JP3209765B2 - Banknote recognition machine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a bill validator and, more particularly, to a bill validator which detects a bill pattern to identify the bill.

[0002]

2. Description of the Related Art Conventionally, a bill validator for detecting a bill pattern and identifying the bill is disclosed in Japanese Patent Application Laid-Open No. 54-13980.
There is an apparatus described in Japanese Patent Laid-Open No. 0-205. This device includes a detecting unit that detects a transmitted light amount of the bill and measures and detects a pattern of the bill, and a pulse generator that generates a pulse signal proportional to the transport speed of the bill, and the detecting unit detects the bill. During the period, the pulse signal generated from the pulse generator is counted to determine the length of the bill, and the pattern of the bill is determined by comparing the pattern of the bill detected by the detecting means with a previously stored reference pattern. In addition, the banknote can be more reliably identified by identifying the banknote based on both the length and the pattern of the banknote.

[0003]

However, even in the above-mentioned conventional bill validator, there is a case where it is difficult to distinguish between denominations of bills. In other words, for bills that differ slightly in size but have very similar patterns, add blank paper to make it the same size as other denominations, or cut off a part of the bill to remove In the case where the size is the same as that of the kind of banknote, the length and the pattern are sometimes determined to be appropriate by simply discriminating the length and the pattern as in the related art. Even in such a case, in order to reliably determine the authenticity, it is necessary to perform fairly strict pattern identification, and if such strict identification is performed, a genuine banknote is discriminated as a fake banknote, and the passability decreases. Could be caused.

[0004] To explain in more detail by giving a specific example, for example, an Italian 100,000 lira bill has an outer shape of 70 m.
mx 156mm, 50,000 lira is 70mm x 149
mm, 10,000 lira is 70mm x 133mm, 5,
The 000 lira is 70 mm × 126 mm, and the shapes are all the same in width, and the length is gradually increased as the bill becomes larger. With regard to the pattern, a human figure is arranged at substantially the same position on the right side of the banknote, and a watermark is arranged at substantially the same position on the left side of the banknote.

Here, the outer shape is reduced to 1 by adding 7 mm blank paper in the length direction of the 50,000 lira bill.
70mm x 156m with the same shape as the 00,000 lira bill
Suppose that a counterfeit banknote is created by setting m. Since the outer shape of this counterfeit banknote is the same as that of the 100,000 lira banknote and the pattern is very similar as a whole, when the pattern matching is performed and the similarity is detected, the reference value is cleared.
There was a problem that it was mistaken for a 000 lira bill.

On the other hand, even in such a case, if pattern matching is performed by setting a strict reference value for reliable identification, there is a problem that the pass rate of bill identification must be considerably reduced. In addition, not only Italian banknotes but also banknotes have similar patterns, and similar problems have occurred between banknotes having slightly different outer shapes.

By the way, in the case of the above-mentioned counterfeit banknote, that is, a counterfeit banknote in which blank paper is added to, for example, 50,000 lira banknotes to have the same outer shape as that of the 100,000 lira banknote, the pattern is 50,000 lira banknotes. Pattern and 10
It resembles both the 000 lira banknote pattern and the intermediate pattern. Therefore, 1
There is a feature that the similarity when comparing with the reference pattern of the 00000 lira bill and the similarity when comparing with the reference pattern of 50,000 lira bill are very similar values. Therefore, by measuring the difference between the similarities, it is possible to identify whether or not the banknote is a counterfeit banknote as described above.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and it is an object of the present invention to reduce the discrimination pass rate even for authenticity of counterfeit bills made of bills having similar sizes and patterns. It is an object of the present invention to provide a bill validator capable of reliably identifying.

