image stored in the read node is stored.
  • the storage location of the image is modified to the last image storage location stored by the read node;
  • Step 7 the storage location of the previous image of the node pointed by the tail pointer of the banknote is modified to be the corresponding node of the storage area where the last banknote image is located;
  • Step 8 Move the authentication image read pointer to the node corresponding to the storage location of the next image recorded by the read node, and continue to step 4.
  • the image storage control method further includes step 0, when the image recognition system is initialized, the memory index list is initialized to each node respectively corresponding to the image storage area, and each node is written adjacent to the node.
  • the image storage area address corresponding to the previous node and the image storage area address corresponding to the next next node, and the set image write pointer and the authentication image read pointer are all directed to the first node, and the banknote rejection image is The tail pointer points to an invalid location.
  • the image storage area controller further sets a counterfeit image head pointer for pointing to the node position corresponding to the first banknote rejecting image storage area, when the number of banknote rejecting images reaches the maximum allowed
  • the counterfeit image is generated again, the node pointed by the identification image read pointer needs to be added to the end of the counterfeit list, and the node pointed by the counterfeit image head pointer is inserted into the position of the identification image read pointer, and the identification list is added.
  • the discriminating image reading pointer moves to the next node
  • the counterfeit image head pointer points to the node corresponding to the storage area where the counterfeit image is located
  • the previous image and the next link linked to the corresponding node stored by the two nodes The storage location of the image is modified.
  • FIG. 1 is a schematic diagram of constituent modules of a conventional image recognition system
  • FIG. 2 is a schematic block diagram of an image recognition system provided by the present invention.
  • FIG. 3 is a schematic diagram showing an organization form of an image storage area when the image recognition system is initialized according to the present invention
  • FIG. 4 is a schematic diagram showing the organization form of a storage area index linked list when the image recognition system is initialized according to the present invention
  • FIG. 5 is a schematic diagram showing an organization form of an image storage area when the image recognition system provided by the present invention collects a third image
  • FIG. 6 is a schematic diagram showing the organization form of a storage area index list when the image recognition system according to the present invention identifies the first image
  • FIG. 7 is a schematic diagram showing the organization of an image storage area when the image recognition system provided by the present invention identifies the first unsuccessful;
  • FIG. 8 is a schematic diagram showing the organization form of the storage area index list when the image recognition system of the present invention identifies the first unsuccessful;
  • FIG. 9 is a schematic diagram showing the organization of an image storage area when the image recognition system of the present invention identifies the second unsuccessful;
  • FIG. 10 is a schematic diagram showing the organization form of a storage area index list when the image recognition system of the present invention identifies the second unsuccessful;
  • FIG. 11 is a schematic diagram showing the organization of an image storage area when the image recognition system provided by the present invention identifies that the number of unrecognizable images reaches a maximum value;
  • FIG. 12 is a schematic diagram of an organization form of a storage area index linked list when the image recognition system provided by the present invention identifies that the number of unrecognizable images reaches a maximum value;
  • FIG. 13 is the image recognition system provided by the present invention, when the number of unrecognizable images reaches a maximum value, the node of the first banknote-rejecting image in the banknote-rejecting image linked list is transferred to the image storage linked list, and the node of the latest banknote-rejecting image in the image storage linked list is rejected.
  • the image recognition system includes an image sensor 1 for collecting an image of the recognition object, and an image memory 2 for dividing the image storage space into N according to the maximum data amount of the captured image.
  • Image storage area, N is a natural number greater than 1, and each storage area stores one image of the collection. In this embodiment, N is equal to 100 without loss of generality; - Image recognition module 3, which discriminates images stored in the image memory
  • An image storage area controller 4 for controlling a storage area of the captured image and a storage area for controlling the unidentifiable image; a main control unit 5 for controlling and coordinating data operation and transmission between the above components.
  • the image storage area controller 4 includes a storage area index linked list 6, the storage area index linked list 6 is divided into N nodes, and N is a natural number greater than 1, and each node has a one-to-one correspondence with the image storage area.
  • the nodes store the storage locations of the previous image and the next image linked to the node, and set three node pointers 7 for the storage area index linked list 6, respectively: a set image write pointer 71, pointing to the image Writing to the corresponding write node position of the storage area; identifying the image read pointer 72, pointing to the read node position corresponding to the image storage area to be authenticated; rejecting the image end pointer 73, pointing to the corresponding reject of the last banknote image storage area The banknote node position, when there is no banknote rejection image, the banknote end pointer points to the invalid position.
  • the storage area index linked list 6 also includes 100 nodes, each node includes two units, and the first unit stores the address of the image storage area where the previous image is linked with the node, the second The unit stores the address of the image storage area of the next image linked with the node.
  • the storage area index list 6 is composed of a 2-dimensional array. Since the array has consecutive storage addresses, the storage area index list 6 Only two units are required to meet the requirements. In addition, since the N value is equal to 100, the first and second units can satisfy the address storage of the image storage area by one byte.
  • each node needs to include three units, one unit for storing the corresponding value of the node and the image storage area, and one for storing the previous one linked with the node.
  • the storage area index list is divided into two 2-dimensional arrays, which are respectively used as an image storage linked list 61 and a counterfeit image linked list 62, and the image storage linked list 61 is allowed to be written into the collected image.
  • the storage area index linked list node corresponding to the storage area is composed of the storage area index linked list node corresponding to the storage area occupied by the rejecting image.
  • the method for controlling image storage includes the following steps:
  • the main control unit 5 according to the image storage area controller 4, the image is written by the pointer 71, the write node writes the captured image into the corresponding image storage area;
  • the captured image write pointer 71 is moved to the node corresponding to the storage location of the next image written by the write node; S 1 or S4 is performed according to the control of the main control unit 5;
  • the image recognition module 3 discriminates the image in the image storage area corresponding to the read node indicated by the authentication image readout pointer 72, and if it is successfully authenticated, executes S5, otherwise, executes S6;
  • the identification image read pointer 72 is moved to the node corresponding to the storage location of the next image recorded by the read node, and continues to S4;
