KR101385358B1 - Apparatus and method for medium recognition, auto teller machine - Google Patents

Abstract

The present invention relates to a medium recognizing apparatus and method for determining the authenticity of a medium by using information obtained from a UV sensor, and to an automated device thereof. This invention judges a banknote winding type. The UV image of the determined banknote is compared with a plurality of pre-stored UV average images, and a predetermined coordinate region is extracted from the UV image. The extracted coordinate region is a coordinate at which the sum of the differences between the pixels becomes the minimum as a result of the comparison. Subsequently, the extracted coordinate area value and the UV average image value corresponding to the determined roll type and bill entry direction are calculated. The result value obtained by the subtraction operation is compared with a preset threshold value, and the authenticity of the banknote is judged according to the comparison result. According to the present invention as described above there is an advantage that the processing speed is improved to determine a large amount of bills accurately and in real time.

Media recognition, UV sensor, UV image, authenticity judgment.

Description

Apparatus and method for media recognition and its automation device {APPARATUS AND METHOD FOR MEDIUM RECOGNITION, AUTO TELLER MACHINE}

The present invention relates to a medium recognizing apparatus, and more particularly, a medium recognizing apparatus and method for identifying the authenticity of a medium by using a UV image obtained from a UV sensor and a pre-stored UV average image after classification of a media type, and automation thereof. It is about a device.

The term " medium " used in this specification refers to, for example, banknotes, checks, tickets, certificates, etc., and may be various in thickness as compared with width or length. For convenience of description of the present embodiment, a banknote will be described as an example of a banknote.

Currently, the most widely used banknotes of currency have various sizes and patterns for each amount of money.In financial automated machines, vending machines, and ticket vending machines that can recognize banknotes, whether the counterfeit is introduced or not, and according to the determined result If it is a counterfeit bill, take it out.

Fig. 1 is a flow chart for discriminating whether or not the banknotes are genuine.

Once the UV sensor provided in the banknote recognition device is composed of a UV light receiving sensor at a position facing the other side of the UV light emitting sensor on one side of the housing, irradiating the ultraviolet rays on the banknotes, that is, the response value of the corresponding banknotes, ie Detect the reflection value and the transmission value of the banknote.

When the bill is input from the outside to pass through the UV sensor, the UV sensor leads the bill image information (s10).

Then, the control unit of the banknote recognition apparatus extracts a predetermined pattern from the read banknote image information (s12), and converts it to a binary image (s14).

Then, a process of comparing the pattern between the converted binary image and the pre-stored reference image is performed (s16), and if it matches the reference image according to the pattern comparison result, it is determined as a normal banknote (Yes in S18). ), If the reference image does not match, it is determined as a counterfeit bill (No in S18).

However, there is the following problem in the process of determining the authenticity of banknotes according to the prior art.

First, since the pattern of the banknote image is compared with the pattern of the reference image, there is a large amount of calculation for matching the patterns, and it is difficult to judge whether the banknotes are authentic in real time.

In addition, although the image is binarized to reduce the amount of comparison in pattern matching, the image may be lost in the binarization process.

In addition, since the fluorescence pattern of the banknote is damaged depending on the degree of contamination of the banknote and the amount of usage, there is a problem in that the fluorescence pattern provided on the banknote does not come out and thus the authenticity cannot be performed. That is, the fluorescent pattern is shown to following (a)-(c). (a) is a nine thousand yuan fluorescent pattern, (b) is a nine thousand yuan fluorescent pattern, and (c) is a nine thousand yuan fluorescent pattern.

       (a) (b) (c)

As described above, although the fluorescent patterns are given differently according to the currency type, if the fluorescent pattern is damaged, the authenticity can not be determined.

If the same material as the fluorescent pattern of the banknote can be obtained by drawing a reactive material corresponding to the light source characteristic on the counterfeit, there is a possibility that the counterfeit banknote is discriminated as a normal banknote.