[0009]

SUMMARY OF THE INVENTION The present invention relates to a bill validator for detecting a bill by detecting a pattern of the bill, and an object of the present invention is to provide a reference data for a plurality of denomination bills to be distinguished. a denomination reference data storing means for storing in each gold type of pixel values for each channel banknotes entire pattern
A pattern reading means for reading as, for each of the channels
Average value of pixel value for each channel from sum and length of pixel values
Average pixel value calculating means for calculating as a channel average value
And from the bill length information and the predetermined number of samples
Zone that calculates the number of zones to be allocated and allocates the zone area
Number calculation means and the image in each zone area of each channel.
The sum of the absolute values of the prime value and the channel average
Blocking means to determine the king value and each channel
Pixel value over the entire length
Calculate the sum of the absolute value of the difference from the average value as the power value
Channel power value calculating means, and
The blocking value determined by the blocking means
Normalization is performed by dividing by the power value.
Normalization calculating means, and the normalized data
The absolute value of the difference from the reference data in the predetermined direction for each denomination
The total sum added over the entire banknote is calculated as the similarity distance
Similar distance calculating means, and the similar distance calculating means
Pattern comparing means for comparing a plurality of similar distances, and a first candidate denomination which is the most similar denomination and a second candidate denomination which is the next similar denomination as a result of comparison by the pattern comparing means. Candidate denomination detecting means for detecting, and a candidate distance detecting means for calculating a difference between the similar distance of the first candidate denomination and the similar distance of the second candidate denomination detected by the candidate denomination detecting means, The banknote to be identified is determined to be the first candidate denomination on the condition that the similarity distance of the first candidate denomination is equal to or less than a predetermined value and the difference detected by the inter-candidate distance detecting means is equal to or more than a predetermined value. This is achieved by providing a denomination determining means.

[0010]

According to the present invention, the first candidate denomination, which is the most similar denomination from the reference data of each denomination, and the next denomination are determined when determining whether or not the identified banknote is a counterfeit banknote. Select a second candidate denomination that is a similar denomination, the similar distance of the first candidate denomination is less than or equal to a predetermined value, and the similar distance of the first candidate denomination and the similar distance of the second candidate denomination Since the difference is greater than or equal to a predetermined value, the identified banknote is determined to be the first candidate denomination, so it is determined whether or not the banknote is a counterfeit banknote without lowering the passage rate of the banknote, and The denomination can be determined.

[0011]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a diagram showing the configuration of one embodiment of the bill validator according to the present invention. The banknote pattern information optically read by the light receiving sensor 1 is converted into an electric signal by the photoelectric conversion unit 2. The converted electric signals are amplified for each channel by the amplifier 3, and all of them become input signals to the multiplexer 4. The multiplexer 4
Input signals are sequentially selected, and the selected signals are converted to A / D (Ana).
log / Digital). The electric signal relating to the pattern information input to the A / D converter 5 is converted into a digital signal, and is temporarily transferred via the bus 6 to the RAM (Ra).
nd Access Memory 7). CPU (Central ProcessingU)
nit) 8 determines the number of normalized zones 8 ₁ , average value calculation 8 ₂ , blocking process 8 ₃ , normalization process 8 ₄ , pattern matching 8 ₅ ,
Candidate denomination detection 8 ₆ , denomination judgment processing 8 ₇ , other control 8
₈ is performed. In the RAM 7, reference data of banknotes of a plurality of denominations to be discriminated are stored in advance for each denomination, and a ROM (Read Only Memory) is stored.
y) 9 stores a control program.

FIG. 2 is a diagram showing a mechanical configuration of a reading portion of the bill validator. In the figure, a bill 41 is conveyed by a conveying roller group 43 while its front side is a light source 4.
Irradiated by 21 and its reflected light is received by the light receiving sensor 11. The back side is illuminated by the light source 422, and the reflected light is received by the light receiving sensor 12.
As shown in FIG. 1, the light receiving sensor 1 is configured to be able to divide into 10 channels and input reflected light. Therefore, as shown in FIG. 3, the banknote pattern information (identification data) is divided into ten channels in the transport direction, and is divided into N data zones (areas) in the direction perpendicular to the transport direction depending on the shape of the medium. To divide. As for the size of the zone, for example, signals are fetched at a pitch of 1 mm in the transport direction, and 10 data thereof are defined as one zone.

Next, a processing procedure for bill identification in the present invention will be described. FIG. 4 is a flowchart illustrating a processing procedure of bill identification. First, when the identification process is started (step S1), it is determined whether or not the bill has been transported and arrived (step S2). If it has not arrived, continue to wait. On the other hand, if it has arrived, the capture of the pattern signal is started from the leading end of the bill (step S3). Here, the bill transport motor (not shown) is a stepping motor, and captures the pixel output of each channel of the light receiving sensor 1 for each synchronization signal. Next, it is determined whether or not the capture of the pattern signal has been completed (step S4). As a result of the determination, if the processing has not been completed, the capture is continued. On the other hand, when the processing is completed, information on the length of the bill is fetched (step S5). Thereafter, an identification process is performed based on the acquired identification data (step S6).

