JPS6365995B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of the Invention The present invention relates to a paper sheet discriminating device that, for example, identifies missing banknotes such as damaged or folded edges during conveyance of banknotes.

(B) Background of the Invention For example, in the case of a circulating deposit/withdrawal device that dispenses banknotes inserted for deposits for payment, the deposited banknotes are genuine banknotes no matter how damaged they are, so the deposited banknotes are However, since the withdrawn banknotes will be recirculated in the market, it is not suitable to use the deficient banknotes for withdrawal, and the system detects the deficient banknotes and removes them from the withdrawn banknotes. I have to do it.

Conventionally, there have been devices that detect the width and length of a banknote, or the area of a banknote based on these, and determine the authenticity and denomination of a banknote. Although disclosed in Publication No. 53-4562,
These were unable to detect the loss of banknotes.

(C) Purpose of the Invention This invention has a unique configuration that sequentially compares the length data of the paper sheet read by scanning with an image sensor from the beginning and the end, so that it is possible to determine whether or not the paper sheet is missing. The purpose is to provide a leaf identification device.

(d) Structure of the Invention The present invention includes a sheet conveying device and a sheet conveying device that scans the sheets conveyed by the conveying device multiple times in a direction perpendicular to the conveying direction, and separates the sheets in each scan. an image sensor that reads the length of the paper sheet; a memory that stores the length data of the paper sheet on each scanning line read by the image sensor; and a memory that stores the length data of the paper sheet on each scanning line read by the image sensor; In response to this, the length data of the paper sheets read from the memory by scanning is sequentially read out from the beginning and end, and compared, and the symmetry of the length data of the paper sheets is checked. The present invention is characterized by a paper sheet discriminating device having a defect detection means for detecting the presence or absence of defects in paper sheets.

(E) Purpose of the Invention This invention scans the paper sheets conveyed by a conveyance device in a direction perpendicular to the conveyance direction of the paper sheets to measure the length of the paper sheets. The data is read and sequentially stored in a memory, and in response to the completion of scanning one paper sheet by the image sensor, the defect detection means reads length data of the paper sheet read from the memory by scanning. The paper sheets are sequentially read out from the beginning and end and compared, and the symmetry of the length data of the paper sheets is checked to detect the presence or absence of defects in the paper sheets.

(F) Effects of the Invention As a result of the above, according to the present invention, there is no need for a reference pattern for comparison, and furthermore, there is no need for complex arithmetic processing such as calculating area.

Furthermore, the system scans one sheet of paper and stores it in memory, and then detects the presence or absence of defects based on the data stored in the memory. Even if there is a defect, the defect can be reliably detected, and the degree of the defect can also be detected.

Furthermore, even if the paper sheet is conveyed at an angle, since the data stored in the memory is used, the length data on the corresponding scanning line can also be obtained as being at the same angle. A defect can be detected by reliably distinguishing between a shortened length on a line and a shortened length due to a defect, and accurate defect detection processing can be performed quickly.

(F) Embodiment of the Invention An embodiment of the invention will be described in detail below based on the drawings.

The drawing shows a banknote discriminating device used in an automatic deposit/payment machine used in banking operations. transaction processing,
The circulating deposit/withdrawal device is configured to circulate deposited banknotes of each denomination accepted during deposit transactions as withdrawn banknotes during payment transactions.

Reference numeral 1 denotes an operation procedure display that sequentially displays the operation procedure, and the customer operates according to the displayed procedure. 2
is a card insertion slot, and a card reader is provided inside, and card data is read by inserting a card given to a customer into this slot. In addition,
A passbook may be used in place of or in addition to the card.

Reference numeral 3 denotes a receipt discharge port, which is provided with a slip issuing device therein, and issues and discharges a receipt on which transaction details are printed at the time of a deposit transaction or a payment transaction.

Reference numeral 4 denotes an amount display that displays the amount at the time of deposit transaction or payment transaction. 5 is a denomination number display which displays the number of denominations of each denomination inserted at the time of deposit transaction.

6 is a deposit key, and 7 is a payment key, which respectively designate a deposit transaction and a payment transaction.

Number 8 is the numeric keypad, and you enter numbers such as your password and payment amount. 9 is a denomination key, which specifies the denomination at the time of payment.

10 is a confirmation key, which is operated when the deposit transaction is approved without canceling the transaction. Reference numeral 11 denotes a cancellation key, which is operated to cancel the above-described deposit transaction.

Reference numeral 12 denotes a deposit slot into which banknotes are deposited during a deposit transaction. Reference numeral 13 denotes a withdrawal port, which returns the banknotes when the banknotes inserted during a deposit transaction are determined to be defective banknotes, when the cancel key 11 is pressed, and discharges banknotes fed out as payment banknotes during a payment transaction. .