Accordingly, it is an object of the present invention to provide a medium recognition apparatus and method, and an automation device thereof, which can accurately identify a medium by only a small amount of data irrespective of the degree of contamination and usage of the medium.

It is another object of the present invention to improve the processing speed of authenticity of media to enable real-time media recognition.

Another object of the present invention is to draw a conventional fluorescent pattern arbitrarily to prevent the possibility that the counterfeit bill is incorrectly identified as a normal bill.

According to a feature of the present invention for achieving the above object, a winding type determination unit for determining the winding type of the introduced banknote; A processing unit for extracting a predetermined coordinate region while comparing the UV image of the banknote with a plurality of UV average images; In addition, the extracted coordinate area value and the UV average image value corresponding to the roll type and the banknote entry direction of the introduced banknote are obtained by the following equation (1), and compared with a threshold value to control whether the authenticity of the banknote is determined. It is configured to include a control unit.

Equation (1) is " UV average image value-calculated coordinate region value. &Quot;
The media recognition device of the present invention, winding type determination unit for judging the winding type of the introduced banknote; A processing unit for extracting a predetermined coordinate region while comparing the UV image of the banknote with a plurality of UV average images; The total sum of the sum of the pixel values of the extracted coordinate region and the sum of the sum of each pixel value of the UV average image corresponding to the roll type and the banknote entry direction of the introduced banknote are subtracted, and the result is thresholded. It may include a control unit for controlling to determine the authenticity of the bill compared to the value.

The bank further comprises a storage unit for storing the UV average image value of each bill type / entry direction, the total value of the UV average image, the threshold value.

The winding type determination unit, the size order of the 'R', 'G', 'B' value for each kind, the range of 'R', 'G', 'B' value for each kind, the size information of the bill and the extracted 'R The scoop type is judged by comparing ',' G 'and' B 'values sequentially.

In the determination of the paper type, an image sensor for sensing a bill image is used, and the UV image is a UV sensor for sensing a UV image of a bill image.

The processing unit compares only the 'B' color among the 'R', 'G', and 'B' colors of the UV image and the UV average image.

The coordinate region is a region in which the sum of the difference of each pixel value between the UV image and the plurality of UV average images is minimum.

The direction which has the smallest value among the sum values of the difference of each pixel value is a banknote entrance direction.

The threshold value is provided differently according to the roll type and the bill entry direction of the bill, and the controller selects a corresponding threshold value when the bill type and the entry direction are determined.

According to another feature of the present invention, there is provided an automatic device comprising a media recognition device configured as described above, and a controller for controlling the normal operation of the bill and stops the operation of the counterfeit bill according to the bill authenticity determination process of the medium recognition device. Is provided.

According to another feature of the invention, the step of determining the bill winding; Comparing the determined UV image of the banknote with a plurality of previously stored UV average images and extracting a predetermined coordinate region from the UV image; And determining the authenticity of the bill by comparing the extracted coordinate area value, the UV average image value corresponding to the determined paper type and bill entry direction by the following equation (1), and comparing the threshold value with a threshold value. . Equation (1) is " UV average image value-calculated coordinate region value. &Quot;

The UV image and the UV average image comparison compares only the 'B' color among the 'R', 'B', and 'G' colors of the image.

The coordinate region is configured to extract an area coordinate at which the sum of the differences between pixels is minimum.

The UV average image value is provided corresponding to the roll type and bill entry direction of the banknote, respectively. When the roll type and entry direction are determined, a corresponding UV average image value is selected.

The threshold is provided corresponding to the winding type and bill entering direction of the banknote, respectively, and when the winding type and entering direction are determined, the corresponding threshold is selected.
The method of recognizing a medium of the present invention comprises the steps of: determining the bill winding type; Comparing the determined UV image of the banknote with a plurality of previously stored UV average images and extracting a predetermined coordinate region from the UV image; The total sum of the sum of the pixel values of the extracted coordinate region and the sum of the sum of each pixel value of the UV average image corresponding to the roll type and the banknote entry direction of the introduced banknote are subtracted, and the result is thresholded. Comparing the value to determine the authenticity of the bill.