FIG. 5 is a detailed flowchart of the identification process. Here, first, the number of zones N is obtained from the length information of the banknote fetched in step S5 and the sampling number n of one zone based on the formula 1 (step S601).

[0015]

N = length / n Next, the total sum T of the pixel outputs for each channel is calculated by the following equation (2).
Obtain OTAL (ch) (Step S602)

[0016]

(Equation 2) According to Equation 3, the total TOTAL (c
h) is divided by the length length to obtain an average value AVE (ch) of the output in each channel (step S603). FIG. 6A shows an image on a certain channel.
It is a figure which shows the prime value D (ch, i) and its pixel average value AVE (ch).

[0017]

AVE (ch) = TOTAL (ch) / length The pixel value D (ch, (zo−1) × n + j), which is the output for each zone , is obtained for each channel by Equation 4 in step S603. Zone data Z (ch, zo), which is the absolute value of the difference from the average value AVE (ch), is obtained (step S604). This is a so-called blocking process. Where zo indicates a zone and j is 1 to N
Indicates the zone number up to.

[0018]

(Equation 4) From Equation 5, for each channel, the power value P which is the sum of the absolute values of the differences between the pixel value D (ch, i) and the channel average pixel value AVE (ch) over the entire length direction.
WR (ch) is obtained (step S605). FIG.
(B) shows a power value PWR (c) in a certain channel.
FIG.

[0019]

(Equation 5) From Equation 6, for each channel, the power of that channel in the zone data Z (ch, zo) of each zone
The normalized data S (ch, zo), which is a ratio to the PWR (ch) , is obtained (step S60).
6). FIG 6 (C) is a diagram showing the normalized zone de <br/> over data S of a channel (ch, zo). The horizontal axis is ZO
The left side is the leading edge of the bill and the right side is the trailing end of the bill
Department. The normalized zone data S (ch, zo) value forms the pattern data. Note that an optimum value is given to the coefficient (256).

[0020]

S (ch, zo) = Z (ch, zo) * 256 / PWR (ch) Next, the normalized data S (ch, zo) obtained in step S 606 and the Z denomination are obtained by the equation 7. The difference KY _ZA from the model data in the A direction is obtained (step S607).

[0021]

(Equation 7) The denomination is determined as the first candidate denomination and the second candidate denomination from the top two denominations with the smallest difference from each denomination (step S).
608).

Hereinafter, it is determined whether the bill is a genuine bill or a fake bill. Therefore,
First, the difference KY1 between the model of the first candidate denomination and the difference KY2 of the model of the second candidate denomination are compared to determine whether or not the difference is equal to or greater than a certain threshold TH2 (step). S609). If the result of the determination is that the bill is less than the threshold value TH2, the bill is determined to be a fake bill (step S6).
14). In other words, it is approximately equivalent to both of the two types of banknotes. On the other hand, if the difference is equal to or greater than the threshold value TH2, the difference KY1 from the model of the first candidate denomination
Is less than or equal to a certain threshold value TH1 (step S610). If the result of the determination is that the bill is equal to or less than the threshold value TH1, the bill is determined to be the first candidate denomination (step S613). On the other hand, the threshold value TH1
If not, whether or not the difference between the difference KY1 with the model of the first candidate denomination and the difference KY2 with the model of the second candidate denomination is equal to or greater than the threshold TH2 plus 100 It is determined (step S611). As a result of the determination, when the bill is not satisfied, it is determined that the bill is a fake bill (step S614). If the condition is satisfied, it is further determined whether or not the difference KY1 from the model of the first candidate denomination is equal to or less than a value obtained by adding 200 to the threshold value TH1 (step S612). If it is determined that the bill is not satisfied, the bill is determined to be a fake bill (step S6).
14). If so, the banknote is determined to be the first candidate denomination (step S613). here,
Steps S611 and S612 are for increasing the pass rate of the bill, and on the condition that the difference between the first candidate and the second candidate is larger, it is determined whether or not the banknote is the first candidate in that case. And the condition is relaxed.

