WO2013127185A1

WO2013127185A1 - Papery medium authentication device and authentication method thereof

Info

Publication number: WO2013127185A1
Application number: PCT/CN2012/083484
Authority: WO
Inventors: 梁添才; 余元超; 王卫峰; 陈�光
Original assignee: 广州广电运通金融电子股份有限公司
Priority date: 2012-02-28
Filing date: 2012-10-25
Publication date: 2013-09-06
Also published as: CN102592351A; CN102592351B

Abstract

A papery medium authentication device (100) and an authentication method thereof. The papery medium authentication device (100) comprises an image data obtaining part (10), a standard template data storing part (20), a broken line detecting and processing part (30), an image cutting part (40) and a judging part (50). The broken line detecting and processing part (30) is used for detecting whether a broken line exists in image data obtained by the image data obtaining part (10), and in the event where a broken line is detected, obtaining the position of the bad line in the whole image and correcting the image according to features of the broken line. After broken line detection and broken line correction are performed for the image, the judging part matches the corrected image against a standard template to obtain an authentication result. Therefore, failure of a light emitting element in a CIS or dirt on a light transmissive sheet in the CIS, which may be encountered by the papery medium authentication device (100), will not affect the authentication, thereby improving the acceptance rate of the papery medium authentication device (100).

Description

Paper type medium identification device and identification method thereof

The present application claims priority to Chinese Patent Application No. 201210049058.6, entitled "Paper Sheet Media Identification Device and Its Identification Method", filed on February 28, 2012, the entire contents of which are incorporated by reference. In this application. Technical field

The invention relates to the technical field of paper sheet medium identification, in particular to a paper sheet medium identification device and a method for identifying the same. Background technique

An authentication device for discriminating sheet-type media (banknotes) is installed on the automatic rejection machine of the financial institution. In the storage unit of the paper type medium discriminating device, there is discriminative paper sheet medium template data. When the sheet-type medium is loaded, the image data acquisition unit of the sheet-type medium discriminating device optically reads the sheet-type medium, and obtains grayscale image data of the sheet-like medium to be discriminated, according to the obtained to-be-identified The pattern of the gray-scale image of the paper-based medium and the shading of the gradation value are matched between the template and the detection result, thereby judging the type (currency) and authenticity of the paper-based medium.

When the paper-based medium discriminating device is used for a long time, there may be a phenomenon that a certain light-emitting element of the image sensor (CIS) is broken or the CIS translucent sheet is dirty. Since the CIS is a line scan, when a CIS light-emitting element fails or the CIS light-transmissive sheet is dirty, a black line with a gray-scale mean value close to 0 appears on the reflection or transmission image collected by the image sensor (hereinafter The black line is abbreviated as "bad line" and this phenomenon is called "CIS bad line phenomenon"). Due to the randomness of a CIS light-emitting element failure or the position where the CIS light-transmissive sheet is dirty, the position of the bad line in the image of the paper-like medium to be tested is also randomly distributed. In this case, if the existing identification process is followed, it will directly affect the acceptance rate of the paper-based media identification device, which will have a huge impact on the performance of the entire paper-based media identification device, affecting the customer experience, and is not conducive to the device. Marketing and use. Therefore, it is necessary to provide a technique capable of effectively handling the CIS bad line phenomenon in a paper sheet type medium discriminating device.

Summary of the invention

SUMMARY OF THE INVENTION One object of the present invention is to provide a paper sheet medium discriminating apparatus which can effectively handle the CIS bad line phenomenon and which has a greatly improved acceptance rate.

The present invention also provides a method of discriminating a sheet-like medium by the above-described sheet-type medium discriminating device.

The paper sheet medium discriminating device includes: a transport path for transporting the sheet-like medium one by one and passing through the image data acquisition unit; and an image data acquisition unit for acquiring the original image data when the sheet medium passes a standard template data storage unit for storing standard template data information required for identifying the paper-based medium; and a bad line detection processing unit for detecting whether the original image data obtained by the image data acquisition unit has a bad line. And detecting the position of the bad line in the entire image when the presence of the bad line is detected, and performing image repair according to the characteristic of the bad line; and the image cutting unit for cutting the identification image information for effectively identifying the paper-type medium; And a determining unit configured to perform template matching recognition on the cut identification image information, and determine a similarity between the identification image information and the standard template data information to obtain an authentication result.

Preferably, the image cutting unit is configured to perform effective recognition region cutting on the image information passing through the bad line detection processing portion to form feature image information.

