WO2015085811A1 - 一种基于厚度信号识别的钞票识别方法及装置 - Google Patents

一种基于厚度信号识别的钞票识别方法及装置 Download PDF

Info

Publication number
WO2015085811A1
WO2015085811A1 PCT/CN2014/087404 CN2014087404W WO2015085811A1 WO 2015085811 A1 WO2015085811 A1 WO 2015085811A1 CN 2014087404 W CN2014087404 W CN 2014087404W WO 2015085811 A1 WO2015085811 A1 WO 2015085811A1
Authority
WO
WIPO (PCT)
Prior art keywords
thickness
area
signal
banknote
recognition
Prior art date
Application number
PCT/CN2014/087404
Other languages
English (en)
French (fr)
Inventor
王晓亮
梁添才
陈�光
陈定喜
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
Priority to US15/021,923 priority Critical patent/US10008065B2/en
Priority to EP14869270.0A priority patent/EP3082112B1/en
Priority to AU2014361530A priority patent/AU2014361530B2/en
Publication of WO2015085811A1 publication Critical patent/WO2015085811A1/zh
Priority to ZA2016/02298A priority patent/ZA201602298B/en

Links

Images

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
    • G07D7/16—Testing the dimensions
    • G07D7/164—Thickness
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • G07D7/2008—Testing patterns thereon using pre-processing, e.g. de-blurring, averaging, normalisation or rotation
    • 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
    • G07D7/2016—Testing patterns thereon using feature extraction, e.g. segmentation, edge detection or Hough-transformation