FIG. 7 shows a table of the similarity distance when the reference data of each genuine bill is compared with the reference data of other genuine bills. Indicates that is large. Also,
FIG. 8 is a diagram showing a table of similar distances of the counterfeit bills to the 100,000 lira bills and the 50,000 lira bills, for example, when the 50,000 lira is forged from the 50,000 lira. This specific data also supports the basic point of the present invention that counterfeiting of another banknote by adding blank paper from one counterfeit banknote is similar to both.

[0024] Incidentally, denomination reference data storage means RAM 7, the light receiving sensor 1 and the photoelectric conversion unit 2 is pattern reading means, the pattern matching 8 ₅ pattern comparing means, the denomination determination processing 8 ₇ candidate distance detection means and gold Each corresponds to a type determining means.

[0025]

As described above, according to the bill validator of the present invention, a blank paper is added to a bill of a slightly different size but very similar in pattern to add another denomination. A banknote that has been counterfeited by making it the same size as that of the original banknote, or by cutting off a portion of the banknote and making it the same size as another banknote of the denomination, such a counterfeit banknote resembles the original banknote. Utilizing the feature of resembling both banknotes, whether or not the identified banknote is a counterfeit banknote based on the similar distance to the most similar denomination and the similar distance to the next similar denomination , Or whether or not the denomination can be determined as the most similar denomination, it is possible to accurately determine the bill without lowering the pass rate.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of one embodiment of a bill validator according to the present invention.

FIG. 2 is a diagram showing a mechanical configuration of a reading part in the bill validator.

FIG. 3 is a diagram showing how the pattern information (identification data) of a bill is divided.

FIG. 4 is a flowchart showing a processing procedure of bill identification.

FIG. 5 is a detailed flowchart of an identification process.

FIG. 6 is a diagram showing a graph based on an example of pattern information of a read bill.

FIG. 7 is a diagram showing a table of similar distances when reference data of each authentic banknote is compared with reference data of another authentic banknote.

FIG. 8 is a diagram showing a table of similar distances of the counterfeit bills to the 100,000 lira bill and the 50,000 lira bill when the 50,000 lira is forged from the 50,000 lira.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Light-receiving sensor 2 Photoelectric conversion part 3 Amplification part 4 Multiplexer 5 A / D conversion part 6 Bus 7 RAM 8 CPU 9 ROM

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-3-17794 (JP, A) JP-A-62-1113293 (JP, A) JP-A-2-108188 (JP, A) JP-A-2- 148383 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G07D 7/00 G07D 7/20

Claims (1)

(57) [Claims] 1. A denomination reference data storage means for storing, for each denomination, reference data of a plurality of denomination bills to be identified; a pattern reading means for reading a pattern of the entire bill as a pixel value for each channel; Sum of pixel values for each
Average value of pixel value for each channel from channel and length
Mean pixel value calculating means for calculating as an average value, length of bill
Zone assigned from the information and a predetermined number of samples
Means for calculating the number of zones and allocating zone areas
And the pixel value in each zone area of each channel and
The sum of the absolute value with the channel average is defined as the blocking value.
What blocking means do you want
Pixel value over the entire length direction and the channel average value
Channel that calculates the sum of the absolute values of the differences from the power value
Power value calculating means;
The blocking value determined by the locking means is the power
A normalization operation that performs normalization by dividing by a value
Calculation means, and the normalized data and each preset gold
Type The absolute value of the difference from the reference data in the predetermined direction
Similarity distance to obtain the sum total over the distance as similarity distance
Calculating means, and a plurality of similar distances calculated by the similar distance calculating means.
As a result of comparison by the pattern comparing means for comparing similar distances and the pattern comparing means , a first candidate denomination which is the most similar denomination and a second candidate denomination which is the next similar denomination are detected. Candidate denomination detecting means, and an inter-candidate distance detecting means for determining a difference between the similar distance of the first candidate denomination and the similar distance of the second candidate denomination detected by the candidate denomination detecting means,
When the similarity distance of the first candidate denomination is equal to or less than a predetermined value, and the banknote to be identified is the first candidate denomination under the condition that the difference detected by the inter-candidate distance detecting means is equal to or more than a predetermined value. A bill discriminating machine comprising a denomination discriminating means for discriminating.