  • the counterfeit image tail pointer 73 moves and the location address stored by the relevant node is modified, the counterfeit image tail pointer 73 is pointed to the read node; and the next node of the node corresponding to the storage location of the previous image of the read node is read.
  • the storage location of the image is modified to the next image storage location of the read node, and the storage location of the previous image of the node corresponding to the storage location of the next image of the read node is modified to be the previous one of the read node Image storage location;
  • the authentication image read pointer 72 is moved to the node corresponding to the storage location of the next image described by the read node, and the process proceeds to S4.
  • the image storage area controller 4 further sets a counterfeit image head pointer 74 for pointing to the node position corresponding to the first banknote repayment image storage area, when the number of banknote rejection images reaches the allowable number
  • the node pointed to by the authentication image readout pointer 72 needs to be added to the end of the counterfeit list, and the node pointed by the counterfeit image head pointer 74 is inserted into the identification image read pointer 72.
  • the identification list is added, the authentication image read pointer 72 moves to the next node, and the counterfeit image head pointer 74 points to the node corresponding to the storage area where the banknote image is located, and the two nodes are linked to the corresponding node.
  • the storage location of one image and the next image is modified.
  • the main control unit 5 controls all image storage areas to divide the image storage space into N image storage areas according to the maximum data amount of the captured image, where N is a natural number greater than 1, not normal
  • the N is 100, and each storage area stores one image.
  • the memory index list 5 is initialized to N nodes.
  • N is 100, and all nodes are in one-to-one correspondence with the image storage area.
  • the fourth node stores 3 and 5, ...
  • the 100 nodes store 99 and 1 to realize the node serial connection, as shown in FIG. 4, at this time, the set image write pointer 71 and the authentication image read pointer 72 all point to the first node, and the counterfeit image head and tail pointers point to invalid. position.
  • the image recognition system starts to perform identification on the value document (here, the banknote), S1, the image of the recognition object is collected by the image sensor 1;
  • the main control unit 5 writes the captured image as the first node according to the image input pointer 71 of the image storage area controller 4, and therefore writes the captured image into the first image storage corresponding to the first node.
  • step S3 moving the captured image write pointer 71 to the second node corresponding to the storage location of the next image described in the first node; and then continuing the image collection S1 step according to the control of the main control unit 5, or The image is identified in step S4;
  • the image recognition module 3 authenticates the image in the first image storage area corresponding to the first node pointed by the authentication image read pointer 72, and if successful, executes S5, otherwise, executes S6;
  • the identification image read pointer 72 is moved to the second node corresponding to the storage location of the next image described in the first node, and the process proceeds to S4.
  • the step S4 is continued, and it is necessary to consider that the pointer cannot be written over the captured image, that is, , the position of the authentication image read pointer 72 is always less than or equal to the position of the image set pointer 71, otherwise an invalid image is read and an identification error is caused;
  • the counterfeit image tail pointer 73 is directed to the first node; and the storage location 1 of the next image stored in the 100 node corresponding to the storage location of the previous image stored in the first node is modified to be stored in the first node.
  • the next image storage location 2 the storage location of the previous image stored in the second node corresponding to the storage location of the previous image stored in the first node is modified to the previous image storage stored in the first node.
  • the position is 100;
  • the storage location of the previous image of the node pointed by the end of the counterfeit image tail 73 is modified to be the corresponding node of the storage area where the previous banknote rejection image is located.
  • the first node is The storage location of the stored previous image and the next storage location are both empty.
  • the storage location of the previous image stored in the node pointed by the counterfeit image tail pointer 73 needs to be modified to be upper.
  • a memory card corresponding to the storage area corresponds to the node; thereby achieving the purpose of releasing the node 1 from the identification list and then concatenating it into the reject list. This process is equivalent to saving a banknote image, but there is no process of copying the image, which saves the copy time of the banknote image.
  • the above-described collection image write pointer 71, the discrimination image read pointer 72, the counterfeit image tail pointer 73, and the counterfeit image head pointer 74 are moved according to the above rules, and are stored in the relevant nodes.
  • the storage location of the previous image and the storage location of the next image Correspondingly modified so that the nodes in the corresponding image storage linked list and the counterfeit image linked list form a tandem organization of nodes.
  • the changing rules of the image storage area and the storage area index list are further shown as the collected image and the authenticated image are continuously performed.
  • the content of the image storage area is as shown in FIG. 5 when the third image is being collected and the image of the first image is identified.
  • the organization of the storage area index list is as shown in FIG.
  • the recognition module cannot successfully identify the first image, then the content of the image storage area is as shown in FIG. 7, and the organization of the storage area index list is formed as shown in FIG. 8.
  • the recognition module continuously authenticates the identification image.
  • the authentication image corresponding to the fifth node is not recognized, the content of the image storage area is as shown in FIG. 9, and the organization of the storage area index list is formed.
  • Figure 10 shows.
  • the number of banknote rejection images is increasing.
  • this embodiment uses more and more When approaching 100 sheets, in order to identify the continued use of the system, it is necessary to empty the image storage area occupied by the counterfeit image and store it in a new image. In addition, a certain access area is needed for the identification module to identify the image. To cooperate with the operation, it is necessary to reserve a certain number of storage areas for the image discrimination operation. This is not general. If it is necessary to reserve 5 storage areas, then when the number of banknote rejection images reaches 95, assume that, at this time, 2.
  • the number of image storage areas is 100.
  • Those skilled in the art can know from the common knowledge in the field of data storage, as the storage space of the memory is different and the image is collected. The data size is different, the number of image storage areas that can be set can be changed at will, and such common sense is known when changing, so it can be recognized
  • the technical solution provided by the present invention can be implemented by a natural number that can be N for image storage and N is greater than 1.
  • the present embodiment assumes the appearance node of the banknote rejecting image, Those skilled in the art should understand that such occurrences are random.
  • the examples in this embodiment are only a hypothetical result, and the nodes of the actual banknote rejection image are not constrained by the contents described in this embodiment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Toxicology (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Quality & Reliability (AREA)
  • Accounting & Taxation (AREA)
  • Strategic Management (AREA)
  • Signal Processing (AREA)
  • Inspection Of Paper Currency And Valuable Securities (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
  • Image Analysis (AREA)