In the automatic deposit/payment machine configured as described above, the user first operates the deposit key 6 or the payment key 7 to specify a transaction, then inserts the card into the card slot 2 and enters the PIN number using the numeric keypad 8. Depending on the situation, the correspondence between the card and the customer is checked. In addition,
This check is not necessarily necessary for deposits.

In the case of a deposit transaction, deposited banknotes are inserted into the deposit slot 12, and the banknotes are taken in one by one to determine the authenticity and denomination of the banknotes. Defective banknotes are discharged to the withdrawal port 13 and returned.

The amount of accepted normal banknotes and the number of each denomination are displayed on the amount display section 4 and the number of denominations display 5, respectively, and if the customer confirms that it is OK, the transaction is completed by pressing the confirmation key 10. The card is returned and a receipt is ejected from the receipt ejection port 3.

When the cancel key 11 is pressed, the transaction is canceled and the accepted banknotes are discharged to the withdrawal port 13 and returned. In addition, in this case, the returned banknotes will be different from the banknotes at the time of insertion, but
This will become clear later in the explanation.

In the case of a payment transaction, the requested payment amount and each denomination are entered using the numeric keypad 8 and the denomination key 9, and these amounts and the number of coins of each denomination are displayed on the amount display 4 and the number of denominations display. 5 is displayed, and if the customer confirms that it is OK, the transaction is completed by pressing the confirmation key 10, the payment bill is discharged to the withdrawal port 13, the card is returned, and the transaction is completed by pressing the confirmation key 10.
A receipt will be issued.

FIG. 2 shows a circulating deposit/withdrawal device configured inside the deposit port 12 and the withdrawal port 13 described above. An eccentric roller 15 and a take-in roller 16 are provided below the eccentric roller 15 facing one surface, a holding plate 17 is provided facing the other surface, and a bottom plate 18 is provided on the bottom surface.
have.

The presser plate 17 presses the bill 19 toward the rollers 15 and 16, but when accepting the bill 19, it is retracted by a suitable cam mechanism to enable acceptance. Then, by detecting acceptance of the banknote 19, it is brought into the pressing position.

A friction roller 20 is provided opposite to the above-mentioned taking-in roller 16, and the above-mentioned eccentric roller 15 is rotated to take in banknotes 19, which have been loosened from overlapping, one by one as the taking-in roller 16 rotates.

In addition, 21 in the figure is the intake empty sensor, and the intake part 1
It is detected that four banknotes 19 are taken in one after another and become empty.

The banknotes taken in by the above-mentioned take-in roller 16 are conveyed by first conveyor belts 22, 22 which are conveyed at a conveyance speed faster than this take-in speed.

A length abnormality detector 23 and a double-bill overlap detector 24 are provided in the conveyance path of the first conveyor belts 22, 22, and the detector 23 is constituted by, for example, a photoelectric sensor and measures the passage time of the banknotes. Then, it is checked whether two bills have been taken out. The detector 24 also checks whether two or more bills overlap by passing the bill between a cam and a roller that has a gap of one and a half the thickness of the bill. Note that this configuration itself is well known.

The first banknote discriminating unit 25 is constituted by, for example, a magnetic head, and reads the printed pattern of the banknote to determine authenticity and denomination.

The banknotes read by the first banknote discrimination section 25 described above are conveyed to the sorting section 29 by a conveyance path formed by the second conveyance belts 26, 26, the drum 27, and the third conveyance belt 28.

Fourth and fifth conveyor belts 3 are attached to the above-mentioned swing section 29.
0, 31, and sixth, seventh, eighth, and ninth conveyor belts 32, 3 that are in contact with these belts 30, 31.
3, 34, 35, and distribution plates 36A, 36B, 3
6C, 36D, and distribution plates 36A to 36D.
The sorting operation is performed based on the determination result of the first bill discriminating section 25 described above. Note that this operation is performed by an appropriate solenoid.

The denominations of banknotes handled by this deposit/withdrawal device are A, B,
The three denominations A, B, and C are mixed as D.

The aforementioned distribution board 36A distributes denominations A, the distribution plate 36B distributes denominations B, the distribution board 36D distributes denominations C, and the distribution board 36D distributes denominations A, B, C, and more. Each distribution board 36A~
36C performs a sorting operation on the banknotes that were not sorted.

The aforementioned conveyor belts 30 to 35 convey banknotes for the aforementioned distribution process.

The conveyance end portions of the fifth conveyor belt 31 and the ninth conveyor belt 35 are connected to the discharge section 37, and defective banknotes determined to be counterfeit by the bill discriminating section 25 described above are conveyed to the discharge section 37 without being sorted. Ru.