According to the media recognition device and method of the present invention configured as described above, and an automated device thereof, the following effects are obtained.

First, since only the 'B' color is compared among the UV image information of the different types of banknotes, the amount of data is reduced during mutual matching with the reference banknote information, and thus the processing speed is improved, thereby enabling real-time processing.

The improvement of the processing speed is also effective in authenticating a large amount of bills.

Moreover, the function which can confirm the entry direction of a banknote can also be acquired.

Hereinafter, an apparatus and method for recognizing a medium according to the present invention, and an automated apparatus thereof will be described in detail with reference to the preferred embodiments shown in the accompanying drawings.

2 is a block diagram of a media recognition apparatus according to an embodiment of the present invention. In the present embodiment, the medium will be described by taking banknotes as an example, and the media recognizing apparatus will also be described as a banknote recognizing apparatus.

Looking at the configuration of Figure 2, the insertion hole (not shown) of the banknote recognition device 100 is provided with an image sensor 102 and the UV sensor 104. The image sensor 102 reads the 'A' image of the banknote A to be input, and the 'R', 'G', and 'B' information of the banknote A is determined by the paper type determining unit 106 to be described later. It plays a role in sensing. UV sensor 104 serves to obtain a UV image (B) of the bill (A). The installation position of the image sensor 102 and the UV sensor 104 may be installed at about the same position in the inlet. Alternatively, after the banknote image is read by the image sensor 102, the UV sensor 104 is installed at the rear end side of the image sensor 102 so as to acquire the UV image B when the banknote A is transferred along the transfer path. May be

A paper type determination unit 106 is provided to determine the paper currency type by referring to the paper money image sensed by the image sensor 102.

In order to judge the paper type, the size order of 'R', 'G', and 'B' values by bill type, the range of 'R', 'G' and 'B' values by kind, and size information of bill (A) are provided. It should be, and the storage unit 108 for storing it is provided. The storage unit 108 is a UV average image (C) of the bills for each kind of bill in four entry directions (that is, the forward direction, front reverse direction, backward forward direction, backward reverse direction) described later, and the overall UV average image of the four directions The sum D and a threshold value E preset for each kind / direction are also stored to determine the authenticity of the bill. Of course, providing a separate storage unit, the order of the size of the 'R', 'G', 'B' value for each bill type, the range of the 'R', 'G', 'B' value for each kind, bill size information And, the average UV image (C) of the bills for each kind of paper in the four directions of entry, the total sum (D) of the UV average image of the four directions, respectively, to store the preset threshold value (E) to determine whether the bill authenticity It may be.

After judging the paper type, the processing unit 110 is configured to search the predetermined area coordinates by comparing the UV image B and the UV average image C, and to determine the entry direction of the banknote A. The processing unit 110 includes a comparison unit 112 for comparing the UV image (B) obtained by the UV sensor 104 and the UV average image (C) in four directions stored in the storage unit 108 with each other. Is provided. As a result of the comparison, a coordinate extraction unit 114 is provided for extracting the region coordinates of the UV image B in which the sum of the differences between the pixel values is the minimum. In addition, an entry direction determination unit 116 is configured to determine a direction having the smallest value among the sums of the differences between the respective pixel values as the banknote entry direction.

The calculation unit 120 is connected to the output terminal of the processing unit 110.

The calculation unit 120, the first operation unit 122 to calculate the sum of the UV image (B) by summing each pixel value with respect to the area coordinates of the UV image (B) extracted by the coordinate extraction unit 114 And, the sum of the UV image (B) (the output coordinate region value) that is the result of the first operation unit 122 and the total sum of the UV average image (C) determined by the entry direction determination unit 116 ( A second operation unit 124 for subtracting the UV average image value) is provided.

That is, "│UV average image value-calculated coordinate area value |" ----- It calculates by Formula (1).