Abstract

一种用于自动拒员机中对有价文件进行图像识别的图像识别***和相应的图像存储控制方法。这种图像识别***，包括：一图像传感器；一图像存储器；一图像识别模块；一图像存储区控制器；一主控单元，所述图像存储区控制器包括一存储区索引链表，该存储区索引链表被划分为N个节点，N为大于1的自然数，每个节点与图像存储区控制器划分的图像存储区一一对应，每个节点存放与该节点链接的上一张图像和下一张图像的存储位置，设置采集图像写入指针，鉴别图像读出指针，拒钞图像头指针，拒钞图像尾指针。存储区索引链表配合4个指针实现将拒钞图像所在图像存储区位置与采集图像存储区位置的软划分，从而实现了不需要拷贝拒钞图像完成保留拒钞图像的目的。

Description

一种图像识别***及图係 储控制方法
本申请要求于 2012 年 08 月 09 日提交中国专利局、 申请号为 201210282874.1、 发明名称为"一种图像识别***及图像存储控制方法"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
本发明涉及一种图像识别***， 特别是一种用于自动拒员机中对有价 文件进行图像识别的图像识别***和相应的图像存储控制方法。 背景技术
在金融票据处理的机具上均需要安装用来鉴别所处理金融票据特征的 鉴别装置。 比如， 现金自动存取款机（ATM )、 清分机及点钞机等金融设备 中都有纸币识别装置， 其中主要用于识别纸币的币种、 面值、 面向、 真伪 等信息的鉴别。 由于纸币在设计和印刷的过程中都加入了可鉴别的特定防 伪标志， 故在上述纸币识别装置中通常都配置不同的传感器来检测和获取 相关的防伪标志的信息， 例如通过不同波长的光照射来获取纸币表面图像 或透射图像， 然后根据所釆集的相关图像信息对相应纸币进行鉴别。
目前， 针对上述光学图像传感器所釆集的图像信息首先是存储在一预 定的识别图像存储器内 , 然后由图像识别模块对存储的每一张有价文件的 图像进行币种、 面值、 面向、 真伪等信息的鉴别， 由于识别模块对所釆集 的图像信息的鉴别存在无法成功鉴别的情况发生， 因此， 为了事后了解和 研究鉴别不成功的原因， 我们需要将无法成功鉴别的图像信息 （下称： 拒 钞图像）进行保存， 根据保存的拒钞图像， 方便开发者发现传感器及硬件 故障， 发现识别模块故障， 发现算法缺陷， 为后续的改进提供依据。 现有 的拒钞图像存储方法如图 1所示， 通过预先设定一个拒钞图像存储器， 将 无法鉴别的图像信息从识别图像存储器拷贝到拒钞图像存储器内。
在单 CPU的情况下，由于拒钞图像拷贝和算法识别图像无法同时进行， 这些操作只能是顺序执行， 所以在识别模块遇到无法成功识别的图像时， 就需要花费大量的时间将该图像拷贝至拒钞图像存储器内 , 而识别模块对 其它图像的鉴别被迫停止， 从而大大影响了识别模块的识别速度。 为了满 足自动拒员机对有价文件的高速处理需求， 如何使 CPU的处理时间最多地 用于对釆集图像信息的鉴别， 尽可能地减少非鉴别运算的时间浪费成为本 领域技术人员一直努力解决的技术问题。 发明内容
本发明的目的之一是提供一种不需要拷贝拒钞图像的图像识别***， 该***可以大大降低非鉴别运算任务对 CPU的处理时间的占用，有效提高 图像识别***的鉴别速度。
这种图像识别***， 包括：
一图像传感器， 用于釆集识别对象的图像；
一图像存储器， 根据釆集图像的最大数据量将图像存储空间划分成 N 个图像存储区， N为大于 1的自然数， 每个存储区存放 1幅釆集图像； 一图像识别模块， 对图像存储器内存储的图像进行鉴别；
一图像存储区控制器， 用于控制釆集图像的存储区和控制无法鉴别图 像的存储区；
一主控单元， 用于控制和协调上述部件之间的数据运算和传输； 其特殊之处于， 所述图像存储区控制器包括一存储区索引链表， 该存 储区索引链表被划分为 N个节点， N为大于 1的自然数， 每个节点与图像 存储区——对应， 每个节点存放与该节点链接的上一张图像和下一张图像 的存储位置， 同时为存储区索引链表设置 3个节点指针， 分别是： 釆集图 像写入指针， 指向釆集图像写入存储区对应的写入节点位置； 鉴别图像读 出指针， 指向待鉴别图像存储区对应的读出节点位置； 拒钞图像尾指针， 指向最后一幅拒钞图像存储区对应的拒钞节点位置， 当没有拒钞图像时， 拒钞图像尾指针指向无效位置。 优选的， 所述存储区索引链表包含 N个节点， 每个节点至少包括两个 单元，第 1个单元存放与该节点链接的上一张图像所在图像存储区的地址， 第 2个单元存放与该节点链接的下一张图像所在图像存储区的地址。
进一步的， 所述存储区索引链表划分为两个链表， 分别作为图像存储 链表和拒钞图像链表， 图像存储链表由允许写入釆集图像的存储区对应的 存储区索引链表节点组成， 拒钞图像链表由拒钞图像所占存储区对应的存 储区索引链表节点组成。
优选的， 图像存储区控制器还设置一拒钞图像头指针， 用于指向第 1 幅拒钞图像存储区对应的节点位置。