Definitions

  • Embodiments of the present invention relate to the field of banknote identification, and in particular, to a banknote recognition method and apparatus based on thickness signal recognition.
  • Variable coinage refers to a new banknote that is reassembled by applying two pieces of different genuine or counterfeit banknotes, applying paste, digging and other methods.
  • the criminals created the coinage to realize the appreciation of the banknotes.
  • the circulation of such coins in the market seriously affected the normal financial order and threatened the national financial security.
  • the coinage does not have the characteristics of counterfeit banknotes on the spectral image, so the application of digital image processing cannot effectively identify the coinage.
  • the damaged currency due to the long circulation time of the banknotes in the market, causes the banknotes to fall off, crack, or tear into two halves, and the tapes are patched and restored to restore the original appearance.
  • the circulation of such banknotes in the market seriously impressed the image of the country, and such banknotes were recycled and concentrated in accordance with the standards of the People's Bank of China.
  • the identification of the two banknotes cannot be effectively realized at the image processing angle, and both of the banknotes have their thickness characteristics caused by physical changes such as patching and patching. There is a significant change, so the application of the thickness identification method can effectively realize the identification of the banknote.
  • Embodiments of the present invention provide a banknote identification method and apparatus based on thickness signal recognition, In combination with the two methods of recognizing the banknotes by the two methods of the upper area recognition method and the downward area identification method, the abnormal banknotes can be easily and efficiently separated.
  • the fusion result is identified to obtain a recognition result.
  • the method further includes:
  • the banknotes are sorted and sent to a position corresponding to the category based on the recognition result.
  • the step of collecting the thickness signal of the banknote includes:
  • the thickness signal is a collection of multiplexed signals.
  • the step of preprocessing the thickness signal includes:
  • An effective signal region determination is performed on the denoising processing signal to obtain an effective region.
  • the step of identifying the thickness signal by using the upward area identification method includes:
  • the area of the thickness feature is defined as: the area of the area formed by the sliding window and the upper threshold line and the curve of the thickness signal; the width of the sliding window and the sliding step are preset values; the upper threshold line is in the The horizontal line above the curve;
  • the sliding window is gradually moved according to a preset sliding step length, and the area of the thickness feature of each step is calculated;
  • the thickness signal is identified according to the area on the minimum thickness feature and the area average value of the thickness characteristics of each step, to obtain an upward processing recognition result.
  • the step of identifying the thickness signal by using the down area identification method includes:
  • the area under the thickness feature is defined as: the area of the area formed by the sliding window and the lower threshold line and the curve of the thickness signal; the width and the sliding step of the sliding window are preset values; the lower threshold line is in the The horizontal line below the peak of the curve;
  • the sliding window is gradually moved according to a preset sliding step length, and the area under each thickness characteristic is calculated;
  • the thickness signal is identified according to the area under the minimum thickness feature and the area average under the thickness characteristic of each step, and the recognition result is processed downward.
  • the step of fusing the upward processing identification result and the downward processing identification result according to the preset rule includes:
  • the abnormal region of the upward processing recognition result and the abnormal region of the downward processing recognition result are merged according to a preset rule to obtain a fusion result.
  • the banknote identification device based on thickness signal recognition comprises: a thickness sensor, a DSP chip, an embedded system and a mechanical motion module;
  • the thickness sensor is connected to the DSP chip for collecting a thickness signal of the banknote
  • the DSP chip is connected to the embedded system, configured to analyze and identify the banknote according to the thickness signal, and send the identification result to the embedded system;
  • the embedded system is connected to the mechanical motion module, and configured to control the mechanical motion module according to the recognition result;
  • the mechanical motion module is configured to classify the banknotes according to a control instruction set of the embedded system and send the banknotes to a position corresponding to the category.
  • the thickness sensor is a multi-way thickness sensor.
  • the DSP chip includes:
  • a preprocessing module which is respectively connected to the up processing module and the down processing module, for preprocessing the thickness signal
  • the upward processing module is connected to the fusion module, and is configured to identify the thickness signal by using an upward area recognition method to obtain an upward processing recognition result;
  • the down processing module is connected to the fusion module, and is configured to identify the thickness signal by using a downward area identification method to obtain a downward processing recognition result;
  • the merging module is connected to the identification module, and configured to fuse the upward processing identification result and the downward processing identification result according to a preset rule to obtain a fusion result;
  • the identification module identifies the fusion result to obtain a recognition result.
  • the thickness signal of the banknote is first collected; then the thickness signal is pre-processed; then the thickness signal is identified by the upward area recognition method to obtain an upward processing recognition result; and the downward area identification method is used to identify the The thickness signal is obtained, and the recognition result is processed downward; then the up process identification result and the down process recognition result are merged according to a preset rule to obtain a fusion result; finally, the fusion result is identified to obtain a recognition result. Since the banknote recognition method and device based on the thickness signal identification according to the embodiment of the present invention, after collecting the thickness signal of the banknote, the method of identifying the thickness signal by combining the two methods of the upper area identification method and the downward area identification method can be simple. Efficiently separate abnormal banknote recognition.
  • FIG. 1 is a flow chart of a first embodiment of a banknote identification method based on thickness signal recognition according to an embodiment of the present invention
  • FIG. 2 is a flow chart of a second embodiment of a banknote identification method based on thickness signal recognition according to an embodiment of the present invention
  • FIG. 3 is a schematic view showing a thickness sensor for collecting a banknote thickness signal according to a second embodiment of the present invention
  • FIG. 4 is a schematic view showing a thickness signal of an i-th road of a normal banknote according to a second embodiment of the present invention.
  • FIG. 5 is a schematic diagram of identifying a thickness signal by using an upward area recognition method according to a second embodiment of the present invention.
  • FIG. 6 is a schematic diagram of identifying a thickness signal by using a downward area recognition method in a second embodiment of the present invention.
  • Figure 7 is a schematic view showing the change of banknotes in the second embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a coin thickness signal according to a second embodiment of the present invention.
  • FIG. 9 is a schematic diagram showing an effective area of a coin thickness signal according to a second embodiment of the present invention.
  • FIG. 10 is a flow chart of a second embodiment of a method for identifying an abnormal thickness banknote according to an embodiment of the present invention.
  • Figure 11 is a schematic view showing the change of banknotes in the second embodiment of the present invention.
  • FIG. 12 is a schematic structural diagram of an embodiment of a banknote identification device based on thickness signal recognition according to an embodiment of the present invention.
  • the embodiment of the invention provides a method and a device for recognizing a banknote based on thickness signal recognition, which can easily and efficiently dispose an abnormal banknote by combining the two methods of the upper area identification method and the downward area identification method for recognizing the banknote. Identify separation.
  • the method and the device for identifying the thickness signal according to the embodiment of the present invention can be used not only for the identification of the banknote, but also for identifying the sheet-like file such as a check, which is not limited herein.
  • the method and apparatus of the embodiment of the present invention will be described below by taking the identification of the banknote as an example. Although the identification of the banknote is only taken as an example, it should not be construed as limiting the method and apparatus of the present invention.
  • a first embodiment of a banknote identification method based on thickness signal recognition in an embodiment of the present invention includes:
  • the thickness sensor can first be used to collect the thickness signal of the banknote.
  • the thickness signal After the thickness signal is obtained, the thickness signal can be preprocessed to input the thickness signal. Line identification.
  • the thickness signal After the thickness signal is preprocessed, the thickness signal can be identified by the upward area identification method, and the recognition result is processed upward.
  • the thickness signal After pre-processing the thickness signal, the thickness signal can be identified by the down-area identification method to obtain a downward processing recognition result.
  • step 103 may be performed simultaneously with the step 104, or may be performed after the step 104, and is not limited to the step 104 before, and is not limited herein.
  • the recognition result is processed upward and the recognition result is processed downward, the recognition result is processed upward according to the preset rule, and the recognition result is processed downward to obtain the fusion result.
  • the fusion result is obtained, and the fusion result can be identified and the recognition result is obtained.
  • the thickness signal of the banknote is first collected; then the thickness signal is preprocessed; then the thickness signal is identified by the upward area recognition method to obtain the upward processing recognition result; and the thickness signal is identified by the downward area identification method to obtain the direction
  • the recognition result is processed downward; then the recognition result is processed upward according to the preset rule and the recognition result is processed downward to obtain the fusion result; finally, the fusion result is identified, and the recognition result is obtained. Since the banknote recognition method based on the thickness signal identification is used in the embodiment of the present invention, after the thickness signal of the banknote is collected, the thickness signal can be identified by combining the two methods of the upper area identification method and the downward area identification method to easily and efficiently The abnormal banknotes are separated.
  • a second embodiment of the banknote identification method based on the thickness signal recognition includes:
  • the thickness sensor can first be used to collect the thickness signal of the banknote.
  • the thickness sensor described above may be a multi-channel thickness sensor, correspondingly, the thickness of the multi-channel thickness sensor is collected.
  • the degree signal is a collection of multiple acquisition signals.
  • each sensor unit will acquire a relatively independent one-dimensional thickness signal, which is called a thickness signal, and M sensors collectively collect M opposite thickness signals, which are called M-path thickness signals, and each thickness signal is The number of sampling points is recorded as N, and the thickness signal collected by the ith sensor is represented by Signal(i, j), which is called the i-th thickness signal.