The total sum of the UV average images (C) is calculated in advance for each bill type and entry direction, which is stored in advance in the storage unit (108). The total sum of the UV average image (C) will be described in Table 2 below. The value calculated by the second operation unit 124 is calculated as an absolute value.

Selecting a predetermined threshold value E according to the winding / entry direction of the banknote A, comparing the selected threshold value E with a result calculated by the second calculation unit 124, and determining whether or not the authenticity is Banknote authenticity determination unit 126 is provided.

Meanwhile, the UV image B and the UV average image C include 'R', 'G', and 'B' colors, but in this embodiment, only the 'B' color is compared. This is because, as shown in FIG. 3, the 'B' color may be distinguished enough to be visually identified from the 'R' and 'G' colors. That is, FIG. 3 illustrates a color pattern for explaining the distinction between counterfeit banknotes and normal banknotes according to the colors 'R', 'G', and 'B'.

In addition, the bill (A) contains a fluorescent material and a solid line, BLB (Black Light Blue) light source is used to detect this.

Next, the process of determining the medium recognition, that is, the authenticity of the banknote by using the medium recognition device having the above-described configuration will be described in detail.

4 is a flowchart illustrating a medium recognition method according to an exemplary embodiment of the present invention.

This embodiment is disclosed by inserting the banknote A into the input hole of the banknote recognition device 100.

Then, the image sensor 102 and the UV sensor 104 installed in the inlet leads the image of the bill (A) to be transferred by the transfer module. The image sensor 102 extracts 'R', 'G', 'B' values and bill size information of the bill A, and the UV sensor 104 obtains a UV image B (S100).

R ',' G ', and' B 'values extracted by the image sensor 102 are transmitted to the denomination determining unit 106. Then, the kind determination unit 106 is the size order of the 'R', 'G', 'B' value for each kind stored in the storage 108, the range of the 'R', 'G', 'B' value for each kind , And compares the size information of the bill (A) and the extracted 'R', 'G', 'B' value sequentially, and determines the bill winding type according to the comparison result (s102). At this time, if one of the methods is used to classify the denomination, the remaining methods can be omitted.

When the bill type of the bill A is determined in the bill sorting process, the comparison unit 112 determines the UV image B of the bill sort and the UV average image C in four directions stored in the storage 108. Are respectively compared (s104).

Herein, a method of obtaining the UV average image C will be described with reference to FIG. 5.

As shown in FIG. 5 (a), the actual banknote partial image is cut out from the UV image B by the contour search. In addition, the image is extracted as shown in FIG. 5 (b) in consideration of the size of the bill portion and the sensing portion when the bill is moved left and right. The extracted image, that is, the sample image, is plural. The sample image is framed, and the average of each pixel is obtained to obtain a UV average image C as shown in FIG.

The UV average image C thus obtained is carried out for each of four directions of the banknote A, and the UV average image C for these four directions is stored in advance in the storage unit 108.

Thus, the comparison unit 112 compares the UV image B, which is the banknote image determined as described above, with the UV average image C in four directions stored in the storage unit 108. In this case, the UV image B applied to the comparison process is also provided with a banknote image extracted after detecting the edge.

As a result of the comparison, the coordinate extraction unit 114 extracts the coordinates of the region where the sum of the difference between the pixels of the UV image B and the UV average image C is minimum from the UV image B (s106). .

The area coordinate extraction will be described with reference to FIG. 6.

As described above, before comparing the images, it is necessary to extract the banknote image from the UV image (B), which is edge-detected by a Sobel mask in the UV image as shown in FIG. This is achieved by extracting only the bill image. The Sobel mask is a second derivative operator of edge detection, and is a second derivative operator, which means to perform differentiation once for each of the x and y axes.

And this is compared with the UV average image (B) of the four directions, as shown in Figure 6 (c) while moving the pixel to calculate the sum of the difference between the pixels, the coordinates of the area where the sum of the difference between the pixels is the minimum Extract The extracted region coordinates are shown in FIG. 6 (d).