进一步的， 当再产生拒钞图像时， 需要把鉴别图像读出指针所指节点 加入拒钞链表末尾， 同时把拒钞图像头指针所指节点***鉴别图像读出指 针所指位置， 加入识别链表， 鉴别图像读出指针移动到下一个节点， 拒钞 图像头指针指向一下张拒钞图像所在存储区对应的节点， 并对两个节点存 放的与相应节点链接的上一张图像和下一张图像的存储位置进行修改。
本发明的另一目的是提供一种相应的图像存储控制方法， 其包括如下 步骤：
步骤 1 , 通过图像传感器釆集识别对象的图像；
步骤 2, 主控单元根据图像存储区控制器的釆集图像写入指针所指写 入节点将釆集图像写入对应的图像存储区内；
步骤 3 , 将釆集图像写入指针移动至写入节点所记载的下一张图像的 存储位置对应的节点； 根据主控单元的控制进行步骤 1或步骤 4;
步骤 4, 图像识别模块对鉴别图像读出指针所指读出节点对应的图像 存储区内的图像进行鉴别，如果成功鉴别，执行步骤 5 ,否则，执行步骤 6; 步骤 5 , 将鉴别图像读出指针移动至读出节点所记载的下一张图像的 存储位置对应的节点， 继续步骤 4;
步骤 6, 拒钞图像尾指针移动及相关节点存放的位置地址修改， 将拒 钞图像尾指针指向读出节点； 并将读出节点存放的上一张图像的存储位置 所对应的节点存放的下一张图像的存储位置修改为该读出节点存放的下一 张图像存储位置， 将读出节点存放的下一张图像的存储位置所对应的节点 存放的上一张图像的存储位置修改为该读出节点存放的上一张图像存储位 置；
步骤 7, 将拒钞图像尾指针所指节点的上一张图像的存储位置修改为 上一张拒钞图像所在存储区对应节点；
步骤 8, 将鉴别图像读出指针移动至读出节点所记载的下一张图像的 存储位置对应的节点， 继续步骤 4。
优选的， 该图像存储控制方法还包括步骤 0, 当图像识别***初始化 时， 存储器索引链表初始化为每个节点分别与图像存储区一一对应， 且为 每个节点写入与该节点相邻的上一节点所对应的图像存储区地址和相邻的 下一节点所对应的图像存储区地址， 并将釆集图像写入指针、 鉴别图像读 出指针均指向第一个节点， 将拒钞图像尾指针指向无效位置。
为了在满足图像识别***的正常识别功能， 图像存储区控制器还设置 一拒钞图像头指针， 用于指向第 1幅拒钞图像存储区对应的节点位置， 当 拒钞图像数量达到允许的最大值时， 如果再产生拒钞图像时， 需要把鉴别 图像读出指针所指节点加入拒钞链表末尾， 同时把拒钞图像头指针所指节 点***鉴别图像读出指针所指位置， 加入识别链表， 鉴别图像读出指针移 动到下一个节点， 拒钞图像头指针指向一下张拒钞图像所在存储区对应的 节点， 并对两个节点存放的与相应节点链接的上一张图像和下一张图像的 存储位置进行修改。
该图像识别***与现有技术对比分析具有如下优点：
由于釆用存储区索引链表的图像存储区控制器， 对于无法鉴别图像的 存储不再需要将图像拷贝至其他存储区，从而将图像拷贝所占用的 CPU的 时间用于对图像的鉴别处理， 因此可以艮好地了图像识别***对有价文件 的处理速度。 附图说明
图 1是现有图像识别***的组成模块示意图；
图 2是本发明所提供的图像识别***的组成模块示意图；
图 3是本发明所提供的图像识别***初始化时图像存储区的组织形式 示意图；
图 4是本发明所提供的图像识别***初始化时存储区索引链表组织形 式示意图；
图 5是本发明所提供的图像识别***釆集第三张图像时图像存储区的 组织形式示意图；
图 6是本发明所提供的图像识别***鉴别第一张图像时存储区索引链 表的组织形式示意图；
图 7是本发明所提供的图像识别***鉴别第一张不成功时的图像存储 区的组织形式示意图；
图 8是本发明所提供的图像识别***鉴别第一张不成功时的存储区索 引链表的组织形式示意图；
图 9是本发明所提供的图像识别***鉴别第二张不成功时的图像存储 区的组织形式示意图；
图 10是本发明所提供的图像识别***鉴别第二张不成功时的存储区 索引链表的组织形式示意图；
图 11 是本发明所提供的图像识别***鉴别无法鉴别图像数量达到最 大值时图像存储区的组织形式示意图；
图 12是本发明所提供的图像识别***鉴别无法鉴别图像数量达到最 大值时存储区索引链表的组织形式示意图；
图 13 是本发明所提供的图像识别***鉴别无法鉴别图像数量达到最 大值时， 拒钞图像链表中最早拒钞图像所在节点向图像存储链表转移， 图 像存储链表中最新拒钞图像所在节点向拒钞图像链表末尾转移的示意图。 具体实施方式
为进一步阐述本发明， 以下结合图示介绍本发明的一个优选实施例。 参阅图 2和图 3所示， 这种图像识别***， 包括一图像传感器 1 , 用 于釆集识别对象的图像； 一图像存储器 2, 根据釆集图像的最大数据量将 图像存储空间划分成 N个图像存储区， N为大于 1的自然数， 每个存储区 存放 1幅釆集图像， 本实施例不失一般性设 N等于 100; —图像识别模块 3 , 对图像存储器内存储的图像进行鉴别； 一图像存储区控制器 4, 用于控 制釆集图像的存储区和控制无法鉴别图像的存储区； 一主控单元 5 , 用于 控制和协调上述部件之间的数据运算和传输。