Further, the feature image information is complete image information of the paper-based medium or at least a portion of image information of the paper-based medium.

The method for discriminating the paper-based medium comprises: step one, inputting a paper-based medium, the paper-based medium enters a transport channel; and step 2, the image sensor scans the image of the paper-based medium through the transport channel to form original image data Step 3: The bad line detection processing unit analyzes the original image data acquired by the image data acquisition unit to detect whether there is a bad line; if the detection does not find a bad line, the process proceeds to step 4; if there is a bad line, the bad line is confirmed. The position of the line in the image, based on the grayscale characteristics of the bad line Applying the broken line repair, and then proceeding to step 4; Step 4, the image cutting unit acquires the paper-like medium image area; that is, detects the boundary of the paper-like medium image area, and cuts the paper-like medium image area from the original image to obtain The identification image information for effective recognition; and step 5, the determination unit calls the standard template data in the storage unit, performs template matching identification, that is, determines the similarity between the identification image information and the standard template data to obtain the identification result.

The method for detecting the presence of a bad line in the third step is: placing the image of the paper-based medium in the xy coordinate system, and setting:

1 The image gray value matrix is ^x . Column, ^. Row;

2 is the gray value of the Xth column position of the first line of the image;

3 is the grayscale mean of the Xth column of the image;

According to the constraint g(x,y) < , (O < y < y ₀ )\ ( A , both greater than o, set the threshold)

Automatic search and detection of the image gray value matrix acquired by the image sensor, if there is a column 图像 = « in the image gray matrix. , that is, the image of the gray matrix of the image. If the column satisfies the above constraints, it is determined that there is a bad line on the image, otherwise there is no bad line.

When there is a bad line, the following model is used to repair the bad line:

g(n ₀ , y) = ^i^(«o _ ² , + ₂ g(« ₀ - y) + ^g(n ₀ + ly) + ^g( ⁿ o + 2, J , where ye( 0,j _Q ) , 63⁄4, ω ₂ . ί3⁄4, " ₄ is the bad line repair gradation fitting linear weighting coefficient, ω _{ι +} ω ₂ +ω, +ω _Α =\ , that is, the gray level of any point on the bad line The value is replaced by a linear weighting of the gray values of the first two points and the last two points of the same line as the point.

Due to bad line detection and bad line repair due to the image of the paper medium to be authenticated, the paper sheet The failure of a CIS light-emitting element or the contamination of the CIS transparent sheet that may occur in the medium identification device has no effect on the identification, thereby improving the acceptance rate of the paper-based medium identification device, which is beneficial to the market promotion and use of the device.

DRAWINGS

1 is a schematic structural diagram of a paper sheet medium discriminating device according to a preferred embodiment of the present invention; FIG. 2 is a flow chart of a paper sheet medium discriminating method according to a preferred embodiment of the present invention;

Figure 3 is a schematic diagram of an image obtained by an image sensor when a CIS has a bad line;

4 is a schematic diagram of matching an image obtained by an image sensor with a standard template when the bad line is not repaired; FIG. 5 is a diagram showing the image obtained by the image sensor in the ^x ^ coordinate system to be repaired by a bad line in a preferred embodiment of the present invention; schematic diagram;

6 is a schematic diagram of an image after repairing a bad line in an image obtained by an image sensor; FIG. 7 is a schematic diagram of matching an image after repairing a bad line with a standard template.

detailed description

In order to further illustrate the invention, specific embodiments of the invention are described in the accompanying drawings.

1 is a schematic structural view of a paper sheet medium discriminating device according to a preferred embodiment of the present invention. The sheet-type medium discriminating device 100 is embedded in a sheet-type medium processing device such as an ATM (Bank Automatic Rejector), and conveys and passes the sheet-like medium (for example, banknotes) to be discriminated one by one through a conveying path. Image data acquisition unit. Specifically, the paper sheet medium identification device 100 includes an image data acquisition unit 10, a standard template data storage unit 20, a bad line detection processing unit 30, an image cutting unit 40, and a determination unit 50, wherein the image data acquisition unit 10 uses The original template data storage unit 20 is configured to store standard template data information required for identifying the paper-based medium; the bad line detection processing unit 30 is configured to detect the image data. Obtained by the acquisition unit 10 Whether the original image data has a bad line, and the position of the bad line in the entire image is obtained in the presence of the detected dead line, and the image is repaired according to the characteristics of the bad line; the image cutting portion 40 is used for cutting the effective identification The identification unit 50 is configured to perform template matching recognition on the cut identification image information, and determine the similarity between the identification image information and the standard template data information to obtain an authentication result. The image cutting unit 40 is configured to perform effective identification area cutting on the image information passing through the bad line detection processing unit 30 to form feature image information, which is the complete image information of the paper-based medium or the paper sheet type. At least a portion of the image information of the medium.