  • FIG. 4 is a schematic diagram of the thickness signal of the i-th road of the normal banknote.
  • the thickness signal can be pre-processed to identify the thickness signal.
  • the pre-processing operation may include: sampling the thickness signal to obtain a sampling processing signal; performing denoising processing on the sampling processing signal to obtain a denoising processing signal; and performing effective signal region determination on the denoising processing signal to obtain an effective region. .
  • the above pretreatment operation is mainly used to reduce the influence of the outside on the thickness signal.
  • the thickness signal After the thickness signal is preprocessed, the thickness signal can be identified by the upward area identification method, and the recognition result is processed upward.
  • the area of the thickness feature is defined as: the area of the area formed by the sliding window x and the upper threshold line T and the curve signal Signal(i, t) UpS(i, x); the width of the sliding window and the step of the sliding step are The preset value; the upper threshold line T is a horizontal line above the curve; referring to FIG.
  • the specific process of identifying the thickness signal by using the upward area recognition method may include: gradually moving the sliding window x according to the preset sliding step step, and calculating The area of the thickness characteristic of each step is UpS(i, x); the area of the minimum thickness feature in the area of the thickness feature of each step is obtained; the area average value of the thickness features of each step is calculated according to the area of the thickness characteristic of each step; The area and the average value of the area on each step of the thickness feature identify the thickness signal and obtain an upward processing recognition result.
  • the thickness signal After pre-processing the thickness signal, the thickness signal can be identified by the down-area identification method to obtain a downward processing recognition result.
  • the area under the thickness characteristic is defined as: the area of the area formed by the sliding window x and the lower threshold line T and the curve signal Signal(i, t), DownS(i, x); the width of the sliding window and the step of the sliding step are Preset value; lower threshold line T is the horizontal line below the peak of the curve; see Figure 6, using the downward
  • the specific process of identifying the thickness signal by the product identification method may include: gradually moving the sliding window x according to the preset sliding step step, and calculating the area under the thickness characteristic of each step DownS(i, x); obtaining the minimum of the area under each thickness characteristic The area under the thickness characteristic; the area average value under each step thickness characteristic is calculated according to the area under each thickness characteristic; the thickness signal is identified according to the area under the minimum thickness feature and the area average value under each step thickness characteristic, and the downward processing recognition result is obtained.
  • step 203 may be performed simultaneously with step 204, or may be performed after step 204, and is not limited to before step 204 in this embodiment, and is not limited herein.
  • the specific process of combining the upward processing of the recognition result and the downward processing of the recognition result according to the preset rule may include: acquiring an upward processing recognition result and processing the abnormal area within the recognition result; and processing the abnormal area of the recognition result upward according to the preset rule.
  • the abnormal regions of the recognition result are processed downward to be fused, and the fusion result is obtained.
  • the fusion result is obtained, and the fusion result can be identified and the recognition result is obtained. If the abnormal region of the fusion result covers the authentication area, the banknote is identified as a coin, and if the area of the abnormal region of the fusion result exceeds a fixed threshold, the banknote is identified as a damaged banknote, otherwise the banknote is identified as a banknote.
  • the above fixed threshold value is set in advance according to the banknote to be tested and the device structure, and is not limited herein.
  • the banknotes are sorted and sent to the position corresponding to the category according to the recognition result, for example, different types of banknotes can be transferred to the predetermined position, thereby completing the banknote recognition.
  • the first step collecting the thickness signal of the banknote
  • FIG. 7 is a schematic diagram of the banknote change of the coin.
  • the foreign matter posting area is covered by four thickness sensors.
  • the second step pre-processing the thickness signal
  • determining the coin thickness signal determines the start point, the end point, and the start channel and the end channel of each channel, and obtains an effective area of the thickness signal, and the effective area is as shown in FIG.
  • the black dotted line is shown in the selection area.
  • the third step identifying the thickness signal by using the upward area recognition method, and obtaining the upward processing recognition result;
  • the area value of the thickness signal corresponding to the nth sliding window is UpS(i, n), and the calculation formula is
  • P start (i) is the starting point of the i-th thickness signal banknote thickness collection area
  • P end (i) is the i-th thickness signal banknote thickness collection area termination point
  • n is the nth sliding window
  • MThkSignal(i , j) is the amplitude of the jth sampling point of the i-th signal after median filtering
  • C start is the signal starting path
  • C end is the signal ending path
  • N(i) is the number of sliding windows of the ith thickness signal ;
  • min() is the minimum function
  • the upper area average is calculated as
  • a sliding window having a size of ⁇ and 3 ⁇ is applied to identify each signal by using an upward area recognition method, and the first thickness signal, the second thickness signal, and the third thickness are detected. There is an abnormality in the signal.
  • the area where the abnormal area is recorded is Area R (1), Area R (2), Area R (3), and the foreign areas of the three areas corresponding to the banknote are S s (1), S s ( 2), S s (3).
  • the fourth step identifying the thickness signal by using the downward area identification method, and obtaining the downward processing recognition result;
  • each parameter in the above formula is similar to the method of identifying the upward area.
  • the lower area value average DownS avg (i) is calculated as
  • the thickness signal of the road is abnormal, and the abnormal area area S s (i) and the abnormal area position Area R (i) are recorded. Otherwise, it is determined that there is no thickness signal abnormality in the road.
  • the size of the sliding window is determined by applying the downward area identification method to identify each signal, and the abnormality of the thickness signal of the i-th path is detected, and the area of the abnormal area is recorded as Area R (4).
  • the area of the foreign object corresponding to the banknote is S s (4).
  • Step 5 merging the recognition result according to the preset rule and processing the recognition result downward to obtain the fusion result
  • the relevant constraints include: the location of the authentication area Area N (this parameter is set according to different currencies and different denominations, for example, the RMB 100 yuan for the watermark area and the national emblem area); the thickness anomaly area threshold T S (this value can be According to different testing standards, for example, ECB ECB currency circulation standard is 4cm 2 ).
  • the abnormal thickness area covers the authentication area Area N , and the banknote is judged to be a coin.
  • the abnormal thickness area does not cover the authentication area, and the abnormal area is larger than Th s , and the banknote is judged to be a damaged banknote.
  • a banknote that does not fall into the above two cases determines that the sheet is a banknote.
  • the thickness anomaly area Area R (1), Area R (2), Area R (3) covers the authentication area, and the abnormal area area S S (1) + S S (2) + S S (3 ) +S S (4) is greater than the threshold Th s , and the banknote is judged to be a coin.
  • the thickness signal of the banknote is first collected; then the thickness signal is preprocessed; Then, the upper area recognition method is used to identify the thickness signal, and the recognition result is processed upwards; and the thickness signal is identified by the downward area recognition method to obtain the downward processing recognition result; then the upward processing recognition result and the downward processing recognition result are merged according to the preset rule. , the fusion result is obtained; finally, the fusion result is identified, and the recognition result is obtained. Since the banknote recognition method based on the thickness signal identification is used in the embodiment of the present invention, after the thickness signal of the banknote is collected, the thickness signal can be identified by combining the two methods of the upper area identification method and the downward area identification method to easily and efficiently The abnormal banknotes are separated.
  • an embodiment of the banknote identification device based on the thickness signal identification in the embodiment of the present invention includes:
  • a thickness sensor 1201 connected to the DSP chip 1202, for collecting a thickness signal of the banknote;
  • the DSP chip 1202 is connected to the embedded system 1203 for analyzing and identifying the banknote according to the thickness signal, and transmitting the recognition result to the embedded system 1203;
  • the mechanical motion module 1204 is configured to classify the banknotes according to the control instruction set of the embedded system 1204 and send them to a location corresponding to the category.
  • the thickness sensor 1201 is a multi-way thickness sensor.
  • the DSP chip 1202 includes:
  • the preprocessing module is respectively connected to the up processing module and the down processing module for preprocessing the thickness signal
  • the upward processing module is connected to the fusion module for identifying the thickness signal by using the upward area recognition method to obtain an upward processing recognition result;
  • the downward processing module is connected to the fusion module for identifying the thickness signal by using the downward area identification method to obtain a downward processing recognition result;
  • the fusion module is connected to the identification module, and is configured to process the recognition result according to the preset rule fusion Processing the recognition result downwards to obtain a fusion result;
  • the identification module identifies the fusion result and obtains the recognition result.
  • the thickness sensor 1201 first collects the thickness signal of the banknote, and transmits the thickness signal to the DSP chip 1202 for analysis and identification. After the DSP chip 1202 obtains the recognition result, the recognition result is transmitted to the embedded system 1203, and is embedded.
  • the system 1203 controls the mechanical movement module 1204 to transfer the banknotes, the damaged banknotes, and the altered coins to different banknotes to realize the classification of different types of banknotes.
  • the banknote identification device based on the thickness signal identification includes: a thickness sensor 1201, a DSP chip 1202, an embedded system 1203, and a mechanical motion module 1204.
  • the thickness sensor 1201 is connected to the DSP chip 1202 for collecting the thickness of the banknote.
  • the DSP chip 1202 is connected to the embedded system 1203 for analyzing and identifying the banknote according to the thickness signal, and transmitting the recognition result to the embedded system 1203; the embedded system 1203, and the mechanical motion module 1204, for identifying The result controls the mechanical motion module;
  • the mechanical motion module 1204 is configured to sort the banknotes according to the control instruction set of the embedded system 1204 and send them to a location corresponding to the category. Since the thickness sensor 1201 collects the thickness signal of the banknote based on the thickness signal identification, the thickness sensor 1201 combines the upper area identification method and the downward area identification method to identify the thickness signal. In this way, abnormal banknote recognition can be separated easily and efficiently.
  • the storage medium may be a read only memory, a magnetic disk or an optical disk or the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Inspection Of Paper Currency And Valuable Securities (AREA)
  • Controlling Sheets Or Webs (AREA)
  • Testing Of Coins (AREA)