其中， 所述图像存储区控制器 4包括一存储区索引链表 6, 该存储区 索引链表 6被划分为 N个节点， N为大于 1的自然数， 每个节点与图像存 储区一一对应， 每个节点存放与该节点链接的上一张图像和下一张图像的 存储位置， 同时为存储区索引链表 6设置 3个节点指针 7 , 分别是： 釆集 图像写入指针 71 , 指向釆集图像写入存储区对应的写入节点位置； 鉴别图 像读出指针 72, 指向待鉴别图像存储区对应的读出节点位置； 拒钞图像尾 指针 73 , 指向最后一幅拒钞图像存储区对应的拒钞节点位置， 当没有拒钞 图像时， 拒钞图像尾指针指向无效位置。 本实施例中， 所述存储区索引链 表 6同样包含 100个节点， 每个节点包括两个单元， 第 1个单元存放与该 节点链接的上一张图像所在图像存储区的地址， 第 2个单元存放与该节点 链接的下一张图像所在图像存储区的地址， 本实施例中存储区索引链表 6 由一个 2维数组构成， 由于数组具有连续的存储地址， 因此， 该存储区索 引链表 6只需要两个单元既可以满足要求， 另外由于 N数值等于 100, 故 第 1和第 2单元分别由一个字节就可以满足图像存储区的地址存放。 当存 储区索引链表 6釆用地址非连续的形成组织时， 每个节点需要包括 3个单 元， 一个单元用来存放节点与图像存储区的对应值， 一个用来存放与该节 点链接的上一张图像所在图像存储区的地址， 最后一个用来存放与该节点 链接的下一张图像所在图像存储区的地址。 参阅图 5 , 当存在拒钞图像时， 所述存储区索引链表划分为两个 2维 数组， 分别作为图像存储链表 61和拒钞图像链表 62, 图像存储链表 61由 允许写入釆集图像的存储区对应的存储区索引链表节点组成， 拒钞图像链 表 62由拒钞图像所占存储区对应的存储区索引链表节点组成。
随着釆集图像数越来越多， 拒钞图像也越来越多时， 为了图像识别系 统能够继续对有价文件的识别工作， 需要保留一小部分存储区给鉴别图像 使用， 因此当拒钞图像的存储数量达到设定的最大值时， 需要将早期的拒 钞图像所占存储区逐步释放出来用来存放新釆集的图像， 因此， 在图像存 储区控制器 4中增设一拒钞图像头指针 74用于指向第 1幅拒钞图像存储区 对应的节点位置，该拒钞图像头指针 64随着拒钞图像所占存储区的释放逐 步向拒钞图像尾指针靠拢。
其中对图像存储的控制方法包括如下步骤：
51 , 通过图像传感器 1釆集识别对象的图像；
52,主控单元 5根据图像存储区控制器 4的釆集图像写入指针 71所指 写入节点将釆集图像写入对应的图像存储区内；
53 ,将釆集图像写入指针 71移动至写入节点所记载的下一张图像的存 储位置对应的节点； 根据主控单元 5的控制进行 S 1或 S4;
54,图像识别模块 3对鉴别图像读出指针 72所指读出节点对应的图像 存储区内的图像进行鉴别， 如果成功鉴别， 执行 S5 , 否则， 执行 S6;
S5 ,将鉴别图像读出指针 72移动至读出节点所记载的下一张图像的存 储位置对应的节点， 继续 S4;
56,拒钞图像尾指针 73移动及相关节点存放的位置地址修改，将拒钞 图像尾指针 73指向读出节点；并将读出节点的上一张图像的存储位置对应 的节点的下一张图像的存储位置修改为该读出节点的下一张图像存储位 置， 将读出节点的下一张图像的存储位置对应的节点的上一张图像的存储 位置修改为该读出节点的上一张图像存储位置；
57 ,将拒钞图像尾指针 73所指节点存放的上一张图像的存储位置修改 为上一张拒钞图像所在存储区对应节点；
S8,将鉴别图像读出指针 72移动至读出节点所记载的下一张图像的存 储位置对应的节点， 继续 S4。
为了在满足图像识别***的正常识别功能， 图像存储区控制器 4还设 置一拒钞图像头指针 74, 用于指向第 1 幅拒钞图像存储区对应的节点位 置， 当拒钞图像数量达到允许的最大值时， 当再产生拒钞图像时， 需要把 鉴别图像读出指针 72所指节点加入拒钞链表末尾，同时把拒钞图像头指针 74所指节点***鉴别图像读出指针 72所指位置， 加入识别链表， 鉴别图 像读出指针 72移动到下一个节点， 拒钞图像头指针 74指向一下张拒钞图 像所在存储区对应的节点， 并对两个节点存放的与相应节点链接的上一张 图像和下一张图像的存储位置进行修改。
下面通过结合图 3-图 13 ,进一步说明本图像识别***的工作原理及工 作过程：
参阅图 3 , 当图像识别***初始化时， 主控单元 5控制将所有图像存 储区根据釆集图像的最大数据量将图像存储空间划分成 N个图像存储区， N为大于 1的自然数， 不是一般性， 本实施例 N为 100, 每个存储区存放 1幅釆集图像； 并且将存储器索引链表 5初始化为 N个节点， 本实施例中 N为 100, 且所有节点与图像存储区一一对应， 并为每个节点存放与该节 点链接的上一张图像和下一张图像的存储位置， 即第一节点存放 100和 2, 第二节点存放 1和 3 , 第三节点存放 2和 4, 第四节点存放 3和 5 , ... ...第
100节点存放 99和 1 , 从而实现节点串联， 如图 4所示， 此时， 釆集图像 写入指针 71、 鉴别图像读出指针 72都指向第一节点， 拒钞图像头尾指针 都指向无效位置。
当图像识别***开始对有价文件 （这里指纸币）进行釆集鉴别时， S1 , 通过图像传感器 1釆集识别对象的图像；