Through this process, the coordinate extractor 114 extracts the region coordinates of the UV image B obtained by the UV average image C. At this time, the entry direction determination unit 116 compares the area coordinates of the UV image (B) with the UV average image (C) in four directions, and determines the direction having the smallest value among the sums of the differences between the four pixels. Judging by the direction of entry.

Subsequently, the first calculator 122 adds each pixel value of the region coordinates extracted by the coordinate extractor 114 to obtain a total sum (s108).

The second operation unit 124 subtracts the sum of the sum of the result value of the first operation unit 122, the identified banknote roll value, and each pixel value of the UV average image C corresponding to the entry direction. In step 110, the result is obtained by subtracting the result by Equation (1) (s112).

The result is provided to the bill authenticity determination unit 126. Then, the bill authenticity determining unit 126 determines a predetermined threshold value E corresponding to the paper type information of the bill determined by the book type determining unit 106 and the entry direction of the bill determined by the entry direction determining unit 116. Is compared with the result (s114).

As a result of the comparison, if the result value falls within the threshold range, the banknote authenticity determining unit 126 determines that the banknote is authentic (YES in s114). no').

As described above in more detail with reference to the following [Table 1] and [Table 2] the process of identifying the bill (A) and whether or not authenticity.

It is assumed that the bill A is judged to be 10,000 won (new) by the winding type determination unit 106.

In addition, when the sum of the differences between the pixels between the UV image B and the UV average image C is as shown in Table 1 below, the entrance direction determining unit 116 determines this from the coordinate extraction unit 114. The direction which is provided and has the smallest value, that is, the backward direction, is determined as the banknote entry direction.

[Table 1]

Forward direction
Forward reverse direction Back Backward Frame processing result 15000 14000 11000 16000 Entry direction O

So banknotes can be seen that the bill of paper (A) is '10, 000 won (new) 'and the entry direction is' backward direction'.

In this state, the first calculator 122 adds each pixel value of the region coordinates extracted by the coordinate extractor 114 to obtain a total sum. This sum is assumed to be '528002'.

Then, the second operation unit 124 subtracts the total sum of the sum of the result value of the first operation unit 122 and each pixel value of the identified banknote roll and the UV average image C corresponding to the entry direction. Performs the operation. In this case, it is assumed that each pixel value of the UV average image C is as shown in Table 2 below.

[Table 2]

Forward direction
Forward reverse direction Back Backward 10,000 won (old) 513500 513500 511896 511896 Thousand won (old) 518123 518123 509252 509252 Ten thousand won (new) 520702 520702 523245 523245 5,000 won (new) 520936 520936 527842 527842 Thousand won (new) 541417 541417 528578 528578

The reason why the sum of the front direction and the rear direction is the same is because the image information sensed is the same whether the forward direction or the reverse direction is received.

Then, the second operation unit 124 may know the value '523245' corresponding to the '10, 000 won (new) ticket 'and the' backward direction 'and the' 528002 'value which is the total value of the area coordinates. The second operation unit 124 absolutely calculates these two values.

The result is '4757'.

The extracted result value '4757' is transmitted to the bill authenticity determination unit 126. The banknote authenticity determination unit 126 determines whether the extracted result value satisfies a threshold range corresponding to '10, 000 won (new) ticket 'and' backward direction '. Assume that the threshold is '15000'. This value is obtained when the coordinate region is obtained by the coordinate extractor 114.

As a result of the determination, if it is set as the normal banknote when the threshold range condition is smaller than the threshold value, since 4757 <15000 is established, the banknote authenticity determination unit 126 determines that the banknote is normal banknote. On the contrary, the counter is considered a counterfeit bill.

On the other hand, such a media recognition device can be applied to automated devices such as banknote recognizer, ATM (Automatic Teller Machine) device, CD (Cash Dispenser) device, vending machine, money exchanger. In this case, the controller of the automated device performs a predetermined function according to whether or not the bill of the medium recognition device is authentic. For example, if it is discriminated as a normal bill, the function is normally performed, whereas if it is discriminated as a counterfeit bill, the function is stopped.