Abstract

一种基于厚度信号识别的钞票识别方法及装置，通过结合向上面积识别法和向下面积识别法两种厚度识别方法对钞票进行识别的方式，能够简单高效地将异常钞票识别分离。所述识别方法包括：采集钞票的厚度信号；对所述厚度信号进行预处理；利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；对所述融合结果进行识别，得到识别结果。

Description

一种基于厚度信号识别的钞票识别方法及装置
本申请要求于2013年12月12日提交中国专利局、申请号为201310681853.1、发明名称为“一种基于厚度信号识别的钞票识别方法及装置”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
本发明实施例涉及纸币识别领域，尤其涉及一种基于厚度信号识别的钞票识别方法及装置。
背景技术
变造币是指将两张来自不同真钞或伪钞的残券，应用粘贴、挖补等方法，重新组合成的一张新钞票。不法分子制造变造币用以实现钞票增值，这种变造币在市面上的流通严重影响了正常的金融秩序，威胁国家金融安全。同时因为变造币为真钞和真钞拼接而成，因此在光谱图像上，变造币不具有假钞特征，故应用数字图像处理的方式不能够有效识别变造币。
残损币，因钞票在市面的流通时间较长，造成钞票掉角、裂痕、或撕为两半，应用胶带粘贴等方法，将这些钞票修补、恢复原貌。这种钞票在市面上的流通严重印象了国家形象，根据人民银行标准需要将此类钞票回收并集中销毁。
由于变造币、残损币图像上的特征与正常的流通钞票一致，因而在图像处理角度不能有效实现两种钞票的识别，而这两种钞票均因粘贴挖补等物理变化导致了其厚度特征发生了明显变化，因此应用厚度识别方法能够有效实现所述钞票的识别。
发明内容
本发明实施例提供了一种基于厚度信号识别的钞票识别方法及装置，通过 结合向上面积识别法和向下面积识别法两种厚度识别方法对钞票进行识别的方式，能够简单高效地将异常钞票识别分离。
本发明实施例提供的基于厚度信号识别的钞票识别方法，包括：
采集钞票的厚度信号；
对所述厚度信号进行预处理；
利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；
利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；
根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；
对所述融合结果进行识别，得到识别结果。
可选的，
在步骤对所述融合结果进行识别，得到识别结果之后还包括：
根据所述识别结果将所述钞票分类并送至与类别相对应的位置。
可选的，
步骤采集钞票的厚度信号包括：
利用多路厚度传感器采集钞票的厚度信号；
所述厚度信号为多路采集信号的集合。
可选的，
步骤对所述厚度信号进行预处理包括：
对所述厚度信号进行抽样处理，得到抽样处理信号；
对所述抽样处理信号进行去噪处理，得到去噪处理信号；
对所述去噪处理信号进行有效信号区域确定，得到有效区域。
可选的，
步骤所述利用向上面积识别法识别所述厚度信号包括：
厚度特征上面积定义为：滑动窗口与上阈值线及所述厚度信号的曲线所形成的区域的面积；所述滑动窗口的宽度和滑动步长为预设值；所述上阈值线为在所述曲线之上的水平线；
根据预设滑动步长逐步移动所述滑动窗口，并计算每步厚度特征上面积；
获取所述每步厚度特征上面积中的最小厚度特征上面积；
根据所述每步厚度特征上面积计算各步厚度特征上面积平均值；
根据所述最小厚度特征上面积和所述各步厚度特征上面积平均值识别所述厚度信号，得到向上处理识别结果。
可选的，
步骤所述利用向下面积识别法识别所述厚度信号包括：
厚度特征下面积定义为：滑动窗口与下阈值线及所述厚度信号的曲线所形成的区域的面积；所述滑动窗口的宽度和滑动步长为预设值；所述下阈值线为在所述曲线峰值之下的水平线；
根据预设滑动步长逐步移动所述滑动窗口，并计算每步厚度特征下面积；
获取所述每步厚度特征下面积中的最小厚度特征下面积；
根据所述每步厚度特征下面积计算各步厚度特征下面积平均值；
根据所述最小厚度特征下面积和所述各步厚度特征下面积平均值识别所述厚度信号，得到向下处理识别结果。
可选的，
步骤所述根据预设规则融合所述向上处理识别结果和所述向下处理识别结果包括：
获取所述向上处理识别结果和所述向下处理识别结果内的异常区域；
根据预设规则将所述向上处理识别结果的异常区域和所述向下处理识别结果的异常区域进行融合，得到融合结果。
本发明实施例提供的基于厚度信号识别的钞票识别装置，包括：厚度传感器、DSP芯片、嵌入式系统和机械运动模块；
所述厚度传感器，与所述DSP芯片相连，用于采集钞票的厚度信号；
所述DSP芯片，与所述嵌入式系统相连，用于根据所述厚度信号对所述钞票进行分析识别，并将所述识别结果发送至所述嵌入式系统；
所述嵌入式系统，与所述机械运动模块相连，用于根据所述识别结果控制所述机械运动模块；
所述机械运动模块，用于根据所述嵌入式系统的控制指令集将所述钞票进行分类并送至与类别相对应的位置。
可选的，
所述厚度传感器为多路厚度传感器。
可选的，
所述DSP芯片包括：
预处理模块，分别与向上处理模块及向下处理模块相连，用于对所述厚度信号进行预处理；
所述向上处理模块，与融合模块相连，用于利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；
所述向下处理模块，与所述融合模块相连，用于利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；
所述融合模块，与识别模块相连，用于根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；
所述识别模块，对所述融合结果进行识别，得到识别结果。
本发明实施例中，首先采集钞票的厚度信号；然后对所述厚度信号进行预处理；接着利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；并利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；然后根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；最后对所述融合结果进行识别，得到识别结果。由于本发明实施例基于厚度信号识别的钞票识别方法及装置，采集钞票的厚度信号之后，通过结合向上面积识别法和向下面积识别法两种厚度识别方法对厚度信号进行识别的方式，能够简单高效地将异常钞票识别分离。
附图说明
图1为本发明实施例中基于厚度信号识别的钞票识别方法第一实施例流程图；
图2为本发明实施例中基于厚度信号识别的钞票识别方法第二实施例流程图；
图3为本发明第二实施例中厚度传感器采集钞票厚度信号的示意图；
图4为本发明第二实施例中正常钞票第i路厚度信号示意图；
图5为本发明第二实施例中利用向上面积识别法识别厚度信号的示意图；
图6为本发明第二实施例中利用向下面积识别法识别厚度信号的示意图；
图7为本发明第二实施例中变造币走钞示意图；
图8为本发明第二实施例中变造币厚度信号示意图；
图9为本发明第二实施例中变造币厚度信号有效区域示意图；
图10为本发明实施例中厚度异常钞票的识别方法第二实施例流程图；
图11为本发明第二实施例中变造币走钞示意图；
图12为本发明实施例中基于厚度信号识别的钞票识别装置实施例结构示意图。
具体实施方式
本发明实施例提供了一种基于厚度信号识别的钞票识别方法及装置，通过结合向上面积识别法和向下面积识别法两种厚度识别方法对钞票进行识别的方式，能够简单高效地将异常钞票识别分离。
需要说明的是，本发明实施例基于厚度信号识别的方法和装置不仅可以用于钞票的识别，还可以用于识别支票等薄片类文件，在此处不作限定。下面以钞票的识别为例对本发明实施例的方法及装置进行说明，虽然仅以钞票的识别为例进行说明，但是不应将此作为本发明方法及装置的限定。
请参阅图1，本发明实施例中基于厚度信号识别的钞票识别方法的第一实施例包括：