Pool portions 38A to 38D are temporarily provided on the distribution direction side of each of the aforementioned distribution plates 36A to 36D,
Banknotes of denominations sorted by each sorting board 36A to 36D are once pooled.

In addition, 39 in the figure is a guide roller.

The above-mentioned discharge section 37 is constructed below the dispensing port 13, and has a movable bottom plate 40, a capstan 41 and a pinch roller 42 arranged oppositely to the upper surface of the movable bottom plate 40. The pinch roller 42 is connected to the capstan 41 by an appropriate solenoid. It is provided to be transferred to the side and pressed.

The above-mentioned capstan 41 and pinch roller 42 are usually separated, and the dispensed banknotes and returned banknotes are fed between them, and when the banknotes are ejected, the pinch rollers 42 move and pinch the banknotes, and the capstan 41 rotates in the ejecting direction. By doing so, the banknotes are released.

The above-mentioned movable bottom plate 40 is provided to open by swinging downward using its left end as a fulcrum,
It is opened and closed by an appropriate solenoid. A recovery conveyance path is formed in the lower part of the movable bottom plate 40 by the tenth conveyance belt 43 that is in contact with the ninth conveyance belt 35, and when collecting the banknotes on the movable bottom plate 40,
The movable bottom plate 40 is opened and the banknotes are placed on the belts 35, 4.
3 and once collected in the pool section 38D.

Each of the above-mentioned pool parts 38A to 38D is formed into a box shape with an open top surface, and the bottom part is provided to open and close downward, and is opened and closed by an appropriate solenoid to move the pooled banknotes downward. can be supplied.

The lower part of each pool part 38A-38D has storage boxes 44A-44D for storing banknotes from each upper stage, and the storage boxes 44A-44C are separated by denomination for storing the corresponding denominations A, B, and C, respectively. The storage box 44D is a storage box that is fixed to the device, and the storage box 44D is a bulk storage box that stores mixed denominations D all at once. The storage capacity is larger than that of the boxes 44A to 44C. Note that if the banknotes to be stored in the storage box 44D are stacked together for each denomination, when sorting and storing the banknotes into the storage boxes 44A to 44C by denomination,
This is preferable because the number of operations of the distribution plates 36A to 36C is reduced.

The lower part of the storage boxes 44A to 44D constitutes a feeding section 45 for feeding out banknotes, and the lower surface of each of the storage boxes 44A to 44D has feeding rollers 46A to 46D, respectively.
The banknotes are fed out by being rotated.

In addition, 47... is an eccentric roller to loosen the weight of each stored banknote, and 48... is a friction roller that is in contact with each of the feeding rollers 46A to 46D.

The above-mentioned feeding section 45 has an eleventh conveyor belt 49, and twelfth, thirteenth, and fourteenth conveyor belts 5 that are in contact with the eleventh conveyor belt 49.
0, 51, 52, and the third conveyor belt 28 mentioned above.
The banknotes are conveyed from each of the storage boxes 44A to 44D, and are conveyed to the discharge section 37 using the conveyance path of the sorting section 29.

Each storage box 44 is placed in the conveyance path of the above-mentioned feeding section 45.
It has length abnormality detectors 53A to 53D that detect the length of banknotes fed out from A to 44D, and a double overlap detector 54 that detects the overlap of two banknotes, and also has a second banknote discrimination unit 55. has been established.

The length abnormality detectors 53A to 53D described above are composed of photoelectric sensors, and the double-sheet overlap detector 54 is constructed in the same manner as the detector 24 described above.

The length abnormality detector is double overlap detector 5.
4, only one may be provided in the common transfer path. Further, the second bill discriminating section 55 inspects whether or not the dispensed bill is an unfit bill, although the details will be described later.

In addition, if the dispensed bills are due to a mistake in dispensing two bills or overlapping two bills, the sorting plate 36D is moved and the bills are temporarily collected in the pool portion 38D.

In the storage boxes 44A to 44D described above, 56A
- 56D are storage box full sensors that detect the fullness of banknotes stored in each storage box; 57A to 57D are storage box remaining sensors that detect the amount of banknotes stored in each storage box; 58A; ~58D is a storage box empty sensor that detects whether each storage box is empty.

In addition, the storage box full sensors 56A to 56C described above
If the amount of banknotes used varies depending on the frequency of use of each denomination, the detection position can be freely set depending on the frequency of use.

In addition, 59A to 59D in the pool parts 38A to 38D mentioned above are pool part empty sensors,
Each time it is detected that the pool section is empty.

And each of the above-mentioned sensors 56, 57, 58, 59
is composed of a photoelectric sensor, but a micro switch or the like may also be used.