In addition, the automated device or the media recognition device is provided with a communication interface (not shown) so that when the bill is determined as a counterfeit bill, it can be reported to a financial institution. For example, if the inputted bill is discriminated as a counterfeit bill, various information such as the device's unique number and the installation location are transmitted to the financial institution so that the counterfeit bill can be quickly recognized and coped with the situation.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and alternative constructions without departing from the spirit and scope of the invention. It will be apparent that other, alternative and equivalent embodiments are possible. Therefore, the true scope of the present invention should be determined by the technical idea of the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a flow chart for discriminating whether a bill is true or false;

2 is a block diagram of a media recognition device according to an embodiment of the present invention.

3 is a color pattern diagram for explaining the distinction between counterfeit and normal banknotes according to the 'R', 'G', 'B' color of the UV image information.

4 is a flowchart illustrating a medium recognition method according to an exemplary embodiment of the present invention.

5 is an explanatory diagram showing a process of obtaining a UV average image (C) in accordance with the present invention.

6 is an explanatory diagram showing a process of extracting region coordinates according to the present invention;

Description of the Related Art [0002]

102: image sensor 104: UV sensor

106: Kwonjong judging unit 108: Storage unit

112: comparison unit 114: coordinate extraction unit

116: direction determination unit 122: first operation unit

124: second calculation unit 126: bill authenticity determination unit

Claims (15)

Winding type determination unit for determining the winding type of the introduced banknote; A processing unit for extracting a predetermined coordinate region while comparing the UV image of the banknote with a plurality of UV average images; And, Subtract the total value of the sum of the pixel values of the extracted coordinate region and the sum of the sum of the pixel values of the UV average image corresponding to the roll type and the banknote entry direction of the introduced banknote, and the result is subtracted from the threshold value. And a control unit for controlling to determine whether the bill is authentic. The method according to claim 1, Media recognition, characterized in that further comprises a storage unit for storing each threshold value, the total sum value of the sum of each pixel value of the UV average image for each paper type and entry direction of the bill, the threshold value Device. 3. The method of claim 2, The winding type determination unit, the size of the 'R', 'G', 'B' value included in the UV image by the 'R', 'G', 'B' value for each kind, 'R', 'G for each kind And at least one or more ranges of 'B' values and bill size information. 3. The method of claim 2, The determination of the paper type is used an image sensor for sensing a bill image, The UV image is a medium recognition device, characterized in that a UV sensor for sensing the UV image of the bill image. 3. The method of claim 2, And the processing unit compares only the 'B' color among the 'R', 'G', and 'B' colors of the UV image and the UV average image. 3. The method of claim 2, And the coordinate area is an area in which the sum of the difference of each pixel value between the UV image and the plurality of UV average images is minimum. The method according to claim 6, And an entrance direction determining unit configured to determine a direction having a smallest value among the sum values of the difference values of the respective pixel values as a banknote entry direction. 3. The method of claim 2, The threshold is provided differently depending on the winding type and bill entering direction of the bill, And the control unit selects a corresponding threshold value when the bill type and the entry direction are determined. A medium recognizing apparatus, comprising any one of claims 1 to 8, And a controller for controlling the bill to be operated normally when the bill is normal and to stop the operation of the counterfeit bill according to the bill authenticity determination process of the medium recognition device. Determining the kind of paper currency introduced; Comparing the determined UV image of the banknote with a plurality of previously stored UV average images and extracting a predetermined coordinate region from the UV image; And, Subtract the total value of the sum of the pixel values of the extracted coordinate region and the sum of the sum of the pixel values of the UV average image corresponding to the roll type and the banknote entry direction of the introduced banknote, and the result is subtracted from the threshold value. A medium recognition method comprising the step of determining the authenticity of the bill by comparing. delete delete delete delete delete

Images