S2,主控单元 5根据图像存储区控制器 4的釆集图像写入指针 71所指 写入节点为第 1节点， 因此将釆集图像写入第 1节点对应的第 1图像存储 区内；
S3 ,将釆集图像写入指针 71移动至第 1节点所记载的下一张图像的存 储位置对应的第 2节点； 然后， 根据主控单元 5的控制进行继续图像釆集 S1步， 或对图像进行鉴别 S4步；
S4, 图像识别模块 3对鉴别图像读出指针 72所指第 1节点对应的第 1 图像存储区内的图像进行鉴别， 如果成功鉴别， 执行 S5 , 否则， 执行 S6;
55 ,将鉴别图像读出指针 72移动至第 1节点所记载的下一张图像的存 储位置对应的第 2节点， 继续 S4, 当然继续执行步骤 S4需要考虑不能越 过釆集图像写入指针， 即，鉴别图像读出指针 72的位置总是小于等于釆集 图像写入指针 71的位置， 否则会读到无效图像而导致识别错误；
56,将拒钞图像尾指针 73指向第 1节点； 并将第 1节点存放的上一张 图像的存储位置所对应的 100节点存放的下一张图像的存储位置 1修改为 该第 1节点存放的下一张图像存储位置 2, 将第 1节点存放的上一张图像 的存储位置所对应的第 2节点存放的上一张图像的存储位置修改为该第 1 节点存放的上一张图像存储位置即 100;
57,将拒钞图像尾指针 73所指节点的上一张图像的存储位置修改为上 一张拒钞图像所在存储区对应节点， 对于第 1张拒钞图像时， 此时， 第 1 节点所存放的上一张图像的存储位置和下一张的存储位置均为空， 当非第 1张时， 需要将拒钞图像尾指针 73所指节点存放的上一张图像的存储位置 修改为上一张拒钞图像所在存储区对应节点； 从而实现了把节点 1从识别 链表中释放， 然后串联到到拒钞链表内的目的。 这个过程相当于保存了一 张拒钞图像，但是没有图像拷贝的过程，从而节约了拒钞图像的拷贝时间。
S&将鉴别图像读出指针 72移动至读出节点所记载的下一张图像的存 储位置对应的节点， 继续 S4。
随着图像的不断釆集和鉴别， 上述釆集图像写入指针 71、 鉴别图像读 出指针 72、拒钞图像尾指针 73以及拒钞图像头指针 74依据上述规则移动， 并对相关节点存放的上一张图像的存储位置和下一张图像的存储位置进行 相应修改， 以使相应图像存储链表和拒钞图像链表中的节点形成节点的串 联组织。
下面结合图 5至图 13 , 进一步展示随着釆集图像和鉴别图像的的不断 进行， 图像存储区与存储区索引链表的变化规律。 当已经釆集了两张图像 之后， 正在釆集第三张图像和鉴别第一张的图像时图像存储区的内容如图 5所示， 存储区索引链表的组织形成如图 6所示， 如果识别模块对第一张 图像无法成功识别， 那么图像存储区的内容如图 7所示， 而存储区索引链 表的组织形成如图 8所示。 随着釆集图像的不断写入， 识别模块对鉴别图 像的不断鉴别， 当第 5节点对应的鉴别图像无法识别时， 图像存储区的内 容如图 9所示， 而存储区索引链表的组织形成图 10所示。
在实际的图像识别***中， 随着图像识别***的长时间使用， 拒钞图 像越来越多， 当拒钞图像数量接近图像存储区的数量时， 本实施例用当拒 钞图像越来越接近 100张时， 为了识别***的继续使用， 需要将拒钞图像 所占图像存储区清空后给新的釆集图像来存储， 另外对于识别模块对图像 的鉴别过程中， 需要一定的存取区来配合运算， 因此需要保留一定数量的 存储区给图像鉴别运算时使用， 这里不是一般性， 假设需要保留存储区为 5个， 那么当拒钞图像达到 95个时， 假设， 此时， 2、 4、 6、 7、 79图像存 储区没有被拒钞图像所占据， 那么此时的图像存储区的内容如图 11所示， 而存储区索引链表的组织形成图 12所示，而当在继续工作的过程中又出现 了无法识别的图像， 假设， 出现在第 2节点， 那么将图像存储链表中的节 点 2追加到拒钞图像链表的后面， 将拒钞图像链表中拒钞图像头指针所指 的节点给回图像存储链表中， 此时的存储区索引链表的组织形成图 13 所 示。
以上仅是本发明的优选实施方式， 而且是不是一般性假设图像存储区 个数为 100, 本领域技术人员根据数据存储领域的常识可以获知， 随着存 储器的存储空间的大小不同以及釆集图像的数据大小不同， 可以设定的图 像存储区个数是可以随便改变的， 这样的改变时公知的常识， 因此可以认 为图像存储区可以为 N个，且 N为大于 1的自然数均可以实现本发明所提 供的技术方案。 而且为了清楚表达随着拒钞图像头尾指针的移动以及相应 节点存放的上一张图像的存储位置和下一张图像的存储位置的修改规律， 本实施例假设了拒钞图像的出现节点， 本领域技术人员应该明白这样的出 现时随机的， 本实施例中所举例子仅是一个假设结果， 实际拒钞图像的出 现的节点并不受本实施例所记载内容所约束。
应当指出的是， 上述优选实施方式不应视为对本发明的限制， 本发明 的保护范围应当以权利要求所限定的范围为准。 对于本技术领域的普通技 术人员来说， 在不脱离本发明的精神和范围内， 还可以做出若干改进和润 饰， 这些改进和润饰也应视为本发明的保护范围。