2 is a flow chart of a method for discriminating a sheet-type medium by using a sheet-type medium discriminating device 100, comprising the following steps: Step 1: Injecting a sheet-type medium, the sheet-like medium enters a conveying passage, and the flow starts; The image sensor scans the sheet-type medium to form the original image data through the image when the transport channel is passed; Step 3, the bad line detection processing unit analyzes the original image data acquired by the image data acquisition unit to detect whether there is a bad line; If no bad line is found, go to step 4. If there is a bad line and confirm the position of the bad line in the image, perform the bad line repair according to the bad line gray feature, and then go to step 4; Step 4, the image cutting part gets the paper a medium-like image area; that is, detecting a boundary of a paper-like medium image area, and cutting a paper-like medium image area from the original image to obtain identification image information for effective recognition; and step 5, the determination unit calls the storage unit Standard template data, template matching identification, that is, the identification image Similar information of the data to obtain a standard template authentication result.

The image data acquisition unit 10 is configured by a projection sensor that detects a transmission image and/or a reflection sensor that detects a reflection image. When the sheet medium is transported through a transport path, the image data acquisition unit 10 acquires the sheet medium. The fluoroscopic image and/or the reflected image, and the obtained fluoroscopic image and/or reflected image are transmitted to the bad line detection processing section 30.

The bad line detection processing section 30 first detects whether or not there is a bad line in the original image acquired by the image sensor, and if there is a bad line in the original image of the input paper-type medium, the positional information of the bad line in the entire image is accurately acquired. Since CIS imaging is a line scan, if there is a bad line, in the paper to be tested A black line with a gray value close to 0 will appear on the prime original image. Due to the uncertainty of the position of the CIS bad line, there is also uncertainty in the position of the bad line in the original image of the paper-like medium to be tested. Assume that when the CIS has a bad line (a bad line), the image obtained by the image sensor is shown in Figure 3.

If image cutting and template matching recognition are directly implemented without taking any measures on the bad line area, when the bad line happens to be in an area to be template-matched, the matching is not successful, as shown in the figure.

4. The identification of the media to be identified is not successful, and eventually the normal paper-based media is rejected by the financial self-service device (such as ATM).

In the present embodiment, it is detected whether or not there is a bad line by using the reflection image of the sheet-type medium taken by the reflection sensor. Place the reflected image of the paper-based medium into the coordinate system as shown in Figure 5, and set:

1 The gray value matrix of the reflected image is ^x . Column, ^. Row;

2 g( ^x , the gray value of the position of the first column of the reflected image;

3 is the grayscale mean of the Xth column of the reflected image.

According to the constraints

g(x,y) < A,(O < y <); (W is greater than o, is the set threshold), ( ii )

For example, A ^{= 2 ()} , ^{= 20} , = 80, = 80 can be used to automatically search and detect the gray value matrix of the reflected image obtained by the image sensor.

If the traversing reflection image gray value matrix does not satisfy the column ^{x of the} constraint condition (1.1), it is determined that there is no bad line on the reflection image acquired by the image sensor, and it is inferred that the paper sheet medium identification device does not have a CIS bad line phenomenon. Directly enter the cutting section of the identification paper sheet medium image area, and finally implement template matching recognition on the cut paper sheet medium image area to obtain the final recognition result.

If there is a column in the gray matrix of the reflected image That is, the image of the gray matrix of the image. If the column satisfies the above constraint condition (1.1), it is determined that there is a bad line on the reflected image, that is, the paper sheet type discrimination device has a CIS bad line phenomenon, and the bad line position is located in the gray image obtained by the image sensor. Enter the bad line repair link.

When the image obtained by the image sensor is ^x = ". The column is judged to be a bad line, the image processing technique is used to perform the bad line repair on the image to be recognized. Because the bad line has normal gray values on both sides, and the gray image has a certain point. The gray value has a certain correlation with the gray value of the point in the neighborhood of the point. In this embodiment, the following model is used to repair the bad line: g( , y) = ^i^(«o _ 2, + ₂ g (" ₀ - y) + ^>3§( + ly) + ^ ₄ s( + y) ( 1.2 ) where ye (0,j.), ω, + ω^ ω^ + ω^ Ι , for example ί3⁄4丄, that is, the gray value of any point on the bad line is replaced by the mean of the sum of the first two and the last two gray values in the same line as the point. The image after the bad line repair is shown in Fig. 6.