FIGS. 3 and 4 show the aforementioned second banknote discriminating section 55, in which the conveyance path composed of the conveyor belts 28 and 49 has an interrupted part, and a guide plate 63 is provided on one side of this interrupted part. The banknotes being conveyed are provided so as to pass through this interruption part via this guide plate 63.

The above-mentioned guide plate 63 is formed to have the width of the conveyance path, and a light-transmitting glass plate 64 such as ground glass or transparent glass is fitted in the middle part, and the lower surface of the plate 64 has a light-transmitting glass plate 64 that emits light evenly with respect to the width of the path. A light source 65 that emits light is provided.

On the upper surface side of the above-mentioned glass plate 64, first and second line image sensors 66 and 67 are arranged on the left and right from the center of the width of the conveyance path, respectively, and lenses 68 and 67 are installed on the upper surface side of the glass plate 64.
68, and these lenses 68, 68
is 1/1 of each banknote conveyed on the glass plate 64.
2 images are formed on line image sensors 66 and 67.

The aforementioned first and second line image sensors 6
6 and 67 are constituted by, for example, a CCD array, and each line (in the path width direction) is configured to read 256 bits, and each sensor 66 and 67 divides this into two and reads 128 bits.

A pulse generator 71 is connected to a shaft 70 of a pulley 69 on which a conveyor belt 49 is hung.

The pulse generator 71 described above includes a disk 73 having slits 72 provided at equal intervals on the circumference, and a photoelectric sensor 74 that detects the slits 72 of the disk 73 and outputs pulses. It is set to output one pulse each time it is conveyed. Note that the relationship between this pulse and bill conveyance is not limited to this.

First and second banknote tilt sensors 75 and 76 are provided on the left and right sides of the glass plate 64 from the center in the width direction and on the upper side of the glass plate 64 in the transport direction, and these sensors 75 and 76 are photoelectric sensors. It consists of

Figure 5 shows the actual size value Y of the tilted banknote 77 in the longitudinal direction.
and the logic of measuring the actual size value V in the width direction. When the distance E between both banknote inclination sensors 75 and 76 is set to 100 mm, for example, and the inclination of the banknote 77 is an angle θ, the length of the sloped side F , the length D of the lagging side can be expressed as the following relational expression.

F ² = E ² + D ² = 100 ² + D ² F = √100 ² + ² The length D of the lag side described above is the length of the time difference in which the banknote 77 passes through both banknote tilt sensors 75 and 76 due to the slope of the banknote 77. It can be calculated by converting to

Furthermore, the measured value W of the banknote 77 in the width direction can also be calculated by converting the time taken by the banknote 77 through the banknote tilt sensors 75 and 76 into length.

Therefore, the actual size value V of the banknote 77 in the width direction is It is calculated by

Furthermore, the length of the banknote 77 in the longitudinal direction is measured by line image sensors 66 and 67, and the actual size value Y in the longitudinal direction is expressed as follows: It is calculated by

As described above, the length Y of the banknote 77 in the longitudinal direction is
and the length V in the width direction are calculated, but in actual use, when D is 1, X-a ₁ , W-a ₂ When D is 2, X-b ₁ , W-b ₂ When D is n, a correction value n1 in the length direction and a correction value n2 in the width direction are set corresponding to the length D of the lagging side, such as X-n ₁ and W-n ₂ , and these correction values The actual size values Y and V are calculated by subtracting n1 and n2 from the respective measured values X and W.

The above correction values n1 and n2 are the length of the delay side D
A correction value table is constructed in correspondence with the above and stored in the memory.

Note that if the banknote 77 is tilted too much, a failure such as an error in ejecting the banknote 77 will occur, so a limit DX of the length D of the lagging side is set, and if the banknote 77 is tilted by a length D greater than this, the banknote will be rejected. do.

FIG. 6 shows the control circuit of the second banknote discriminating section 55, in which the CPU 78 controls each circuit device, performs various calculations, and also outputs various reset signals.

The width direction length/inclination detection circuit 79 includes banknote inclination sensors 75 and 76 and a pulse generator 71,
The above-mentioned length W in the width direction and length D of the lagging side are counted and the respective measured values W and D are output to the CPU 78, and the CPU 78 corrects the measured value W to obtain the actual size value V in the width direction. Calculate.

The longitudinal length/unfit note detection circuit 80 includes line image sensors 66 and 67, and detects the length X in the longitudinal direction described above and the unfit note banknote described later, and calculates the measured length value X. Line image sensor 66, 67
The banknote detection count value of one line is calculated by CPU78.
The CPU 78 corrects the measured value X to calculate the actual size value in the length direction, and detects unfit bills based on the banknote detection count value of one line.