Claims

权 利 要 求
1、 一种图像识别***， 其包括：
一图像传感器， 用于釆集识别对象的图像；
一图像存储器， 根据釆集图像的最大数据量将图像存储空间划分成 N 个图像存储区， N为大于 1的自然数， 每个存储区存放 1幅釆集图像； 一图像识别模块， 对图像存储器内存储的图像进行鉴别；
一图像存储区控制器， 用于控制釆集图像的存储区和控制无法鉴别图 像的存储区；
一主控单元， 用于控制和协调上述部件之间的数据运算和传输； 其特征在于， 所述图像存储区控制器包括一存储区索引链表， 该存储 区索引链表被划分为 N个节点， N为大于 1的自然数， 每个节点与图像存 储区一一对应， 每个节点存放与该节点链接的上一张图像和下一张图像的 存储位置， 同时为存储区索引链表设置 3个节点指针， 分别是： 釆集图像 写入指针， 指向釆集图像写入存储区对应的写入节点位置； 鉴别图像读出 指针， 指向待鉴别图像存储区对应的读出节点位置； 拒钞图像尾指针， 指 向最后一幅拒钞图像存储区对应的拒钞节点位置， 当没有拒钞图像时， 拒 钞图像尾指针指向无效位置。
2、 如权利要求 1所示的图像识别***， 其特征在于， 所述存储区索引 链表包含 N个节点， 每个节点至少包括两个单元， 第 1个单元存放与该节 点链接的上一张图像所在图像存储区的地址， 第 2个单元存放与该节点链 接的下一张图像所在图像存储区的地址。
3、 如权利要求 2所示的图像识别***， 其特征在于， 所述存储区索引 链表划分为两个链表， 分别作为图像存储链表和拒钞图像链表， 图像存储 链表由允许写入釆集图像的存储区对应的存储区索引链表节点组成， 拒钞 图像链表由拒钞图像所占存储区对应的存储区索引链表节点组成。
4、 如权利要求 1所示的图像识别***， 其特征在于， 图像存储区控制 器还设置一拒钞图像头指针， 用于指向第 1幅拒钞图像存储区对应的节点 位置。
5、 如权利要求 4所示的图像识别***， 其特征在于， 当拒钞图像数量 达到允许的最大值时， 当再产生拒钞图像时， 需要把鉴别图像读出指针所 指节点加入拒钞链表末尾， 同时把拒钞图像头指针所指节点***鉴别图像 读出指针所指位置，加入识别链表，鉴别图像读出指针移动到下一个节点， 拒钞图像头指针指向一下张拒钞图像所在存储区对应的节点， 并对两个节 点存放的与相应节点链接的上一张图像和下一张图像的存储位置进行修 改。
6、一种如权利要求 1至 5所述图像识别***的图像存储控制方法， 包 括如下步骤：
步骤 1 , 通过图像传感器釆集识别对象的图像；
步骤 2, 主控单元根据图像存储区控制器的釆集图像写入指针所指写 入节点将釆集图像写入对应的图像存储区内；
步骤 3 , 将釆集图像写入指针移动至写入节点所记载的下一张图像的 存储位置对应的节点； 根据主控单元的控制进行步骤 1或步骤 4;
步骤 4, 图像识别模块对鉴别图像读出指针所指读出节点对应的图像 存储区内的图像进行鉴别，如果成功鉴别，执行步骤 5 ,否则，执行步骤 6; 步骤 5 , 将鉴别图像读出指针移动至读出节点所记载的下一张图像的 存储位置对应的节点， 继续步骤 4;
步骤 6, 拒钞图像尾指针移动及相关节点存放的位置地址修改， 将拒 钞图像尾指针指向读出节点， 并将读出节点存放的上一张图像的存储位置 所对应的节点存放的下一张图像的存储位置修改为该读出节点存放的下一 张图像存储位置， 将读出节点存放的下一张图像的存储位置所对应的节点 存放的上一张图像的存储位置修改为该读出节点存放的上一张图像存储位 置；
步骤 7, 将拒钞图像尾指针所指节点存放的上一张图像的存储位置修 改为上一张拒钞图像所在存储区对应节点； 步骤 8 , 将鉴别图像读出指针移动至读出节点所存放的下一张图像的 存储位置对应的节点， 继续步骤 4。
7、 如权利要求 6所述图像存储控制方法， 其特征在于， 包括步骤 0, 当图像识别***初始化时， 存储器索引链表初始化为每个节点分别与图像 存储区——对应， 且为每个节点写入与该节点相邻的上一节点所对应的图 像存储区地址和相邻的下一节点所对应的图像存储区地址， 并将釆集图像 写入指针、 鉴别图像读出指针均指向第一个节点， 将拒钞图像尾指针指向 无效位置。
PCT/CN2013/073932 2012-08-09 2013-04-09 一种图像识别***及图像存储控制方法 WO2014023109A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
EP13828011.0A EP2884467B1 (en) 2012-08-09 2013-04-09 Image identification system and image storage control method
US14/416,626 US9838598B2 (en) 2012-08-09 2013-04-09 Image identification system and image storage control method
AU2013302149A AU2013302149B2 (en) 2012-08-09 2013-04-09 Image identification system and image storage control method
ZA2015/00981A ZA201500981B (en) 2012-08-09 2015-02-11 Image identification system and image storage control method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210282874.1A CN102831694B (zh) 2012-08-09 2012-08-09 一种图像识别***及图像存储控制方法
CN201210282874.1 2012-08-09

Publications (1)

Publication Number Publication Date
WO2014023109A1 true WO2014023109A1 (zh) 2014-02-13

Family

ID=47334805

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/073932 WO2014023109A1 (zh) 2012-08-09 2013-04-09 一种图像识别***及图像存储控制方法

Country Status (6)

Country Link
US (1) US9838598B2 (zh)
EP (1) EP2884467B1 (zh)
CN (1) CN102831694B (zh)
AU (1) AU2013302149B2 (zh)
WO (1) WO2014023109A1 (zh)
ZA (1) ZA201500981B (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 (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102831694B (zh) 2012-08-09 2015-01-14 广州广电运通金融电子股份有限公司 一种图像识别***及图像存储控制方法
CN103606221B (zh) * 2013-12-04 2016-01-20 广州广电运通金融电子股份有限公司 清分机故障自动诊断方法以及装置
CN107085885A (zh) * 2017-03-22 2017-08-22 深圳怡化电脑股份有限公司 一种验钞数据的采集方法、存储方法及其装置
WO2020237480A1 (zh) * 2019-05-27 2020-12-03 西门子股份公司 基于图像识别的控制方法与装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101082999A (zh) * 2006-06-02 2007-12-05 日立欧姆龙金融***有限公司 纸叶类识别装置
CN101290689A (zh) * 2007-04-19 2008-10-22 蔡永权 一种钞票成色的检测与识别方法
EP2124204A2 (en) * 2008-05-19 2009-11-25 Emmequadro S.r.l. Control apparatus and method for an automatic dispenser
CN101833815A (zh) * 2010-01-20 2010-09-15 上海古鳌电子机械有限公司 一种纸币清分机的图像识别***及其控制方法
CN102142168A (zh) * 2011-01-14 2011-08-03 哈尔滨工业大学 纸币清分机高速高分辨率号码采集装置及其识别方法
CN102831694A (zh) * 2012-08-09 2012-12-19 广州广电运通金融电子股份有限公司 一种图像识别***及图像存储控制方法