Then, the image cutting unit 40 performs paper sheet medium image area detection and cutting on the image that has been subjected to the bad line repair, and cuts the effective recognition area from the entire image acquired by the image sensor to obtain template matching recognition. The image is shown in Figure 7.

The determination unit 50 performs template matching recognition on the sheet-type medium image area image by using the standard template stored in the standard template data storage unit 20, and obtains the final recognition result.

Due to the bad line detection and the bad line repair on the image of the paper-based medium, the CIS light-emitting element failure or the CIS transparent sheet may appear to have no effect on the identification of the paper-based medium identification device, thereby improving The acceptance rate of the paper-based media identification device is beneficial to the marketing and use of the device.

The above is only a preferred embodiment of the present invention, and it should be noted that the above-described preferred embodiments are not to be construed as limiting the scope of the invention, and the scope of the invention should be determined by the scope defined by the claims. It will be apparent to those skilled in the art that various modifications and improvements may be made without departing from the spirit and scope of the invention.

Claims

Rights request

A paper sheet medium discriminating device comprising:

a transport channel for transporting the sheet-like medium one by one and passing through the image data acquisition unit; an image data acquisition unit for acquiring raw image data when the sheet-type medium passes; and a standard template data storage unit for storing the identification Standard template data information required for the paper-based medium;

An image cutting unit, configured to cut a paper-like medium image area from the original image, to obtain identification image information that effectively identifies the paper-type medium;

a determining unit, configured to perform template matching identification on the cut identification image information, and determine a similarity between the identification image information and the standard template data information to obtain an authentication result;

The paper sheet medium discriminating device further includes:

The bad line detection processing unit detects whether the original image data acquired by the image data acquisition unit has a bad line, and if a bad line is detected, obtains a position of the bad line in the entire image, and according to the The features of the bad line are image repaired.

The paper sheet medium discriminating device according to claim 1, wherein the image cutting unit is configured to cut the image information of the bad line detecting processing unit into the effective recognition area to form the feature image information.

The paper sheet medium discriminating device according to claim 2, wherein the feature image information is complete image information of the sheet type medium or at least a part of image information of the sheet type medium.

4. A paper sheet medium identification method, characterized in that:

Step one, inputting a paper-type medium, the paper-type medium enters a conveying passage;

Step two, the image sensor scans the image of the sheet medium through the conveying channel to form original image data;

Step 3: The bad line detection processing unit divides the original image data acquired by the image data acquisition unit If there is no bad line, if it does not find a bad line, go to step 4. If there is a bad line and confirm the position of the bad line in the image, perform the bad line repair according to the bad line gray feature, and then enter Step four;

Step 4: The image cutting unit acquires a paper-like medium image area; that is, detects a paper-sheet medium image area boundary, and cuts the paper-type medium image area from the original image to obtain identification image information for effective recognition;

Step 5: The determining unit calls the standard template data in the storage unit to perform template matching identification, that is, determining the similarity between the identification image information and the standard template data to obtain the authentication result.

The paper sheet medium identification method according to claim 4, wherein the method for detecting whether there is a bad line in the third step is:

Place the image of the paper-like medium in the xy coordinate system.

Assume:

1 The image gray value matrix is ^x . Column, ^. Row;

2 is the gray value of the Xth column position of the first line of the image;

3 is the gray level mean of the image column;

According to the constraints

g(x, y) < , (O < y <); , both greater than 0, is the threshold)

Automatic search detection of the image gray value matrix acquired by the image sensor, if the image gray matrix exists in a certain column x = «. , that is, the image of the gray matrix of the image. If the column satisfies the above constraints, it is determined that there is a bad line on the image, otherwise there is no bad line.

6. The paper sheet medium identification method according to claim 4, wherein the step 3 is Repair the bad line with the following model:

g(n ₀ , y) = ^i^(«o _ 2, + i3⁄4g(" ₀ - y) + ^g(n ₀ + ly) + ^g( ⁿ o + ² , y) where ye(0 , j _Q ) , ω . ω ₂ , ω ₃ , « ₄ is the bad line repair gradation fitting linear weighting coefficient, ω _γ + ω ₂ + ω, +ω _Α =\ , that is, the gray level of any point on the bad line The value is replaced by a linear weighting of the gray values of the first two points and the last two points of the same line as the point.