101、采集钞票的厚度信号；
在对钞票进行识别之前，首先可以利用厚度传感器采集钞票的厚度信号。
102、对厚度信号进行预处理；
得到厚度信号之后，可以对厚度信号进行预处理操作，以便对厚度信号进 行识别。
103、利用向上面积识别法识别厚度信号，得到向上处理识别结果；
对厚度信号进行预处理之后，可以利用向上面积识别法识别厚度信号，得到向上处理识别结果。
104、利用向下面积识别法识别厚度信号，得到向下处理识别结果；
对厚度信号进行预处理之后，可以利用向下面积识别法识别厚度信号，得到向下处理识别结果。
需要说明的是，步骤103可以与步骤104同时执行，也可以在步骤104之后执行，而不限于本实施例中在步骤104之前，在此处不作限定。
105、根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果；
得到向上处理识别结果及向下处理识别结果之后，可以根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果。
106、对融合结果进行识别，得到识别结果。
得到融合结果，可以对融合结果进行识别，并得到识别结果。
本发明实施例中，首先采集钞票的厚度信号；然后对厚度信号进行预处理；接着利用向上面积识别法识别厚度信号，得到向上处理识别结果；并利用向下面积识别法识别厚度信号，得到向下处理识别结果；然后根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果；最后对融合结果进行识别，得到识别结果。由于本发明实施例基于厚度信号识别的钞票识别方法，采集钞票的厚度信号之后，通过结合向上面积识别法和向下面积识别法两种厚度识别方法对厚度信号进行识别的方式，能够简单高效地将异常钞票识别分离。
上面简单介绍了本发明基于厚度信号识别的钞票识别方法的第一实施例，下面对本发明基于厚度信号识别的钞票识别方法的第二实施例进行详细的描述，请参阅图2，本发明实施例中基于厚度信号识别的钞票识别方法的第二实施例包括：
201、采集钞票的厚度信号；
在对钞票进行识别之前，首先可以利用厚度传感器采集钞票的厚度信号。上述的厚度传感器可以是多路厚度传感器，相应的，多路厚度传感器采集的厚 度信号为多路采集信号的集合。
请参阅图3，每个传感器单元将采集一条相对独立的一维厚度信号，称为一路厚度信号，M个传感器总计采集M条对立的厚度信号，称为M路厚度信号，每路厚度信号的采样点数记为N，用Signal(i,j)表示第i个传感器采集到的厚度信号，称为第i路厚度信号。请参阅图4，为正常钞票第i路厚度信号示意图。
202、对厚度信号进行预处理；
得到厚度信号之后，可以对厚度信号进行预处理操作，以便对厚度信号进行识别。上述的预处理操作具体可以包括：对厚度信号进行抽样处理，得到抽样处理信号；对抽样处理信号进行去噪处理，得到去噪处理信号；对去噪处理信号进行有效信号区域确定，得到有效区域。上述的预处理操作主要用于降低外界对厚度信号的影响。
203、利用向上面积识别法识别厚度信号，得到向上处理识别结果；
对厚度信号进行预处理之后，可以利用向上面积识别法识别厚度信号，得到向上处理识别结果。
其中厚度特征上面积定义为：滑动窗口x与上阈值线T及厚度信号的曲线Signal(i,t)所形成的区域的面积UpS(i,x)；滑动窗口的宽度和滑动步长step为预设值；上阈值线T为在曲线之上的水平线；请参阅图5，利用向上面积识别法识别厚度信号的具体过程可以包括：根据预设滑动步长step逐步移动滑动窗口x，并计算每步厚度特征上面积UpS(i,x)；获取每步厚度特征上面积中的最小厚度特征上面积；根据每步厚度特征上面积计算各步厚度特征上面积平均值；根据最小厚度特征上面积和各步厚度特征上面积平均值识别厚度信号，得到向上处理识别结果。
204、利用向下面积识别法识别厚度信号，得到向下处理识别结果；
对厚度信号进行预处理之后，可以利用向下面积识别法识别厚度信号，得到向下处理识别结果。
其中厚度特征下面积定义为：滑动窗口x与下阈值线T及厚度信号的曲线Signal(i,t)所形成的区域的面积DownS(i,x)；滑动窗口的宽度和滑动步长step为预设值；下阈值线T为在曲线峰值之下的水平线；请参阅图6，利用向下面 积识别法识别厚度信号的具体过程可以包括：根据预设滑动步长step逐步移动滑动窗口x，并计算每步厚度特征下面积DownS(i,x)；获取每步厚度特征下面积中的最小厚度特征下面积；根据每步厚度特征下面积计算各步厚度特征下面积平均值；根据最小厚度特征下面积和各步厚度特征下面积平均值识别厚度信号，得到向下处理识别结果。
需要说明的是，步骤203可以与步骤204同时执行，也可以在步骤204之后执行，而不限于本实施例中在步骤204之前，在此处不作限定。
205、根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果；
得到向上处理识别结果及向下处理识别结果之后，可以根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果。根据预设规则融合向上处理识别结果和向下处理识别结果的具体过程可以包括：获取向上处理识别结果和向下处理识别结果内的异常区域；根据预设规则将向上处理识别结果的异常区域和向下处理识别结果的异常区域进行融合，得到融合结果。
206、对融合结果进行识别，得到识别结果；
得到融合结果，可以对融合结果进行识别，并得到识别结果。若融合结果异常区域覆盖鉴伪区域，则识别该张钞票为变造币，若融合结果异常区域的面积超过固定阈值，识别该张钞票为残损钞，否则识别该张钞票为流通钞。
需要说明的是，上述的固定阈值为根据被测钞票和装置结构预先进行设定的，在此处不作限定。
207、根据识别结果将钞票分类并送至与类别相对应的位置。
得到识别结果之后，根据识别结果将钞票分类并送至与类别相对应的位置，例如可以将不同种类钞票传送至预定仓位，进而完成钞票识别。
下面以一个具体实例对本发明实施例的工作过程进行详细的描述：
第一步：采集钞票的厚度信号；
请参阅图7，为一张变造币的走钞示意图，在走钞过程中，异物张贴区域被四个厚度传感器覆盖。厚度信号的具体采集图样请参阅图8，选择具有M个传感器单元的厚度传感器，每路厚度信号的采样点数为N。
第二步：对厚度信号进行预处理；
相关约束条件包括：起始路终止路判断阈值Thnotethk＝ηTHK，η∈[0.4,0.6]；起始点终止点判断阈值Thnotethk＝ηTHK，η∈[0.4,0.6]；本部分的作用是确定钞票区域厚度信号，去除背景区域厚度信号。
首先确定各路起始点，对于信号MThkSignal(i,j)满足如下列不等式组
Figure PCTCN2014087404-appb-000001
时
判断第i路的信号起始点Pstart(i)＝j；
然后确定各路终止点，对于信号MThkSignal(i,j)满足如下不等式组
Figure PCTCN2014087404-appb-000002
时
判断第i路的信号终止点Pend(i)＝j；
接着确定信号的起始路，当第i路的厚度均值ThkAvg(i)满足如下不等式
Figure PCTCN2014087404-appb-000003
时
判断第i路信号起始路，记Cstart＝i,
然后确定信号的终止路，当第i路的厚度均值ThkAvg(i)满足如下不等式
Figure PCTCN2014087404-appb-000004
时
判断第i路为信号终止路，记Cend＝i；
请参阅图9，针对图7情况，应用上面的方法，确定变造币厚度信号确定各通道的起始点、终止点和起始通道和终止通道，获得厚度信号的有效区域，有效区域如示意图9中黑色虚线框选区域所示。
第三步：利用向上面积识别法识别厚度信号，得到向上处理识别结果；
相关约束条件包括：△为1cm长度对应的信号采样点数(由走钞速度和DSP采样频率确定)，滑动窗宽度x△为xcm宽度对应的信号采样点数，下面滑 动窗的宽度x△随x的变化而灵活变化；上阈值UpTh＝THKnote+ηthk，η∈[1.5,2]；滑动步长step＝δx△，δ∈(0,0.5)；滑动窗x△对应的异常区域判断阈值ThUpSx△＝η1thkx△，η1<η-0.7。