The memory 81 has a table of the above-mentioned width direction correction value n2, a table of the length direction correction value n1, an allowable error table for determining unfit notes to be described later, and a plurality of points for storing other data.

FIG. 7 shows the aforementioned width direction length/inclination detection circuit 79.
As described above, the pulse generator 71 outputs one pulse for every 1 mm of conveyance of the banknote 77, and this pulse signal is output to the CPU 78 and the AND gate 82.

The first and second banknote tilt sensors 75 and 76 output "L" (low level signal) when the banknote 77 is detected, and output "H" when the banknote 77 is not detected.
(high level signal) is output from the respective sensors 75 and 76 via inverters 82 and 83.

The AND gate 81 is operated by the pulse generator 7 while the first bill tilt sensor 75 detects the bill 77.
1 and the first and second counters 84 and 85 count the pulses and output the counted values. Note that the second counter 85 is the same as the first counter 84.
This is for carrying.

Therefore, the first and second counters 84, 85
The count value output from the first banknote tilt sensor 75
Since the number of pulses is counted while detecting the banknote 77, the measured value W of the length in the width direction of the banknote 77
becomes.

Further, first and second banknote tilt sensors 75, 76
The output via inverters 82 and 83 is input to an exclusive OR gate 86 and an AND gate 87, and the output of the exclusive OR gate 86 is input to an AND gate 88, and the output of this AND gate 88 is further input to an AND gate 89. is input.

The output of the AND gate 87 mentioned above is input to the input terminal of FF (abbreviation for flip-flop) 90, and this
The output from the output end of the FF 90 is input to the AND gate 88 mentioned above.

The output of the AND gate 89 is counted by third and fourth counters 91 and 92. Note that the fourth counter 92 is for carrying up the third counter 91.

The aforementioned exclusive OR gate 86 detects the length D of the tilted side when one of the first and second banknote tilt sensors 75 and 76 is detecting the banknote 77, that is, when the banknote 77 is tilted. When both sensors 75 and 76 are not detecting the bill 77 at the same time and when they are detecting the bill 77, no output is produced. Furthermore, the detection between the length D due to the inclination of the banknote 77 is performed twice: when the banknote 77 rushes into both the above-mentioned sensors 75 and 76, and when it passes through both the sensors 75 and 76, so the above-mentioned exclusive The OR gate 86 will output twice, once when the banknote 77 enters and when it passes.

On the other hand, the FF90 detects the AND gate 87 when both the sensors 75 and 76 detect the banknote 77 at the same time.
Since it is set at the output of , there will be no output at that output terminal, and the AND gate 88 will be inhibited.

As a result, the AND gate 88 outputs a detection signal between the length D of the inclined side of the bill 77 at the time of entry, the AND gate 89 outputs a pulse during this period, and the counters 91 and 92 count this.

Therefore, the third and fourth counters 91, 92
The count value output from is the measured value D of the length of the slope side
becomes.

Furthermore, the output from the AND gate 88 is the banknote 77
It is input to the CPU 78 as a bill intrusion detection signal that detects that the bill has entered either the first or second bill tilt sensor 75 or 76, and the output from the Q output terminal of the FF90 is used to measure the length of the tilted side. The CPU 78 sends a tilt measurement end signal when it detects that the
is input.

Note that the first to fourth counters 84, 85, 9
1, 92 and FF90 are reset by a program from the CPU 78.

Figure 8 shows the length direction length/unfit note detection circuit 8 mentioned above.
0, and FIG. 9 shows a time chart of the main parts.

First and second line image sensors 66, 67
are driven by sensor drive circuits 93 and 94, respectively.

A program start signal from the CPU 78 is input to the start terminal of the sensor drive circuit 93 on the first line image sensor 66 side, and this circuit 93 outputs an EOS signal indicating the end of reading.
Since the terminal is connected to the start terminal of the sensor drive circuit 94 on the second line image sensor 67 side, when the first line image sensor 66 finishes reading 128 bits, the EOS signal is transmitted to the second line image sensor 67. As a start signal, this second line image sensor 67 continues reading 128 bits.

When the reading of the second line image sensor 67 is completed, the EOS signal of the sensor drive circuit 94 is
It is input to the CPU 78.

The video output of each sensor drive circuit 93, 94 outputs "L" when the line image sensors 66, 67 detect the banknote 77, and outputs "H" when the line image sensor 66, 67 does not detect the banknote 77.

Therefore, the video outputs are the above-mentioned signals corresponding to the respective bits of the line image sensors 66 and 67.

The discrimination circuits 95 and 96 are connected to each sensor drive circuit 93,
The "L" video output output from 94 is taken out, that is, each line image sensor 66, 67
extracts the output of the bit position where the banknote 77 is detected, converts it into a pulse signal, and outputs it.