Family Cites Families (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4464787A (en) * 1981-06-23 1984-08-07 Casino Technology Apparatus and method for currency validation
GB2284293B (en) * 1993-11-30 1998-06-03 Mars Inc Article classifying method and apparatus
US6088777A (en) * 1997-11-12 2000-07-11 Ericsson Messaging Systems, Inc. Memory system and method for dynamically allocating a memory divided into plural classes with different block sizes to store variable length messages
WO2000011603A2 (en) * 1998-08-20 2000-03-02 Apple Computer, Inc. Graphics processor with pipeline state storage and retrieval
US6499028B1 (en) * 1999-03-31 2002-12-24 International Business Machines Corporation Efficient identification of candidate pages and dynamic response in a NUMA computer
US6427195B1 (en) * 2000-06-13 2002-07-30 Hewlett-Packard Company Thread local cache memory allocator in a multitasking operating system
US7113516B1 (en) * 2000-11-28 2006-09-26 Texas Instruments Incorporated Transmit buffer with dynamic size queues
US7245304B2 (en) * 2001-11-16 2007-07-17 Lockheed Martin Corporation System and method for managing compressed graphic image data
US7111289B2 (en) * 2001-12-21 2006-09-19 Agere Systems, Inc. Method for implementing dual link list structure to enable fast link-list pointer updates
US6785674B2 (en) * 2003-01-17 2004-08-31 Intelitrac, Inc. System and method for structuring data in a computer system
JP4349849B2 (ja) * 2003-06-17 2009-10-21 日立オムロンターミナルソリューションズ株式会社 預金自動預け払い機及び紙幣の追跡方法
US7310719B2 (en) * 2003-07-24 2007-12-18 Sap Aktiengesellschaft Memory management tile optimization
JP4422515B2 (ja) * 2004-03-11 2010-02-24 日立オムロンターミナルソリューションズ株式会社 紙葉類鑑別装置
HUP0401870A2 (en) * 2004-09-17 2006-04-28 3D Histech Kft Method and computer program product for storing high volume coherent primarily 3d information assuring fast retrieval and transmission
CN1834919A (zh) * 2005-03-18 2006-09-20 株式会社东芝 纸张处理设备和纸张处理方法
US7975149B2 (en) * 2006-06-30 2011-07-05 Microsoft Corporation Software protection using data structures
EP2183671B1 (en) * 2007-08-29 2018-08-29 Lifescan Scotland Limited A data management system and method
CN100498740C (zh) * 2007-09-11 2009-06-10 腾讯科技（深圳）有限公司 一种数据缓存处理方法、***及数据缓存装置
US20090294244A1 (en) * 2008-05-30 2009-12-03 Harold Charych Currency Validator with Rejected Bill Image Storage
US8682056B2 (en) * 2008-06-30 2014-03-25 Ncr Corporation Media identification
CN101452588B (zh) * 2008-12-09 2012-07-04 广州广电运通金融电子股份有限公司 钞票识别***及具有该***的钞票识别方法
CN102117278B (zh) * 2009-12-31 2016-10-05 联想(北京)有限公司 链表的创建方法及***、数据的查找方法及***
US9262554B1 (en) * 2010-02-16 2016-02-16 Pmc-Sierra Us, Inc. Management of linked lists within a dynamic queue system
JP5584028B2 (ja) * 2010-07-07 2014-09-03 日立オムロンターミナルソリューションズ株式会社 紙幣処理装置、紙幣処理システム、紙幣情報管理サーバ、紙幣情報管理プログラムおよび紙幣処理方法
CN102253898B (zh) * 2011-07-22 2013-10-30 杭州海康威视数字技术股份有限公司 一种图像数据的内存管理方法及装置
CN102521278B (zh) * 2011-11-24 2014-03-05 广州广电运通金融电子股份有限公司 图像数据快速存储方法、有价文件识别方法及其识别装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101082999A (zh) * 2006-06-02 2007-12-05 日立欧姆龙金融***有限公司 纸叶类识别装置
CN101290689A (zh) * 2007-04-19 2008-10-22 蔡永权 一种钞票成色的检测与识别方法
EP2124204A2 (en) * 2008-05-19 2009-11-25 Emmequadro S.r.l. Control apparatus and method for an automatic dispenser
CN101833815A (zh) * 2010-01-20 2010-09-15 上海古鳌电子机械有限公司 一种纸币清分机的图像识别***及其控制方法
CN102142168A (zh) * 2011-01-14 2011-08-03 哈尔滨工业大学 纸币清分机高速高分辨率号码采集装置及其识别方法
CN102831694A (zh) * 2012-08-09 2012-12-19 广州广电运通金融电子股份有限公司 一种图像识别***及图像存储控制方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP2884467A4 *

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
US9838598B2 (en) 2017-12-05
EP2884467B1 (en) 2020-08-19
US20150271397A1 (en) 2015-09-24
CN102831694A (zh) 2012-12-19
EP2884467A1 (en) 2015-06-17
AU2013302149B2 (en) 2016-05-12
CN102831694B (zh) 2015-01-14
ZA201500981B (en) 2016-05-25
AU2013302149A1 (en) 2015-02-12
EP2884467A4 (en) 2015-08-19

Similar Documents

Publication Publication Date Title
WO2014023109A1 (zh) 一种图像识别***及图像存储控制方法
JP4362339B2 (ja) 紙幣取扱装置
US8155423B2 (en) System and method for tracing bank notes
RU2596591C2 (ru) Автоматическое устройство для транзакций
CN103377507B (zh) 纸张类处理装置以及纸张类处理方法
JP3611295B2 (ja) コンピュータシステム、メモリ管理方法及び記憶媒体
CN104169979B (zh) 纸币处理装置以及纸币处理方法
TW480395B (en) Record regenerating device
CN106407128A (zh) 用于向持久性存储器多播数据的方法和***
CN104167049A (zh) Atm机纸币处理装置
CN103366166B (zh) 纸张号码读取装置和方法以及***、纸张处理装置和方法
US20030051103A1 (en) Shared memory system including hardware memory protection
JP3988340B2 (ja) 画像処理装置およびそれを備えた複写装置
JP2012123696A (ja) 識別装置の損券レベルの自動設定
CN108470404B (zh) 用于读卡器的闸门控制方法、装置、金融设备及存储介质
US10403078B2 (en) Medium processing apparatus and method thereof
CN104573134B (zh) 数据采集方法和数据采集设备
CN204087326U (zh) 一种具备币种识别的纸币处理装置
CN110136320A (zh) 一种钞票清分处理***及方法
KR101384906B1 (ko) 이미지 처리 장치 및 방법
KR101364880B1 (ko) 매체처리장치, 매체처리방법 및 이를 이용한 금융기기
CN106652175A (zh) 一种控制钞票识别单元采集信息的方法和装置
CN104134270A (zh) 一种具备币种识别的纸币处理装置
CN116311664A (zh) 出钞方法、出钞设备和计算机可读存储介质
CN105139517A (zh) 一种交易处理装置

Legal Events

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

Ref document number: 13828011

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 14416626

Country of ref document: US

Ref document number: 2013828011

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2015000275

Country of ref document: CL

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2013302149

Country of ref document: AU

Date of ref document: 20130409

Kind code of ref document: A