首先计算各滑动窗口的上面积值：第n个滑动窗口对应的厚度信号上面积值为UpS(i,n)，其计算公式为
Figure PCTCN2014087404-appb-000005
其中i∈[Cstart,Cend]，n∈[0,N(i)]，
Figure PCTCN2014087404-appb-000006
上式中Pstart(i)为第i个厚度信号钞票厚度采集区起始点，Pend(i)为第i个厚度信号钞票厚度采集区终止点，n为第n个滑动窗口，MThkSignal(i,j)为中值滤波后第i个信号第j个采样点的幅值，Cstart为信号起始路，Cend为信号终止路，N(i)为第i个厚度信号的滑动窗口数目；
接着计算各滑动窗口上面积值的最小值
上面积最小值为
Figure PCTCN2014087404-appb-000007
其中min()为取最小函数；
然后计算各滑动窗口的上面积平均值
该上面积平均值的计算公式为
Figure PCTCN2014087404-appb-000008
最后根据上面积最小值与上面积平均值做出判定，记录厚度异常区域。
判断不等式UpSavg(i)-UpSmin(i)≥ThUpSx△是否成立，若成立则该信号存在厚度异常，记录信号异常区域的面积大小Ss(i)与异常区域位置AreaR(i)，否则认为该信号不存在厚度异常。
请参阅图10，针对图7情况，设置大小为△和3△的滑动窗口应用向上面积识别方法对各路信号进行识别，检测到第1路厚度信号、第2路厚度信号、第3路厚度信号存在异常情况，记录异常区域的为AreaR(1)，AreaR(2)，AreaR(3)，,三个区域对应于钞票的异物粘贴面积分别为Ss(1)，Ss(2)，Ss(3)。
第四步：利用向下面积识别法识别厚度信号，得到向下处理识别结果；
相关约束条件包括：下阈值DownTh＝ηTHKnote，η∈[0.4,0.6]；滑动步长step＝δx△，δ∈(0,0.5)；滑动窗x△对应的异常区域下面积判断ThDownSx△＝ηTHKnote+η1thkx△，η∈[0.4,0.6]，η1∈[0.8,2]；下面积最小值阈值ThDownSminx△＝ρηTHKnotex△，ρ∈[0.8,1)。
首先计算各滑动窗口的下面积值：第n个滑动窗口对应的厚度信号下面积 值为Down(i,n)，则
Figure PCTCN2014087404-appb-000009
上式中各参数的定义方式与向上面积识别方法类似。
接着计算各滑动窗口下面积值的最小值：下面积值的最小值DownSmin(i)的计算公式为
Figure PCTCN2014087404-appb-000010
然后计算各滑动窗口的下面积均值
下面积值平均值DownSavg(i)的计算公式为
Figure PCTCN2014087404-appb-000011
最后根据下面积最小值与下面积平均值做出判定，厚度异常区域记录。
判断下列不等式组是否成立
Figure PCTCN2014087404-appb-000012
若成立，则该路厚度信号异常，记录异常区域面积Ss(i)与异常区域位置AreaR(i)，否则判断该路不存在厚度信号异常。
请参阅图11，针对图7情况，设置大小为△滑动窗口应用向下面积识别方法对各路信号进行识别，检测到第i路厚度信号存在异常情况，记录异常区域的为AreaR(4)，该区域对应于钞票的异物粘贴面积为Ss(4)。
第五步：根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果；
相关约束条件包括：鉴伪区域位置AreaN(此参数根据不同币种、不同面值进行设置，例如人民币100元鉴伪区域为水印区域和国徽区域)；厚度异常区域面积阈值TS(此数值可根据不同检测标准设置，例如ECB欧洲央行货币流通标准为4cm2)。
首先判断厚度异常区域位置是否覆盖鉴伪区域：
厚度异常区域覆盖鉴伪区域AreaN，判断该张钞票为变造币。
接着判断厚度异常区域面积是否较大：
厚度异常区域没有覆盖鉴伪区域，异常区域面积大于Ths，判断该张钞票为残损钞。不属于上述两种情况的钞票判断该张为流通钞。
针对图7情况，厚度异常区域AreaR(1)，AreaR(2)，AreaR(3)覆盖鉴伪区域，且异常区域面积SS(1)+SS(2)+SS(3)+SS(4)大于阈值Ths，判断该张钞票为变造币。
本发明实施例中，首先采集钞票的厚度信号；然后对厚度信号进行预处理； 接着利用向上面积识别法识别厚度信号，得到向上处理识别结果；并利用向下面积识别法识别厚度信号，得到向下处理识别结果；然后根据预设规则融合向上处理识别结果和向下处理识别结果，得到融合结果；最后对融合结果进行识别，得到识别结果。由于本发明实施例基于厚度信号识别的钞票识别方法，采集钞票的厚度信号之后，通过结合向上面积识别法和向下面积识别法两种厚度识别方法对厚度信号进行识别的方式，能够简单高效地将异常钞票识别分离。
上面对本发明基于厚度信号识别的钞票识别方法的第二实施例作了详细描述，特别是对利用向上面积识别法识别厚度信号和利用向下面积识别法识别厚度信号的过程，下面介绍本发明基于厚度信号识别的钞票识别装置实施例，请参阅图12，本发明实施例中基于厚度信号识别的钞票识别装置实施例包括：
厚度传感器1201、DSP芯片1202、嵌入式系统1203和机械运动模块1204；
厚度传感器1201，与DSP芯片1202相连，用于采集钞票的厚度信号；
DSP芯片1202，与嵌入式系统1203相连，用于根据厚度信号对钞票进行分析识别，并将识别结果发送至嵌入式系统1203；
嵌入式系统1203，与机械运动模块1204，用于根据识别结果控制机械运动模块；
机械运动模块1204，用于根据嵌入式系统1204的控制指令集将钞票进行分类并送至与类别相对应的位置。
可选的，
厚度传感器1201为多路厚度传感器。
可选的，
DSP芯片1202包括：
预处理模块，分别与向上处理模块及向下处理模块相连，用于对厚度信号进行预处理；
向上处理模块，与融合模块相连，用于利用向上面积识别法识别厚度信号，得到向上处理识别结果；
向下处理模块，与融合模块相连，用于利用向下面积识别法识别厚度信号，得到向下处理识别结果；
融合模块，与识别模块相连，用于根据预设规则融合向上处理识别结果和 向下处理识别结果，得到融合结果；
识别模块，对融合结果进行识别，得到识别结果。
本发明实施例中，厚度传感器1201首先采集钞票的厚度信号，并将上述的厚度信号传输给DSP芯片1202进行分析识别，DSP芯片1202得到识别结果之后，将识别结果传输给嵌入式系统1203，嵌入式系统1203控制机械运动模块1204将流通钞、残损钞和变造币传送至不同出钞仓位，实现不同种类钞票的分类。
本发明实施例中，基于厚度信号识别的钞票识别装置包括：厚度传感器1201、DSP芯片1202、嵌入式系统1203和机械运动模块1204；厚度传感器1201，与DSP芯片1202相连，用于采集钞票的厚度信号；DSP芯片1202，与嵌入式系统1203相连，用于根据厚度信号对钞票进行分析识别，并将识别结果发送至嵌入式系统1203；嵌入式系统1203，与机械运动模块1204，用于根据识别结果控制机械运动模块；机械运动模块1204，用于根据嵌入式系统1204的控制指令集将钞票进行分类并送至与类别相对应的位置。由于本发明实施例基于厚度信号识别的钞票识别装置，厚度传感器1201采集钞票的厚度信号之后，通过DSP芯片1202结合向上面积识别法和向下面积识别法两种厚度识别方法对厚度信号进行识别的方式，能够简单高效地将异常钞票识别分离。
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，这些步骤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基于厚度信号识别的钞票识别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 (10)