And this read pulse signal is the OR gate 9
7 to the fifth and sixth counters 98 and 99 for counting. Note that the sixth counter 99 is for carrying up the fifth counter 98.

Note that clock pulses are input from the CPU 78 to the sensor drive circuits 93, 94 and the discrimination circuits 95, 96, and the counters 98, 99 are input from the CPU 78.
It is reset by the program.

The banknote tilt sensors 75 and 76 described above are connected to the banknote 77.
When detecting the entry of pulse generator 7 (output signal of AND gate 88), CPU 78 starts the next pulse generator 7.
Both counters 98, 99 based on the pulse from 1
After that, a start signal is output to the sensor drive circuit 93.

The sensor drive circuit 93 described above outputs the reading state of each bit of the first line image sensor 66 as a video output, and when the reading of all bits is completed, the sensor drive circuit 94 outputs the reading state of each bit of the second line image sensor 67. It outputs the reading status as a video output and outputs the EOS signal when the reading of the gold bits, that is, the reading of a total of 256 bits, is completed.
Output to CPU78.

The counters 98 and 99 then count the pulses from the start signal described above until the EOS signal is output, and this count value is stored in a predetermined pointer in the memory 81 based on the output of the EOS signal.

Line image sensors 66, 67 as described above
Reading one line by pulse generator 7
This is repeated based on one pulse.

FIG. 10 shows the reading state, and when the bill 77 is read in an inclined state, the first reading line X1 is a slight length of the corner of the bill 77, and the second reading line The length of the reading line becomes longer each time it is read, and the reading line at the end is Xn−2, Xn
-1, xn becomes shorter sequentially.

Therefore, since the regular banknote 77 is rectangular, the above-mentioned reading lines are placed first on the X1 side and last on the X1 side.
When compared from the Xn side, the banknotes 77 are graphically symmetrical on the top and bottom on the center line, so they have the same length.

If the banknote 77 has a folded end, a scrap 100, or a damage 101, the symmetry will be disrupted and the length will appear different. By detecting this difference in length, unfit bills can be detected.

In the above length comparison, there may be a large difference between the first X1 side and the last Xn side due to the sampling position shift, and it is thought that the difference is small in the middle part. To this end, the unfit note tolerance table in the memory 81 described in FIG. 6 stores tolerable error values by increasing the error due to the initial comparison and decreasing it thereafter.

Furthermore, the length X of the banknote 77 in the longitudinal direction is measured by reading out the measured value X of three lines Xe, Xe+1, and One of them is selected and this is set as the measurement value X.

Detection of unfit banknotes by the second banknote discriminating section 55 configured as described above will be described in detail with reference to the flowchart of FIG.
9 and enters the second bill discriminating section 55 . When either one of the bill tilt sensors 75 and 76 detects the intrusion of the bill 77 , the bill intrusion detection signal is input to the CPU 78 .

Based on this signal, the CPU 78 clears the data area of the previously stored read data in the memory 81, and then stores the address at which the read data is stored in the pointer 1 of the memory 81.

Next, it is determined whether a pulse is being output from the pulse generator 71, and when this output is confirmed, a start signal is output to the line image sensor 66. The counters 98 and 99 count by reading the length X, and an EOS signal is output.

At the initial stage when the banknote 77 enters the banknote tilt sensors 75 and 76, the banknote 77 has not yet reached the line image sensors 66 and 67, so the counter 98,
99 count value, that is, the measured value of length in the longitudinal direction
is 0 and is not measured, so the read flag is not set.

Therefore, a measurement value of 0 is not stored in memory 81.

When the reading of one line by the line image sensors 66 and 67 is repeated using the start signal and the EOS signal, and the first counted value, that is, the measured value X1 is output from the counters 98 and 99,
The read flag is set and the measured value X1 is stored in the area at the address indicated by pointer 1.

Then, +1 is added to pointer 1 to measure the next line.

By repeatedly reading such lines, the measured values X1 are stored in the memory 81 in sequence.

If the measured value X becomes 0, this indicates that the banknote 77 has passed through the line image sensors 66, 67, so line reading is finished and the reading flag is reset.

At this time, the address of the read data (measured value Store address.

Next, the measured value X is sequentially compared from the first X1 side and the last Xn side to check their symmetry.

That is, store the first stored start address of the measured value X and its measured value X1 in pointer 2,
Furthermore, the start address of the allowable error table for determining unfit notes and its allowable error value are stored in pointer 3.

Then, the first measured value X1 of pointer 2 is subtracted from the last measured value Xn of pointer 1, and it is determined whether the difference is smaller than the allowable error value.