  1. 一种基于厚度信号识别的钞票识别方法，其特征在于，包括：
    采集钞票的厚度信号；
    对所述厚度信号进行预处理；
    利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；
    利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；
    根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；
    对所述融合结果进行识别，得到识别结果。
  2. 根据权利要求1所述的基于厚度信号识别的钞票识别方法，其特征在于，在步骤对所述融合结果进行识别，得到识别结果之后还包括：
    根据所述识别结果将所述钞票分类并送至与类别相对应的位置。
  3. 根据权利要求1所述的基于厚度信号识别的钞票识别方法，其特征在于，步骤采集钞票的厚度信号包括：
    利用多路厚度传感器采集钞票的厚度信号；
    所述厚度信号为多路采集信号的集合。
  4. 根据权利要求1至3中任一项所述的基于厚度信号识别的钞票识别方法，其特征在于，步骤对所述厚度信号进行预处理包括：
    对所述厚度信号进行抽样处理，得到抽样处理信号；
    对所述抽样处理信号进行去噪处理，得到去噪处理信号；
    对所述去噪处理信号进行有效信号区域确定，得到有效区域。
  5. 根据权利要求1至3中任一项所述的基于厚度信号识别的钞票识别方法，其特征在于，步骤所述利用向上面积识别法识别所述厚度信号包括：
    厚度特征上面积定义为：滑动窗口与上阈值线及所述厚度信号的曲线所形成的区域的面积；所述滑动窗口的宽度和滑动步长为预设值；所述上阈值线为在所述曲线之上的水平线；
    根据预设滑动步长逐步移动所述滑动窗口，并计算每步厚度特征上面积；
    获取所述每步厚度特征上面积中的最小厚度特征上面积；
    根据所述每步厚度特征上面积计算各步厚度特征上面积平均值；
    根据所述最小厚度特征上面积和所述各步厚度特征上面积平均值识别所述厚度信号，得到向上处理识别结果。
  6. 根据权利要求1至3中任一项所述的基于厚度信号识别的钞票识别方法，其特征在于，步骤所述利用向下面积识别法识别所述厚度信号包括：
    厚度特征下面积定义为：滑动窗口与下阈值线及所述厚度信号的曲线所形成的区域的面积；所述滑动窗口的宽度和滑动步长为预设值；所述下阈值线为在所述曲线峰值之下的水平线；
    根据预设滑动步长逐步移动所述滑动窗口，并计算每步厚度特征下面积；
    获取所述每步厚度特征下面积中的最小厚度特征下面积；
    根据所述每步厚度特征下面积计算各步厚度特征下面积平均值；
    根据所述最小厚度特征下面积和所述各步厚度特征下面积平均值识别所述厚度信号，得到向下处理识别结果。
  7. 根据权利要求1至3中任一项所述的基于厚度信号识别的钞票识别方法，其特征在于，步骤所述根据预设规则融合所述向上处理识别结果和所述向下处理识别结果包括：
    获取所述向上处理识别结果和所述向下处理识别结果内的异常区域；
    根据预设规则将所述向上处理识别结果的异常区域和所述向下处理识别结果的异常区域进行融合，得到融合结果。
  8. 一种基于厚度信号识别的钞票识别装置，其特征在于，包括：厚度传感器、DSP芯片、嵌入式系统和机械运动模块；
    所述厚度传感器，与所述DSP芯片相连，用于采集钞票的厚度信号；
    所述DSP芯片，与所述嵌入式系统相连，用于根据所述厚度信号对所述钞票进行分析识别，并将所述识别结果发送至所述嵌入式系统；
    所述嵌入式系统，与所述机械运动模块相连，用于根据所述识别结果控制所述机械运动模块；
    所述机械运动模块，用于根据所述嵌入式系统的控制指令集将所述钞票进行分类并送至与类别相对应的位置。
  9. 根据权利要求8所述的基于厚度信号识别的钞票识别装置，其特征在于，所述厚度传感器为多路厚度传感器。
  10. 根据权利要求8或9所述的基于厚度信号识别的钞票识别装置，其特征在于，所述DSP芯片包括：
    预处理模块，分别与向上处理模块及向下处理模块相连，用于对所述厚度信号进行预处理；
    所述向上处理模块，与融合模块相连，用于利用向上面积识别法识别所述厚度信号，得到向上处理识别结果；
    所述向下处理模块，与所述融合模块相连，用于利用向下面积识别法识别所述厚度信号，得到向下处理识别结果；
    所述融合模块，与识别模块相连，用于根据预设规则融合所述向上处理识别结果和所述向下处理识别结果，得到融合结果；
    所述识别模块，对所述融合结果进行识别，得到识别结果。
PCT/CN2014/087404 2013-12-12 2014-09-25 一种基于厚度信号识别的钞票识别方法及装置 WO2015085811A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
US15/021,923 US10008065B2 (en) 2013-12-12 2014-09-25 Method and device for banknote identification based on thickness signal identification
EP14869270.0A EP3082112B1 (en) 2013-12-12 2014-09-25 Method and device for banknote identification based on thickness signal identification
AU2014361530A AU2014361530B2 (en) 2013-12-12 2014-09-25 Method and device for banknote identification based on thickness signal identification
ZA2016/02298A ZA201602298B (en) 2013-12-12 2016-04-06 Method and device for banknote identification based on thickness signal identification