When the above-mentioned difference is equal to or larger than the allowable error value, the banknote 77 is determined to be a damaged banknote. Since the unfit banknotes are not suitable as banknotes for withdrawal, they are collected in the storage box 44D.

Further, when the aforementioned difference is smaller than the allowable error value, it is determined that the symmetry of those lines is OK, and then the measured values X2 and Xn-1 are determined.

That is, the address of pointer 1 is subtracted by 1 and the second-to-last address and its measured value Xn-1 are stored, and the address of pointer 2 is added +1 and the second address and its measured value X2 are stored. Furthermore, +1 is added to the address of pointer 3, and the tolerance value at address No. 2 is stored.

Then, it is determined whether the address of pointer 1 is smaller than the address of pointer 2.

In other words, as mentioned above, if you repeatedly compare the symmetry of the first line and the last line,
In the end, the lines in the middle of the banknote 77 will be compared. When the address of pointer 2 becomes smaller than the address of pointer 1, it can be assumed that the banknote 77 is in the middle, although there is a difference depending on whether the number of lines is odd or even. Finish the inspection.

When the symmetry of all lines is recognized, it is determined that the banknote 77 is not an unfit banknote but is a banknote suitable as a banknote for withdrawal.

Also, if a mismatch in line symmetry is found midway through, the line will be rejected.

In this way, unfit banknotes are determined.

Note that the measured value X of the address finally stored in the pointer 1 of the memory 81 becomes the measured value Xe of the middle part of the banknote 77.

Next, with reference to the flowchart in FIG. 12, the line image sensors 66 and 67 determine the measured value X of the length of the banknote 77 in the longitudinal direction, and the lengths Y and V of the banknote 77 in the longitudinal and width directions. Let's talk about the judgment.

In the case of a banknote that is determined to be suitable as a banknote for withdrawal in the above-mentioned determination of damaged banknotes, the address of the intermediate position of the banknote and its measured value Xe are stored in the pointer 1 of the memory 81. The measured values Xe, Xe+1, and Xe+2 may have mutual errors within the tolerance range.

The measurement value X in the length direction is determined using the measurement value Xe described above.
An intermediate measured value is determined from among the measured value Xe+1 of the next line and the measured value Xe+2 of the next line.

That is, the measured value Xe stored in the aforementioned pointer 1 is compared with the measured value Xe+1 at the address of the next line, which is obtained by adding +1 to this address.

When the measured values Xe and Xe+1 are equal, the measured value Xe is determined to be the measured value in the longitudinal direction of the banknote 77, and the measured value Xe at point 1 is stored in the pointer 4.

However, when the measured values Xe and Xe+1 are not equal, the comparison process is performed as follows.

In other words, if Xe>Xe+1, pointer 1
+2 is added to the address, the third address is read from pointer 1, and the measured value Xe+2 is compared with the first measured value Xe.

In this comparison, if Xe≦Xe+2, the relationship is Xe+1<Xe≦Xe+2, and the measured value Xe is an intermediate value, so
This measured value Xe is stored in pointer 4.

However, if the aforementioned comparison is Xe>Xe+3, the magnitude relationship between the second measured value Xe+1 and the third measured value Xe+2 is unclear, so these measured values Xe+1 and Xe+2 are compared.

In this comparison, if Xe+2≦Xe+1, then the relationship is Xe+2≦Xe+1<Xe, and the second measured value Xe+1 is an intermediate value, so this measured value Xe+1 is stored in the pointer 4.

However, if Xe+1<Xe+2, then the relationship is Xe+1<Xe+3<Xe, and the third measured value Xe+2 is an intermediate value, so this measured value
Stored in

Furthermore, the above-mentioned first and second measured values Xe, Xe+
If the comparison of 1 is Xe<Xe+1, the first and third measured values Xe and Xe+2 are also compared.

In this comparison, if Xe+2≦Xe, then the relationship is Xe+2≦Xe<Xe+1, and the first measured value Xe is stored in pointer 4 as an intermediate value.

However, if the above comparison is Xe<Xe+2, the magnitude relationship between the second measured value Xe+1 and the third measured value Xe+2 is unclear, so these measured values Xe+1 and Xe+2 are compared.

In this comparison, if Xe+2≦Xe+1, then the relationship is Xe<Xe+2≦Xe+1, and the third measured value Xe+2 is stored in pointer 4 as an intermediate value.

However, if the above comparison is Xe+1<Xe+2, then the relationship is Xe<Xe+1<Xe+2, and the second measured value Xe+1 is stored in pointer 4 as an intermediate value.

In this way, the measured value of the banknote 77 in the longitudinal direction is
Xe is determined and stored in pointer 4.

Next, the actual size value is determined from the above-mentioned measured value Xe and it is determined whether it has the specified length. A measured value D is required.

Since the counters 91 and 92 described above stop when they finish counting, the measured value D on the lagging side may be read at this point, but the interrupt routine determines whether the measured value D is output or not, and if there is an output, this value is Measured value D
All you have to do is read it, store it in pointer 5, and return.

Next, the limit value of the length D of the side lagging behind the memory 81
The DX is read out and compared with the measured value D of the pointer 5 to determine whether the bill 77 is tilted too much.

If it is determined that the banknote is too tilted, the banknote is collected in the storage box 44D because there is a risk of an error in ejecting the banknote.

If the inclination of the banknote 77 is within the allowable range, then a correction value n1 whose address is the measured value D of the lagging side is read out from the longitudinal correction value n1 table in the memory 81 and stored in the pointer 6.

Then, the above-mentioned correction value n1 is subtracted from the measured value X in the length direction stored in the pointer 4 to calculate the actual size value Y in the length direction.

It is determined whether the calculated actual size value Y is appropriate as a banknote for withdrawal by determining whether it is within an allowable length range for determining a banknote for withdrawal stored in the memory 81 in advance.

As banknotes get older, they become shorter than new banknotes due to wrinkles, etc., so a tolerance range for the length of the banknotes is set, and if the banknotes are outside this tolerance range, the actual length value Y is short and the banknotes are determined to be old banknotes. and will be rejected.

If it is within the allowable range, then the length V in the width direction is determined in the same way.

That is, a correction value n2 whose address is the measured value D of the delayed side is read out from the width direction correction value n2 table and stored in the pointer 7.

Next, the measured value W of the counters 84 and 85 in the width direction
is read out and stored in pointer 8, and this measured value W
The actual size value V in the width direction is calculated by subtracting the correction value n2 from.

Then, it is determined whether or not this actual size value V is within the permissible length range for determining a banknote for withdrawal, and if it is outside the permissible range, it is rejected, and if it is within the permissible range, it is determined that the banknote is suitable for a banknote for withdrawal. Ru.

In this way, the banknote for withdrawal is OK if both its length and width are within the permissible length range.
, and other cases will be rejected.

After that, each counter 84, 85, 91, 92, 9
8, 99, FF99, further resets the data storage area and ends the process.

Regarding the correspondence between the structure of this invention and the structure of the above-mentioned embodiment, the paper sheet of this invention corresponds to the banknote 77 of the embodiment, and the conveyance device corresponds to the conveyance belts 28 and 49 in the same manner. , the image sensor is line image sensor 6
6 and 67, the memory corresponds to the memory 81, and the defect detection means corresponds to the longitudinal length/unfit note detection circuit 80 and the CPU 78, but the present invention is limited to the configuration of the above-described embodiment. It is not limited to.

[Brief explanation of drawings]

The drawings show an embodiment of the present invention, in which Fig. 1 is an external perspective view of an automatic deposit support machine, Fig. 2 is a schematic side view of a circulating deposit/withdrawal device, and Fig. 3 is a side view of a second bill discriminating section. , FIG. 4 is a sectional view taken along the line AA in FIG. 3,
Fig. 5 is an explanatory diagram of banknote length measurement, Fig. 6 is a control circuit block diagram of the second banknote discriminating section, Fig. 7 is a control circuit block diagram of the width direction length/inclination detection circuit, and Fig. 8 is a control circuit block diagram of the width direction length/inclination detection circuit.
The figure is a control circuit block diagram of the longitudinal length/unfit note detection circuit, Figure 9 is a time chart of the main part, Figure 10 is an explanatory diagram showing the reading state of the line image sensor, and Figure 11 is an unfit note. FIG. 12 is a flowchart for banknote detection and banknote length determination. 55... Second banknote discrimination unit, 66, 67... Line image sensor, 78... CPU, 80... Length direction length/unfit note detection circuit, 81... Memory, 93, 94...
Sensor drive circuit, 98, 99...counter.

Claims (1)

[Scope of Claims] 1. A sheet conveying device, and a sheet conveying device that scans the sheet conveyed by the conveying device multiple times in a direction perpendicular to the conveying direction and calculates the length of the sheet for each scan. an image sensor for reading; a memory for storing length data of a sheet on each scanning line read by the image sensor; and a memory for storing length data of a sheet on each scanning line read by the image sensor; Then, the length data of the paper sheets read from the memory by scanning is sequentially read out from the beginning and the end and compared, and the symmetry of the length data of the paper sheets is checked to detect missing paper sheets. A paper sheet discriminating device comprising a defect detection means for detecting presence or absence.