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201310681853.1A CN103679914B (zh) 2013-12-12 2013-12-12 一种基于厚度信号识别的钞票识别方法及装置
CN201310681853.1 2013-12-12

Publications (1)

Publication Number Publication Date
WO2015085811A1 true WO2015085811A1 (zh) 2015-06-18

Family

ID=50317345

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/087404 WO2015085811A1 (zh) 2013-12-12 2014-09-25 一种基于厚度信号识别的钞票识别方法及装置

Country Status (8)

Country Link
US (1) US10008065B2 (zh)
EP (1) EP3082112B1 (zh)
CN (1) CN103679914B (zh)
AU (1) AU2014361530B2 (zh)
CL (1) CL2016000519A1 (zh)
TR (1) TR201901673T4 (zh)
WO (1) WO2015085811A1 (zh)
ZA (1) ZA201602298B (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105184953A (zh) * 2015-08-03 2015-12-23 深圳怡化电脑股份有限公司 一种纸币处理的方法及装置

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN103617671B (zh) * 2013-12-12 2016-08-17 广州广电运通金融电子股份有限公司 一种厚度异常钞票的识别方法及系统
CN103679914B (zh) * 2013-12-12 2016-06-15 广州广电运通金融电子股份有限公司 一种基于厚度信号识别的钞票识别方法及装置
CN104408819B (zh) * 2014-12-05 2017-06-13 深圳怡化电脑股份有限公司 一种纸币上胶带的检测方法
CN105354913B (zh) * 2015-09-14 2018-07-17 深圳怡化电脑股份有限公司 一种检测纸币的方法及装置
CN106296971A (zh) * 2016-08-31 2017-01-04 东方通信股份有限公司 一种纸币表面粘贴物高精度检测识别系统及方法
CN106997635A (zh) * 2017-03-03 2017-08-01 深圳怡化电脑股份有限公司 一种纸币表面异物的检测方法及其装置
CN108932788B (zh) * 2017-05-22 2020-10-20 深圳怡化电脑股份有限公司 一种纸币厚度异常等级的检测方法、装置及设备
CN109737859B (zh) * 2018-12-26 2021-01-12 广州国瀚计算机通讯科技有限公司 薄片类介质的粘贴异物检测方法及其装置
CN111599080B (zh) * 2019-02-20 2022-04-12 深圳怡化电脑股份有限公司 拼接纸币的检测方法、装置、金融机具设备及存储介质
CN113160480A (zh) * 2020-01-21 2021-07-23 深圳怡化电脑股份有限公司 钞票厚度的检测方法、装置、计算机设备及存储介质
CN113256870B (zh) * 2020-12-19 2024-02-02 深圳市怡化时代科技有限公司 薄片介质厚度异常检测方法及装置、存储介质及设备
CN115147977B (zh) * 2021-03-30 2023-12-22 深圳怡化电脑股份有限公司 纸币的鉴伪方法、装置、电子设备和存储介质
CN116644281B (zh) * 2023-07-27 2023-10-24 东营市艾硕机械设备有限公司 一种游艇船体偏移检测方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH01209309A (ja) * 1988-02-17 1989-08-23 Oki Electric Ind Co Ltd 紙葉類判別装置
JP2007072583A (ja) * 2005-09-05 2007-03-22 Toshiba Corp 紙葉類の厚さ検知装置および紙葉類の厚さ検知方法
CN1987935A (zh) * 2005-12-19 2007-06-27 株式会社东芝 纸张鉴别设备、纸张处理设备以及纸张鉴别方法
CN101754919A (zh) * 2007-07-26 2010-06-23 富士通株式会社 纸张厚度检测装置
CN101872501A (zh) * 2009-04-27 2010-10-27 日立欧姆龙金融系统有限公司 介质厚度检测装置
CN203133923U (zh) * 2013-03-29 2013-08-14 深圳贝斯特机械电子有限公司 纸币点钞机
CN103617671A (zh) * 2013-12-12 2014-03-05 广州广电运通金融电子股份有限公司 一种厚度异常钞票的识别方法及系统
CN103679914A (zh) * 2013-12-12 2014-03-26 广州广电运通金融电子股份有限公司 一种基于厚度信号识别的钞票识别方法及装置

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS61270609A (ja) * 1985-05-27 1986-11-29 Toshiba Corp 紙葉類の厚み異常検出装置
WO2001039139A1 (fr) 1999-11-25 2001-05-31 Fujitsu Limited Processeur de billets
JP3693993B2 (ja) 2002-01-28 2005-09-14 日立オムロンターミナルソリューションズ株式会社 紙葉類確定装置
JP4387176B2 (ja) * 2003-12-12 2009-12-16 日立オムロンターミナルソリューションズ株式会社 紙幣の鑑別
CA2510943A1 (en) * 2005-06-28 2006-12-28 Cashcode Company Inc. Method and apparatus for detecting overlapped substrates
US20100052237A1 (en) * 2006-10-18 2010-03-04 Lars Karoly Herczeg Document handling apparatus
CN201594293U (zh) * 2008-12-26 2010-09-29 上海古鳌电子机械有限公司 高速纸币清分机
CN101788280B (zh) * 2010-02-11 2012-01-18 浙江金利电子有限公司 超声波鉴别纸币厚度异常装置及其方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH01209309A (ja) * 1988-02-17 1989-08-23 Oki Electric Ind Co Ltd 紙葉類判別装置
JP2007072583A (ja) * 2005-09-05 2007-03-22 Toshiba Corp 紙葉類の厚さ検知装置および紙葉類の厚さ検知方法
CN1987935A (zh) * 2005-12-19 2007-06-27 株式会社东芝 纸张鉴别设备、纸张处理设备以及纸张鉴别方法
CN101754919A (zh) * 2007-07-26 2010-06-23 富士通株式会社 纸张厚度检测装置
CN101872501A (zh) * 2009-04-27 2010-10-27 日立欧姆龙金融系统有限公司 介质厚度检测装置
CN203133923U (zh) * 2013-03-29 2013-08-14 深圳贝斯特机械电子有限公司 纸币点钞机
CN103617671A (zh) * 2013-12-12 2014-03-05 广州广电运通金融电子股份有限公司 一种厚度异常钞票的识别方法及系统
CN103679914A (zh) * 2013-12-12 2014-03-26 广州广电运通金融电子股份有限公司 一种基于厚度信号识别的钞票识别方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3082112A4 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105184953A (zh) * 2015-08-03 2015-12-23 深圳怡化电脑股份有限公司 一种纸币处理的方法及装置
CN105184953B (zh) * 2015-08-03 2017-12-08 深圳怡化电脑股份有限公司 一种纸币处理的方法及装置

Also Published As

Publication number Publication date
EP3082112A1 (en) 2016-10-19
CN103679914B (zh) 2016-06-15
TR201901673T4 (tr) 2019-02-21
ZA201602298B (en) 2017-06-28
EP3082112B1 (en) 2018-11-07
US10008065B2 (en) 2018-06-26
AU2014361530A1 (en) 2016-03-31
CL2016000519A1 (es) 2016-08-12
US20160225216A1 (en) 2016-08-04
CN103679914A (zh) 2014-03-26
EP3082112A4 (en) 2016-12-07
AU2014361530B2 (en) 2017-07-13

Similar Documents

Publication Publication Date Title
WO2015085811A1 (zh) 一种基于厚度信号识别的钞票识别方法及装置
WO2015085815A1 (zh) 一种厚度异常钞票的识别方法及系统
WO2018019041A1 (zh) 一种粘贴钞检测方法和装置
US8965068B2 (en) Apparatus and method for discriminating disguised face
CN109409315B (zh) 一种atm机面板区域遗留物检测方法及系统
CN103984951B (zh) 一种磁粉检测缺陷自动识别方法及系统
US20130287264A1 (en) Defect categorisation
EP3252710B1 (en) Banknote crack detection method and apparatus
US9014458B2 (en) Valuable document identification method and system
WO2016078455A1 (zh) 纸币褶皱的识别装置和方法
CN103279735A (zh) 一种金融票据识别模块中积灰检测方法及系统
CN102176262A (zh) 薄片类介质厚度检测方法及检测装置
CN106447897A (zh) 一种检测纸币磁性特征的方法及装置
Dhar et al. Paper currency detection system based on combined SURF and LBP features
CN117197700A (zh) 智能化无人巡检接触网缺陷识别系统
Destruel et al. Color noise-based feature for splicing detection and localization
US9047723B2 (en) Defect categorization
CN110060703A (zh) 一种检测并定位语音片段内的平滑处理的方法
CN104794443B (zh) 一种基于图像处理的自助存取款机实时人数检测方法
CN105006058B (zh) 一种识别纸币变造信息的方法及装置
TWI559261B (zh) Identify methods of forging coins
CN106780958A (zh) 检测纸币在厚度传感器的检测范围上越界的方法和装置
Zacharias et al. Singularity detection in fingerprint image using orientation consistency
Neelambika et al. Currency Recognition System
Wang et al. Invariant features extraction for banknote classification

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: 14869270

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2014869270

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2014869270

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 15021923

Country of ref document: US

ENP Entry into the national phase

Ref document number: 2014361530

Country of ref document: AU

Date of ref document: 20140925

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE